organizational behaviour comportement organisationnel - Library

Volume 27, No. 5, 2006Edited by ~ Édité parLouise TourignyUniversity of Wisconsin-WhitewaterORGANIZATIONAL BEHAVIOURCOMPORTEMENT ORGANISATIONNELProceedings of the AnnualConference of the AdministrativeScience Association of Canada,Organizationl Behaviour DivisionBanff, AlbertaJune 3-6, 2006Copies of these proceedings are availableonline.Actes de Congres annuel de laSection Comportement Organisationnelde l’associationdes sciences administrative du CanadaBanff, Alberta3-6 Juin 2006Des exemplaires de ces Actes peuvent êtreconnectés.

ACKNOWLEDGEMENTS ~ REMERCIEMENTSJe remercie et félicite toutes et tous lesmembres de la section ComportementOrganisationnel de leur contribution aucongrès de l’ASAC. Je remercie tous ceuxqui ont soumis leurs communications ànotre section et ceux qui ont servi commeprésident ou commentateur de séance.Aussi, je remercie tous les évaluateurs (voirla liste ci-jointe), qui ont su faire de solidesévaluations constructives dans des délaisserrés.It is my pleasure to thank all the membersof the Organizational Behaviour Divisionof ASAC for their contribution to thisconference. I especially thank everyonewho submitted their work to the divisionand those who served as session chairs anddiscussants.I also extend a sincere thank you to all theindividuals who served as ad hoc reviewersfor providing such thoughtful reviewswithin a short time frame.Louise TourignyAcademic reviewer/responsable de l’évaluationUniversity of Wisconsin-WhitewaterDenis MorinProgram Chair/responsable du programmeUniversité du Québec à MontréalCéleste BrotheridgeDivision Chair/responsable de la divisionUniversité du Québec à Montréal

REVIEWERS ~ ÉVALUATEURSCaroline AubéVishwanath V. BabaDaniel BeaupréNicole BérubéKathleen BoiesJames BronsonCéleste BrotheridgeRonald BurkeJulie CloutierNina ColeDaniel F. ColemanMagda DoniaWilliam L. DouganWilliam DragoLaura ErskineJoan FineganMarylène GagnéBella L. GalperinIan GellatlyYezdi GodiwallaSylvie GuerreroTracy HechtTamiko HynesI.M. JawaharJoe KrasmanKai LemertzScott LesterTerri LituchySanjay MenonSamantha MontesDenis MorinPeter MudrackSteven MurphyFilotheos NtalianisPraveen K. ParboteeahRobert ParentBruce PrinceMarcia PulichVincent RousseauNiti SinghAmy SommerFei SongRichard SoparnotRick TallmanRowan WagnerÉcole des Hautes Études Commerciales, MontréalMcMaster UniversityUniversité du Québec à MontréalConcordia UniversityConcordia UniversityUniversity of Wisconsin-WhitewaterUniversité du Québec à MontréalYork UniversityUniversité du Québec à MontréalRyerson UniversityUniversity of New BrunswickConcordia UniversityUniversity of Wisconsin-WhitewaterUniversity of Wisconsin-WhitewaterUniversity of Southern CaliforniaUniversity of Western OntarioConcordia UniversityUniversity of TampaUniversity of AlbertaUniversity of Wisconsin-WhitewaterUniversité du Québec à MontréalConcordia UniversityYork UniversityIllinois State UniversityYork UniversityConcordia UniversityUniversity of Wisconsin-Eau ClaireConcordia UniversityState University of Louisiana- ShreveportWilfrid Laurier UniversityUniversité du Québec à MontréalKansas State UniversityCarleton UniversityConcordia UniversityUniversity of Wisconsin-WhitewaterUniversité de SherbrookeKansas State UniversityUniversity of Wisconsin-WhitewaterUniversité de MontréalNarsee Monjee Institute of Management Studies, Mumbai, IndiaUniversity of Western OntarioRyerson UniversityGroupe ESCEM Tours-PoitiersUniversity of Northern British ColumbiaInternational Business School Kilajak Ilmi, Uzbekistan

Xiaoyun WangJohn WernerMichael WitheyUniversity of ManitobaUniversity of Wisconsin-WhitewaterMemorial University of NewfoundlandJuges des meilleures communications / Best Paper JudgesVishwanath V. BabaWilliam L. DouganPeter MudrackPraveen K. ParboteeahMcMaster UniversityUniversity of Wisconsin-WhitewaterKansas State UniversityUniversity of Wisconsin-Whitewater

TABLE OF CONTENTS ~ TABLE DES MATIÈRESPRIX D’EXCELLENCE– BEST PAPER AWARDDevelopmental Antecedents of Leader-Follower Relationships and Trust ........................ 1Stephanie Grosvenor (JMSB, Concordia University)Kathleen Boies (JMSB, Concordia University)MENTION D’HONNEUR- HONOURABLE MENTIONContinuing Employment Discrimination amidst the Need to EmbraceDiversity: The Role of Organizational Recruitment Climates........................................... 19Chetan Joshi (University of Western Ontario)Joerg Dietz (University of Western Ontario)Victoria Esses (University of Western Ontario)Crafting Job Satisfaction: The Interplay between Volunteer and Paid WorkExperiences ............................................................................................................................. 34Kevin Hill (student, University of Toronto)J. Mark Weber (University of Toronto)How Do Work Motivation and Social Support Predict Job Burnout Over Time?.......... 51Claude Fernet (Ph.D. candidate, Université Laval)Marylène Gagné (JMSB, Concordia University)Can Organizational Support and Political Skill Reduce Burnout?................................... 65I.M. Jawahar (Illinois State University)Thomas H. Stone (Oklahoma State University)Multiple Identifications in Organizations............................................................................ 81Peter Lee (student, York University)Tacit Knowledge Integration in Organizations: A Knowledge-AssemblyPerspective .............................................................................................................................. 91A.R. Elangovan (University of Victoria)Jia Lin Xie (University of Toronto)Kevin Hill (student, University of Toronto)Perceptions of Leadership as a Determinant of Legitimate and IllegitimateUse of Sick Leave.................................................................................................................. 107Morris B. Mendelson (University of New Brunswick Saint John)Jeff Frooman (University of New Brunswick Saint John)J. Kevin Murphy (MBA student, University of New Brunswick Saint John)Straight From the Source’s Mouth: Exploring the Implications of FeedbackSeeking On the Source ......................................................................................................... 124Christopher Fredette (student, York University)Joseph Krasman (student, York University)

Organizational Structure and Organizational Commitment: A ContingencyApproach............................................................................................................................... 142Ivy Kyei-Poku (student, University of Western Ontario)Communities of Practice and Individual Performance .................................................... 156Zeying Wan (student, University of Western Ontario)Uncertainty, Threat, and Work Attitudes in Anticipation of a Merger.......................... 170D. Lance Ferris (University of Waterloo)Jeanetta Vena (Export Development Canada)John L. Michela (University of Waterloo)Contexte De Travail Et Engagement Envers L’organisation: Le RôleMédiateur Du Soutien Du Supérieur Hiérarchique.......................................................... 181Jacques Barrette (University of Ottawa)Louise Lemyre (University of Ottawa)Wayne Corneil (University of Ottawa)Nancy Beauregard (University of Ottawa)Louise Legault (University of Ottawa)Individual Determinants of Threshold Models of Collective Action:A Laboratory Simulation .................................................................................................... 198Emily Block (student, University of Illinois at Urbana-Champaign)Laura Erskine (student, University of Southern California)Dispositional and Contextual Determinants of Teamwork BehaviorIn Self-Managing Teams: A Longitudinal Cross-Level Analysis .................................... 220Kevin Tasa (McMaster University)Greg Sears (Public Service Commission of Canada)Aaron Schat (McMaster University)The Mediating Role of Trust in Psychological Contract Breach Effects:Clarifying the Relations....................................................................................................... 236Samantha Montes (Wilfrid Laurier University)Bridging and Bonding Social Capital in Organizational Groups: TheDevelopment of Assessment Tools ...................................................................................... 243Paul Varella (University of Calgary)The Influence of Relational Age on Older Employees’ PerceivedTreatment.............................................................................................................................. 253Marjorie Armstrong-Stassen (University of Windsor)Seung Hwan Lee (student, University of Windsor)

ABSTRACTS/RESUMESLeadership during Critical Times: Transformational Leadership ................................... 266Influences Newcomers’ Union Attitudes Eleven Years LaterAnthony E. Carroll, Queen’s School of BusinessJulian Barling, Queen’s School of BusinessDaniel G. Gallagher, James Madison UniversityPaul F. Clark, Pennsylvania State UniversityClive J. Fullagar, Kansas State UniversityDispositional, Relational, and Attitudinal Correlates of OrganizationalCitizenship Behavior: A Test of an Integrative Model ...................................................... 267Laurent M. Lapierre, University of OttawaRick D. Hackett, McMaster UniversityPourquoi Les Pratiques De Gestion Des Ressources Humaines Ne Peuvent Se Passer DuSupport Et De L’engagement Dans La Mobilisation Des Employés? .............................. 268Nadine Quenneville (étudiante), Université du Québec à MontréalGilles Simard, Université du Québec à MontréalKathleen Bentein, Université du Québec à MontréalTransformational Leadership in India: Developing and Validating a New Scale UsingGrounded Theory Approach ................................................................................................ 269Niti Singh (Doctoral Student), Institute of Management Studies, IndiaNarsee Monjee, Institute of Management Studies, IndiaVenkat R. Krishnan, Xavier Labour Relations Institute, IndiaHow Far Does Leadership Travel? Distance as a Multidimensional Concept ................. 270Laura Erskine, University of Southern CaliforniaLa Coopération Interindividuelle: Une Perspective Par La ThéorieDe L’apprentissage, Le Cas De La Gestion De Projet Dans Le Secteur Automobile...... 271Richard Soparnot, Professeur de Management Stratégique ESCEMLe Travail Émotionnel: Analyse Critique Et Implications Pratiques............................... 272Céleste M. Brotheridge, ESG-UQAMDu Courage D’agir Au Courage D’être Profils Du Courage Managérial ......................................273Michelle Harbour (étudiante), HEC MontréalLe Rôle Du Supérieur Dans La Reconstruction Du Lien Social Entre L’employé EtL’organisation......................................................................................................................... 274Olivier Doucet (étudiant), ESG-UQAMMichel Tremblay, HEC MontréalGilles Simard, ESG-UQAM

The Affective Underpinnings of Psychological Contract Fulfillment ............................... 275Sylvie Guerrero, Université du Québec à MontréalOlivier Herrbach, Université Montesquieu – Bordeaux IVEffects of Problem Structure on Behaviour and Performance in Group ProblemSolving: Theory and Experiment.......................................................................................... 276Gbemisola A. Abimbola (student), University of WaterlooZhehui Zhong (student), University of WaterlooP. Robert Duimering, University of WaterlooLe Rôle De L’autonomie Au Travail Et Du Lieu De Contrôle En Regard De La SantéPsychologique : Effets Principaux Et D’interaction ........................................................... 277Vincent Rousseau, Université de MontréalCaroline Aubé, HEC MontréalEstelle M. Morin, HEC MontréalBalancing the Roles of Student, Family Member and Paid Employee:An Investigation of Coping, Conflict, Facilitation, and Satisfaction................................. 278Tracy Hecht, Concordia UniversityWork Motivation and Commitment: Multidimensional and Multifoci Perspectives ...... 279Marylène Gagné, JMSB Concordia UniversityPerceptions of Group Potency and Personality on Volunteer Group Member SellingBehavior .................................................................................................................................. 280Mitchell J. Neubert, Baylor UniversitySimon Taggar, Wilfrid Laurier University

ASAC 2006Banff, AlbertaStephanie GrosvenorKathleen BoiesJohn Molson School of BusinessConcordia UniversityDEVELOPMENTAL ANTECEDENTS OF LEADER-FOLLOWERRELATIONSHIPS AND TRUSTThis study attempted to determine the role of developmental antecedents in explainingleadership style and trust in the leader. The hypotheses were tested in the field withquestionnaires. Transformational leaders were securely attached, had high-quality LMXrelationships, and were perceived as benevolent and trustworthy by their followers. Somemediation effects were observed.As today’s organizations merge and restructure, effective leadership is becoming increasinglyimportant. Moreover, with the growing diversity of today’s workforce, it is becoming more and moreimportant to find ways for leaders and followers to engage in trusting, give-and-take relationships. Inorder to have positive relationships with followers, leaders must foster trust. Trust in leadership has beenfound to lead to many positive organizational outcomes such as increased performance, organizationalcitizenship behaviors (OCBs), and job satisfaction (Dirks & Ferrin, 2002). Thus, it is imperative to findout what factors explain trust in the leader.Recent research in the management field has focused on transformational leadership (Bass, 1985;1998) as an effective way to inspire and motivate followers. The transformational leadership theoryfocuses on the behaviors of the leader himself. Leadership, however, would not exist without followersand thus, leader-follower relationships must be understood. The leader-member exchange theory (LMX)takes a relationship approach to leadership, which takes into account the dynamic relationship betweenleaders and their followers (Graen & Uhl-Bien, 1995).It is important to understand what factors contribute to this relationship. Previous research haslooked at the developmental antecedents of transformational leaders (Popper, Mayseless, & Castelnovo,2000) in an attempt to understand how transformational leaders develop their relational qualities. Thesedevelopmental antecedents are called “attachment styles” (Ainsworth, Blehar, Waters, & Wall, 1978;Bartholomew, 1990; Hazan & Shaver, 1987). Essentially, attachment styles form during infancy, througha child’s relationship with his/her primary caregiver. These attachment styles act as blueprints for the waythat children grow to interact and form relationships with others in their adult life. Attachment styles cantherefore offer a potential explanation as to how transformational leaders develop and form relationshipswith their followers that are characterized by trust.The present study will attempt to answer two research questions: (1) What is the role ofdevelopmental antecedents in explaining leadership style and trust in the leader? (2) What specific factorsinfluence the decision to trust the leader? This study will expand on Mayer, Davis, and Schoorman’s(1995) model of trust, in conjunction with the developmental theory of attachment (Bartholomew, 1990),transformational leadership theory (Bass, 1985; 1998) and leader-member exchange theory (e.g., Graen &Uhl-Bien, 1995).1

AttachmentAccording to attachment theory, learning how to interact with others begins at infancy(Ainsworth et al., 1978; Bowlby, 1973). Infants form mental models of themselves and others byinternalizing interactions with their primary caregivers over time (Bowlby, 1973). The child’s mentalmodel of others is based on how available the attachment figure is when needed, and the child’s mentalmodel of the self is based on whether the child judges him or herself to be worthy of the attachmentfigure’s attention. Following Bowlby’s research, Ainsworth et al. conducted a series of laboratory studieswith infants and their mothers. Basically, each child had a different reaction to separation and reunionwith his or her mother. These reactions were analyzed and classified into three types of infant attachment:secure, anxious/ambivalent, and avoidant. The secure style is characterized by a child who has a closerelationship with his/her mother and receives consistent love and support. This child grows into an adultwho is confident and trusting of others. The anxious/ambivalent child receives ambiguous andinconsistent responses from his/her mother. This child grows into an adult who desires close relationshipswith others, but fears rejection. Finally, the avoidant style is characterized by a mother who is cold andnon-responsive. This child grows up to rely solely on him/herself with insecurity regarding the intentionsof others (Ainsworth et al., 1978; Hazan & Shaver, 1987). The attachment styles provide the groundworkfor the way children will grow to interact with others later in life (Bowlby, 1973; Hazan & Shaver, 1987).Hazan and Shaver (1987) found that the childhood attachment patterns identified by Ainsworth et al.(1978) continued into adult romantic relationships. Thus, the relationship or attachment style that a childdevelops through interactions with his/her primary caregiver has a large effect on how an adult interactswith significant others.This tri-partite definition of attachment was expanded by Bartholomew in 1990 to include afourth attachment style. Bartholomew renamed Hazan and Shaver’s anxious/ambivalent style to be calledpreoccupied and separated the avoidant style into two distinct styles called fearful and dismissing. Afearful adult is described as being uncomfortable with the lack of close relationships in his/her life, whilethe dismissing adult is uncomfortable being in close relationships altogether. The present study is basedon Bartholomew’s (1990) four-category model of attachment, which describes attachment styles inadulthood.Transformational LeadershipTransformational leaders are unique in that they have passion and vision and articulate thosecharacteristics and beliefs to their followers. They motivate followers to perform beyond what theynormally would in a simple exchange relationship. According to Bass (1985; 1998) who expanded Burns’(1978) version of transformational leadership, there are four components: individualized consideration(the leader pays individual attention to each subordinate), intellectual stimulation (the leader challengessubordinates’ assumptions), inspirational motivation (the leader formulates an inspiring vision of thefuture), and idealized influence, or charisma (the leader invokes feelings of admiration, respect, and trustin his/her followers).Leader-Member ExchangeThe transformational leadership theory focuses on the characteristics and behaviors of the leaderalone. On the other hand, the leader-member exchange theory is described as a “relationship-basedapproach” (e.g., Graen & Uhl-Bien, 1995). Thus, it provides us with further insight into the complexrelationship between leaders and their followers (Graen & Uhl-Bien, 1995).Leader-member exchange (LMX) is based on the original theory called Vertical Dyad Linkage(VDL: Dansereau, Graen, & Haga, 1975), which posited that leaders can have different relationships witheach of their followers. In other words, leaders do not adopt the same leadership style with all of their2

followers. A leader might be transformational, but depending on the particular follower, he/she may havea transformational or transactional relationship with them. A high-quality LMX relationship ischaracterized by trust, respect, and mutual obligation (Graen & Uhl-Bien, 1995).Gerstner and Day (1997) point out that LMX is more strongly related to member affectiveoutcomes than to objective measures, such as turnover and productivity. Gomez and Rosen (2001) foundthat the higher the managerial trust, the more likely the member rated his/her relationship with his/herleader as high quality. Given the relational nature of LMX, it is not surprising that outcomes such asorganizational citizenship behaviors (Deluga, 1994), satisfaction with supervision (Gerstner & Day,1997), and managerial trust (Gomez & Rosen, 2001) have been found to be correlates of LMX inprevious research. The present study expands on the previous research by not only looking at trust as anoutcome of LMX relationships, but also in answering a call from Gerstner and Day (1997) for morestudies examining the antecedents of leader-member exchange.TrustTrust in organizational relationships has been the topic of recent research (e.g., Mayer, Davis, &Schoorman, 1995). Mayer et al. developed an integrative model of organizational trust, which asserts thattrust is a combination of the characteristics of the trustor (i.e., the person doing the trusting) and thecharacteristics of the trustee (i.e., the person being trusted).Trustees are judged on their ability, benevolence and integrity by the trustor. Ability is a “group ofskills, competencies and characteristics that enable a party to have influence within some specificdomain” (p. 717). Benevolence is “the extent to which a trustee is believed to want to do good to thetrustor, aside from an egocentric profit motive” (p. 718). Integrity “involves the trustor’s perception thatthe trustee adheres to a set of principles that the trustor finds acceptable” (p. 719). According to Mayer etal. (1995), trust is defined as a willingness to be vulnerable to the trustee, based on the belief that he/shewill act in the trustor’s best interest, regardless of whether or not the trustee’s behavior can be monitoredor controlled. Thus, if the trustor perceives the trustee to have high ability, benevolence and integrity, thetrustor will likely trust the trustee. Furthermore, once trust exists between a given dyad, risk-takingbehavior can occur. This is the behavioral manifestation of the feeling of trust in the dyad, where the twofeel that they can depend on each other to do what they ask each other to do.Being able to trust co-workers and leaders has many positive benefits in the organizationalenvironment. Trust in leadership has been found to lead to outcomes such as increased performance,organizational citizenship behaviors (OCBs), and job satisfaction (Dirks & Ferrin, 2002). Specifically, ina study by Boies and Corbett (2005), transformational leadership was found to be positively related tosubordinate perceptions of ability, benevolence, and integrity. Ability and integrity were found to besignificant predictors of trust and mediated the relation between transformational leadership and trust.Benevolence was not found to be related to trust in their study. However, in a study exploring Mayer atal.’s (1995) model, Gill, Boies, Finegan, and McNally (2005) found that co-workers perceived as havinghigh ability, benevolence, and integrity were more likely to be trusted than co-workers perceived ashaving low levels of these three antecedents of trust. Furthermore, in a quasi-experiment testing their ownmodel, Mayer and Davis (1999) found that ability, benevolence, and integrity played a mediating role inthe relation between perceptions of a new appraisal system and trust in top management. Given thepositive outcomes related to trust, the present study delves deeper into the nature of trusting relationshipsbetween leaders and followers.In the present study, the Mayer et al. (1995) model was chosen because this model distinguishesbetween perceptions of trustworthiness (i.e., ability, benevolence, and integrity) and the construct of trustitself. The separation of the perceptions from trust allows for an understanding of the mechanism through3

which trust occurs. It is the mechanism through which trusting leader-follower relationships are formedthat will be examined in greater detail in the present study. Of the three perceptual variables in Mayer etal.’s model, benevolence is the only factor that can be considered relational. Given its relevance,therefore, the focus of the current study will be on benevolence.A model linking the constructs defined previously is outlined below. The proposed modelsuggests that secure attachment is related to both transformational leadership and LMX, and that it isthrough positive leader-member relations that followers come to perceive their leaders as benevolent andtrustworthy. These relationships are illustrated in Figure 1. The links between each of these constructs andsupport for the hypotheses are now discussed in turn.Figure 1Graphical Representation of the Hypothesized Direct RelationsSecureAttachmentH1aH2aLeader-MemberExchangeH3BenevolenceH4TrustTransformational LeadesrhipH2bThe Developmental Antecedents of LeadershipAttachment styles provide insight into the developmental antecedents of how adults interact andform relationships, and have therefore great relevance for the study of leader- follower relationships.According to Bartholomew and Horowitz (1991), securely attached people are generally comfortable inrelationships with others and believe that others are trustworthy. As well, securely attached people tend tohave healthy relationships characterized by happiness, friendliness, and trust (Hazan & Shaver, 1987).These studies have prompted researchers to investigate how attachment influences the way people behaveand interact in the work setting.Hazan and Shaver (1990) examined the relation between attachment styles and on-the-jobbehaviors and attitudes. In two studies that supported their hypotheses, secure respondents were morelikely to be satisfied with their jobs, and they were more likely to feel that they were good workers andthat their co-workers evaluated them positively. As well, secure workers were not worried about beingrejected by co-workers, had positive views about themselves, and felt confident in completing workrelatedtasks. As expected, anxious/ambivalent (i.e., preoccupied) participants reported a preference towork with others rather than alone, but felt underappreciated and unrecognized. They also reported beingpreoccupied with wanting approval from others and felt that this interfered with their productivity.Finally, avoidant participants (i.e., fearful/dismissing) dissociated themselves from co-workers andpreferred to work alone. However, they seemed to be satisfied with their jobs. This is typical of theavoidant relationship style in that the person avoids relating to others by burying him/herself in his/herwork. Schmidt and Bell (2005) related attachment style to psychological contract and organizational4

commitment. They found that preoccupied individuals had higher perceived psychological contractviolations and lower organizational commitment than individuals exhibiting the other attachment styles.Sumer and Knight (2001) examined the relation between attachment and work/family spillover effects.Preoccupied individuals were more likely to experience negative spillover from the home/family to thework domain, whereas secure individuals were more likely to experience positive spillover in bothdomains.Some of the above evidence can be applied to the leadership domain. For example, leaders whofeel confident and have positive relationships with others (i.e., securely attached) are likely to betransformational, whereas those who would prefer to work alone (i.e., fearful/dismissing) are probablynot. Previous research has shown that secure attachment is positively related to the four components oftransformational leadership (i.e., inspirational motivation, idealized influence (charisma), intellectualstimulation, and individual consideration) (Popper et al., 2000). Perhaps being securely attachedfacilitates individualized consideration, given that secure attachment reflects a positive relationship stylewith others. Secure attachment is likely to have more of an influence on certain components oftransformational leadership (e.g., individualized consideration), and perhaps this can provide oneexplanation as to how secure attachment is related to transformational leadership. As further evidence forthis relation, it has also been shown that the insecure attachment styles (i.e., preoccupied, fearful, anddismissing) are negatively related to transformational leadership and its components (Popper et al., 2000).Thus, a positive relationship style is likely to influence perceptions of transformational leadership,whereas negative relationship styles associated with insecure attachment are likely to detract fromperceptions of transformational leadership.One study examined the link between transformational leadership and attachment directly. Popperet al. (2000) conducted a series of three studies in order to test their hypothesis that secure attachment iscorrelated with transformational leadership. The first sample consisted of six teams from an officer’scourse in the Israeli Police, each led by a commander. The cadets in each team were being trained tobecome leaders and had been previously selected based on their aptitude and personality. In this case, thecommanders were asked to assess both the attachment and leadership styles of their trainees. The othertwo studies were conducted in the Israeli Defense Forces (IDF). In the second study, soldiers in anofficer’s course were asked to report on their own attachment style and two commanders rated eachtrainee on their leadership style. The third study used more conventional methods of measurement in thatit was now the soldiers who were asked to assess their commanders’ leadership style and the commandersrated their own attachment style.All three of the studies obtained similar results supporting the hypothesis that secure attachmentcorrelates highly with transformational leadership. The third study looked at whether secure attachmentwas correlated with any of the other leadership styles, such as contingent reward and management-byexception,to assess discriminant validity. Secure attachment did not show significant correlations witheither of the two. This finding further supports the decision to focus only on transformational leadershipin the present study. In a more recent study by Popper, Amit, Gal, Mishkal-Sinai, and Lisak (2004),secure attachment, along with internal locus of control, low level of anxiety, high self-efficacy andoptimism were defined as psychological capacities to lead. Leaders were found to possess these capacitieswhereas non-leaders were not. All of these important findings lead to the first two hypotheses:Hypothesis 1a: Secure attachment style will be positively related to transformational leadership.Hypothesis 1b: Fearful, preoccupied, and dismissing attachment styles will be negatively related totransformational leadership.Inherent in transformational leadership is the notion that these leaders establish high-qualityrelationships with their followers. Certain transformational leadership characteristics such as5

individualized consideration and charisma facilitate the development of high-quality LMX relationships.This suggests that there may be a link between transformational leadership and LMX. The positiveassociation between transformational leadership and LMX has been established empirically (Howell &Hall-Merenda, 1999).A link can also be established between LMX and attachment styles in that they both essentiallydescribe the way people relate to each other. A high-quality LMX relationship is dominated by trust,respect, and mutual obligation (Graen & Uhl Bien, 1995) and a securely attached person relates positivelyto others and is trusting in the workplace (Hazan & Shaver, 1990). Thus, it is expected that both secureattachment and transformational leadership would be related to LMX. Two more hypotheses can bederived from these findings:Hypothesis 2a: Secure attachment will be positively related to the quality of leader-member exchange.Hypothesis 2b: Transformational leadership will be positively related to the quality of leader-memberexchange (LMX).Outcomes of Positive Leader-Follower RelationshipsOne of the key aspects of LMX, according to Graen and Uhl Bien (1995), is that there is a cyclethrough which the relationship goes in order to reach high-quality status. This cycle begins at theacquaintance phase and ends at the partnership phase. Throughout the cycle, leaders and members aregetting to know each other better and are able to make better assessments about each other in the process.One of the propositions from Mayer et al.’s (1995) integrative model of trust states that “the effect ofperceived benevolence on trust will increase over time as the relationship between the parties develops”(p. 722). Ability and integrity are proposed to be more salient early on in the relationship between the twoparties, perhaps when information on benevolence is less available or less easily obtained. Thus, becausethis study focuses on transformational leaders and high-quality LMX relationships (that by definition taketime to form), benevolence appears highly relevant.Among the three antecedents of trust according to Mayer et al. (1995), benevolence is the onlyrelational factor in that it actually involves the trustee’s personal relationship with the trustor. In this case,it involves how well intentioned the leader is towards his/her followers. A follower can only judgewhether or not his/her leader is benevolent based on personal interactions and experiences over time. Infact, high-quality LMX relationships are characterized by mutual obligation, which is conceptuallysimilar to benevolence in that they both involve positive intentions and actions towards others.According to Mayer et al. (1995), perceptions of benevolence lead to trust. Some previousfindings have supported the positive relation between benevolence and trust (e.g., Gill et al., 2005; Mayer& Davis, 1999) while others have not (e.g., Boies & Corbett, 2005). No research to date has focusedsolely on benevolence as a predictor of trust. However, because this study pertains to leader-followerrelationships and how trust is fostered, it appears likely that followers would base their assessments oftheir leaders’ trustworthiness on perceptions of benevolence. The preceding discussion leads to thefollowing hypotheses:Hypothesis 3: The quality of leader-member exchange (LMX) will be positively related to followerperceptions of leader benevolence.Hypothesis 4: Follower perceptions of leader benevolence will be positively related to trust in the leader.Mediation HypothesesIn addition to the hypothesized direct relations, three mediating hypotheses can be derived.Firstly, a person who is securely attached is both trusting and confident in relationships (Hazan & Shaver,6

1990), and is comfortable depending on others and having others depend on him/her (Hazan & Shaver,1987). It seems logical that someone who relates to others in such a positive way would be consideredbenevolent, in that they would look out for the best interests of others. In the context of leadership, secureattachment may not determine perceptions of benevolence directly. However, perhaps it is through theestablishment of high-quality LMX relationships that securely attached leaders are perceived asbenevolent. More specifically, since secure attachment reflects a positive relationship style, leaders whoare securely attached may have greater facility establishing high-quality relationships with their followers,which may, in turn, influence follower perceptions of leader benevolence. In high-quality relationships,which form over time, securely attached leaders may have more occasions to display acts of benevolence,and thus influence their followers’ perceptions of their desire to act in their best interest. Thus, thefollowing mediating hypothesis is proposed:Hypothesis 5a: The relation between secure attachment and follower perceptions of benevolence will bemediated by LMX.Secondly, transformational leadership has been found to be positively related to followers’perceptions of benevolence in previous research (Boies & Corbett, 2005). Since transformational leaderspay close attention to individuals and are supportive (i.e., through individualized consideration), it seemslogical that they would be perceived as benevolent. The same relation is proposed in the present study,however it is believed that transformational leadership acts on followers’ perceptions of benevolencethrough positive leader-follower relationships. LMX has been previously established as a mechanism fortransformational leaders to influence followers’ performance and organizational citizenship behaviors(Wang, Law, Hackett, Wang, & Chen, 2005), a construct conceptually similar to benevolence. In thiscase, LMX is believed to act on followers’ perceptions of benevolence. Transformational leaders, bydefinition, establish high-quality relationships with their followers. As previously stated, these highqualityrelationships develop over time, allowing followers the opportunity to witness their leaders actingin their best interest. Thus, the following mediating hypothesis is proposed:7

Hypothesis 5b: The relation between transformational leadership and follower perceptions ofbenevolence will be mediated by LMX.Finally, LMX relates to trust because high-quality LMX relationships are characterized by trust(Graen & Uhl-Bien, 1995). However, before high-quality LMX relationships can lead to trust, LMX mayact on followers’ perceptions of benevolence in the manner described previously. Benevolence hasalready been found to be correlated to trust (Gill et al., 2005) and thus, a logical extension is thatbenevolence could be the mechanism through which high-quality LMX relationships are perceived astrustworthy. Furthermore, ability, benevolence, and integrity have already been established as mediatorsin the relation between perceptions of a new performance appraisal system and trust in top management(Mayer & Davis, 1999). Specifically, the introduction of the new appraisal system influenced trust in topmanagement through employees’ perceptions of ability, benevolence, and integrity. The final hypothesisis as follows:Hypothesis 6: The relation between LMX and trust in the leader will be mediated by benevolence.In sum, secure attachment is expected to be positively related to transformational leadership andLMX, given that a secure attachment style is the basis for positive relationships with other people (Hazan& Shaver, 1987). It is expected that through these positive relationships, followers will perceive theirleader as benevolent and trustworthy. In the sections that follow, the method for testing the proposedmodel is described, the results of the study are presented, and finally, the results are interpreted andimplications are discussed.Sample and ProcedureMethodThe sample consisted of a total of 121 employees from three different organizations in theMontreal area: a large local hospital affiliated with a university health center, a medium-size investmentbanking firm and a large manufacturing company. The 121 employees rated 20 leaders. The number ofpeople per leader ranged from 2 to 25 with an average of 6.61 individuals per leader. Three hundred andeighteen questionnaires were distributed, and 121 were returned for a response rate of 38%. Age of theparticipants ranged from 21 to 60 (M=38.9, SD=9.5). Seventy-eight percent of the participants werefemale and 22% were male. Most participants had undergraduate or graduate degrees (76%). They hadbeen working with their current manager/supervisor for an average of 4.8 years (SD=4.6).The organizations used to collect data in this study were recruited through personal contacts.Each of the contacts in the three organizations agreed to distribute the questionnaires to managers orsupervisors who had previously agreed to participate in the study. Each manager/supervisor was asked tofill out a questionnaire regarding his/her own leadership and attachment styles. The leader was also askedto distribute the follower questionnaires to all of his/her subordinates. Because there were too few leaderswho returned questionnaires, all analyses are based exclusively on follower information, which is in linewith previous research (e.g., Popper et al., 2000). Subordinates were informed that the survey askedquestions about their immediate supervisor/manager and their relationship with them. There were twoversions of the questionnaire to assess order effects. The questionnaires were also distributed in bothEnglish and French. All scales were previously validated in English. The MLQ was previously translatedand validated in French (Cacciatore, Faulk, Perret, & Antonakis, 2003). The benevolence and trust scales,the Leader-Member Exchange questionnaire, and the Relationship Questionnaire were translated by aprofessional translator and verified by the researchers (one French-speaking and one English-speaking)for accuracy.8

MeasuresLeadership style. Leadership style was assessed using the Multifactor Leadership QuestionnaireForm 5X (MLQ: Bass & Avolio, 1995). Four items measure each of the following subscales: inspirationalmotivation, idealized influence (behavior), idealized influence (attributed), intellectual stimulation, andindividualized consideration. These subscales form a higher-order factor, which is transformationalleadership (e.g., Avolio, Bass, & Jung, 1999). According to Bass and Avolio (2000), each of the subscalesreceived adequate internal consistency reliability and have acceptable validity. Each item was rated on a5-point scale where participants were asked to assess the frequency of occurrence of each item (0 = not atall to 4 = frequently, if not always).LMX. Leader-member exchange was measured by the LMX-7 (Graen & Uhl-Bien, 1995). Sevenitems with 5-point anchored rating scales assessed the quality of the exchange. A sample item was “Howwell does your supervisor recognize your potential?” LMX was measured from the follower’sperspective, as recommended by Gerstner and Day (1997).Attachment style. Attachment style of the leader was measured from the follower’s point ofview using Griffen and Bartholomew’s (1994b) Relationship Questionnaire (RQ). This measure containsfour paragraphs, rated on a 7-point scale (1 = not at all like him/her to 7 = very much like him/her). Eachparagraph describes one attachment style: secure, fearful, preoccupied, and dismissing. An example of thesecure description was: “It is easy for my supervisor to become close to others. He/she is comfortabledepending on others and having others depend on him/her. He/she doesn’t worry about being alone orhaving others not accept him/her.” It is recommended that this be used as a continuous measure ratherthan a categorical one (Griffen & Bartholomew, 1994b) because it assesses the degree to which a personis secure, fearful, dismissive and/or preoccupied; it was not originally intended to be used a way tocategorize participants into specific types.Benevolence. Benevolence was measured by five items developed by Mayer and Davis (1999).Each item was rated on a 5-point scale (1 = strongly disagree to 5 = strongly agree). A sample item was:“My supervisor is very concerned with my welfare.”Trust. Participants rated the degree to which they trusted their supervisor or manager. Four itemsdeveloped by Mayer and Davis (1999) were used to measure trust in the leader. Each item was rated on a5-point scale (1 = strongly disagree to 5 = strongly agree). A sample item was: “I would be willing to letmy supervisor have complete control over my future in this company.” Although Cronbach’s alpha forthis scale was slightly lower than .7 (α = .67), this value is considered marginally acceptable, thus thescale was included. It should be noted that similar Cronbach’s alphas for this scale were obtained inprevious studies (e.g., Mayer & Davis, 1999).Data AnalysisLevel of analysis. The sample consisted of 20 groups of employees each rating their respectiveleader. Analyses were conducted at the individual level because there were not enough groups to be ableto conduct analyses at the leader (or group) level. Furthermore, although individuals rated the sameleader, their work was not organized around teams and, therefore, the individual level of analysisappeared appropriate. However, to account for variance that may be attributed to having the same leader,all regression analyses controlled for group membership.T-tests. An independent-samples t-test was conducted to verify whether there were significantdifferences between the two counterbalanced versions of the questionnaire. The independent variable was9

the version of the questionnaire and the dependent variables were the scores on each of the scales (i.e.,transformational leadership, LMX, benevolence, trust, and attachment). The analyses showed that therewere no significant differences between scale scores for those who answered version 1 and those whoanswered version 2, except for the secure scale, which was significant (results are shown in Table 1).Because of this difference, hierarchical regression analyses testing for hypotheses including secureattachment (Hypotheses 1a, 1b, 2a, and 5a) were conducted with version as a control variable. This wasdone in order to verify whether the significant difference for secure attachment on the different versionsof the questionnaire made a difference in the results. The same pattern of results was obtained whencontrolling for group and version as when controlling for group alone, showing that the difference insecure attachment for version did not have an effect on the results. Thus, all regression analyses presentedin the results which included secure attachment controlled for group only.An independent-samples t-test was also conducted to verify whether there were significantdifferences between the French and English questionnaires. The independent variable was the language ofthe questionnaire and the dependent variables were the scores on each of the scales. The analyses showedthat there were no significant differences between those who answered the questionnaires in French andthose who answered the questionnaires in English for all scales except the preoccupied scale. In order todetermine whether the significant difference for the preoccupied scale on the different languages of thequestionnaire had an effect on the results, hierarchical regression, controlling for group and language, wasconducted on Hypothesis 1b. The same pattern of results was obtained when controlling for group andlanguage of the questionnaire as when controlling for group alone. This indicates that the difference in thepreoccupied scale was not meaningful, and thus the regression analysis presented for Hypothesis 1bcontrolled for group only.Analysis of variance. A one-way ANOVA was conducted to verify whether there weresignificant differences between the three organizations. The independent variable was the organizationand the dependent variables were the scores on each of the scales. Only one significant difference wasfound for the fearful scale. No significant differences were found for any of the other scales. Hierarchicalregression, controlling for group and organization, was conducted on Hypothesis 1b, in order to verifywhether the significant difference for the fearful scale in the different organizations had an effect on theresults. The pattern of results obtained was the same when controlling for group and organization as whencontrolling for group alone, showing that the significant difference did not have a meaningful effect onresults. Thus, as previously mentioned, the regression analysis presented for Hypothesis 1b controlled forgroup only.10

ResultsMeans, standard deviations, Cronbach’s alphas, and intercorrelations between variables can befound in Table 1.Table 1Means, Standard Deviations, Cronbach’s Alphas, and Intercorrelations among VariablesM SD 1 2 3 4 5 6 7 81. Secure 4.35 1.5 --2. Fearful 2.84 1.7 -.55** --3. Preoccupied 2.89 1.6 -.14 .32** --4. Dismissing 3.98 1.7 -.36** .43** -.12 --5. TFL 2.39 .75 .40** -.47** -.38** -.21* .946. LMX 3.46 .82 .38** -.43** -.26** -.18 .84** .917. Benevolence 3.56 .89 .40** -.41** -.29** -.19* .78** .80** .928. Trust 3.4 .75 .39** -.35** -.35** -.12 .69** .70** .72** .67Note. TFL: Transformational leadership. Diagonal elements in boldface represent coefficient alphas for each scale. Off-diagonalelements are correlations between measures. Hypothesized correlations are one-tailed. Non-hypothesized correlations are twotailed.* p < .05, ** p < .01.Hypothesis 1a predicted that secure attachment style would be positively related totransformational leadership and this relation was found (r(120) = .40, p < .01, one-tailed). Hypothesis 1bpredicted a negative relationship between the three insecure attachment styles (fearful, preoccupied, anddismissing) and transformational leadership. Correlations for fearful, preoccupied, and dismissing werenegative and significant (r(120)= -.47, p < .01, one-tailed; r(116) = -.38, p < .01, one-tailed; r(116) = -.21,p < .05, one-tailed) respectively. After controlling for group, transformational leadership was regressed onall four attachment styles. This yielded a significant change in the squared multiple correlation (∆R 2 =.21, ∆F(4,91) = 9.14, p < .01). The beta weight associated with secure attachment style was marginallysignificant (p < .10), and fearful and preoccupied attachment styles were significant predictors oftransformational leadership. Thus, Hypotheses 1a and 1b were supported.Hypothesis 2a predicted that secure attachment would be positively related to LMX. Based onPearson’s r, this relation was found (r(119) = .38, p < .01, one-tailed). As well, this relation wassupported by a hierarchical regression analysis, controlling for group membership (∆R 2 = .06, ∆F(1,98) =9.62, p < .01). Thus, this hypothesis was supported. Hypothesis 2b predicted a positive relation betweentransformational leadership and LMX and this relation was found (r(120) = .84, p < .01, one-tailed).Regression analysis also showed a significant change in the squared multiple correlation (∆R 2 = .47,∆F(1,99) = 197.89, p < .01) after controlling for group, and thus the hypothesis was supported.Hypothesis 3 predicted a positive relation between LMX and followers’ perceptions of theirleaders’ benevolence, which was supported (r(120) = .80, p < .01, one-tailed). When benevolence wasregressed onto LMX, there was a significant change in the squared multiple correlation (∆R 2 = .46,∆F(1,99) = 154.57, p < .01) after controlling for group, showing further support for this hypothesis.Hypothesis 4 predicted a positive relation between follower perceptions of leader benevolenceand trust in the leader. This relation was also highly significant and positive (r(120) = .72, p < .01, one-11

tailed). Regression analysis also showed a significant change in the squared multiple correlation (∆R 2 =.31, ∆F(1,99) = 75.13, p < .01) after controlling for group. Thus, Hypothesis 4 was supported.Hypothesis 5a predicted that LMX would mediate the relation between secure attachment andfollower perceptions of leader benevolence. Using the Baron and Kenny (1986) method, all preconditionsto mediation were met. In Step 1 of a hierarchical regression analysis, group membership was introduced.Then, benevolence was regressed on secure attachment and the relation was significant (β = .29, p < .01).Finally, LMX was added to the relation to assess the mediation effect. Once LMX was added, secureattachment became non significant, while LMX was significant (β = .78, p < .01). Thus, according to theBaron and Kenny method (1986), full mediation was found (Sobel z=4.32, p

concerning biodata from transformational and transactional leaders. Of particular relevance to the presentstudy, Bass and Avolio found that transformational leaders had parents who were strict but fair, hadhappy rather than unhappy childhoods, were praised as children and had mothers who took interest intheir schooling. Laissez-faire leaders on the other hand, had parents who seemed indifferent about thingsthat their children did. Bass asserts that transformational leaders do not simply emerge, but rather they areshaped by their early experiences with their families and in their personal lives. This is consistent with thelogic in the present study and lends further support to the importance of seeking to understand howtransformational leaders develop.In support of Mayer et al.’s (1995) model of trust, leaders perceived as benevolent were alsoperceived as trustworthy. Previous research has also found this link (e.g., Gill et al., 2005; Mayer &Davis, 1999). Although LMX was positively related to benevolence and benevolence was positivelyrelated to trust, the hypothesis that benevolence would mediate the relation between LMX and trust onlyreceived partial support. This mediation effect seemed intuitive especially since according to Gerstner andDay (1997), LMX is highly related to affective outcomes (such as benevolence and trust). One possibleexplanation for the partial mediation that was found could be due to the fact that in this study,benevolence was the only predictor used from Mayer et al.’s model. Although the predictors can beseparated, it is really perceptions of ability, benevolence, and integrity combined that should determinetrust in the leader. Thus, perhaps ability, benevolence, and integrity would have fully mediated therelation between LMX and trust.LMX fully accounted for the relation between secure attachment and benevolence, which wasconsistent with the prediction. As stated previously, all three of these are relational constructs. Secureattachment dictates the positive style in which adults relate to others; leaders rated highly on the LMXscale have positive relationships with their followers, which are characterized by trust, respect, andmutual obligation; and benevolence is associated with wanting to do good unto others. Thus, it is likelythat leaders who are securely attached have greater facility establishing high-quality relationships withtheir followers, which, in turn, influences follower perceptions of leader benevolence. Furthermore, ifthese positive relationships are established over time, it is likely that followers will have more opportunityto witness benevolent behavior on the part of their leader.It was also expected that LMX would explain the relation between transformational leadershipand followers’ perceptions of their leader’s benevolence. In fact, LMX only partially mediated thisrelation. This was somewhat surprising since LMX has been found to mediate the relation betweentransformational leadership and organizational citizenship behaviors (Wang et al., 2005), which isconceptually similar to benevolence. Furthermore, the present study showed that transformational leaderstend to be perceived as benevolent, that LMX was positively related to benevolence and that LMX andtransformational leadership were highly related. Thus it is surprising that full mediation was not found.It is possible, however, that other variables that were not measured may have played a role.Perhaps the amount of opportunities to interact as well as the types of opportunities leaders and followershad to interact could explain the relation between transformational leadership and perceptions ofbenevolence. For example, followers who do not have much contact with their leaders might not be ableto make judgments about their leader’s benevolence because they have little opportunity to experiencebenevolent interactions. Furthermore, followers who only have formal interactions or exchanges withtheir leaders may see their leaders as less benevolent than followers who interact with their leaders on amore personal level. Similar to these suggestions, previous research has shown that distance moderatesleadership-performance relationships (Howell & Hall-Merenda, 1999). Specifically, it was found thattransformational leaders had stronger effects on followers’ performance in close versus distantrelationships. This indicates that organizational context might play a role in determining the effects oftransformational leaders on their followers (e.g., Berson, Shamir, Avolio, & Popper, 2001). Thus, perhaps13

future research could focus more on other mechanisms through which followers come to perceive theirleaders as benevolent.Limitations and Future DirectionsAlthough the hypotheses in this study received substantial support, there were limitations. Firstly,since this study was correlational, causation could not be assessed. For example, a leader who has highqualityLMX relationships may cause followers to perceive their leader as trustworthy. On the other hand,the fact that the leader is perceived as trustworthy to begin with could cause followers to evaluate theirrelationship with their leader as high quality (high LMX). In this case, one can only speculate about thedirection of the relation. However, field experiments have confirmed the directions of some of therelations in this model, such as the impact of perceptions of benevolence on the decision to trust (e.g., Gillet al., 2005; Mayer & Davis, 1999). Furthermore, longitudinal studies could also help establish thedirection of some of these relations, especially in dealing with developmental antecedents. This type ofresearch has been called for by a number of authors (e.g., Kunhert & Lewis, 1987; Popper et al., 2000).Secondly, this study was conducted in three very different organizations. Although eachorganization was structured in teams or departments and had leaders who were easily identifiable, it ispossible that context effects could have come into play. For example, employees in the investmentbanking firm worked in small teams and thus most likely had a lot of personal contact with their leaders.Perhaps due to the high level of interaction, employees may have had more opportunities to seetransformational behaviors on the part of their leaders. On the other hand, employees working in largedepartments in the hospital may have had less contact with their leaders, which could have made followerassessments of their leader’s attachment styles less accurate. More importantly, the amount of interactionand length of the relationship could change perceptions of benevolence and trust. The effects of contexton leadership have been examined (e.g., Berson et al., 2001) and it has been suggested that atransformational leader may not be able to fully express his/her transformational style in certain contexts.However, in the current study, with one exception, there were no significant differences betweenorganizations, thus minimizing the possibility of contextual influences on the results.Thirdly, although this study dealt with relationships between leaders and followers, analyses wereconducted at the individual level, using follower data only. Perhaps if the sample had been larger,analyses could have been done at the leader level. Furthermore, if leadership and attachment styleinformation were assessed by the leader him/herself as well as the follower, comparisons between the twosources could have been made and agreement could have been determined. However, much research inleadership is conducted at the individual level. In fact, leadership style is generally assessed bysubordinates in most studies, and Gerstner and Day (1997) suggest that LMX is most accurately measuredfrom the follower’s perspective. Peer ratings of attachment are also widely used (e.g., Hazan & Shaver,1987; Popper et al., 2000; Schmidt & Bell, 2005; Sumer & Knight, 2001) and have been found toaccurately measure attachment styles (Griffen & Bartholomew, 1994a). Thus, it is likely thatmeasurement from the followers’ perspective was suitable in assessing the variables in this model.In continuing with the research on attachment and leadership, it may be interesting to assessattachment styles of both leaders and followers. It has been suggested that perhaps a fit between leadersand followers may lead to optimal performance and effectiveness (Kunhert & Lewis, 1987). This notionof fit has been examined more closely by Keller (2003) in her analysis of leader and follower attachmentstyles. She looked at implicit leadership theories and claimed that individuals with different attachmentstyles will have different assumptions and expectations (implicit theories) of what a leader-followerrelationship should entail. She made theoretical interpretations and propositions as to which type of leaderwould fit better with which type of follower. One of her propositions was that a secure follower would fitbest with a secure leader because secure followers might portray a positive regard for the leader, and the14

leader would likely respond with sensitivity and support. On the other hand, anxious-ambivalent (i.e.,preoccupied) and avoidant (i.e., fearful/dismissing) followers would clash with secure leaders because theformer would be too clingy, and the latter would be too independent. Thus, the leader would be turned offand would not provide the support that he/she could provide in other circumstances. In combining thesepropositions with the results from the present study, one could speculate that perhaps the followers, whoperceived their leaders as benevolent and trustworthy, were, in fact, securely attached themselves. On theflip side, it is also possible that followers who did not find their leaders to be benevolent and trustworthywere either anxious-ambivalent (i.e., preoccupied) or avoidant (i.e., fearful/dismissing). Future researchcan test these propositions and the results can have important practical implications, which will bediscussed later.ImplicationsThis study has a number of practical implications for organizations. Trust in leaders is becomingcrucial as organizations face increasing complexity and change. By understanding exactly what factorsform the basis of trusting relationships, leaders may be more effective at fostering trust placed in them bytheir followers. Perhaps more importantly, it may help identify factors that could prevent the erosion oftrust. There has been little research done to date on the developmental antecedents of trustingrelationships between leaders and followers and the mechanisms through which trust occurs. This studyhas expanded on past research by delving into these issues. With the knowledge that secure attachment isa blueprint for positive interpersonal relationships, it might be possible to train leaders on how to becomemore effective at relating to their followers.Instead of looking at attachment styles as developmental patterns of interpersonal relationships,perhaps they can be used as the basis for instructing leaders on how to better relate to their followers. Forexample, since the secure style dictates trusting, mutually dependent relationships, it might be possible totrain leaders to be more open with their followers and trust that they will fulfill their duties andresponsibilities. On the other hand, the insecure styles (i.e., fearful, preoccupied, and dismissing) could beused as examples of behaviors a leader should avoid. Leaders who ignore their followers (i.e., dismissingstyle), or smother them (i.e., preoccupied style), may negatively impact their relationships. It has alreadybeen established that leaders can be trained to be perceived as more transformational by their followers(Barling, Weber, & Kelloway, 1996; Kelloway, Barling, & Helleur, 2000). Given these findings, it seemslikely that leaders can also be trained to relate more positively to their followers using the secureattachment style as a basis.This study has also identified certain characteristics of effective leaders which may haveimplications for leader selection. Many studies have looked at the relation between transformationalleadership and personality traits, as evidenced in Bono & Judge’s (2004) meta-analysis. Personality testsare already used in some organizations for selection purposes and it has been found that some personalitytraits are correlated with increased performance on the job (e.g., Tett, Jackson, & Rothstein, 1991).Attachment styles can be considered a part of a person’s personality, given that they represent the waypeople interact and form relationships with others. Thus, attachment styles, like certain personality traits,can perhaps be used as criteria for selecting leaders in organizations. In the present study, secure leaderswere perceived as benevolent and trustworthy, and thus selecting leaders based on positive relationalqualities (i.e., secure attachment) might contribute to the frequency and intensity of trusting leaderfollowerrelationships in organizations.As previously mentioned, followers may play a role in shaping their relationships with theirleaders. For example, a securely attached follower may have more positive relationships with a securelyattached leader than a preoccupied, fearful, or dismissing follower. Furthermore, recent theoreticalpropositions about the charismatic leader-follower relationship have been made (Howell & Shamir,15

2005). Specifically, it has been suggested that perhaps followers play a role in empowering their leaders,thus having an effect on the type of relationship formed between a leader and follower. It therefore seemsimportant to establish an appropriate fit between leaders and followers who will form the most positiveand effective relationships. If organizations are able to establish matches between leaders and followers,perhaps optimal working environments can be established.Perhaps the most important practical contribution of this study lies in the discovery thatrelationships, and more specifically, positive and trusting relationships between leaders and followers, aresalient in the work environment. Although many believe that business is not personal, the findings of thisstudy beg to differ. This study has shown that positive personal relationships between leaders andfollowers are essential in determining trust in the leader, and therefore, that there is in fact a personal sideof business. This personal side should not be neglected as it has the potential to lead to positiveorganizational outcomes.ConclusionThe results of this study offer possible answers to the research questions posed at the beginning ofthis paper: (1) What is the role of developmental antecedents in explaining leadership style and trust inthe leader? (2) What specific factors influence the decision to trust the leader? In answering the firstquestion, this study showed that secure attachment can explain leadership style by acting as a blueprintfor the positive relational qualities that a transformational leader exhibits. Furthermore, secure attachmentcan explain trust in the leader by acting on followers’ perceptions of their leaders’ benevolence. Inanswering the second question, this study showed that high-quality LMX relationships and followers’perceptions of their leaders’ benevolence influenced followers’ decision to trust their leaders. Thesefindings have important practical implications and should stimulate future research on how best to usethis knowledge in the field. Facilitating the establishment of positive relationships between leaders andtheir subordinates could ultimately play a critical role in the success of individuals, teams, andorganizations as a whole.16

ReferencesAinsworth, M.D., Blehar, M., Waters, E. & Wall, S., Patterns of attachment: A psychological study ofstrange situation, Hillsdale NJ: Erlbaum, 1978.Avolio, B.J., Bass, B.M. & Jung, D.I., “Re-examining the components of transformational andtransactional leadership using the multifactor leadership questionnaire,” Journal of Occupationaland Organizational Psychology, 72, (1999), 441-462.Barling, J., Weber, T. & Kelloway, E.K., “Effects of transformational leadership training on attitudinaland financial outcomes: A field experiment,” Journal of Applied Psychology, 81, (1996), 827-832.Baron, R.M. & Kenny, D.A., “The moderator-mediator variable distinction in social psychologicalresearch: Conceptual, strategic, and statistical considerations,” Journal of Personality and SocialPsychology, 51, (1986), 1173-1182.Bartholomew, K., “Avoidance of intimacy: An attachment perspective,” Journal of Social and PersonalRelationships, 7, (1990), 147-178.Bartholomew, K. & Horowitz, L.M., “Attachment styles among young adults: A test of a four-categorymodel,” Journal of Personality and Social Psychology, 61, (1991), 226-244.Bass, B.M., Leadership and performance beyond expectations, New York: Free Press, 1985.Bass, B. M., Transformational leadership: Industrial, military, and educational impact, Mahwah NJ:Lawrence Erlbaum, 1998.Bass, B. M. & Avolio, B. J., Multifactor Leadership Questionnaire. Redwood City CA: Mind Garden,1995.Bass, B.M. & Avolio, B.J., Multifactor Leadership Questionnaire: Technical Report, Redwood City CA:Mind Garden, 2000.Berson, Y., Shamir, B., Avolio, B.J. & Popper, M., “The relationship between vision strength, leadershipstyle, and context,” Leadership Quarterly, 12, (2001), 57-73.Boies, K. & Corbett, R., The relation between leadership styles and subordinates’ perceptions of leaders’trustworthiness. In R. Sinclair (Chair), Correlates of Leader Trust and Morality. Interactiveposter session conducted at the Society for Industrial/Organizational Psychology, Los AngelesCA, 2005.Bono, J. E. & Judge, T. A., “Personality and Transformational and Transactional Leadership: A Meta-Analysis,” Journal of Applied Psychology, 89(5), (2004), 901-910.Bowlby, J., Attachment and loss: Volume II. London: The Hogarth Press, 1973.Burns, J.M., Leadership, New York: Harper & Rowl, 1978.Cacciatore, S., Faulk, S., Perret, F. & Antonakis, J., Multifactorelle de Leadership Formulaire. Palo AltoCA: Mind Garden, 2003.Dansereau, F., Graen, G.B. & Haga, W., “A vertical dyad linkage approach to leadership in formalorganizations,” Organizational Behavior and Human Performance, 13, (1975), 46-78.Deluga, R.J., “Supervisor trust building, leader-member exchange and organizational citizenshipbehavior,” Journal of Occupational & Organizational Psychology, 67, (1994), 315-326.Dirks, K. T. & Ferrin, D.L., “Trust in leadership: Meta-analytic findings and implications for research andpractice,” Journal of Applied Psychology, 87(4), (2002), 611-628.Gerstner, C.R. & Day, D.V., “Meta-analytic review of leader-member exchange theory: Correlates andconstruct issues,” Journal of Applied Psychology, 82(6), (1997), 827-844.Gill, H., Boies, K., Finegan, J. E. & McNally, J., “Antecedents of trust: Establishing a boundary conditionfor the relation between propensity to trust and trust,” Journal of Business and Psychology, 19(3),(2005), 287-302.Gomez, C. & Rosen, B., “The leader-member exchange as a link between managerial trust and employeeempowerment,” Group & Organization Management, 26(1), (2001), 53-69.17

Graen, G.B. & Uhl-Bien, M., “Relationship-based approach to leadership: Development of leadermemberexchange (LMX) theory of leadership over 25 years: Applying a multi-level, multidomain perspective,” Leadership Quarterly, 6(2), (1995), 219-247.Griffen, D.W. & Bartholomew, K., “Models of self and other: Fundamental dimensions underlyingmeasures of adult attachment,” Journal of Personality and Social Psychology, 67, (1994a), 430-445.Griffen, D.W. & Bartholomew, K., The metaphysics of measurement: The case of adult attachment. In K.Bartholomew & D. Perlman (Eds.), Advances in personal relationships, London: JessicaKingsley, 5, (1994b), 17-52.Hazan, C. & Shaver, P., “Romantic love conceptualized as an attachment process,” Journal of Personalityand Social Psychology, 52, (1987), 511-524.Hazan, C. & Shaver, P., “Love and work: An attachment-theoretical perspective,” Journal of Personalityand Social Psychology, 59, (1990), 270-280.Howell, J.M. & Hall-Merenda, K.E., “The ties that bind: The impact of leader-member exchange,transformational and transactional leadership, and distance on predicting follower performance,”Journal of Applied Psychology, 84(5), (1999), 680-694.Howell, J.M. & Shamir, B., “The role of followers in the charismatic leadership process: Relationshipsand their consequences”, Academy of Management Review, 30, (2005), 96-112.Keller, T., “Parental images as a guide to leadership sensemaking: An attachment perspective on implicitleadership theories,” Leadership Quarterly, 14, (2003), 141-160.Kelloway, E.K., Barling, J. & Helleur, J., “Enhancing transformational leadership: The roles of trainingand feedback,” Leadership and Organizational Development Journal, 21(3), (2000), 145-149.Kuhnert, K.W. & Lewis, P., “Transactional and transformational leadership: A constructive/developmental analysis,” Academy of Management Review, 12(4), (1987), 648-657.Mayer, R.C. & Davis, J.H., “The effect of the performance appraisal system on trust for management: Afield quasi-experiment,” Journal of Applied Psychology, 84(1), (1999), 123-136.Mayer, R.C., Davis, J.H. & Schoorman, F.D., “An integrative model of organizational trust,” Academy ofManagement Review, 20(3), (1995), 709-734.Popper, M., Amit, K., Gal, R., Mishkal-Sinai, M. & Lisak, A. “The capacity to lead: Major psychologicaldifferences between leaders and non-leaders,” Military Psychology, 16(4), (2004), 245-263.Popper, M., Mayseless, O. & Castelnovo, O., “Transformational leadership and attachment,” LeadershipQuarterly, 11(2), (2000), 267-289.Schmidt, G. & Bell, B.S., “Attachment style as a predictor of individual-organizational attachment,”Paper presented at the Society for Industrial and Organizational Psychology Conference, LosAngeles, CA, 2005.Sumer, H.C. & Knight, P.A., “How do people with different attachment styles balance work and family?A personality perspective on work-family linkage,” Journal of Applied Psychology, 86(4),(2001), 653-663.Tett, R.P., Jackson, D.N. & Rothstein, M., “Personality measures as predictors of job performance: Ameta-analytic review,” Personnel Psychology, 44, (1991), 703-742.Wang, H., Law, K.S., Hackett, R.D., Wang, D. & Chen, Z.X., “Leader-member exchange as a mediator ofthe relationship between transformational leadership and followers’ performance andorganizational citizenship behavior,” Academy of Management Journal, 48(3), (2005), 420-432.18

ASAC 2006Banff, AlbertaChetan Joshi (student)Joerg DietzVictoria EssesUniversity of Western OntarioCONTINUING EMPLOYMENT DISCRIMINATION AMIDST THE NEED TO EMBRACEDIVERSITY: THE ROLE OF ORGANIZATIONAL RECRUITMENT CLIMATESThe current study illustrates that organizational recruitment climates influence hiring relatedbias against immigrant applicants. Specifically, hiring related bias against immigrants ismaintained/exacerbated/attenuated in recruitment climates silent-on/biased-against/inclusiveofdiversity. Further, differentials in hiring related evaluations are explained by recruiters’considerations of applicants’ fit with the profiles of customers these potential hires areexpected to serve.IntroductionThe changing workforce demographics have resulted in an increasing availability of nontraditionalworkforce talent (e.g., women, people of color, immigrants) for the North American workforce (Judy &D’Amico, 1997). Building domain knowledge that helps organizations embrace this inevitable diversity (Cox& Blake, 1991) has been of concern to social scientists (Arthur Jr. & Doverspike, 2005; Robinson & Dechant,1997; Thomas, 1990) and public policy makers alike (Camarota, 2002). Evidence from employmentdiscrimination research suggests, however, that nontraditional workforce members continue to be facediscrimination in hiring related decisions (Bhardwaj, Dietz, & Esses, 2004; Brief, Dietz, Cohen, Pugh, &Vaslow, 2000). The present study examines how organizational recruitment climates might explain thecontinuing employment discrimination when, in fact, organizations are expected to embrace and reap thebenefits of diversity.We focused on employment outcomes for immigrants, a particularly salient social group (Dietz &Pugh, 2004) that has shaped the socio-economic structure of both the U.S. and Canada (Phinney, Horenczyk,Liebkind, & Vedder, 2001) and is increasingly in demand for maintaining the global competitiveness ofWestern economies (e.g., Independent Commission on Migration to Germany, 2001; Citizenship andImmigration Canada, 2001). Organizational research has rarely accorded audience to the workplace accessoutcomes for immigrants, more so failing to document evidence on strategies that can improve organizationalaccess for them in receiving societies; we attempt to fill this void in the literature. In this regard, we borrowfrom research on organizational climate (Schneider, 2000) to understand how organizational social norms fordiversity reflected in recruitment related policies affect hiring related outcomes for immigrant job applicants.In investigating the issue of employment discrimination faced by immigrants in Canada, wepotentially contribute to the literature on ingroup favoritism (Sherif, Harvey, White, Hood, & Sherif, 1961)and also contribute to research (Dietz & Pugh, 2004) that has drawn attention to immigrants as a meaningfulsocially salient category. Moreover, as a constructive replication (Lykken, 1968) of prior research by Brief etal. (2000) and (Zeigert & Hanges, 2005), the present study extends work on employment discrimination in thefollowing ways: First, in addition to the two organizational climate conditions used in previous research (i.e.presence or absence of a climate for biased hiring), we included a third climate condition that indicatedpreference for hiring a diverse workforce. Second, we included a measure of fit with the customer profile19

evaluations to test whether the commonplace customer demographics matching assertions made byorganizations account for the differential evaluations of job applicants across different organizational climateconditions. In summary, the present study enriches our understanding of the role of organizational recruitmentclimates in maintaining/contributing to/combating employment discrimination.Conceptual BackgroundGroup Identities and Employment Discrimination against ImmigrantsResearch on intergroup relations has illustrated that difference in group identities (e.g., race, gender,sexual orientation, national origin) seems to matter both inside and outside the workplace (Thomas &Chrobot-Mason, 2005). Group identities are useful social markers that allow individuals to compare similarothers favorably over those who share a different social identity (e.g., Sherif et al., 1961). Social identitytheory (Tajfel & Turner, 1986) and self-categorization theory (Tajfel, 1982) posit that individuals categorizeothers into ingroup and outgroup members and make positive attributions to ingroup members at the expenseof the outgroup.While making hiring decisions, individual decision makers (hereafter ‘recruiters’) are confronted withjob applicants from the same socially salient group as themselves and others who do not belong to the ingroup(Lewis & Sherman, 2003). In order to maintain high self-esteem, recruiters would be motivated to seemembers of their ingroup relatively positively and be biased towards ingroup members. Of particularsignificance would be recruiters from dominant social groups (e.g. U.S. or Canadian citizens) who would“achieve a sense of their own identity by highlighting the differences between themselves and others …[resulting in] negative behaviors toward disempowered others [immigrants]” (Thomas & Chrobot-Mason,2005, p. 67). Indeed, it has been documented that immigrants, more so than non-immigrants, continue to berelegated to subordinate positions in organizations (Catanzarite, 2002) and receive lower wages for the samework (Iceland, 1999). In hiring related decisions, we expect the typical ingroup favoritism effect to emergewhen Canadian participants evaluate equally well-qualified Canadian and immigrant job applicants.Specifically, we predict that:Hypothesis 1: When evaluating equally well-qualified Canadian and immigrant job applicants,Canadian participants would evaluate the overall suitability of well-qualified Canadian applicantsmore favorably than that of comparably qualified immigrant applicants. That is, we expect a maineffect of job applicants’ citizenship status on overall suitability evaluations.Organizational Recruitment ClimatesSchneider (1990) defined climate as organizational members’ “perception of the events, practices,and procedures and the kinds of behaviors that get rewarded, supported, and expected” in any organization (p.384). Critical to Schneider’s conceptualization was the existence of different kinds of climates (e.g., climatefor service, safety climate) for focal organizational facets (e.g., customer service, work safety). The facetspecificapproach to climate research has gained currency among organizational behavior scholars in currenttimes (Zohar & Luria, 2004) and we propose that formal and informal policies, procedures and practices thatconvey to recruiters organizational expectations regarding new hires constitute a specific organizationalrecruitment climate. Consistent with Schneider’s (1990) notion of climate (see also Rentsch, 1990), a specificorganizational recruitment climate serves to send strong signals to recruiters on what is rewarded, supportedand expected in the organization with respect to hiring related decision making.Consistent with the above conceptualization, evidence suggests the role of different recruitmentclimates in hiring related decisions (Zeigert & Hanges, 2005; study 2, Brief et al., 2000). In these studies,20

participants completed an in-basket exercise in which a hiring recommendation task involving evaluation ofequally well-qualified Black and White applicants was embedded. While participants in the biasedrecruitment climate condition received a memo from the organization’s president that contained a statementexpressing preference for the new hire to be White, others in the absence of such a bias condition did notreceive the statement regarding racial preference in hiring. The findings indicated that Black applicants werenegatively affected in the biased recruitment climate condition. Although studies by Brief et al. (2000) andZeigert & Hanges (2005) involve laboratory simulations, the organizational climates manipulated in thesestudies do reflect organizational practices at some organizations. Consider, for example, the evidenceregarding the shared perception at Shoney’s of its Chief Executive Officer’s tacit rule that “Blacks should notbe employed in any position where they would be seen by customers” (Watkins, 1993, p.424) and its impacton the segregation of Blacks in low paying, noncustomer contact positions.Although Zeigert and Hanges (2005) suggested that participants who did not receive a biased-climatemanipulation were in a climate for equality condition, their findings indicated that the bias against Blackapplicants, albeit smaller in magnitude, continued in the climate for equality condition. We argue that thepurported climate for equality condition in prior research more adequately reflects a climate that is silent-ondiversity.In such a climate, minorities do fare relatively better compared to a biased climate; yet, the majoritygroup continues to enjoy preference in personnel selection decisions.Notwithstanding that some organizations might be biased against minority applicants or silent, theremay be others (e.g., organizations with identity-conscious HRM structures) – with an organizationalrecruitment climate for diversity - that realize the need to respond to increasing diversity in the labor forcepool, customer demographics and stakeholder expectations (Konrad & Linnehan, 1995). The practices,procedures and rewards at these organizations place value on actively recruiting, retaining individuals fromdiverse backgrounds (Thomas, 1991). The diversity conscious recruitment climates at such organizations aresensitive and committed to fostering diversity (Cox, 1994) and eliminating employment discrimination(Nishii & Raver, 2003).We concur that organizations recruitment climates that are silent-on-diversity will continue tomaintain ingroup bias against nontraditional potential hires. Even while the organizational climate remainssilent-on-diversity, majority group recruiters are “not blind to group identity when making human resourcedecisions” (Konrad & Linnehan, 1995, p. 789). Not surprisingly, mounting research evidence indicates thatdecisions regarding nontraditional workers are prone to bias (Bhardwaj et al., 2004; Nieva & Gutek, 1980)and remain even when nontraditional workers have credentials comparable to the dominant social groupcounterparts (Dietz, Esses, Bhardwaj, & Joshi, 2005). Additionally, replicating previous employmentdiscrimination research (Brief et al., 2000; Zeigert & Hanges, 2005), we expect that the ingroup bias tendencywill be significantly pronounced in organizations with biased recruitment climates. Importantly, and central toour study is the assertion that ingroup bias tendency would be attenuated in organizational recruitmentclimates for diversity. These climates reinforce organizational expectations regarding the qualities diverseemployees bring to organizational functioning without giving the appearance of preferential treatment fornontraditional group members. As such they are less likely to suffer backlash at the hands of members ofdominant social group. Instead, we expect that organizational recruitment climates for diversity might allowfor fair treatment of nontraditional employees. Taken together, the above arguments suggest:Hypothesis 2: Job applicant citizenship status and organizational recruitment climate will interactsuch that Canadian participants’ tendency to favorably evaluate the overall suitability of wellqualifiedCanadian applicants in comparison with that of equivalently qualified immigrant applicants(hypothesis 1) would be exacerbated/maintained/ attenuated in biased-climate/climate silent-ondiversity/diversity-climatecondition, respectively.Tailoring Recruitment to Match Customer Demographics21

Arguments for a representational workforce that is uniquely placed to understand the requirements ofthe markets being served abound in scholarly assertions on the business case for diversity (Cox & Blake,1991; Robinson & Dechant, 1997). Contemporary workplaces that reflect the consumer base just make goodbusiness sense (Robinson et al., 1997) and such considerations might surface in organizational recruitmentpractices. Organizations might be potentially interested in new hires that presumably share customerdemographics and can “understand certain aspects of the world view” of the customers (Cox & Blake, 1991,p. 49). Consequently, organizational recruitment climates might indicate organizational expectations on newhires’ (e.g., those who fit well with the customer demographics of the marketplace being served) fitrequirements. Ironically, while diversity research portrays the importance of addressing fit arguments as anopportunity and an imperative (Cox et al., 1991), evidence on employment discrimination indicates that thematching assertion lowers perceptions on minorities’ ability to serve in majority markets (Watkins, 1993).The current study explores whether recruiters’ evaluations of job applicants’ fit with the customer profilemight be a potential underlying mechanism for the effects proposed in hypothesis 2. Specifically:ParticipantsHypothesis 3: Canadian participants’ differential evaluations of equally well-qualified Canadian andimmigrant job applicants’ fit with customer profile will mediate the job applicant citizenship status Xorganizational recruitment climate interaction effect on overall suitability evaluations.MethodOne hundred sixty five (130 female, 35 male) students enrolled in an introductory undergraduatepsychology course at a Canadian University participated in this study in exchange for course credit.Participants’ ages ranged from 17 to 61 with a mean of 20.77 (SD = 5.24) years. Given the purpose of thestudy to assess bias against immigrants, all participants were Canadian citizens with White Canadians as themajority (White = 81.8%; Asian = 9.7%; African-American = 1.8%; Hispanic = 1.2%; Native Americans =.6%; and Other = 4.8%). There were no significant differences among participants from different ethnicgroups for any of the study variables.DesignWe employed a 2 (Job applicant citizenship status: Canadian or immigrant) by 2 (Organizationalclimate: biased-climate condition or diversity-climate condition or silence on climate condition) mixedfactorial design with job applicant citizenship status as a within participant factor.ProcedureParticipants were recruited through a message posted on the online sign up system (used for anintroductory psychology course) at the Department of Psychology at a Canadian University for a study titled“Study on managerial decision making.” The study was conducted during the semester where participants, insmall group sessions, independently completed an in-basket exercise. The in-basket exercise, adapted fromprior research (Brief et al., 2000), required participants to make a series of managerial decisions in the role ofChris Meyer, Head of Human Resources of a restaurant chain (Captain Cook) with about 25 units spreadacross Ontario, Canada.Specifically, the in-basket contained twelve memoranda requiring Chris Meyer to make decisions ona number of issues (e.g., decision on a request from the Executive Vice President – Marketing for getting apersonal salary advance of $5000). Embedded among the 12 tasks was a memorandum from the President of22

Captain Cook that required participants to make a hiring recommendation decision for a new restaurant inOntario:I understand that you will start screening applicants for the new restaurant manager position inWindsor, Ontario. Our Vice President – Recruitment and Selection has already started the process andwould send you brief profiles of eight pre-screened applicants. These profiles include scores on twotests (management potential test and communication skills test) which were conducted by ourrecruitment and selection department during the pre-screening process. Let me reiterate the followingcriteria, which you absolutely have to keep in mind when selecting the restaurant manager:(1) The candidate has to have experience in managing restaurants. I do not want someone whom wehave to train from the ground up, even if he or she may be a very promising candidate.(2) The job market is saturated with qualified applicants, which allows us to elevate the educationalprofile of the restaurant manager position. Hence, the candidate has to have a university degree inbusiness or commerce, food and hospitality management, or service management.In addition, participants in the biased-climate condition received the following instruction from the President:The vast majority of our customers in Windsor are average Canadian, local, middle-class familieswith household incomes around $70,000. The new restaurant manager will do better if he/she matchesthis demographic profile. Mary Copeland, the EVP Marketing, also endorses the notion of matchingon employee and customer demographics because it makes it easier to run local advertising andpromotion campaigns.Elsewhere, participants in the diversity-climate condition received the following instruction:The vast majority of our customers in Windsor are middle-class families with household incomesaround $70,000, who come from a rich diversity of ethnic backgrounds. The new restaurant managerwill do better if he/she matches this demographic profile. Mary Copeland, the EVP Marketing, alsoendorses the notion of matching on employee and customer demographics because it makes it easierto run local advertising and promotion campaigns.Participants in the climate silent-on-diversity condition did not receive any instruction from thePresident and had to consider only the two criteria about applicant’s experience in managing restaurants andeducational qualification while making their decisions.After seeing the above mentioned instructions from the President of the company, participantsreceived pre-tested summary profiles of nine applicants (five Canadian citizens and four immigrants fromEuropean nations) to complete the hiring recommendation task. The summary profiles, purportedly prescreenedby the Recruitment and Selection department of Captain Cook, included information about eachapplicant’s education, relevant work experience, management potential test score, oral and writtencommunication skills score, language fluency, gender, age, citizenship status, marital status and hobbies. Forthe Canadian applicants, education and work experience were shown as acquired at a location in Canada. Forimmigrant applicants, both education and work experience were shown as acquired in their home country (aEuropean country). To ensure equivalence, all applicants were married males with at least one child, hadsimilar hobbies and were at least fluent in English.The profiles were manipulated such that four (two Canadian citizens and two immigrants) of the nineapplicants were qualified for the position and the other five (three Canadians and two immigrants) were not.The four qualified applicants met the first two criteria set by the President in terms of having relevant23

education and work experience and had higher scores on the management potential test and the verbal andwritten sections of the communications skill test. The other five unqualified applicants did not meet the twoprofessional requirements stated by the President of Captain Cook and also had lower scores on themanagement potential test and the verbal and written sections of the communications skill test.Participants, upon arrival for the in-basket materials, were randomly assigned to one of theorganizational climate conditions. Additionally, the order of the nine applicant profiles was randomizedbetween participants. After completing the in-basket exercise, participants completed a manipulation checktask and a demographic questionnaire. Following this, participants returned the experimental materials to theexperimenter (a White Canadian graduate student) and were debriefed.MeasuresEvaluation of fit with customer profile and overall suitability. While completing the hiringrecommendation task embedded in the in-basket exercise, participants read each applicant’s profile and rated“applicant’s fit with the customer profile” and “overall suitability for job” on 7-point Likert-type scalesranging from 1 (“extremely poor”) to 7 (“extremely good”).The evaluation of fit with customer profile measure for qualified Canadians/immigrants was createdby averaging each participant’s ratings of the evaluation of fit with customer profile of the two qualifiedCanadian/immigrant applicants. Similarly, the measure for overall suitability for job of qualifiedCanadians/immigrants was created by averaging each participant’s ratings of the overall suitability for job ofthe two qualified Canadians/immigrants.Manipulation check. In order to establish that participants correctly followed the instruction fromthe President of Captain cook, participants were asked to recall the stated hiring preferences of the Presidentregarding the professional experience, education and demographic profile for new hires. Participants recalledwhether the President stated (or not) that the new restaurant manager had to have experience in managingrestaurants. Participants also recalled if the President stated (or not) that the new restaurant manager had tohave a university degree in business, commerce, food & hospitality management, or service management.Moreover, participants were asked to recall whether the president (a) preferred to hire an applicant whomatched the average demographic profile of customers that were local, Canadian, middle class families, or (b)preferred to hire an applicant who matched the average demographic profile of customers who came from arich diversity of ethnic backgrounds, or (c) did not state any preference about the demographic profile of thecustomers. For each manipulation check item, participants responded by choosing the most correct choice.PretestResultsThe pretest of the stimulus materials involved an independent sample of 23 undergraduate psychologystudents who rated the qualifications of twelve applicant descriptions (from which we chose our ninedescriptions for the main study) that excluded citizenship status information but included information onprofessional experience and education. As expected, the quality of education, quality of work experience andoverall suitability for the job evaluations for the group of the seven qualified applicants were rated as morequalified than the group of the remaining five unqualified applicants, t(22) = 9.26, p

espectively. Additionally, the quality of education, quality of work experience and overall suitability for thejob evaluations for the group of the qualified immigrant applicants were rated higher than the group of theremaining unqualified immigrant applicants, t(22) = 9.70, p

To test for this last step, we first formed two climate condition dummies as we had three levels oforganizational climate conditions. For the first dummy variable (‘biased climate dummy’), the biased-climatecondition was coded 1 and the other two conditions were coded 0. For the second dummy variable (‘silentclimate dummy’), the climate silent-on-diversity condition was coded 1 and the other two conditions 0.Test of hypotheses. Table 1 shows the means, standard deviations, and intercorrelations among thevariables. A repeated-measures ANOVA with overall suitability evaluations of qualified Canadians andqualified immigrants as criterion indicated a main effect of job applicant citizenship status. Canadianparticipants evaluated the overall suitability of qualified Canadian applicants more favorably (M = 5.99, SD =.77) than that of qualified immigrant applicants (M = 5.69, SD = .84), F (1, 162) = 27.58, p < .001, η 2 = .15.Furthermore, consistent with hypothesis 2, the interaction between job applicants’ citizenship status andorganizational recruitment climate conditions was significant, F (2, 162) = 4.01, p < .05, η 2 = .05. Additionalplanned comparisons showed that in the biased-climate condition and the climate silent-on-diversityconditions, the overall suitability of qualified Canadian applicants was evaluated more favorably relative toqualified immigrant applicants (M = 6.08 [SD = .68] vs. 5.56 [SD = .83]; t(56) = 6.09, p < .001 and M = 6.03[SD = .69] vs. 5.76 [SD = .80]; t(52) = 2.73, p < .01, respectively) , whereas in the diversity- climatecondition the overall suitability evaluations of qualified Canadian applicants did not differ significantly fromthose of qualified immigrant applicants (M = 5.88 [SD = .92] vs. 5.76 [SD = .88]; t(54) = 1.09, ns; See figure1).Support for hypothesis 1 fulfilled the first step for mediation analysis in within-participant designs.For the second step, a repeated-measures ANOVA using fit with Customer profile evaluations of qualifiedCanadians and qualified immigrants as criterion indicated a main effect of job applicants’ citizenship status, F(1, 162) = 36.35, p < .001, η 2 = .18, and an interaction between job applicants’ citizenship status andorganizational recruitment climate conditions, F (2, 162) = 13.27, p < .001, η 2 = .14. Additional plannedcomparisons showed that in the biased- climate condition and the climate silent-on-diversity conditions, thefit with customer profile evaluations of qualified Canadian applicants were more favorable relative toqualified immigrant applicants (M = 5.97 [SD = .69] vs. 5.04 [SD = 1.06]; t(56) = 7.57, p < .001 and M = 5.87[SD = .80] vs. 5.55 [SD = .89]; t(52) = 3.02, p < .01, respectively), whereas in the diversity-climate conditionthe fit with customer profile evaluations of qualified Canadian applicants did not differ significantly fromthose of qualified immigrant applicants (M = 5.56 [SD = .98] vs. 5.51 [SD = .98]; t(54) = .38, ns). Takentogether, the results of the first and second steps provide support for the third step of mediation analysis.26

Table 1Means, Standard Deviations, and Intercorrelations among the variablesVariables M SD 1 2 3 4 51. Sex a .21 .412. Age b 20.76 5.24 .17*3. Evaluation of overall suitability of Qualified5.99 .77 -.09 -.06Canadian applicants c4. Evaluation of overall suitability of Qualified Immigrant 5.69 .84 -.08 -.06 .57**applicants c5. Evaluation of fit with customer profile of Qualified5.80 .85 -.13 -.07 .79** .44**Canadian applicants c6. Evaluation of fit with customer profile of Qualified5.36 1.00 -.09 -.05 .42** .75** .43**Immigrant applicants cp < .05, **p < .01, N = 165a. 0 = Female, 1 = Maleb. Age in yearsc. Values range from 1 to 7, with higher values indicating more favorable evaluationsFurther, results of a regression analysis also provided support for step 4 of the mediation analysis. For the biased-climate condition, the fitwith customer profile evaluations of qualified Canadians were related to the overall suitability evaluations of qualified Canadians, β = .74, p

Figure 1Evaluation of Overall Suitability of Qualified Immigrant and Qualified Canadian Job Applicants asa Function of the Organizational Recruitment Climate6.2Overall Suitability Evaluations6.165.95.85.75.65.55.765.56QualifiedImmigrants6.086.035.88QualifiedCanadiansBiased-climateconditionDiversity-climateconditionClimate silent-ondiversityconditionJob ApplicantsFor step 5 of mediational analysis, when the difference in the fit with customer profile evaluationsof qualified immigrants and qualified Canadians was controlled for in regressing the difference in theoverall suitability evaluations of qualified immigrants and qualified Canadians on organizational climateconditions, results indicated that the difference in the overall suitability evaluations of qualifiedimmigrants and qualified Canadians for the biased climate dummy and the silent climate dummy nolonger remained significant, β = -.04, ns and β = -.00, ns, respectively; while the difference in the fit withcustomer profile evaluations of qualified immigrants and qualified Canadians was still significant, β =.67, p < .001. Overall, support for steps 1 through 5 indicated the job applicant citizenship status Xorganizational climate interaction on overall suitability evaluation was completely mediated by the fitwith customer profile evaluations.DiscussionIn a scenario based laboratory study, we attempted to constructively replicate and extend priorresearch on employment discrimination (e.g., Brief et al., 2000; Zeigert & Hanges, 2005). Our findingssuggest that Canadian participants exhibit ingroup bias in evaluating equally well-qualified Canadian andimmigrant job applicants. Furthermore, this ingroup favoritism tendency varies across differentorganizational recruitment climate conditions. In particular, the tendency is pronounced in organizationalclimates biased against diversity, lingers on unabated in organizational recruitment climates silent ondiversity and is attenuated in organizational recruitment climates inclusive of diversity. Importantly,28

evaluations of job applicants’ fit with the customer demographics of the organization accounted for thedifferential overall evaluations across climate conditions. We acknowledge the experimental context ofour study; as such the findings reflect what might occur and not what will occur in “real world” personnelselection settings. Our interest was in constructively replicating and extending prior research onemployment discrimination and in identifying the role of specific recruitment climates, and as discussedin detail below, we believe that our study did so.First, the present study provides evidence for Dietz & Pugh’s (2004) assertion that immigrantsconstitute a socially salient group and illustrates that immigrants continue to be negatively impacted inreceiving societies. Whereas most organizational research has remained particularly oblivious to theeffects of citizenship status, we believe our study would elevate among organizational scholars the needto address workplace issues faced by immigrants.Second, we argued that recruitment specific organizational climates i.e. practices and policies onrecruitment convey to recruiters organizational expectations on potential hires. Herein, we discussed threevariations of organizational recruitment climates – biased climate, climate for diversity and climate silenton diversity. Critical to enriching our understanding of employment discrimination, immigrant applicantsas compared to the equally well-qualified Canadian counterparts were evaluated equally favorably/considerably negatively in the recruitment climates for diversity/silent-on diversity and biased againstdiversity respectively. Our findings raise, among diversity scholars, the visibility of the role oforganizational recruitment climates in stemming (or not) discrimination faced by nontraditional workers’entry to contemporary workplaces. Further, we assert that an organizational recruitment climate silent ondiversity does not equate as a climate for equality but continues to sustain majority group’s biasedevaluations of minorities. Taken together, these findings suggest that realizing the benefits of the diversityimperative is not independent of organizational efforts, for organizational climates, in themselves, can actas champions or constraints for diversity.Third, our study provided support that the matching assertion potentially accounts for thedifferential evaluations of job applicants of different national origin across climate conditions. Takentogether, these findings replicate Brief et al.’s (2000) and Zeigert and Hanges’ (2005) research with asample in a different national context (Canada versus the U.S.), an additional organizational climatecondition, and different targets of discrimination (immigrants versus Blacks).In terms of practical contributions, our results suggest that workplace discriminatory outcomesagainst nontraditional workers are potentially organizationally controllable. Organizational climatesmirror organizational policies, practices and behaviors that are rewarded, supported and expected in theworkplace setting. It is time that organizational leaders and human resource managers looked inside theorganizations to structure practices that prepare organizational settings to be inclusive of diversity.Organizations that stress totally on meritocratic selections and are silent-on-diversity might indirectly bedoing equal harm in negatively influencing organizational entry for immigrants. The message fororganizations wanting to embrace diversity is clear – include diversity related expectations in recruitmentpractices.We also understand that support for the received view of employee-customer profile matchingrepresents a problem for immigrants. Immigrants received comparable evaluations of fit with customerprofile in diverse recruitment climate condition and not in the other two conditions. We speculate thatapprehensions related to an immigrant applicant’s potential for a specific customer market might possiblybe reduced for immigrants who can provide information that reflects their potential to serve any customerbase (e.g., immigrants with Canadian work experience or prior work experience in diverse markets mightbe better off). Here again, human resource managers should stress for such information to be included in29

ecruitment related applicant description summaries. Future research should attempt to address theseexpectations.It is also important to acknowledge the limitations of our study. The experimental nature of thisstudy, including the use of undergraduate students to assess “paper” job applicants, raises external validityconcerns. At the same time, the laboratory setting provided a high degree of control so that Canadianparticipants’ behavior could be assessed accurately under experimentally manipulated organizationalrecruitment climate settings. We do believe that conducting the current study in the province of Ontario,the state with the largest immigrant population in Canada (Ontario Ministry of Finance, 2005), puts us ina good position to study discrimination against immigrants. Consistent with the conceptual rationales weadvanced, we do believe and expect that the climate manipulations we employed do realistically simulateelements of recruitment climates in the “real world”. While limiting the generalizability of our findings,the present study does further our understanding of employment discrimination.ConclusionChanging demographic trends are shaping contemporary Western societies in significant ways.Arguably, ability to recruit nontraditional workers makes good business sense for organizations(Robinson & Dechant, 1997), but organizational access related barriers continue to restrict realization ofthe diversity imperative in most organizations (Dipboye & Colella, 2005). Our study attempts to shedlight on how specific organizational recruitment climates might play a role in influencing hiring relatedoutcomes for nontraditional workers. We hope it would help to stimulate future research into thecontextual and intervening process variables that influence organizational access outcomes fornontraditional workers.30

ReferencesArthur Jr., W., & Doverspike, D., “Achieving diversity and reducing discrimination in the workplacethrough human resource management practices: Implications for research and theory for staffing,training, and rewarding performance,” In R. Dipboye, & A. Colella (Eds.), Discrimination atwork: The psychological and organizational bases. Hillsdale, NJ: Lawrence Erlbaum Associates,2005.Bhardwaj, A., Dietz, J., & Esses, V. M., “The role of prejudice in the discounting of immigrant skills,”Paper presented at the 19th Annual Conference of the Society of Industrial and OrganizationalPsychology, Chicago, IL, 2004.Brief, A. P., Dietz, J., Cohen, R. R., Pugh, S. D., & Vaslow, J. B., “Just doing business: Modern racismand obedience to authority as explanations for employment discrimination. OrganizationalBehavior and Human Decision Processes, 81(1), (2000), 72-97.Camarota, S. A., “Immigrants in the United States - 2002: A snapshot of America's foreign bornpopulations,”1-24, Washington, DC: Center for Immigration Research, 2002.Catanzarite, L., “Dynamics of segregation and earnings in brown-collar occupations. Work andOccupations,” An International Sociological Journal, 29(3), (2002), 300-345.Citizenship & Immigration Canada, “Planning now for Canada’s future: Introducing a multiyearplanning process and the immigration plan for 2001-2002,” Ottawa, Canada: Government ofCanada, 2001.Cox, T. H., Jr., & Blake, S., “Managing cultural diversity: implications for organizationalcompetitiveness,”Academy of Management Executive, 5(3), (1991), 45-56.Cox, T. J., “Cultural diversity in organizations: Theory, research and practice,” San Francisco: Berrett-Koehler, 1994.Dietz, J., Esses, V. M., Bhardwaj, A., & Joshi, C., “Employment discrimination against ethnicimmigrants: The role of foreign credentials,” Paper presented at the Society for Industrial andOrganizational Psychology, Los Angeles, CA, 2005.Dietz, J., & Pugh, S. D., “I say tomato, you say domate: Differential reactions to English-only workplacepolicies by persons from immigrant and non-immigrant families,” Journal of Business Ethics,52(4), (2004), 365-379.Dipboye, R., & Colella, A., “The dilemmas of workplace discrimination,” In R. Dipboye, & A. Colella(Eds.), Discrimination at work: The psychological and organizational bases: 425-462. Hillsdale,NJ: Lawrence Erlbaum Associates, 2005.Iceland, J., “Earnings returns to occupational status: Are Asian Americans disadvantaged?” SocialScience Research, 28(1), (1999), 45-65.Independent Commission on Migration to Germany, “Structuring Immigration Fostering Integration,”Bundesministerium des Innern Offentlichkeitsarbeit, Berlin, 2001.http://www.eng.bmi.bund.de/Annex/en_14626/Download_Summary.pdfJudd, C. M., Kenny, D. A., & McClelland, G. H., “Estimating and testing mediationand moderation inwithin-subjects designs,” Psychological Methods, 6(2), (2001), 115-134.Judy, R. W., & D'Amico, C., “Workforce 2020: Work and workers in the 21st century,” Indianapolis, IN:Hudson Institute, 1997.Konrad, A. M., & Linnehan, F., “Race and sex differences in line managers’ reactions to equalemployment opportunity and affirmative action interventions,” Group & OrganizationManagement, 20(4), (1995), 409-439.Lewis, A. C., & Sherman, S. J., “Hiring you makes me look bad: Social-identity based reversals of theingroup favoritism effects,” Organizational Behavior and Human Decision Processes, 90, (2003),262-276.Lykken, D. T., “Statistical significance in psychological research,” Psychologcial Bulletin, 70, (1968),151-159.31

Nieva, V. F., & Gutek, B. A., “Sex effects on evaluation,” Academy of Management Review, 5(2), (1980),267-276.Nishii, L. H., & Raver, J. L., “Collective climates for diversity: Evidence from a field study,” Paperpresented at the Annual conference of the Society for Industrial and Organizational Psychology,Orlando, FL, 2003.Ontario Ministry of Finance, “Immigration to Ontario,”http://www.fin.gov.on.ca/english/demographics/cenhi5.html [Accessed January 29, 2006]Phinney, J. S., Horenczyk, G., Liebkind, G., & Vedder, P., “Ethnic Identity, Immigration, and Well-Being: An Interactional Perspective,” Journal of Social Issues, 57(3), (2001): 493-510.Robinson, G., & Dechant, K., “Building a business case for diversity,” Academy of ManagementExecutive, 11(3), (1997), 21-31.Rentsch, J. R., “Climate and culture: Interaction and qualitative differences in organizational meanings,”Journal of Applied Psychology, 75, (1990), 668-681.Schneider, B., “The psychological life of organizations,” In N. M. Ashkanasy, C. P. M. Wilderom, & M.F. Peterson (Eds.), Handbook of organizational culture and climate: (pp. xvii-xxii). ThousandOaks: Sage, 2000.Schneider, B., “The climate for service: An application of the climate construct,” In B. Schneider (Ed.),Organizational climate and culture: 383-412. San Francisco: Jossey-Bass, 1990.Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., & Sherif, C. W., “Intergroup conflict andcooperation: The robbers cave experiment,” Norman: University of Oklahoma Book Exchange,1961.Tajfel, H., “Social identity and intergroup relations,” Cambridge,UK: Cambridge University Press, 1982.Tajfel, H., & Turner, J., “The social identity theory of intergroup behavior,” In S. Worchel, & W. G.Austin (Eds.), The social psychology of intergroup relations: 7-24. Chicago: Nelson-Hall, 1986.Thomas, K. M., & Chrobot-Mason, D., “Group-level explanations of workplace discrimination,” In R.Dipboye, & A. Colella (Eds.), Discrimination at work: The psychological and organizationalbases. Mahwah, NJ: Lawrence Erlbaum Associates, 2005.Thomas, R. R., Jr., “From affirmative action to affirming diversity,”Harvard Business Review, 68(2),1990, 107-117.Thomas, R. R., Jr., “Beyond race and gender, New York: American Management Association, 1991.Watkins, S., “Racism du jour at Shoney's,” The Nation, (1993, October 18), 424-427.Zeigert, J. C., & Hanges, P. J., “Employment discrimination: The role of implicit attitudes, motivation,and a climate for racial bias,”Journal of Applied Psychology, 90(3), 2005, 553-562.Zohar, D., & Luria, G., “Climate as a Social-Cognitive Construction of Supervisory Safety Practices:Scripts as Proxy of Behavior Patterns,” Journal of Applied Psychology, 89(2), (2004), 322-333.32

ASAC 2006Banff, AlbertaKevin Hill (student)J. Mark WeberJoseph L. Rotman School of ManagementUniversity of TorontoCRAFTING JOB SATISFACTION: THE INTERPLAY BETWEENVOLUNTEER AND PAID WORK EXPERIENCESEmployees’ level of activity in the voluntary sector and their (paid) job satisfaction arepositively related through two intervening constructs: life satisfaction and job autonomy.This “spillover” effect between volunteering and job satisfaction suggests that jobenrichment and job crafting theories should be broadened to consider possible growthneeds satisfaction through volunteering.Integrating Job and Life ExperiencesThe field of organizational behaviour concerns itself, in part, with the impact of the formalcharacteristics of jobs on their incumbents’ experiences. However, the job experience of any job holderforms only part of a larger, integrated context of experiences related to the multiple roles assumed by thatthat individual in life. Indeed, the ways people work in and experience their paid jobs has strikinglyrelevant, yet often ignored, parallels to work engaged in outside of the paid work sphere. Studs Terkel hasmade the following observation: “Learning is work. Caring for children is work. Community action iswork. Once we accept the concept of work as something meaningful – not just as the source of a buck –you don’t have to worry about finding enough jobs” (1990; xxviii).This paper demonstrates why established ways of thinking regarding the motivational propertiesof jobs, like the job characteristic model (JCM; Hackman & Oldham, 1975) and job crafting model(Wrzesniewski & Dutton, 2001) could productively be expanded to consider the role of other life domainsthat are highly similar to paid work. Specifically, we suggest that volunteer work provides an alternatesetting where needs-fulfillment consistent with the underlying drivers of these more job-focused theoriesis often found. Upon this need-based rationale, and drawing from literature that suggests the possibility ofreciprocal relationships between the ways life is enacted (and evaluated) both at and away from the job,we identify two potential intervening variables – life satisfaction and job autonomy – that may assist inunderstanding how the two sectors are linked. In opening the door to the world outside of the job, thispaper addresses longstanding views and debates concerning the interplay between the job and the rest oflife, including whether these two activity spheres are more usefully modelled as segmented oroverlapping (Champoux, 1981; cf. Dubin, 1956, and Wilensky, 1960).33

Spillover versus Compensation across Life SpheresWhether the nature of the correlation between job and life (i.e., extra-job) spheres is positive ornegative is a question that has received considerable research attention over the last 50 years. Wilensky(1960) offered that work and extra-work characteristics and evaluations could exhibit positive correlationswith each other (what has been labelled a ‘spillover’ between domains) or negative correlations (what hasbeen labelled the ‘compensation’ in one domain for deficits in the other). For instance, according to thespillover hypothesis, people who are dissatisfied with their lives (or jobs) are more likely to view theirjobs (or lives) in similarly unfavourable ways. Another proposition consistent with the spilloverhypothesis suggests that people in enriched jobs (i.e., growth-needs satisfying) are more likely to engagein enriching pursuits in their spare time. Alternatively, the compensation hypothesis suggests that peoplewho are dissatisfied with their non-work lives would be spurred to seek and find satisfaction in their jobs,or vice versa. As it concerns task characteristics, people in impoverished job settings (i.e., growth-needsfrustrating) would seek more stimulating settings in which to spend their spare time. In addition to thesetwo main arguments, a third suggestion has been offered that posits no relation between the two lifespheres, which remain effectively segmented. The segmentation hypothesis is actually framed in the null(i.e., predicting no relationship), making it a particularly thorny hypothesis for empirical validation. Giventhe dual theoretical-empirical focus of the present investigation, and given that there are multiple reasonsfor failing to uncover a relationship between variables (including methodological concerns), we will focusprimarily upon the spillover and compensation distinctions here.Much research has been conducted to test the competing notions of spillover and compensation.A meta analysis of some of this research has demonstrated that evaluations of job and life satisfaction arepositively related to one another (Tait, Padgett, & Baldwin, 1989), which parallels the greater support forthe spillover hypothesis noted in reviews of work/extra-work variable comparisons (Staines, 1980; Rain,Lane, & Steiner, 1991). Yet, Near, Rice and Hunt (1980) suggested the importance of distinguishing theclasses of variables that have been examined in spillover/compensation research, particularly differencesbetween objective variables (e.g., structural characteristics of job and extra-job activity) and subjectivevariables (e.g., affective reactions like satisfaction). Tait et al.’s finding provides convincing evidence forthe likelihood of spillover of satisfaction across spheres. However, the evidence for the nature of therelationship between objective activity characteristics is mixed (Champoux, 1981), and reviewers of theliterature in this area have noted that underlying mechanisms or causal explanations for both the spilloverand compensation effects are wanting (Staines, 1980; Rain et al., 1991). Indeed, although the existingempirical findings regarding these competing hypotheses have been informative, we are reminded ofSutton and Staw’s (1995) observation that hypotheses do not constitute a theory. In addition, althoughTait et al.’s (1989) meta analysis suggests that “…work should not be examined in isolation fromextrawork concerns” (504), we feel that a correlation alone cannot further our understanding of the driversof this association, the potential causal story contained within it, or whether and how traditional predictoror outcome variables in either the job or extra-job domains have implications for traditional predictors oroutcomes in the other. Toward a more theoretical understanding of the interplay between volunteeringand job experience, we now turn our attention to possible intervening links between the two spheres,starting with the experience of satisfaction itself before considering the experience of job autonomy. Inour discussion of job autonomy, we raise the suggestion that the investigations into the perceivedcharacteristics of jobs may effectively blend Near et al’s (1980) objective-structural and subjectivereactionvariable distinction, but may do so in a substantively meaningful way.34

The Experience of SatisfactionMuch of the early research on job and life satisfaction proceeded under the untested assumptionthat the paid work sphere is the primary origin for people’s evaluations of their global life satisfaction(Rain, Lane, & Steiner, 1991; see for e.g., Kornhauser, 1965). Empirical findings, however, suggest thatthe actual relationship between the “two satisfactions” is reciprocal (Schmitt & Bedeian, 1982); andwhether job satisfaction appears to exert the greater influence (Chacko, 1983), or vice versa (Schmitt &Mellon, 1980) is unclear. While the logic of those who believe in the central importance of the job inpredicting an overall sense of life satisfaction is intuitive since full-time work is a person’s most timeconsumingactivity, a strong alternative argument recognizes the multiple, and distinct sources ofsatisfaction experiences, many of them outside of paid jobs (Argyle, 1987).In considering the relationship as potentially reciprocal, the range of possible predictors of life orjob satisfaction naturally expands. Some limited research has shown that job-related variables are able topredict life satisfaction and that demographic (e.g., marital status) and socioeconomic indicators (e.g.,neighbourhood quality) sustain strong influence upon both life and job satisfaction evaluations (Near,Rice, & Hunt, 1978; Rice, Near, & Hunt, 1979). Exploration beyond these demographic or socioeconomicindicators has been limited, although some have examined the possibly constructive role that involvementin community activities has in relation to job satisfaction (Kirchmeyer, 1992). In a more theoreticallyorientedexamination along dimensions taken from the JCM, Rousseau (1978) found that people’s ratingsof the level of challenge and responsibility in their non-work lives were more strongly related to someworkplace outcomes than were equivalent ratings of the work itself. It would appear, then, that non-workvariables are highly salient to evaluations of multiple facets of life.In spite of suggestive cross-domain correlations, a review of the job/extra-job literature makes itclear that most research has neglected the role of intervening links in explaining these correlations (Nearet al., 1980). Staines (1980) noted “…the absence of any compelling causal mechanisms…”(p.117) forpositive relationships (in either direction) between the characteristics of activities in one domain andsubjective reactions in the other. Some research explored the possibility that job satisfaction mediates theeffect of work characteristics and overall mental health (Kornhauser, 1965). In addition, and a briefempirical test of job satisfaction as a mediator of the effect of objective workplace variables on lifesatisfaction was explored by Rice and colleagues (1979). Though job satisfaction was rejected as apossible mediator in their study, Rice et al. issued the caveat that future tests would be required due tomethodological concerns. Nonetheless, we feel that it is even more interesting to explore the alternativepossibility (i.e., that life satisfaction might serve as a mediator of relationships between non-work factorsand job satisfaction). Rice et al. failed to do this in spite of the potential to do so with their data. In ouropinion, what appears to be suggested by the strong positive correlation between the “two satisfactions”and by the possibly reciprocal relationship between them is that either life satisfaction or job satisfactionmay represent intervening links in the causal chains between the characteristics and evaluations of lifeacross domains. Therefore, activities in peoples’ lives outside of paid work that have been demonstratedto lead to higher levels of life satisfaction should be considered reasonable predictors of job satisfaction.Volunteering is an ideal activity for such a consideration, since evidence suggests it is positivelyand reciprocally related to well-being. Longitudinal studies have shown that not only are happier and welladjusted people more likely to volunteer, but also that the act of volunteering predicts levels ofsatisfaction and well-being over time controlling for initial levels (Thoits & Hewitt, 2001; Wilson, 2000).While research perspectives concerning the drivers of volunteering vary, a micro-level explanation forthis life satisfaction effect is that volunteering satisfies underlying needs (Wilson, 2000). We return todiscuss needs in more detail in the second half of this paper. For now, we feel that, regardless of the selfselectionof generally happier people into volunteer work situations, it is important to draw attention to35

the possible role of volunteering’s residual effect on life satisfaction as it pertains to our interest inbroadening the range of job satisfaction predictors. Therefore, our first set of hypotheses suggests howvolunteering might affect people’s satisfaction with the work they do for pay by incorporating itsdemonstrated impact on life satisfaction into the present model (see upper branch of Figure 1).Hypothesis 1: The extent of volunteering is positively related to job satisfaction.Hypothesis 2: The extent of volunteering is positively related to life satisfaction.Hypothesis 3: The relationship between volunteering and job satisfaction is mediated by life satisfaction,such that people who volunteer more are more satisfied with life. This higher satisfactionwith life, in turn, predicts higher levels of job satisfaction.Figure 1.Linking Volunteering and Job Satisfaction through Life Satisfaction and AutonomyLife SatisfactionVolunteeringJobSatisfactionJob AutonomyIn presenting Figure 1, we have an opportunity to reemphasize our primary objective in thepresent paper: the identification of important intervening variables between the experience ofvolunteering outside of the paid work sphere and the experience of job satisfaction. Identifying the causaldirection of the relationships in this model is not our main interest here. As already noted, the reciprocalnature of the relationships contained in the upper branch of the figure has been empirically demonstrated.As we turn our attention toward the lower branch of the model, we also note the multiple perspectivesregarding relationship directionality. That is, although the JCM (Hackman & Oldham, 1975) hastraditionally suggested that job characteristics lead to job satisfaction, its biggest challenge has come fromtheories that suggest that job characteristic perceptions are socially constructed on the basis of contextualcues (e.g., Salancik & Pfeffer, 1977; 1978), and that people who feel satisfied in their jobs may imbuetheir jobs with favourable characteristics (regardless of their objective presence) in order to maintaincognitive consistency (Berlinger, Glick & Rodgers, 1988). In the section to follow, we explore thepossible relationship between volunteering and the perception of job autonomy. Existing empiricalresearch relevant to this branch of the model tends to view occupational characteristics as causally prior tovolunteering activity. While we feel that this is reasonable, we will offer our own theoretical rationale forthe reverse-causality position toward proposing how this relationship might also be reciprocal.The Experience of Job CharacteristicsIn spite of the more common support for the spillover hypothesis alluded to earlier, negative (i.e.,compensatory) relationships between the characteristics of jobs and other activities have been found(Staines, 1980). In fact, the assorted findings are less clearly supportive of either spillover orcompensation when considering relationships between job and extra-job characteristics, as compared tothe findings regarding the evaluations of satisfaction across these spheres (see Champoux, 1981). In terms36

of the relationship between objective occupational characteristics and volunteering, most researchers haveadopted a macro focus, suggesting that occupations containing opportunities for employees to developtheir human capital (e.g., civic skills that are valuable in the voluntary sector), and those that connectemployees to large social networks generate more volunteers (Voydanoff, 2001; Wilson, 2000).Specifically, Wilson and Musick (1997) have shown that the level of volunteering among employeeswhose occupations score highly on objective occupational measures self-directedness is higher than thelevel in other occupations. Although these studies appear convincing in support of spillover effects, theyroutinely use objective occupational codes (sometimes taken from the dictionary of occupational titles) aspredictor variables without accounting for employees’ actual perceptions of their jobs’ characteristics. Infact, Champoux (1978, 1980) has found that, in comparing respondents’ own ratings of their workplacesand their lives outside of work (e.g., family, community, hobbies, etc…), compensation as well asspillover patterns emerged. Similarly, Kabanoff and O’Brien (1980) found compensation and spilloverpatterns in different peoples’ ratings of their work and leisure activities in a large-scale survey. Champouxsuggested that his findings demonstrated that caution is warranted when drawing conclusions from joband extra-job comparison research in which the researchers’ own notions of equivalence or differencebetween life spheres had been used as part of the test of spillover/compensation hypotheses.Given the mixed nature of the findings in terms of job characteristics and characteristics ofpeoples’ non-job lives (volunteering as well as leisure), the spillover or compensation hypotheses couldbe equally plausible, but there is a need for a theoretical explanation for either hypothesis (Rain et al.,1991). We feel the findings of macro occupational studies (e.g., Wilson & Musick, 1997) are convincingin describing a spillover effect by way of an “opportunity to volunteer” mechanism provided by way ofcertain objective job characteristics (i.e., variations in human and social capital development opportunitiesby occupation). However, Champoux’s (1980) finding that “blue collar” workers exhibited compensatoryevaluations of the work and non-work spheres, while “white-collar” workers more often exhibitedspillover in their evaluations of the two spheres, suggests that spillover or compensation may occuraccording to the nature of the needs that are being satisfied in the work sphere (Rain et al., 1991). In spiteof this suggestion, research has not examined this needs-based mechanism in any depth. Thus, we nowturn to elaborate this possibility by highlighting the underlying mechanisms of existing job characteristicstheories – primarily the JCM (Hackman & Oldham, 1975) and the job crafting model (Wrzesniewski &Dutton, 2001). In so doing we provide another possible intervening link between volunteering and jobsatisfaction through the perceived job characteristic of job autonomy (see lower branch of Figure 1).The job characteristics model. The relationship between job characteristics and job satisfaction iswell established by meta-analytic research on the JCM (Fried & Ferris, 1987). The JCM proposes threeunderlying mechanisms, which serve as the mediators of the impact of job characteristics:meaningfulness, responsibility, and knowledge of results. These psychological states may emerge as aresult of job characteristics as follows (Hackman, 1977): 1) Meaningfulness emerges from the extent towhich the job involves a variety of skills, an identifiable or “whole” piece of work, and has a significantimpact on the lives or work of others; 2) Responsibility stems from job autonomy, the freedom todetermine the way the work is undertaken; and 3) Knowledge of results is facilitated by the degree towhich the job permits employees to gauge how well or poorly they are performing while working. Theseintervening variables are needs-satisfying states, which drive favourable psychological and behaviouraloutcomes at work depending on employees’ growth needs strength (GNS). Indeed, Hackman and Lawler(1971) drew from the need theories of Maslow (1943) and Alderfer (1971) in developing the JCM, andarticulated one of its primary principles as this: “individuals who are capable of higher-order needsatisfaction will in fact experience such satisfaction when they learn that they have, as a result of theirown efforts, accomplished something that they personally believe is worthwhile and meaningful” (1971,p.262).37

Job crafting theory. Wrzesniewski and Dutton’s (2001) concept of job crafting is also clearlyneed-based due to what they define as the drivers of the activity: peoples’ needs for control over theirenvironment (Adler, 1930), positive interpersonal connections (Baumeister & Leary, 1995), and positiveself-image (Tajfel, 1981), all of which are suggested to be universal needs. Specifically, the theorysuggests that, when jobs fail to meet these underlying needs, incumbents themselves will redesign the (a)task boundaries, (b) relational boundaries, or (c) cognitive boundaries of their jobs in ways that remedysuch deficiencies. Indeed, job crafting succeeds when employees manage to positively change themeaning of their work and their work identities (Wrzesniewski & Dutton, 2001). The only boundaryconditions placed on the model, which would prevent people from crafting their jobs toward meeting theirneeds, are limitations that reside in the jobs themselves (e.g., high degrees of monitoring andinterdependence); or else situations in which people have already satisfied these needs outside of work.Underlying Needs as the Drivers of Work Experiences across SpheresThus, previous research tells us that engaging in meaningful activity is the primary characteristicof enriched jobs and the driver of job satisfaction by way of needs-satisfaction mechanisms (Berlinger,Glick, & Rodgers, 1988). The large body of research reviewed by Fried and Ferris (1987) supports thenotion that, by satisfying growth needs, jobs can be redesigned toward improving the psychologicalexperience of workers. Moreover, Wrzesniewski and Dutton (2001) have drawn attention to the role ofjob holders as proactive pursuers of certain needs in their work. In both theories, the driving factor behindany job change initiative is the motivational potential of employees’ needs. From our perspective, thecommon theme in this research regards the motivating potential of growth and development opportunities,which have been represented by the extent to which the work provides opportunities for personal agency(i.e., autonomy or control). The identification of self-identity concerns as a separate need is notable in thejob crafting theory. Yet, it could be said that identity-related needs are captured in the JCM to extent thatit emphasizes personal agency. Indeed, Maslow suggested that identity-relevant experiences are mostsalient during growth experiences, particularly those that are self-determined: “…that people in peakexperiences are most their identities, closest to their real selves…” (Maslow, 1962, p.97). Aside from theirshared emphasis on agency, the need for interpersonal connection figures more prominently in the jobcrafting model than the JCM. That being said, in an early formulation of the JCM, Hackman and Lawler(1971) examined interpersonal components of jobs. By no means downplaying the importance of thesesocial needs, Hackman and Lawler simply doubted the long-term motivational implications of “relatingmeaningfully to others” as a job characteristic since relatedness needs (Alderfer, 1972), once satisfied,would cease to provide ongoing motivational potential.Indeed, both Maslow (1943) and Alderfer (1972) suggested that satisfied needs generally fail tomotivate. What are the implications of this postulate for the way job characteristics and volunteeringpotentially relate to each other? We have to assume that people are motivated to volunteer, since extrinsicjustification is largely absent. Thus, if the needs that spur job crafting (or else enhance the motivationalproperties of job characteristics) are satisfied in the job, then the motivation to seek needs-satisfaction inthe voluntary sphere should be negligible. The same rationale would also apply to this relationship in theopposite direction. That is, to the extent that people are able to satisfy their needs outside of the job (e.g.,through volunteering), they would not be motivated to satisfy them in their jobs, making the motivatingpotential of job enrichment interventions negligible. Wrzesniewski and Dutton’s (2001) argument seemsto be in line with this thinking since they suggest that, rather than job crafting, “…people may find thatthese needs are met elsewhere in their lives” (p.183). Phenomenologically, this need dynamic – coupledwith the framing of volunteering as a substitute activity for needs-satisfaction – would lead to a pattern ofpaid work and volunteer work that is consistent with the compensation hypothesis.38

However, this postulate of accepted need-based theories does not pertain to the motivationalproperty of higher-order needs (e.g., growth needs; Alderfer, 1972). Maslow, a strident advocate for theidea that satisfaction subsumes motivation, explained the exception this way: “Growth is instead acontinued, more or less steady upward or forward development. The more one gets, the more one wants,so that this kind of wanting is endless and can never be attained or satisfied” (1962, p.31). Lewin (1975),too, appeared to agree with this characteristic of growth need satisfaction: “Satiation occurs only if theactivity has, psychologically, the character of an actual repetition, of marking time, as opposed to makingprogress. If the character of making progress can be maintained, the usual symptoms of satiation will notappear” (p.282). The satisfaction of higher-order growth needs, then, potentially serves as a potentmotivational base for subsequent satisfaction of these needs in other contexts. Stemming from this type ofdynamic, it could in fact be the perceived quality (rather than deficiency) of peoples’ jobs that provides alevel of motivation toward future growth. The extent to which this motivation for growth translates intohigher levels of volunteer activity in the extra-job sphere would be consistent with the spilloverhypothesis. Similarly, those employees who satisfy their higher-order growth needs through volunteeringmay be more motivated to job craft and/or may benefit more from job enrichment interventions to theextent that their appetites for growth needs satisfaction are whetted by their volunteer experiences.In order to pursue this line of thinking further, we need to consider whether, compared to theirjobs, peoples’ volunteering can be considered a substitute activity sphere for needs satisfaction and,additionally, whether the needs satisfied through volunteering are growth-related. Wilson definedvolunteering as “…any activity in which time is given freely to benefit another person, group ororganization” (p.215). However, the compensation and spillover mechanisms we have just outlinedsuggest that the degree of similarity between the activities in question is a relevant consideration.Regarding the compensation hypothesis, Lewin (1935) proposed that substitute activities often arise whenpeople’s goals are frustrated in an initial activity sphere, and that substitute acts would be effective partlyto the degree that they were similar in content, but primarily in terms of their ability to satisfy thefrustrated goal. The similarity between activities in the paid and voluntary sectors is also relevant to thepotential for the spillover mechanism. For instance Breer and Locke (1965) suggested that peoples’beliefs are largely created by engaging in tasks, tasks including (but not exclusive to) their jobs. In fact,they argued that beliefs developed in one task situation would tend to generalize to similar task situations.For these reasons, we feel it important to specify our interest in volunteering to unpaid work in the formalvoluntary sector, a close cousin to the domain of paid work. Indeed, Peter Drucker (1989) has offered thatin well-organized volunteer organizations “…there are no more ‘volunteers.’ There is only unpaid staff”(p.203). It is worth noting that, if it is accepted that the underlying similarity of tasks is an importantcondition for either compensation or spillover effects, this calls into question the adequacy of pastresearch that has compared the occupational sphere with broadly defined leisure or “non-work” spheres.In addition to formal volunteering’s apparent similarity to the world of paid work, the truemeasure of these spheres’ commensurability for the present argument comes down to their needssatisfactionpotential. We feel it is safe to say that, given it is freely chosen, work in the voluntary sectoris an inherently meaningful activity that satisfies important intrinsic needs or else people would cease toengage in it. Moreover, volunteering is a primary medium for the expression of human agency followinga similar deductive argument, which has been noted by others (Thoits & Hewitt, 2001). This descriptionon the basis of the inherent nature of volunteering in general seems to fit with research that has focusedon the motives people tend to report concerning their volunteering. For example, Clary and Snyder (1999)found that people volunteer to express important values, to learn, and to develop psychologically, andeven to gain career-related experience for potential future employment. These are particularly congruentwith the notion of growth needs-satisfaction. Clary and Snyder also found that people report that theyvolunteer to develop interpersonal connections, which is consistent with relatedness needs satisfaction.Obviously, not all aspects of volunteer work are perfectly congruent with the satisfaction of higher-orderneeds, but it is encouraging for the present line of thinking that volunteers appear to be motivated by the39

growth-related potential of this work since the present paper explores the relationship between the extentof volunteering and the perception of a particular job characteristic, autonomy. We feel that, not only isautonomy in the job a central feature of job enrichment in both the JCM and job crafting literatures as asource of higher-order needs satisfaction, but it is also inherent to the choice to volunteer and the level ofvolunteering activity pursued. To the extent that people engage in volunteer work for more groups at anygiven time, we can say that they are behaving autonomously in this sphere.Up to this point, the implications of volunteering on job autonomy (or vice versa), have beenmade in terms of the nature of volunteering as both a possible substitute for and an extension of needssatisfaction. Using growth needs satisfaction is a possible mechanism, for linking the experiences in thepaid and voluntary sectors, we have made the case for how either the frustration or the satisfaction ofgrowth needs at work could both plausibly lead people to seek this satisfaction in their volunteering.However, it is worth noting the difference between the temporal dynamics involved in these twocompeting arguments. At the critical juncture when people choose to volunteer, either frustration orsatisfaction of growth needs could serve as the catalyst for this choice. However, as people concurrentlyengage in the two spheres over time, we suggest that the satisfaction of growth needs throughvolunteering would serve to maintain the motivational potential of growth needs across work-relatedspheres, resulting in generalization. In other words, growth needs satisfaction in the voluntary sectorcould lead people to craft their jobs to accommodate future growth needs (by assuming more autonomy intheir paid jobs, among other means, for instance). Since our focus here is primarily on identifying jobautonomy as a linking variable and, only secondarily, on testing the compensation/spillover hypothesis,and given our data do not permit the teasing apart of these two distinct dynamics, we propose that ageneral spillover hypothesis will characterise the interplay between volunteering and job autonomy. Thatis, no matter what condition it was that led employees to choose volunteering (or led volunteers to chooseemployment), we expect the generalization dynamics that play out over time to overwhelm anycompensation-oriented entry process effects. In addition, since the positive relationship betweenautonomy and job satisfaction is established in the job design literature, we expect it to positively predictjob satisfaction, and serve as a link between job satisfaction and volunteering.Hypothesis 4: The extent of volunteering will positively relate to the perception of autonomy in paid workHypothesis 5: The influence of volunteering on job satisfaction (hypothesis 1) will be mediated byperceived autonomy in the job such that people who engage in more volunteer work willperceive greater autonomy. In turn, this greater level of autonomy will be associated withhigher levels of job satisfaction.The case presented thus far has appealed largely to the potential influence of engaging in growthneedssatisfying work in the voluntary sector on the experience in the paid sector. We acknowledge theequally plausible case for the reverse-causality argument, whereby the experience of autonomy at workmotivates and/or facilitates the degree of involvement in the voluntary sector. However, our emphasis onvolunteering as a potential source of influence has been meant to counterbalance the neglect of thispossibility in existing studies of occupation/community involvement research (Voydanoff, 2001). Weagree with Breer and Locke (1965), who noted: “beliefs, preferences, and values developed in one tasksituation will generalize to all others (depending on similarity and familiarity)…[and]…it is not necessaryto this way of thinking that task experience in the formal work setting constitute the source from which allgeneralization must flow” (p. 17). Since the activities and tasks involved in paid jobs and volunteeringboth constitute forms of work, generalizing from one of these specific task “constellations” to the other isequally plausible (Breer & Locke, 1965). Indeed, in some ways, an even stronger case could be made forvolunteering’s influence. Given it involves a voluntary choice to act, and thus external or extrinsicjustifications are comparatively absent, people will be more committed to the actions they undertake inthe voluntary sector (Salancik, 1977). As a result, the impact on beliefs could be even more powerful than40

the impact of behaviours conducted in the more constrained job context, where extrinsic justifications forbehaviour are also present. Indeed, we agree that “…behavior which cannot be justified by externalrewards is justified by constructs of self-motivation (Salancik & Pfeffer, 1978, p.234).Controlling for occupational characteristics and intrinsic/extrinsic work valuesIt has been pointed out that those interested in relationships between job and non-job spheres should becareful to consider sources of spuriousness (Edwards & Rothbard, 2000). We are concerned that anyvariable that independently influences both the extent of volunteering and the level of job autonomy, if itwere not statistically accounted for, would bias the tests of our hypotheses. Here, we are concerned withcontrolling occupational variables and stable work preferences. First, as already discussed above, muchwork has suggested that certain occupations provide the resources (human and social capital) thatfacilitate involvement in volunteer work (e.g., Voydanoff, 2001). However, we are interested in a needbasedrationale for the relationship between volunteering and job autonomy. Statistically accounting foroccupational classification will permit us to control for the opportunity to volunteer, so that we can betterinvestigate the motive to volunteer. Indeed, we are interested in peoples’ perceptions of their jobs overand above objective occupational distinctions, yet this type of statistical control was absent in previousinvestigations into the role of community work on people’s job-relevant reactions (e.g., Kirchmeyer,1992). We also control for an objective level of job responsibility (number of people supervised) in ourtest of volunteering as an influence on job autonomy. Secondly, Kohn (1977) has demonstrated that notonly do occupational differences predict worker values, but that people’s general orientation toward workpossibly influences their selection into occupations (Kohn & Schooler, 1973). In order to rule out thepossible spuriousness of a “selection effect” by which people who innately value work experiences thatare characterised by higher-order need satisfaction both select themselves into enriched jobs and intovolunteer work, we will control for general work value preferences (intrinsic and extrinsic). Together, theinclusion of such controls enables a more refined examination of the job characteristics that are perceivedby the individual job holder, and their relationships with the degree of volunteer work involvement.Method and MeasuresWe test these hypotheses using data drawn from respondents to the World Values Survey(Inglehart et al., 2002), focusing exclusively on those who were employed at the time of surveycompletion (n= 19422), and to whom questions were posed regarding their evaluation of their jobs(European Values Survey sub sample).Demographic measures. Age and gender were included in the analyses as they have intuitivelinks to work and non-work experience (cf .Tait et al., 1989). A indicator of whether the respondent wasin a stable relationship with a partner (a more inclusive relational categorization than marital status) isincluded to capture a dynamic similar to marriage, which is related to satisfaction (Near et al., 1980). Thisvariable was measured through a one-item question to which respondents indicated their (dis)agreementdichotomously: “Whether you are married or not: Do you live in a stable relationship with a partner?”Occupation. The World Values Survey data contains the International Labour Organisation’sISCO88 codes (ILO, 1990) for each respondent. Using respondents’ 2-digit code entry, dummy variableswere developed for the nine occupational clusters making up the major civilian occupational groups inISCO88: 1) Legislators, senior officials and managers; 2) Professionals; 3) Technicians and associateprofessionals; 4) Clerks; 5) Service workers and shop and market sales workers; 6) Skilled agriculturaland fishery workers; 7) Craft and related trades workers; 8) Plant and machine operators and assemblers;9) Elementary occupations. The ninth category, ‘elementary occupations’ serves as the comparison group.41

Supervisory responsibility. The number of people for whom the respondent was responsible in asupervisory capability was obtained from the dataset. This numeric response ranges from zero to 1000.Work Values. As part of the World Values Survey, respondents were asked to identify theattributes of a job that they personally think are important from a list of eleven items. We ran a principlecomponents analysis with oblique rotation in order to identify people’s underlying work values. Of theseitems, ten loaded unambiguously onto one of two factors (five items on each). The eleventh item loadedequally on both factors and, thus, was removed from consideration. Two five-item scales were created bysumming the remaining items according to their assignment to one or the other factor. The interpretationof these items is highly consistent with an intrinsic and extrinsic value interpretation. Sample items are asfollows: intrinsic values (e.g., an opportunity to use initiative, a job that meets one’s abilities), andextrinsic values (e.g., good pay, generous holidays). Internal consistency reliability estimates for thesetwo scales are contained in Table 1 along the diagonal. Of note, these items appear to be highly similar tothe items used by Kohn & Schooler (1973) as a similar statistical control.Volunteering. Respondents were asked to indicate whether they were currently doing any unpaidvoluntary work for 14 organizations of various types (e.g., social welfare agencies, religiousorganizations, professional associations, etc…), or any another unspecified organization. An index wasconstructed by summing the distinct types of volunteer work activity. Responses ranged from zero to 15.Autonomy. Autonomy was measured by a single-item particularly focused on respondents’freedom to make decisions in their paid jobs. Respondents indicated their agreement on a ten-point scale,ranging from 1 – ‘None at all’ to 10 – ‘A great deal’, to the following question: “How free are you tomake decisions in your job?”Job Satisfaction. Respondents indicated their agreement on a ten-point scale, ranging from 1 –‘Dissatisfied’ to 10 – ‘Satisfied’, to the following question: “0verall, how satisfied or dissatisfied are youwith your job?”Life Satisfaction. Respondents indicated their agreement on a ten-point scale, ranging from 1 –‘Dissatisfied’ to 10 – ‘Satisfied’, to the following question: “All things considered, how satisfied are youwith your life as a whole these days?”These single items for autonomy, job satisfaction, and life satisfaction do not permit theestimation of measurement reliability. The failure of large-scale surveys to use multi-item measures (dueto time constraints) is a common problem when using archival data of this sort. However, to the extentthat significant results are found, these measures should provide a conservative means for exploring themediation hypotheses. Moreover, a recent meta-analysis suggests the correlation between single-itemmeasures of satisfaction and multi-item scales is quite respectable (Wanous, Reichers, & Hudy, 1997).Table 1 (see Appendix A) presents the means, standard deviations and zero-order intercorrelationsbetween the studies main variables. It should be noted that correlations between theoccupational dummies themselves (i.e., mutually exclusive categorical variables) are not interpretable insuch a matrix. Any correlation between them would merely be a statistical artefact since there are morethan two categories and, thus, all cases that “share” a ‘0’ coding for any two of the dummy variablesoverlap, resulting in a correlation. For this reason, we have removed these meaningless correlations fromthe matrix. However, Table 1 provides a descriptive break down in terms of the percentage of samplerespondents represented in each occupational group, and looking at their correlations with other variablesprovides insight into the occupational variation exhibited by the main variables of interest. For instance,the table shows that supervisory responsibility, perceptions of autonomy, satisfaction levels, and intrinsic42

work values all appear to be higher than average among the more “white collar” occupations and lowerthan average among the more “blue collar” categories. This reaffirms the importance of controlling foroccupation in all of the following statistical analyses.ResultsThe hypothesized relationships were tested through a series of multiple regressions. Table 2contains two hierarchical regression analyses, the first of life satisfaction on its hypothesized predictorsets, and the second of autonomy on its hypothesized predictor sets. In each regression, the volunteeringvariable is added separately to show its incremental validity. From the regression of life satisfaction, wesee that males, younger people, and those involved in stable relationships report higher levels of lifesatisfaction, ceteris paribus. Regarding occupational group, clerks and service workers exhibit levels oflife satisfaction that exceed those of elementary occupations, while trade workers and machine operatorsexhibit lower levels than this comparison group. It is interesting to note that the levels of life satisfactionof those in “white collar” occupations does not differ significantly from the life satisfaction of those inelementary occupations. Job autonomy and job satisfaction also predict life satisfaction. Yet, volunteeringpositively predicts life satisfaction when accounting for these job-related variables, which lends support toArgyle’s (1987) notion that life satisfaction can originate in spheres other than the job. This regressiondemonstrates that, accounting for all these control variables, volunteering significantly predicts lifesatisfaction, supporting hypothesis 2.Table 2.Hierarchical Regressions of Job Autonomy and Life Satisfaction on VolunteeringLife satisfactionAutonomyAge -.098 ** -.101 ** .043 ** .041 **Gender .029 ** .026 ** .089 ** .086 **Legislators, senior officials and managers .002 .003 .200 ** .196 **Professionals .015 .006 .214 ** .207 **Technicians and associate professionals .007 .001 .155 ** .150 **Clerks .055 ** .051 ** .117 ** .114 **Service workers and shop and market sales workers .030 ** .028 * .126 ** .125 **Skilled agricultural and fishery workers -.001 -.002 .055 ** .055 **Craft and related trades workers -.021 * -.021 * .058 ** .058 **Plant and machine operators and assemblers -.018 * -.018 * .004 .005Supervisory responsibility .027 ** .026 **Intrinsic work values .169 ** .163 **Extrinsic work values -.148 ** .144 **Stable Relationship .082 ** .082 **Job satisfaction .394 ** .392 **Autonomy .127 ** .123 **Volunteering .072 ** .061 **R 2 .237 .241 .108 .111F 449.1 428.64 156.82 ** 151.11 **(d.f.) 18839 18839 16896 16896∆ R 2 .005 ** .004 **p

From the regression of autonomy on its hypothesized predictors in Table 2, we note that theoccupational classifications afforded by the ISCO codes appear to suggest that autonomy is lowest amongelementary occupations (i.e., the comparison group) and among machine operators and assemblers. Theremaining occupational groups report higher levels of autonomy, and the upward rise in autonomy from“blue collar” to “white collar” categories is consistent with past research (Wilson & Musick, 1997). Thisalso supports the notion that meaningful variation in autonomy is being controlled for by including theseoccupational variables. Controlling for work values also demonstrates that a selection effect is likely to beminimal as an explanation for the relationship between volunteering and autonomy. Volunteering predictsthe perception of autonomy over and above all of these control variables. As such, the relationshipbetween volunteer activity and the autonomy variable can be interpreted more directly, and this positiverelationship provides support for hypothesis 4.Table 3.Hierarchical Regressions Predicting the Mediating Effects of Autonomy and Life Satisfaction in theVolunteering-Job Satisfaction RelationshipJob satisfactionAge .045 ** .026 ** .080 ** .058 **Gender .034 ** -.016 * .010 -.023 **Stable Relationship .025 ** .009 -.017 ** -.021 **Legislators, senior officials and managers .138 ** .004 .100 ** .005Professionals .153 ** .038 ** .110 ** .031 **Technicians and associate professionals .128 ** .042 ** .094 ** .036 **Clerks .090 ** .033 ** .045 ** .010Service workers and shop and market sales workers .096 ** .026 ** .058 ** .013Skilled agricultural and fishery workers .015 -.038 ** .008 -.032 **Craft and related trades workers .057 ** .018 .053 ** .024 *Plant and machine operators and assemblers .033 ** .029 ** .036 ** .032 **Volunteering .061 ** .030 ** .014 * .000Autonomy .453 ** .347 **Life Satisfaction .447 ** .351 **R 2 .03 .213 .222 .321F 48.43 ** 394.28 ** 416.18 ** 635.43 **(d.f.) 19114 18933 19010 18839* p

esearch on job satisfaction (Fried & Ferris, 1987; Tait et al., 1989). We tested hypotheses 3 and 5 inseparate regressions, the results of which are reported in the second and third numerical columns of Table3. When autonomy alone was added to the baseline regression equation, the beta coefficient forvolunteering was reduced by half, indicating that autonomy partially mediated the relationship betweenvolunteering and job satisfaction. Thus, hypothesis 5 is partially supported. When life satisfaction wasadded to the baseline model, the coefficient for volunteering dropped nearly to zero, but remainedsignificant at the p

fact, we believe this to be true, but only partly so. Since, reciprocal relationships characterise research inthis area, we feel it would be overly simplistic to assume a unidirectional argument in the relationshipbetween the job and the rest of life. In spite of any of its limitations, we have offered a theoretical guidefor subsequent research, which may be able to clarify the causal nature of volunteering by testing forchanges in job satisfaction resulting from organizational interventions that encourage employeeinvolvement in volunteering. Although we have controlled for spuriousness to the extent possible giventhe nature of the data, the implications of controlling for more specific occupational characteristics andmore pervasive personality traits should be also be examined in future studies.The results of this study are not only suggestive of potential HR applications, but futuretheoretical extensions as well, especially concerning the relationships between job autonomy andvolunteering. As they pertain to job crafting, our results suggest that higher-order needs satisfactionthrough volunteering might not deflate the peoples’ motivation to craft their jobs, but may actually serveas an extension, or parallel stream for peoples’ crafting of work life in general. While the spilloverhypothesis is more consistent with the results here, we by no means declare it to be the “winner” in thelongstanding debate, since an established literature regarding the “secondary control” provided by extrajobdomains suggests that “…lack of job control may be compensated for by various means” (see Xie &Schaubroeck, 2001, p.41). We would suggest, however, that the long term effects of engaging involunteer work as a form of compensatory control may, in fact, be the development of more pervasiveautonomous orientations concerning job control through the motivating effects of growth needssatisfaction. In fact, we have presented a theoretical rationale for why the spillover effect is much morelikely characterize the relationship between these variables over time. By extension, our rationale maypartly help to explain why cross-sectional studies have been so unsuccessful in uncovering theoreticallyplausible compensation effects. A more robust test of the competing spillover and compensationhypotheses concerning autonomy and volunteering would involve testing a process model longitudinally,and would include more specific comparisons between the characteristics of peoples’ jobs and theirvolunteer work over time, as well as direct assessment of people’s needs satisfaction.We have suggested that autonomy and identity are intimately related, yet have focused more onautonomy in the present paper. Future research would benefit from investigating the role of identitydevelopment across these two spheres as well. Indeed, the voluntary sector may provide employees with a“job-like” setting where they may experiment with “provisional selves” (cf. Ibarra, 1999), beforechoosing which work identities are applicable in the paid sphere. Similar to the rationale provided forgrowth needs satisfaction, the importance of the voluntary sector as it pertains to identity may be itssimilarity to the paid work sphere. However, more research into the domain-specificity of certain needs isrequired in order to extend our findings to other non-job settings. For instance relatedness needs may besatisfied through volunteer work, but also through other activities (joining a book club or chatting online,for instance). Growth needs, on the other hand, may be more specific to domains which are categoricallywork-oriented (e.g., raising children, pursuing education, as well as various forms of volunteer work).In sum, our study has provided an important test of the intervening role played by life satisfactionand job autonomy in explaining volunteers’ greater satisfaction with their jobs. In seeking to expand therealm of relevant predictors of organizational behaviour to include extra-job variables, our study is similarto theoretical expansions that have developed in the absenteeism literature (Johns & Nicholson, 1982) andthe job stress literature (Bhagat, McQuaid, Lindholm, & Segovis, 1985). We hope that these findings,together with the needs-based theoretical mechanism that we have outlined, serve to spur future researchinto the role played by volunteering in people’s experience of paid work.46

Appendix ATable 1. Means, standard deviations, and zero-order intercorrelations of study variables ° (N ranges from 16966 to 19422 due to missing data)M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 181 0.49 1.02 -2 39.7 11.7 .05 ** -3 0.53 0.50 .03 ** .04 ** -4 0.74 0.44 .02 ** .22 ** .06 ** -5 0.09 0.29 .05 ** .09 ** .08 ** .06 ** -6 0.17 0.38 .01 ** .04 ** -.09 ** .00 - -7 0.15 0.36 .03 ** -.01 -.08 ** .00 - - -8 0.10 0.30 .01 -.04 ** -.15 ** -.02 ** - - - -9 0.13 0.33 -.04 ** -.10 ** -.15 ** -.05 ** - - - -10 0.03 0.17 .00 .05 ** .06 ** .00 - - - - - -11 0.15 0.36 -.05 ** -.03 ** .25 ** .02 ** - - - - - - -12 0.08 0.28 -.05 ** -.01 .16 ** .02 ** - - - - - - - -13 0.09 0.28 -.06 ** .03 ** -.06 ** -.03 ** - - - - - - - - -14 4.01 26.2 .05 ** .06 ** .06 ** .04 ** .17 * .03 ** .00 -.03 ** -.03 ** -.02 * -.03 ** -.03 ** -.04 ** -15 4.16 2.52 .09 ** -.05 ** .03 ** .00 .06 ** .09 ** .04 ** .00 -.02 * -.01 -.03 ** -.06 ** -.10 ** .04 ** .7216 2.64 1.48 -.05 ** -.06 ** -.02 * -.03 ** -.09 ** -.05 ** -.03 ** .02 * .01 .00 .06 ** .04 ** .05 ** -.04 ** .40 ** .6517 7.17 2.17 .08 ** .06 ** .02 ** .04 ** .08 ** .07 ** .04 ** .00 -.01 -.03 ** -.04 ** -.05 ** -.09 ** .03 ** .12 ** -.05 ** -18 6.94 2.24 .12 ** -.05 ** .03 ** .08 ** .05 ** .04 ** .02 * .05 ** .02 ** -.02 * -.05 ** -.05 ** -.07 ** .02 ** .12 ** -.05 ** .45 ** -19 6.41 2.68 .11 ** .08 ** .09 ** .07 ** .20 ** .10 ** .02 ** -.03 ** -.02 ** .05 ** -.07 ** -.12 ** -.15 ** .08 ** .15 ** -.11 ** .45 ** .31 *** Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).° 1. Volunteering; 2. Age; 3. Gender (0=female, 1=male); 4. Stable Relationship; 5. Legislators, senior officials and managers; 6. Professionals; 7.Technicians and associate professionals; 8. Clerks; 9. Service workers and shop and market sales workers; 10. Skilled agricultural and fisheryworkers; 11. Craft and related trades workers; 12. Plant and machine operators and assemblers; 13. Elementary Occupations; 14. Supervisoryresponsibility; 15. Intrinsic work values; 16. Extrinsic work values; 17. Job Satisfaction; 18. Life Satisfaction; 19. Autonomy.47

ReferencesAdler, Alfred, “Individual Psychology,” in C. Murchison (Ed.), Psychologies of 1930, (pp. 395-405),Worcester, MA: Clark University Press, 1930.Alderfer, Clayton P., Existence, Relatedness, and Growth; Human Needs in Organizational Settings, NewYork: Free Press, 1972.Argyle, Michael, The Psychology of Happiness, New York: Methuen & Co. Ltd., 1987.Baumeister, Roy F. & Leary, Mark R., "The Need to Belong: Desire for Interpersonal Attachments as aFundamental Human Motivation," Psychological Bulletin, 117(3), (May 1995), 497-529.Baron, Reuben M. & David A. Kenny, “The Moderator-Mediator Variable Distinction in SocialPsychological Research: Conceptual, Strategic, and Statistical Considerations,” Journal ofPersonality and Social Psychology, 51(6), (December 1986), 1173-1182.Berlinger, Lisa R., Glick, William H. & Rodgers, Robert C., “Job Enrichment and PerformanceImprovement,” in John P. Campbell and Richard J. Campbell (Eds.), Productivity inOrganizations: New Perspectives from Industrial and Organizational Psychology, (pp. 219-254),San Francisco, CA: Jossey-Bass, 1988.Bhagat, Rabi. S., McQuaid, Sara J., Lindlom, Hal & Segovis, James, “Total Life Stress - a MultimethodValidation of the Construct and its Effects on Organizationally Valued Outcomes and WithdrawalBehaviors,” Journal of Applied Psychology, 70(1), (February 1985) 202-14.Breer, Paul E. & Edwin A. Locke, Task Experience as a Source of Attitudes. Homewood, Ill.: DorseyPress, 1965.Chacko, Thomas I., “Job and Life Satisfactions - a Causal-Analysis of their Relationships,” Academy ofManagement Journal, 26(1), (March 1983), 163-169.Champoux, Joseph E., “Perceptions of Work and Nonwork: A Reexamination of the Compensatory andSpillover Models,” Sociology of Work & Occupations, 5(4), (November 1978), 402-22.Champoux, Joseph E., “The World of Nonwork: Some Implications for Job Re-Design Efforts,”Personnel Psychology, 33 (Spring 1980), 61-75.Champoux, Joseph E., “A Sociological Perspective on Work Involvement,” International Review ofApplied Psychology, 30 (January 1981), 65-86.Clary, E. Gil & Mark Snyder “The Motivations to Volunteer: Theoretical and Practical Considerations,”Current Directions in Psychological Science, 8(5) (October 1999), 156-9.Drucker, Peter F., The New Realities: In Government and Politics, in Economics and Business, in Societyand World View. New York: Harper & Row, 1989.Dubin, R.. Industrial workers' worlds: A study of the central life interests of industrial workers. SocialProblems, 3, (January 1956), 131-142.Edwards, Jeffrey R. & Nancy P. Rothbard, “Mechanisms Linking Work and Family: Clarifying theRelationship between Work and Family Constructs,” Academy of Management Review, 25(1),(January 2000), 178-99.Fried, Yitzhak & Ferris, Gerald R., "The Validity of the Job Characteristics Model: A Review and Meta-Analysis," Personnel Psychology, 40(2), (Summer 1987), 287-322.Hackman, J. Richard, “Designing Work for Individuals and for Groups,” in J. Richard Hackman, EdwardE. Lawler, & Lyman W. Porter (Eds.), Perspectives on Behavior in Organizations, (pp. 242-257),New York: McGraw-Hill Book Company, 1977.Hackman, J. Richard & Lawler, Edward E., “Employee Reactions to Job Characteristics,” Journal ofApplied Psychology, 55(3), (June 1971), 259-286.Hackman, J. Richard & Oldham, Greg R. “Development of Job Diagnostic Survey,” Journal of AppliedPsychology, 60(2), (April 1975), 159-170.Inglehart, R., et al. World Values Surveys and European Values Surveys, 1999-2001 [Computer file].ICPSR version. Ann Arbor, MI: Institute for Social Research [producer], 2002. Ann Arbor, MI:Inter-university Consortium for Political and Social Research [distributor], 2004.48

ILO, International Standard Classification of Occupations: ISCO-88, Geneva: ILO, 1990.Johns, Gary & Nicholson, Nigel, “The Meanings of Absence: New Strategies for Theory and Research,”in Barry M. Staw and L. L. Cummings (Eds.), Research in Organizational Behavior, (Vol. 4,pp.127-172), Greenwich, Conn: JAI Press, 1982.Kabanoff, Boris & O’Brien, Gordon E., “Work and Leisure - a Task Attributes Analysis,” Journal ofApplied Psychology, 65(5), (October 1980), 596-609.Kirchmeyer, Catherine, “Nonwork Participation and Work Attitudes: A Test of Scarcity Vs. ExpansionModels of Personal Resources,” Human Relations, 45(8), (August 1992), 775-795.Kohn, Melvin L., Class and Conformity: A Study in Values, with a Reassessment, Chicago: University ofChicago Press, 1977.Kohn, Melvin L. & Schooler, Carmi, “Occupational Experience and Psychological Functioning: AnAssessment of Reciprocal Effects,” American Sociological Review, 38(1), (February 1973), 97-118.Kornhauser, Arthur William, Mental Health of the Industrial Worker, New York: Wiley, 1965.Lewin, Kurt, A Dynamic Theory of Personality; Selected Papers, New York: McGraw-Hill, 1935.Lewin, Kurt, Field Theory in Social Science; Selected Theoretical Papers, Westport, Conn: GreenwoodPress, 1975.Maslow, Abraham H., Toward a Psychology of being. Princeton, N.J: Van Nostrand, 1962.Maslow, Abraham H., “A Theory of Human Motivation,” Psychological Review, 50, (1943), 370-396.Near, Janet. P., Rice, Robert W. & Hunt, Raymond, G., “Work and Extra-Work Correlates of Life andJob Satisfaction,” Academy of Management Journal, 21(2), (June 1978) 248-64.Near, Janet P., Robert W. Rice, and Raymond G. Hunt (1980), “The Relationship between Work andNonwork Domains: A Review of Empirical Research,” Academy of Management Review, 5(3),(July 1980), 415-431.Oldham, Greg R. & Hackman, J. Richard, “Work Design in the Organizational Context,” in Barry M.Staw and L. L. Cummings (Eds.), Research in Organizational Behavior, (Vol 2, pp. 247-278),Greenwich, CT: JAI Press, 1980.Rain, Jeffrey. S., Lane, Irving M. & Steiner, Dirk D. (1991), “A Current Look at the Job-Satisfaction LifeSatisfaction Relationship - Review and Future Considerations,” Human Relations, 44(3) (March1991), 287-307.Rice, Robert W., Near, Janet P. & Hunt, Raymond, G., “Unique Variance in Job and Life SatisfactionAssociated with Work-Related and Extra-Workplace Variables,” Human Relations, 32(7), (July1979), 605-623.Rousseau, Denise M., “Relationship of Work to Nonwork,” Journal of Applied Psychology, 63(4),(August 1978), 513-517.Salancik, Gerald R., “Commitment and the control of organizational behavior and belief,” in Barry M.Staw and Gerald R. Salancik (Eds.), New Directions in Organizational Behavior, (pp. 1-54),Chicago: St.-Clair Press, 1977.Salancik, Gerald R. & Pfeffer, Jeffrey, “Examination of Need-Satisfaction Models of Job Attitudes,”Administrative Science Quarterly, 22(3), (September 1977), 427-456.Salancik, Gerald R. & Pfeffer, Jeffrey, “A Social Information Processing Approach to Job Attitudes andTask Design,” Administrative Science Quarterly, 23(2), (June 1978), 224-254.Schmitt, Neal & Bedeian, Arthur G., “A Comparison of LISREL and Two-Stage Least Squares Analysisof a Hypothesized Life-Job Satisfaction Reciprocal Relationship. Journal of Applied Psychology,67(6), (December1982), 806-817.Schmitt, Neal & Mellon, Phyllis M., “Life and job satisfaction: Is the job central?” Journal of vocationalBehavior, 16(1), (February 1980), 51-58.Staines, Graham L., “Spillover Versus Compensation: A Review of the Literature on the Relationshipbetween Work and Nonwork,” Human Relations, 33(2) (February 1980), 111-129.Sutton, Robert I. & Barry M. Staw, “What Theory is Not,” Administrative Science Quarterly, 40(3)(September 1995), 371-384.49

Tait, Marianne, Padgett, Margaret Y. & Baldwin, Timothy T., “Job and Life Satisfaction: A Reevaluationof the Strength of the Relationship and Gender Effects as a Function of the Date of the Study,”Journal of Applied Psychology, 74(3), (June, 1989), 502-507.Tajfel, Henri, Human Groups and Social Categories: Studies in Social Psychology, Cambridge:Cambridge University Press, 1981.Terkel, Studs, Working; People Talk about what they do all Day and how they Feel about what they do.New York: Ballantine Books, 1990.Thoits, Peggy A. & Hewitt, Lyndi N., “Volunteer Work and Well-being,” Journal of Health and SocialBehavior, 42(2) (June 2001), 115-131.Voydanoff, Patricia (2001), "Incorporating Community into Work and Family Research: A Review ofBasic Relationships," Human Relations, 54(12), (December 2001), 1609-1637.Wanous, John P., “Individual-Differences and Reactions to Job Characteristics,” Journal of AppliedPsychology, 59(5), (October 1974), 616-622.Wanous, John P., Reichers, Arnon E. & Hudy, Michael J., "Overall Job Satisfaction: How Good areSingle-Item Measures?" Journal of Applied Psychology, 82(2), (April, 1997), 247-252.Wilensky, Harold. L., “Work, Careers, and Social Integration,” International Social Science Journal, 12,(1960), 543-560.Wilson, John, “Volunteering,” Annual Review of Sociology, 26(1), (2000), 215-240.Wilson, John & Musick, Marc A., (1997), “Work and Volunteering: The Long Arm of the Job,” SocialForces, 76(1), (September, 1997), 251-272.Wrzesniewski, Amy & Dutton, Jane E., “Crafting a Job: Revisioning Employees as Active Crafters oftheir Work,” Academy of Management Review, 26(2), (April 2001), 179-201.Xie, Jia Lin & Schaubroeck, John, “Bridging Approaches and Findings across Diverse Disciplines toImprove Job Stress Research,” in Pamela L Perrewe & Daniel C. Ganster (Eds.), ExploringTheoretical Mechanisms and Perspectives, (Vol. 1, pp1-61), New York: Elsevier Science Ltd.,2001.50

ASAC 2006Banff, AlbertaClaude Fernet (Ph.D. Candidate)Université LavalMarylène GagnéConcordia UniversityHOW DO WORK MOTIVATION AND SOCIAL SUPPORT PREDICTJOB BURNOUT OVER TIME?We tested whether autonomous motivation moderated the relationship between socialsupport and burnout. College employees participated in a longitudinal study. Resultsindicate main effects of social support and autonomous motivation on cross-sectionalsymptoms of burnout. Results also reveal an interaction between social support andautonomous motivation predicting burnout over time. High social support reducedburnout over time only for employees with low autonomous motivation.During the last decades, many studies have shown that supportive interpersonal relationships arean important job resource at work that has a profound impact on individuals’ propensity to react to workstress and to burnout. Burnout is defined as a symptom of emotional exhaustion, depersonalization, andreduced personal accomplishment at work due to work activity (Maslach, 1982). Emotional exhaustionrefers to the depletion of one’s emotional resources. Depersonalization refers to a detached attitude thatemployees use toward others in order to protect themselves from the psychological stress coming frompeople with whom they interact. Reduced personal accomplishment refers to a decrease in the feeling ofcompetence and productivity at work. Thus, in addition to being marked by a loss of emotional energy,burnout implies a negative assessment of the self (reduced personal accomplishment) and of others(depersonalization). So far, studies have suggested that the three dimensions of burnout are associatedwith different aspects of work environment, in particular job demands and job resources (Bakker,Demerouti, & Euwema, 2005). For instance, reviews of Lee and Ashforth (1996) and Schaufeli andEnzmann (1998) suggested that emotional exhaustion is primarily predicted by job demands (e.g.,overload) and lack of job resources (e.g., social support), whereas depersonalization and personalaccomplishment are related, albeit less convincing, with job resources (e.g., social support). In addition, ithas been suggested that burnout dimensions are related with certain personal characteristics, such as selfesteem(Janssen, Schaufeli, & Houkes, 1999), self-efficacy at work (Salanova, Peiro, & Schaufeli, 2002),work self-determination (Fernet, Guay, & Senécal, 2004), and coping style (De Rijk, LeBlanc, Schaufeli,& De Jonge, 1998).Burnout has also been associated with lack of social support (Cordes & Dougherty, 1993;Greenglass, Fiksenbaum, & Burke, 1994). Individuals who possess a high level of social support at workfeel less exhausted and show less detached attitude toward their work (Lee & Ashforth, 1996).Explanations of this effect include the buffering effect that social support may have on stress, so thatindividuals who receive social support may be better able to cope with workplace stressors. Accordingly,social support has been considered as an important job resource at work that reduces the impact of jobdemands when predicting job burnout. This buffering pattern is supposed to reflect an underlying processby which people use their social resources to cope with demands, so that demands do not necessarily leadto burnout. Several recent reviews have shown that the support for the moderating effect is inconclusive(see Van der Doef & Maes, 1999).51

The inconsistent pattern of results suggests that social support at work may function throughmultiple processes. According to Hobfoll (2002), social resources are not only necessary to deal with jobdemands and to ‘get things done’, but they also are important in their own right. Then, job resources, suchas social support, may play either an intrinsic motivational role because they foster employee’sdevelopment and well being, or they may play an extrinsic motivational role because they areinstrumental in achieving work goals. Self-determination theory (SDT; Deci & Ryan, 1985) wouldpropose that social support fulfills basic psychological needs, such as the needs for autonomy (deCharms,1968), competence (White, 1959), and relatedness (Baumeister & Leary, 1995). Interpersonalrelationships at work that support psychological needs enhance the well-being of employees (Ryan &Deci, 2001). For instance, constructive feedback from a colleague may foster learning thereby increasingjob competence. In contrast, unsupportive interpersonal relationships at work that thwart the satisfactionof psychological needs can lead to maladaptive consequences, such as feelings of emotional exhaustion,depersonalization, and reduced personal accomplishment.The question addressed in the present research is whether every worker equally needs socialsupport in order not to feel burned out. Several studies in occupational health psychology indicated thatcertain people are more vulnerable to the influence of work-related resources than others (Parkes, 1994).We contend that people’s motivation for doing their job, that is, why they work, may change the impactthat social support may have as a buffer against stress. There may be several reasons for this moderatingeffect. First, people with different motives may not experience the same amount of work stress. Second,social support may not be as important to people with certain motives compared to people with othermotives. To examine this hypothesis, we used the only work motivation theory that differentiates betweendifferent types of motivation to work, self-determination theory (Deci & Ryan, 1985, 2000; Gagné &Deci, 2005).Self-Determination TheorySelf-determination theory (Deci & Ryan, 1985, 2002) distinguishes between intrinsic motivation(i.e., doing something for its own sake) and extrinsic motivation (i.e., doing something to obtain a rewardor avoid a punishment). The theory also proposes that internalization is the process involved in theacquisition and acceptance of new values or goals, which leads people to become autonomouslymotivated to engage in behaviors that express these values and goals (Ryan, 1995), and this is used topropose different types of extrinsic motivation that can be relatively controlled by external factors, or thatcan be relatively autonomous, that is, regulated through a person’s acquired goals and values. At the lowend lies external regulation, which refers to doing an activity to obtain rewards or to avoid punishments,and is therefore completely externally controlled. Introjected regulation refers to self-regulation throughself-worth contingencies like ego-involvement and guilt, and is therefore only partially internalized andremains controlling, not volitional. Identified regulation refers to doing an activity because one identifieswith its value or meaning, and accepts it as one’s own, which means that it is autonomously regulated.Identification differs from intrinsic motivation in that the activity is done not so much for its own sake(because it is interesting and fun), but for the instrumental value it represents. In sum, when putting theseforms of motivation on the continuum of internalization, external regulation and introjection representcontrolled motivation, whereas identification and intrinsic motivation represent autonomous motivation.Research in different domains, such as education (Williams & Deci, 1996), sports (Vallerand & Fortier,1998), work (Blais, Lachance, Briere, Riddle, & Vallerand, 1993), and health care (Williams, Grow,Freedman, Ryan, & Deci, 1996), has supported that the types of motivation form a quasi-simplex patternthat represents variation in underlying internalization, which means that each subscale correlates mostpositively with the subscales closest to it and less positively or more negatively with subscales fartherfrom it. Many researchers use a composite score that represents how one is relatively autonomouslymotivated. This index is calculated by summing intrinsic and identified scores and substracting theintrojected and external scores (e.g., Sheldon and Elliot, 1999; Williams et al., 1996). Positive scores52

indicate that one is more autonomously motivated, whereas negative scores indicate that one is morecontrollingly motivated.Work motivation fits into job burnout process in two different ways. In that case, it may act eitheras a direct determinant or as a moderator of the relations between job resources (i.e., social support) andjob burnout. According to the first point of view, employees with low self-determination at work wouldbe more likely to develop burnout, or at least to experience the symptoms more intensively, and this,regardless of other factors. For instance, Blais et al. (1993) showed that multiple indicators of workautonomous motivation are significantly related to emotional exhaustion (mean r = -.26),depersonalization (mean r = -.31), and reduce personal accomplishment (mean r = -.48), while indicatorsof work controlled motivation are positively associated with those symptoms, mean r = .32, mean r = .27,mean r = .14, respectively. The second point of view suggests that those with initially high degree ofwork autonomous motivation will adapt favorably in the work context, while those with lowerautonomous motivation will respond to job constraints/resources in a non adaptive manner. For instance,in their study Fernet et al. (2004) showed that when autonomously motivated workers faced job demands,job control (which is a job resource allowing opportunities for control and decision) tends to reduceemotional exhaustion and depersonalization, in addition to fostering feeling of personal accomplishmentat work. Conversely, for low autonomously motivated employees, job control appears to have little valueas a stress-reducing factor when they are facing job demands. These results indicated that workautonomous motivation plays a role in an employee’s adjustment to the constraints/resources of the workenvironment.But our current proposition instead suggests that social support may be especially important tothose employees with lower autonomous motivation. Black and Deci (2000) found support in theeducational domain for this proposition: College students who had a supportive instructor in an organicchemistry course increased in their autonomous motivation over the semester but also obtained bettercourse grades after controlling for SAT scores and GPAs. What particularly interesting is that this findingwas stronger for students with initially low levels of autonomous motivation. This proposition isconsistent with several models of job burnout (e.g., Demerouti, Bakker, Nachreiner, & Schaufeli, 2001).Generally speaking, such models hypothesize that job demands and job resources are importantdeterminants of job burnout by acting either in a direct or an interactive way. Independently of the processproposed by these models, our proposition expands at least the resources part of these models, byclaiming that work motivation can act as a moderator between social support and burnout. Thisproposition agrees with Baumeister and Leary’s (1995) theoretical proposition that interpersonalrelationships be differentially important for individual well-being.Overview and HypothesisThe aim of the present study is to test the moderating cross-sectional and longitudinal influence of workautonomous motivation on the relationship between social support and each of the three dimensions ofburnout (i.e., emotional exhaustion, depersonalization, and personal accomplishment). The followinghypothesis is proposed.Hypothesis: Work autonomous motivation will moderate the effect of social support on each ofthe three burnout dimensions, such that employees who are less autonomously motivated will be morevulnerable to burnout when they perceive low social support at work than employees who are moreautonomously motivated.We expect to find this effect both cross-sectionally and longitudinally.53

MethodProcedure and ParticipantsParticipants were French-Canadian employees from a college in Quebec City. Employees were asked tocomplete a questionnaire and a consent form and return them in a pre-paid envelope on two occasionsover a two-year period. It was emphasized that participation was entirely voluntary and confidentialitywas guaranteed. Employees were also asked to put their name on the questionnaire so that their responsescould be tracked over time. The results of the surveys were fed back to all participants throughpresentations and written reports.At time-1 (T1), 376 employees (159 men and 217 women) completed the questionnaire, giving a responserate of 43%. Participants’ mean age was 45.1 years (SD=8.28) and years of experience mean was 16 years(SD=9.72). 61% of the participants were teachers, 27% were support services employees, 7% wereeducational professionals, and 5% were administrative employees. A total of 69% held a permanentposition and 86% were full-time. At time-2 (T2), 273 employees (98 men, 172 women, and 3 withoutgender identification) completed the questionnaire with a response rate of 31%. Participants’ mean agewas 44.3 years (SD=9.15) and years of experience mean was 14.1 years (SD=10.1). 60% of theparticipants were teachers, 26% were support services employees, 7% were educational professionals, and7% were administrative employees. A total of 69% held a permanent position and 86% were full-time.Finally, a total of 123 employees (39 men, 81 women, and 3 without gender identification) completed thequestionnaire at both time points. Participants’ mean age was 46 years (SD=7.84) and years of experiencemean was 16.4 years (SD=9.23). 55% of the participants were teachers, 25% were support servicesemployees, 11% were educational professionals, and 9% were administrative employees. A total of 81%held a permanent position and 89% were full-time.MeasuresSocial support. Perceptions of social support at work were assessed by the 4-item Quality ofInterpersonal Relationships Scale (EQRI; Senécal, Vallerand, & Vallières, 1992). Participants indicatedthe extent to which each of the item (e.g., harmonious, motivating, satisfying, and lead me to trust them)corresponds to their current relationships with their colleagues on a 0 (not at all) to 4 (extremely) pointscale.The items follow the stem “Currently, my relationship with my colleagues at work are...”.Cronbach’s alphas were .93 at T1 and .92 at T2.Work motivation. Work autonomous motivation was assessed by the short version of the BWMI(Blais et al., 1993), which was developed in French. The short version assesses four motivationaldimensions of three items each on a 1 (do not agree at all) to 7 (agree completely) point-scale. The itemsfollow the stem “Why do I do this job?” with each item representing a reason for working. The subscalesassess intrinsic motivation (e.g., “Because I experience satisfaction when my job provides me withinteresting challenges”), identified regulation (e.g., “Because this is the type of work that I preferregarding my career aspirations”), introjected regulation (e.g., “Because I absolutely want to be good andif I’m not, I’ll be disappointed”), and external regulation (e.g., “For the salary”). We used themotivational index described earlier to obtain the degree to which employees were autonomouslymotivated. Cronbach’s alphas were .85 and .83.Job burnout. The 22-item French version (Dion & Tessier, 1994) of the Maslach BurnoutInventory (MBI; Maslach & Jackson, 1986) was used to assess the three dimensions of job burnout:emotional exhaustion, depersonalization, and personal accomplishment on a 0 (never) to 6 (every day)point-scale. The psychometric properties of the French version of the MBI are similar to those of theoriginal version (Maslach, Schaufeli, & Leiter, 2001). Emotional exhaustion consisted of nine items (e.g.,54

“I feel emotionally drained from my work”; Cronbach’s alphas .88 at T1 and .91 at T2). Five itemsassessed depersonalization (e.g., “I’ve become more callous toward people since I took this job”;Cronbach’s alphas .70 at T1 and .71 at T2). Personal accomplishment consisted of eight items (e.g., “Ihave accomplished many worthwhile things in this job; Cronbach’s alphas .75 at each time points).Statistical analysesHierarchical moderated multiple regression analyses (Cohen & Cohen, 1983) were performed totest our hypothesis. For the cross-sectional tests, the predictor variables were entered within threesuccessive steps. In the first step, following research on job burnout (see Maslach et al., 2001 for areview), gender and age were introduced as control variables. In the second step, main effects (socialsupport and work autonomous motivation) were entered. In the third step, the two-way interaction term(i.e., social support*work autonomous motivation) was entered. For the longitudinal tests, the predictorsvariables were entered in four steps. In these analyses, each burnout variable measured at T2 wasconsidered as the dependent variable. In the first step, gender and age were introduced as controlvariables. In the second step, the corresponding burnout variable at T1 was entered. In the third step, maineffects at T1 (social support and work autonomous motivation) were entered. Finally, in the fourth step,the two-way interaction term at T1 (social support * work autonomous motivation) was entered. To avoidmulticollinearity between the predictors and the interaction terms, interaction terms were computed usingcentered scores (Aiken & West, 1991; Kline, 1998). With this approach, the incremental variance (∆R 2 )accounted for by the interaction term reflects the effect size of the interaction. To interpret interactioneffects, which involved continuous variables, simple slopes were derived for high (+1 SD), and low levels(-1 SD) of the moderator, work autonomous motivation (Aiken & West, 1991).ResultsPreliminary analysesCorrelations between all variables are presented in Table 1. As expected, social support andautonomous motivation correlated significantly with emotional exhaustion, depersonalization, andpersonal accomplishment at each time point.55

Table 1Correlations for time-1 and time-21 2 3 4 5 6 7 8 9 10Time 11. Social support --2. Autonomous motivation .27** --3. Emotional exhaustion -.33** -.20** --4. Depersonalization -.39** -.26** .41** --5. Personal accomplishment .40** .34** -.29** -.34** --Time 26. Social support .46** .22* -.17* -.29** .23* --7. Autonomous motivation .19* .72** -.08 -.20* .39** .27** --8. Emotional exhaustion -.30** -.03 .60** .28** -.25** -.34** -.25** --9. Depersonalization -.36** -.25** .23* .51** -.27** -.38** -.19** .40** --10. Personal accomplishment .36** .33** -.22* -.39** .55** .45** .38** -.28** -.29** --* p

Table 3Hierarchical multiple regression analyses fordepersonalization (cross-sectional)StepsTime-1Time-2Variables entered 1 2 3 1 2 31. GenderAge.02.12*.02.07.02.07.18*-.14.22**-.11.23**-.112. Social supportAutonomous motivation-.34**-.17**-.36**-.18**-.45**-.19*-.44**-.18*3. Support * Motivation -.04 .06R 2 .02 .19** .20 .05 .34** .34∆R 2 .02 .18** .01 .05 .29** .00*p < .05 **p < .01.Table 4Hierarchical multiple regression analyses forpersonal accomplishment (cross-sectional)StepsTime-1Time-2Variables entered 1 2 3 1 2 31. GenderAge.00.13*.00.18**.00.17**.01.04-.03.01-.04.002. Social supportAutonomous motivation.35**.26**.34**.25**.42**.28**.40**.28**3. Support * Motivation -.06 -.08R 2 .02* .25** .25 .00 .33** .34∆R 2 .02* .23** .00 .00 .33** .01*p < .05 **p < .01.Longitudinal analyses. Three sets of regression analyses were performed to test our secondhypothesis based on the interaction effect of social support and autonomous motivation when predictingburnout over time (i.e., one for each burnout dimension). Results based on emotional exhaustion arepresented in Table 5. After controlling for gender and age, as well as exhaustion at time-1, social supportand autonomous motivation at time-1 were not significantly related with emotional exhaustion at time-2.However, the interaction of these predictors was positively associated with emotional exhaustion overtime, explaining 3% of its variance. Simple effects show that social support was negatively associatedwith emotional exhaustion over time only for people with low autonomous motivation (β = -.34, p

autonomous motivation (β = .48, p < .001, vs. β = .04, ns. for people with high autonomous motivation,see Figure 3). Overall, results of longitudinal analyses support our hypothesis.Table 5Hierarchical multiple regression analyses foremotional exhaustion (time-2)StepsVariables entered 1 2 3 41. GenderAge.02-.11.14-.13.14-.15*.15*-.17*2. Emotional exhaustion (time-1) .62** .58** .59*3. Social support (time-1)Autonomous motivation (time-1)-.10.01-.10-.024. Support * Motivation (time-1) .17*R 2 .01 .39 .39 .42*∆R 2 .01 .37** .01 .03**p < .05. *p < .01.**Table 6.Hierarchical multiple regression analyses fordepersonalization (time-2)StepsVariables entered 1 2 3 41. GenderAge.18*-.14.24**-.18*.24**-.20**.25**-.22**2. Emotional exhaustion (time-1) .54** .43** .37**3. Social support (time-1)Autonomous motivation (time-1)-.18*-.10-.20*-.154. Support * Motivation (time-1) .18*R 2 .05 .33** .37* .40*∆R 2 .05 .29** .04* .03**p < .05. *p < .01.**Table 7Hierarchical multiple regression analysesfor personal accomplishment (time-2)StepsVariables entered 1 2 3 41. GenderAge.01.04-.04-.09-.05-.04-.06-.002. Emotional exhaustion (time-1) .60** .50** .46**3. Social support (time-1)Autonomous motivation (time-1).21**.11.21**.154. Support * Motivation (time-1) -.16*R 2 .00 .34** .40** .42*∆R 2 .00 .34** .06** .02**p < .05. *p < .01.**58

10,80,6High AMLow AMLinear (High AM)Linear (Low AM)Emotional exhaustion Time-20,40,20-0,2-0,4y = 0,113x - 0,026y = -0,345x + 0,026-0,6-0,8-1-1 0 1Social support Time-1Figure 1Two-way interaction between social support and work autonomousmotivation predicting emotional exhaustion over time.59

10,80,6High AMLow AMLinear (High AM)Linear (Low AM)Depersonalization Time-20,40,20-0,2-0,4y = 0,003x - 0,147y = -0,427x + 0,147-0,6-0,8-1-1 0 1Social support Time-1Figure 2Two-way interaction between social support and work autonomousmotivation predicting depersonalization over time.60

10,80,6High AMLow AMLinear (High AM)Linear (Low AM)Personal accomplishment Time-20,40,20-0,2-0,4y = 0,54x - 0,168y = 0,036x + 0,168-0,6-0,8-1-1 0 1Social support Time-1Figure 3Two-way interaction between social support and work autonomouamotivation predicting personal accomplishment over time.DiscussionThis study aimed evaluating the moderating cross-sectional and longitudinal influence of workautonomous motivation on the relationship between social support at work and each of the threedimensions of burnout (i.e., emotional exhaustion, depersonalization, and personal accomplishment). Weexpected to find that social support would have a stronger negative influence over burnout for people whohad low autonomous work motivation. After controlling for gender and age, we found no support for ourhypothesis at the cross-sectional level: social support and work autonomous motivation both had negativeindependent effects over burnout. Thus, the more people perceived social support at work and the moreautonomously motivated they were, the least likely they were to burnout. We did, however, find supportfor our hypothesis at the longitudinal level: social support had a strong effect on all three burnoutsymptoms 2 years later only for people with initially low autonomous motivation.These results support the idea that not all people in the same way to social support. Social supportseems especially important to people who are not autonomously motivated, that is, people who do notexperience a lot of pleasure or meaning from their work, but that instead are more motivated throughinternal and external pressure. Social support may thus buffer the effects of that pressure, or job demands,on stress and burnout. A proposition in line with job burnout models (e.g., Demerouti et al., 2001). Forpeople with low autonomous work motivation, positive relationships at work with colleagues tend toreduce exhaustion and depersonalization over time, in addition to fostering feelings of personal61

accomplishment at work. Conversely, having poor interpersonal relationships with colleagues at work canmake people more vulnerable to emotional exhaustion and depersonalization over time without theopportunity to develop their feelings of personal accomplishment. But for people with high autonomouswork motivation, positive interpersonal relationships with colleagues appear to have little value to reducesymptoms of burnout. In general, people with high autonomous work motivation do not score high onburnout symptoms. This implies that their psychological long-term adaptation at work is not contingenton the amount of social support they receive from colleagues. For them, the intrinsic value of achievingmeaningful and interesting goals at work may be sufficient to buffer them against the negative effect ofstress.The results of the present study should be interpreted in light of certain limitations. First, thisstudy relies on self-report data. This could give rise to the common problem of common method variancethat may artifactually inflate the relationships between constructs. However, the existence of nonsignificant zero-order correlations between time-1 outcomes and time-2 predictors (e.g., emotionalexhaustion at time-1 and work autonomous motivation at time-2), sheds doubt on this possibility. Second,there was attrition from time-1 and time-2. The panel group in our study consisted of 14% of totalemployees. However, we conducted some post-hoc analyses to compare non-respondents to our panelrespondents, and found no differences between them. Finally, results cannot be generalized to allemployees on the labor market because the participants consisted only of college employees.Nevertheless, the results of this study indicate that work autonomous motivation plays a role inemployees’adjustments to job demands as well as to their need for job resources, such as social support.62

ReferencesAitken, L. S., & West, S. G. (1991). Multiple Regression: Testing and interpreting interaction. NewburyPark, CA: Sage.Black, A. E., & Deci, E. L. (2000). The effects of instructors’ autonomy support and students’autonomous motivation on learning organic chemistry: A self-determination theory perspective.Science Education, 84, 740-756.Bakker, A. B., Demerouti, E., & Euwema, M. C. (2005). Job Resources Buffer the Impact of JobDemands on Burnout. Journal of Occupational Health Psychology, 10(2), 170-180.Baumeister, R., & Leary, M.R. (1995). The need to belong: Desire for interpersonal attachments as afundamental human motivation. Psychological Bulletin, 117, 497-529.Blais, M. R., Lachance, L., Brière, N. M., Riddle, A. S., & Vallerand, R. J. (1993). L'inventaire desmotivations au travail de Blais. Revue Québécoise de Psychologie, 14, 185-215.Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioralsciences. Hillsdale, NJ: Erlbaum.Cordes, C. L., & Dougherty, T. W. (1993). A review and an integration of research on job burnout.Academy of Management Review, 18(4), 621-656.deCharms, R. (1968). Personal causation: The internal affective determinants of behavior. New York:Academic Press.Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. NewYork: Platinum Press.Deci, E.L., & Ryan, R.M. (2002). Handbook of Self-Determination Research. New York: The Universityof Rochester Press.Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-resourcesmodel of burnout. Journal of Applied Psychology, 96(3), 499-512.De Rijk, A. E., LeBlanc, P. M., Schaufeli, W. B., & de Jonge, J. (1998). Active coping and need forcontrol as moderators of the job demand-control model: Effects on burnout. Journal ofOccupational and Organizational Psychology, 71, 1-18.Dion G, & Tessier R. (1994). Validation and translation of the burnout inventory of Maslach and Jackson.Canadian Journal of Behavior Science, 26, 210–27.Fernet, C., Guay, F., & Senecal, C. (2004). Adjusting to job demands: The role of work selfdeterminationand job control in predicting burnout. Journal of Vocational Behavior, 65, 39-56.Gagné, M., & Deci, E. L. (2005). Self-determination theory as a new framework for understandingorganizational behavior. Journal of Organizational Behavior, 26, 331-362.Greenglass, E. R., Fiksenbaum, L., Burke, R. J. (1994). The relationship between social support andburnout over time in teachers. Journal of Social Behavior and Personality, 9(2), 219-230.Hobfoll, S.E. (2002). Social and psychological resources and adaptation. Review of General Psychology,6, 307-324.Janssen, P. M., Schaufeli, W. B., & Houkes, I. (1999). Work-related and individual determinants of thethree burnout dimensions. Work and Stress, 13, 74-86.Kline, P. (1998). The new psychometrics: Science, psychology and measurement. Florence, KY: Taylorand Francis/Routledge.Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the threedimensions of job burnout. Journal of Applied Psychology, 81, 123-133.Maslach, C. (1982). Burnout: The cost of Caring. New York: Prentice-Hall.Maslach, C., & Jackson, S. E. (1986). Manual Maslach Burnout Inventory. Palo Alto, CA: ConsultingPsychologists Press.Maslach, C., Schaufeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology,52,397-422.Parkes, K. R. (1994). Personality and coping as moderators of work stress processes: Models, methodsand measures. Work and Stress, 8, 110-129.63

Ryan, R. M. (1995). Psychological needs and the facilitation of integrative processes. Journal ofPersonality, 63(3), 397-427.Ryan, R. M., & Deci, E. L. (2001). On happiness and human potentials: A review of research on hedonicand eudaimonic well-being. Annual Review of Psychology, 52, 141-166.Salanova, M., Peiro, J., & Schaufeli, W. B. (2002). Self-efficacy specificity and burnout amonginformation technology workers: An extension of the demand-control model. European Journal ofWork and Organizational Psychology, 11(1), 1-25.Schaufeli, W.B., & Enzmann, D. (1998). The burnout companion to study and research: A criticalanalysis. London: Taylor and Francis.Senécal, C. B., Vallerand, R. J., & Vallières, É. F. (1992). Construction et validation de l'Échelle de laqualité des Relations Interpersonnelles (EQRI). Revue Européenne de Psychologie Appliquée,42(4), 315-322.Sheldon, K. M. & Elliot, A. J. (1999). Goal striving, need-satisfaction, and longitudinal well-being: TheSelf-Concordance Model. Journal of Personality and Social Psychology, 76, 482-497.Vallerand, R. J., & Fortier, M. S. (1998). Measures of intrinsic and extrinsic motivation in sport andphysical activity: A review and critique. In J. Duda (Ed.), Advancements in sport and exercisepsychology measurement (pp. 83-100). Morgantown, WV: Fitness Information Technology.Van der Doef, M., & Maes, S. (1999). The job demand-control (-support) model and psychological wellbeing:A review of 20 years of empirical research. Work and Stress, 13, 87-114.White, R.W. (1959). Motivation reconsidered: The concept of competence. Psychological Review, 66,297-333.Williams, G. C., & Deci, E. L. (1996). Internalization of biopsychosocial values by medical students: Atest of self-determination theory. Journal of Personality and Social Psychology, 70, 767-779.Williams, G. C., Grow, V. M., Freedman, Z. R., Ryan, R. M., & Deci, E. L. (1996). Motivationalpredictors of weight loss and weight-loss maintenance. Journal of Personality and SocialPsychology, 70, 115-126.64

ASAC 2006Banff, AlbertaI. M. JawaharDepartment of ManagementIllinois State UniversityThomas H. StoneDepartment of ManagementOklahoma State UniversityCAN ORGANIZATIONAL SUPPORT AND POLITICALSKILL REDUCE BURNOUT?Role conflict was associated with emotional burnout among 120 software professionalsand managers. Organizational support was associated with less emotional exhaustion anddepersonalization, and moderated the role conflict-emotional exhaustion relationship.Political skill was associated with less depersonalization and feelings of reduced personalaccomplishment, and moderated the role conflict-reduced personal accomplishmentrelationship.Job stress including burnout is on the rise (e.g., Kahn & Langlieb, 2003). Because job stress is ofmajor significance to both employees and organizations, it is important to continue searching formechanisms that can reduce the negative effects of job stress. Drawing on theoretical frameworks of thestress-strain relationship (Karasek, 1979; Lazarus, 1991), we examined the moderating role of perceivedorganizational support and political skill in the relationship between role conflict and burnout.This study makes several contributions to the literature. For instance, this study responds to thecall for more research on “organizational and interpersonal factors that might serve as moderators,buffers, or even antidotes to stresses and their effects” by Kahn and Byosiere (1992, p. 572) and others(e.g., Stamper & Johlke, 2003). Second, to our knowledge, this study is the first to examine the direct andmoderating effects of an individual difference variable and an organizational variable on burnoutdimensions in a single study. Third, results of our study contribute not only to the literature on burnoutbut also to the large body of research on perceived organizational support and to the emerging stream ofresearch on political skill. Finally, results of our study have practical implications for reducing burnoutexperienced by increasing number of employees in contemporary organizations.Prevalence and Consequences of StressChanges in work environments, such as elimination of middle management layers, the adoptionof team-based work structures, and the increasing proportion of service-oriented jobs in the Americaneconomy have brought about a shift from hierarchical and bureaucratic organizational designs to flatter,more networked, and flexible designs (Capelli, Bassi, Katz, Knoke, Osterman, & Unseem, 1997; Cascio,1995; Cooper, Dewe & O’Driscoll, 2001). Consequently, jobs are no longer static but are in a constantstate of flux, and require employees to adapt and perform core functions in the midst of changing goals,expectations, and resources. In order to maintain competitiveness, many contemporary organizationsrequire their employees to assume greater and more widely prescribed work responsibilities.Indeed, role demands are endemic in today’s fast-paced organizational environments (Cryer,McCraty & Childre, 2003). For instance, in one study, 70% of the 150 human resource executivessurveyed indicated that employees in their firms were overburdened with work (HR Agenda, 2002). Inanother survey, substantial proportions of respondents reported that their job demands had prevented them65

from exercising regularly (72.2%), caused them to put on weight (69.2%), caused them to become ill(58.7%), harmed their marriage or a significant relationship (48.8%), contributed to a long-term healthcondition (36.6%), caused them to smoke or drink more alcohol (32.9%), and prevented them fromfinding a significant other (18.4%) (Cummings, 2001, p .46). Other studies report that job stress hascontinued to grow, costing organizations billions of dollars in employee disability claims, employeeabsenteeism, and lost productivity (Spector, Chen & O’Connell, 2000; Xie & Schaubroeck, 2001). Suchhigh levels of chronic stress could lead to employee burnout. Burnout is an extreme form of job stress andthere is evidence that employees are experiencing increasing levels of burnout (Kahn & Langlieb, 2003)BurnoutBurnout refers to the draining of mental resources caused by chronic job stress and is considereda work-related indicator of psychological health (Schaufeli & Enzmann, 1998). The negativeconsequences of burnout are well documented. For instance, burnout decreases organizationalcommitment (e.g., Lee & Ashforth, 1996; Leiter & Maslach, 1988), job performance (Tubre & Collins,2000; Wright & Bonett, 1997; Wright & Cropanzano, 1998), contextual performance (Cropanzano, Rupp& Byrne, 2003) and increases turnover intentions (Cropanzano, Rupp & Byrne, 2003; Wright &Cropanzano, 1998). Because job stress is of major significance to both employees and organizations, it isimportant to continue searching for mechanisms that can reduce the negative effects of job stress(Perrewé , Zellars, Ferris, Rossi, Kacmar & Ralston, 2004; Perrewé, Zellars, Rossi, Ferris, Kacmar, Liu,Zinko & Hochwarter, 2005).Burnout is most commonly conceptualized as a reaction to chronic role stress which takes theform of emotional exhaustion, depersonalization of others, and lack of felt accomplishment in workingwith others (Maslach, 1982). Emotional exhaustion (EE) involves feelings of being depleted of energyand drained of sensation. Depersonalization (DP) is the tendency to de-inviduate and de-humanize othersthrough cynical, callous and uncaring attitudes and behaviors. Reduced personal accomplishment (RPA)involves repeated efforts that fail to produce results, leading to a feeling of inefficacy and reducedmotivation. Maslach’s tripartite model of burnout has received strong support (Boles, Dean, Ricks, Short& Wang, 2000; Lee & Ashforth, 1996) and the most widely used measure of the burnout constructcontinues to be the Maslach Burnout Inventory (Boles et al, 2000; Maslach & Jackson, 1981, 1986).Role Conflict and BurnoutWhile many factors can lead to burnout, a large body of research has identified role demands asan antecedent of burnout (Cooper, Dewe & O’Driscoll, 2001; Lee & Ashforth, 1996; Peeters,Montgomery, Bakker & Schaufeli, 2005). Role conflict, a type of role demand, occurs when two or moresets of role pressures exist in an individual’s workspace, and the compliance with one of these pressuresimpedes the accomplishment of another (Kahn, Wolfe, Quinn, Snoek & Rosenthal, 1964). Role conflictintroduces uncertainty because the employee is uncertain whether he or she is successfully balancing allof his or her role requirements.Experiencing incompatible or irreconcilable expectations associated with multiple roles, or with asingle role, is presumed to be psychologically uncomfortable for individuals and to generate negativeemotional reactions (Schaubroeck, Cotton & Jennings, 1989). For instance, role conflict has been linkedto a number of dysfunctional outcomes, including job dissatisfaction and psychological strain (e.g., Rizzo,House & Lirtzman, 1970; Schaubroeck et al., 1989). Indeed, in their meta-analysis, Lee and Ashforth(1996) reported a corrected correlation of .53 between role conflict and emotional exhaustion, .37between role conflict and depersonalization, and .21 between role conflict and reduced personalaccomplishment.66

Although the relationship between perceived role conflict and burnout dimensions is wellestablished, we offer it as an hypothesis for two reasons. First, while we don’t expect sample-specificdifferences in the support of this relationship, it is worth noting that the preponderance of research onburnout has been conducted with participants in the human service occupations, such as nurses, teachers,policemen, and social service workers (cf. Boles et al., 2000, p. 13). Second, and more importantly, thishypothesis serves as a foundation for our subsequent hypotheses.Hypothesis 1: Perceived role conflict will be positively related to emotional exhaustion(Hypothesis 1a), depersonalization (1b), and reduced personal accomplishment (1c).Theoretical Frameworks for Studying Psychological StrainTransactional Theory of StressAccording to the transactional theory of stress (Lazarus, 1991, 1999; Lazarus & Folkman, 1984),stress is the result of a psychologically mediated process, such that the impact of any given stressordepends on the way that it is construed by the person who is exposed to it. More specifically, stress isseen as the outcome of a two-phase appraisal process. Primary appraisal involves the perceiver assessingthe degree to which a particular stressor poses a threat to self. In the event that any stimulus event iscategorized as a threat, secondary appraisal involves the perceiver’s assessment of available resources tocope with the threat. In this way, stress only develops into a long-term problem when a threat to self isperceived to exist in conjunction with insufficient coping resources.Job Demands-Control Model of StressKarasek (1979) proposed the job demands-control model and suggested that psychological strainis the result of joint effects of the demands of a work situation and the range of decision-making freedomor control available to an employee. When the individual perceives a lack of control, demands generatestress, manifested as psychological and physiological strain (Fox, Dwyer & Ganster, 1993). Alternatively,perceived control to deal with high job demands attenuates the relationship between high jobs demandsand strain. Thus, the central implication of Karasek’s model is that perceived control positively influencesone’s ability to cope with job demands, which in turn, alleviates the negative effects of job demands onstress.Both the transactional theory of stress (Lazarus, 1991) and the job demands-control model(Karasek, 1979) suggest that demands that are perceived to exceed one’s ability to cope elicitpsychological distress and strain. On the other hand, the perception of adequate resources to cope withdemands mitigates the negative effects of demands on psychological strain.Clearly, resources for coping with stressors could be internal (i.e., individual characteristics) orexternal (i.e., situational factors) to the individual. In fact, in their extensive summary of research onworkplace stress, Kahn and Byosiere (1992, p. 572) concluded that “organizational theory and researchhave been too little concerned with organizational and interpersonal factors that might serve asmoderators, buffers, or even antidotes to stresses and their effects,” and that empirical evidence in thisarea is essentially nonexistent. We assert that an individual’s political skill and perceptions of67

organizational support have the potential to mitigate the effects of role demands, such as role conflict, onburnout.Political SkillIn his original proposal, Karasek (1979) suggested that individual differences might play a role inthe function of job control. Since then, several other researchers have also called for an exploration of therole of individuals’ perceived control in the stress-strain relationship (Ganster, 1989; Smith, Tisak, Hahn& Schmieder, 1997). Researchers have empirically tested this possibility, examining a variety ofindividual difference characteristics including locus of control (Rahim & Psenicka, 1996), proactivepersonality (Parker & Sprigg, 1999), self-efficacy (Schaubroeck, Jones & Xie, 2001), Type A (Froggatt &Cotton, 1987), and political skill (Perrewé et al, 2004, 2005).Political skill is the ability to effectively understand others at work, and to use such knowledge toinfluence others to act in ways that enhance one’s personal and/or organizational objectives (Perrewé etal, 2004). It is characterized by social perceptiveness and the ability to adjust one’s behavior to differentand changing situational needs (Ferris, Berkson, Kaplan, Gilmore, Buckley, Hochwarter, & Witt, 1999;Ferris, Treadway, Kolodinsky, Hochwarter, Kacmar, Douglas, & Frink, 2005).Politically skilled individuals view interpersonal interactions as opportunities versus threats,facilitating the establishment of friendships, connections, and alliances, which in turn ensures a favorableidentity in their network (Perrewé, et al., 2000). A positive identity in one’s social network generatessignificant and tangible benefits (Baron & Markman, 2000). For example, politically skilled individualsmay enjoy network centrality, enabling them to gain access to important information. Moreover, theconfidence and trust placed in politically skilled individuals likely increase cooperation from others in thenetwork. Therefore, in addition to contributing to a greater sense of perceived control, political skill, andthe accompanying interpersonal influence, may enable individuals to gain actual control over events.Indeed, the construct of political skill is particularly relevant as Fox et al. (1993, p.291) suggested that theconstruct of most importance in the JD-C model is “an individual’s personal belief in his or her controlover a work situation.”Ferris et al (1999) argued that politically skilled individuals enjoy a sense of personal security andself-confidence from prior experience in, and mastery over, their work environments and the individualswith whom they interact. Such personal security and self-confidence could contribute to politically skilledindividuals experiencing lower levels of burnout. Results of two studies are suggestive of such arelationship. In one study conducted with 230 full-time employees, Perrewé et al (2004) reported thatpolitical skill was negatively related to psychological anxiety and to somatic complaints. In another study,politically skilled individuals reported lower levels of job tension and general anxiety (Perrewé et al.,2005). Results of studies by Perrewé and her colleagues indicate that political skill is related to lowerlevels of anxiety and somatic complaints. Burnout is a form of psychological strain that is more severethan psychological anxiety and somatic complaints, and this study is the first to examine if political skillis related to burnout.Hypothesis 2: Political skill will be negatively related to emotional exhaustion (2a),depersonalization (2b), and reduced personal accomplishment (2c).As previously noted, politically skilled individuals have a distinctive interpersonal style thataffords them a greater sense of personal control and interpersonal influence (Ferris et al., 1999). Througheffective regulation of interpersonal interactions, political skill might act as a coping resource, therebyattenuating the negative effects of job stress. According to Perrewé, Ferris, Frink and Anthony (2000), the68

negative effects arising from a stressor should be reduced for individuals high in political skill because oftheir increased confidence and sense of control.In the study mentioned earlier, Perrewé et al (2004) also found that political skill moderated theinfluence of role conflict on psychological anxiety, somatic complaints and physiological strain, such thatthe relationship between role conflict and the negative effects weakened with increasing levels of politicalskill. In a second study, Perrewé et al. (2005) reported that political skill moderated the relationshipbetween perceived role overload and three measures of strain, such that the relationship between roleoverload and measures of strain (job tension, job dissatisfaction and general anxiety) weakened atincreasing levels of political skill. These results suggest the possibility that political skill has the potentialto moderate the effects of role stressors on burnout. This expectation is consistent with predictions of theJD-C model (Karasek, 1979) and the transactional theory of stress (Lazarus, 1991).Hypothesis 3: Political skill will moderate the relationship between perceived roleconflict and burnout, such that less politically skilled individuals will experience moreemotional exhaustion (3a), depersonalization (3b), and reduced personal accomplishment(3c) than politically skilled individuals.Perceived Organizational SupportConsiderable research has focused on social support as a means of reducing the harmful effects ofstressors (see Pearlin, 1993). In general, when people feel that they have social support from others, theyreport less psychological distress or strain (e.g., Carver, Scheier & Weintraub, 1989). In theorganizational stress literature, several researchers (e.g., Carlson & Perrewé. 1999; La Rocco, House &French, 1980; Parsuraman, Greenhaus & Granrose, 1992) have argued that social support reduces thenegative effects of role stressors by helping employees cope with stress. However, with few exceptions(e.g., Stamper & Johlke, 2003), most studies focused on coworker or supervisor support (e.g., Ganster,Fusilier & Mayes, 1996; La Rocco et al., 1980), not organizational support.We assert that perceptions of organizational support (POS) is an important resource that couldbolster employees’ confidence in their ability to cope with role demands. Perceived organizationalsupport (POS) is defined as the extent to which employees perceive that their contributions are valued bytheir organization and that the organization cares about their well-being (Eisenberger, Fasolo & Davis-LaMastro, 1990; Eisenberger, Huntington, Hutchison & Sowa, 1986; Rhoades & Eisneberger, 1992).POS is thought to develop over time through multiple interactions between employees and theiremployers and to reflect the degree to which employees perceive that their work organizations arecommitted to them (Rhoades & Eisenberger, 1992: Stamper & Johlke, 2003).Theory and research suggest that POS could directly reduce role stress as well as buffer thenegative effects of role stress. For instance, high levels of POS may be negatively associated with rolestress as organizations that care about their employees’ well-being are more likely to reduce unnecessarywork complications and distractions for their workers, such as conflicting job requirements. In addition,such organizations are likely to specify and clarify job expectations and norms for their employees, inorder to better prepare them for work assignments. Indeed, in a study conducted with sales personnel,Stamper and Johlke (2003) reported POS to be negatively related to role stressors, role conflict and roleambiguity.Employees who perceive that their organization cares about them and is genuinely interested intheir well-being are likely to experience lower levels of burnout. At least two studies have examined therelationship between POS and burnout. In a study conducted with part-time employees, Cropanzano,69

Howes, Grandey and Toth (1997, Study 2) reported POS to be negatively related to an overall measure ofburnout. In a recent study, Armstrong-Stassen (2004) reported that POS was negatively related to burnoutin two different samples, but only the emotional exhaustion subscale of burnout was used in her study.This study is the first to relate POS to all three dimensions of burnout.Hypothesis 4: Perceived organizational support will be negatively related to emotionalexhaustion (4a), depersonalization (4b), and reduced personal accomplishment (4c).Support could also moderate the stressor-strain relationship by acting as a buffer that reduces theeffects of the stressors on the outcomes or strain (Viswesvaran, Sanchez & Fisher, 1999). A review of theliterature uncovered three studies in which the potential of POS to moderate the stress-strain relationshipwas examined. In one study, Leather, Lawrence, Beale, Cox and Dickson (1998) reported that POSmoderates the negative effect of workplace violence on both job satisfaction and organizationalcommitment. Casper, Martin, Buffardi and Erdwins (2002) found that POS moderated the relationshipbetween work-family conflict and organizational commitment. Stamper and Johlke (2003) found thatPOS moderated the relationship between role ambiguity and job satisfaction, and role conflict andturnover intentions. Thus, previous research has not examined the potential moderating role of POS in therelationship between role conflict and burnout.Perceptions of organizational support should bolster one’s confidence to cope with stressors, andconsistent with job demands-controls model (Karasek, 1979) and Lazarus’ (1991) transaction theory ofstress, enhanced confidence to cope with stressors should result in reduced levels of burnout. To ourknowledge, this study is the first to examine the moderating role of POS on the relationship betweenperceived role conflict and burnout dimensions.Hypothesis 5: Perceived organizational support will moderate the relationship between perceived roleconflict and burnout, such that individuals who perceive lower levels of support will experience moreemotional exhaustion (5a), depersonalization (5b), and reduced personal accomplishment (5c) thanindividuals who perceive higher levels of support.Sample and ProcedureMethodParticipants were employed in a software development organization located in the West Coast ofthe United States. Surveys were distributed to 171 employees. The study was part of a comprehensiveemployee survey. A letter from the researcher accompanied each survey. The letter informed employeesthat only the researcher would have access to their responses and that only summary statistics would beshared with the organization. Employees were assured that no one in the organization would have accessto any individual's responses and were requested to complete the survey and return it directly to theresearcher. One hundred and twenty employees completed the survey (response rate 70%). Theparticipants were professionals and were employed as programmers, software and hardware engineers,marketing specialists, and managers. Average age of participants was 37.92 years (SD = 7.93). Onaverage, participants had 16.07 years of work experience (SD = 8.12), and 3.67 years of job tenure (SD =1.84).MeasuresPerceived role conflict. Perceived role conflict was measured with Rizzo, House and Lirtzman’s(1970) 8-item scale. This scale has been widely used to measure role conflict (see Lee & Ashforth, 1996).Sample items include “I receive incompatible requests from 2 or more people,” and “I work under70

incompatible policies and guidelines,” Participants used a 7-point Likert scale with anchors 1- Stronglydisagree to 7- Strongly agree to respond to the items.Perceived organizational support. The 17-item scale developed by Eisenberger et al. (1986)was used to measure POS. Sample items include “My organization values my contribution to it’ssuccess,” and “My organization really cares about my well-being.” Participants used a 7-point scale withanchors, 1- Strongly disagree to 7 - Strongly agree to respond to the items.Political Skill. Political skill was measured with the 18-item Political Skill Inventory (PSI)developed by Ferris, Treadway, Kolodinsky, Hotchwarter, Kacmar, Douglas and Frink (2005). Ferris andhis colleagues have reported a substantial body of construct validity evidence for the political skillconstruct (e.g., Ferris et al., 1999; Perrewe et al., 2004; Perrewe et al., 2005) and the PSI measure (e.g.,Ferris et al., 2005). “I am good at getting people to like me” and “It is easy for me to develop rapport withmost people” are sample items. A seven-point response format was used (1 - Strongly disagree to 7 -Strongly agree).Burnout. The original Maslach Burnout Inventory (MBI) (Maslach & Jackson, 1981, 1986) has22 items. Using structural equation modeling techniques, Boles, Dean, Ricks, Short and Wand (2000)revised the MBI and reduced it to 19 items (items 2, 12, and 16 from MBI were dropped). The 19-itemrevised MBI (Boles et al., 2000) was used to measure burnout. A sample item used to measure emotionalexhaustion is “I feel emotionally drained from my work,” A sample item used to measuredepersonalization is “I feel I treat others (employees) as if they were impersonal objects.” A sample itemused to measure reduced personal accomplishment is “I feel that I am positively influencing otherpeoples’ lives through my work (reverse scored).” Participants used a 7-point scale with anchors, 1-Almost never to 7 - Almost always to respond to the items.ResultsMeans, standard deviations and correlations between study variables are reported in Table 1. Thepattern of correlations is consistent with hypothesized relationships. To test our hypotheses, we conductedmoderated hierarchical regression analyses (Cohen & Cohen, 1983). In the first step, age, sex and level inthe organization were entered as control variables. In the second step, the main effects, perceived roleconflict, political skill, and perceived organizational support were entered. The centering technique wasused to properly test for the interaction terms. Centering a variable involves simply replacing it by itsdeviation from the mean. The variables, perceived role conflict, political skill and perceivedorganizational support were centered; and the product of centered role conflict and centered political skill,and the product of centered role conflict and centered organizational support were used to test forinteractions (Cohen & Cohen, 1983, p. 238). These interaction terms were entered in Step 3 of eachhierarchical regression equation.Table 1Means, standard deviations and correlations between study variablesM SD 1 2 3 4 5 6 71. Perceived role conflicit 3.67 1.14 (.86)2. Political skill 5.32 0.89 -.16+ (.93)3. Perceived organizational 4.51 1.18 - .14 (.94)71

support .41**4. Emotional exhaustion 3.16 1.26 .50** -.17+ - (.87).43**5. Depersonalization 1.92 0.99 .17+ - - .48** (.86).26** .36**6. Reduced personal2.76 0.94 .09 - -.08 .13 .24* (.80)accomplishment7. Overall burnout 2.68 0.78 .41** -.46**.65**-.43**.82** .73** .60** (.78)Note: Scale reliabilities are noted on the diagonal. ** p < .01, * p < .05, + p < .10In the first hierarchical regression (see Table 2), the control variables entered in Step 1 failed toexplain any variance in emotional exhaustion (R 2 = .02, F 3,115 = 0.68, p = .57). Perceived role conflict,political skill and perceived organizational support entered in Step 2 collectively explained 29% ofadditional variance in emotional exhaustion (adjusted R 2 = .29, F 3,112 = 16.83, p = .001). Of thesevariables, only perceived role conflict (β = .39, t = 4.45, p < .001; sR 2 = .12) and perceived organizationalsupport (β = -.27, t = -3.13, p < .01; sR 2 = .06) explained unique variance in emotional exhaustion. sR 2 ,the squared semi-partial correlation, was used to ascertain the unique contribution of each variable to thecriterion. sR 2 indicates the incremental change in R 2 for a given variable beyond all other variables.72

Table 2Results of hierarchical regression for burnout dimensionsEE DP RPA BurnoutVariables β sR 2 β sR 2 β sR 2 β sR 2Age -.04 .07 -.03 -.01Sex .06 -.01 -.02 .02Organizational Level -.07 -.05 .15 .01Perceived role conflict (PRC) .39*** .12 .02 .01 .25*** .05Political skill (PS) -.09 -.21* .04 -.58*** .32 -.38*** .14Perceived organizational support(POS)-.27** .06 -.33** .09 .03 -.27*** .06PRC X PS .07 .05 .13* .02 .03PRC X POS .13* .02 -.04 .03 .07Step 1Adjusted R 2 .02 .01 .03 .01Step 2Adjusted R 2 .29*** .18*** .33*** .35***∆R 2 .29*** .18*** .33*** .35***Step 3Adjusted R 2 .31*** .18*** .35*** .35***∆R 2 .02 .01 .02* .01Note. Beta weights are controlled for all other variables. Control variables, age, sex and organizational level enteredin Step 1, PRC, PS, POS in Step 2, and interaction terms in Step 3. EE – emotional exhaustion, DP –depersonalization, and RPA – reduced personal accomplishment. * p < .05, ** p < .01, *** p < .001The interaction terms entered in Step 3 explained 2% of additional variance in emotionalexhaustion (∆R 2 = .02, F 2, 110 = 2.32, p < .05) but only the cross-product of perceived role conflict andperceived organizational support was significant (β = .13, t = 1.93, p < .05; sR 2 = .02). Because theinteraction was significant, we performed follow-up split-group analyses by taking a median-split onperceived organizational support and then regressing emotional exhaustion on perceived role conflict atlow and high levels of perceived organizational support (Aiken & West, 1991). Role conflict wassignificantly related to emotional exhaustion at low levels of perceived organizational support (R 2 = .02, p< .05, β = .21, t = 1.93, p < .05) but not at high levels of perceived organizational support (R 2 = .00, p=.51). These results indicate support for Hypotheses 1a, 4a and 5a but not for Hypotheses 2a and 3a.In the second hierarchical regression (see Table 2), the control variables entered in Step 1 failedto explain any variance in depersonalization (R 2 = .01, F 3,115 = 0.15, p = .93). Perceived role conflict,political skill and perceived organizational support entered in Step 2 collectively explained 18% ofadditional variance in depersonalization (adjusted R 2 = .18, F 3,112 = 8.23, p < .001). Of these variables,only perceived organizational support (β = -.33, t = - 3.50, p < .001; sR 2 = .09) and political skill (β = -73

.21, t = -2.46, p < .05; sR 2 = .044) explained unique variance in depersonalization. The interaction termsentered in Step 3 failed to explain any additional variance in depersonalization (∆R 2 = .01, F 2, 110 = 0.39,p = .82). Thus, only Hypotheses 2b and 4b were supported.The control variables entered in Step 1 of another hierarchical regression (see Table 2) failed toexplain any variance in reduced personal accomplishment (R 2 = .03, F 3,115 = 1.33, p = .27). Perceived roleconflict, political skill and perceived organizational support entered in Step 2 collectively explained 33%of additional variance in reduced personal accomplishment (adjusted R 2 = .33, F 3,112 = 19.43, p < .001).Of these variables, only political skill (β = -.58, t = -7.59, p < .001; sR 2 = .32) explained unique variancein reduced personal accomplishment. The interaction terms entered in Step 3 explained an additional 2%of the variance in reduced personal accomplishment (∆R 2 = .02, F 2, 110 = 1.97, p < .05) and the crossproductof perceived role conflict and political skill was significant (β = .13, t = 1.93, p < .05; sR 2 = .02).Because the interaction was significant, we performed split-group analyses by taking a medianspliton political skill, and regressing reduced personal accomplishment on perceived role conflict at lowand high levels of political skill (Aiken & West, 1991). Role conflict was significantly related to reducedpersonal accomplishment at low level of political skill (R 2 = .07, p < .001, β = .26, t = 1.98, p < .05) butnot at high level of political skill (R 2 = .00, p =.90). These results provide strong support for Hypotheses2c and 3c.The items used to measure the three dimensions of burnout were averaged to create an overallindex of burnout. A final hierarchical regression was conducted with this measure of overall burnout. Thecontrol variables failed to explain any variance in burnout (R 2 = .01, F 3,115 = 0.41, p = .74). Perceived roleconflict, political skill and perceived organizational support entered in Step 2 collectively explained 35%of additional variance in burnout (adjusted R 2 = .35, F 3,112 = 22.06, p < .001). Political skill (β = -.38, t = -4.95, p < .001; sR 2 = .14) explained the highest amount of unique variance in burnout followed byperceived organizational support (β = -.27, t = -3.29, p < .001; sR 2 = .06) and perceived role conflict (β =.25, t = 2.92, p < .001; sR 2 = .05). The interaction terms entered in Step 3 failed to explain any additionalvariance in burnout (∆R 2 = .01, F 2, 110 = 0.59, p = .56).DiscussionThe high levels of job stress experienced by employees cost organizations billions of dollars inemployee disability claims, employee absenteeism, and lost productivity (Spector, Chen & O’Connell,2000; Xie & Schaubroeck, 2001). Drawing on previous research on the effect of role conflict on burnout(e.g., Lee & Ashforth, 1986) and models of the stress-strain relationship (e.g., Karasek, 1979; Lazarus,1991), we proposed that individual differences in political skill and perceptions of organizational supportwill be negatively related to burnout and will also moderate the relationship between perceived roleconflict and burnout.As expected, perceptions of role conflict were positively related to burnout, and political skill andperceptions of organizational support were negatively related to overall perceptions of burnout. However,these variables exhibited differential relationships with the different dimensions of burnout. For instance,perceived role conflict and perceived organizational support interacted to influence emotional exhaustion,such that the relationship between role conflict and emotional exhaustion was weaker at higher levels ofperceived organizational support. That is, high levels of support had a buffering effect and mitigated thenegative effects of role conflict on emotional exhaustion.Both perceived organizational support and political skill were negatively related withdepersonalization. However, only political skill was negatively related to reduced personal74

accomplishment. In addition, it also interacted with role conflict, such that high levels of political skillmitigated the negative effects of role conflict on reduced personal accomplishment. These results indicatethat perceived organizational support influences emotional exhaustion but not reduced personalaccomplishment whereas political skill exhibits the opposite pattern.Theoretical ImplicationsThis study responds to the call for more research on “organizational and interpersonal factors thatmight serve as moderators, buffers, or even antidotes to stresses and their effects,” by Kahn and Byosiere(1992, p. 572) who lamented that “empirical evidence in this area is essentially nonexistent.” Others haveexpressed similar views (Stamper & Johlke, 2003). For instance, this study is the first to examine theeffect of an individual difference variable and an organizational variable on burnout dimensions in asingle study. It is the first study to relate political skill to burnout, and consequently also the first toinvestigate the moderating effect of political skill on the role conflict – burnout relationship. We are alsounaware of any study that has examined the moderating effects of POS on the role conflict – burnoutrelationship. Consequently, results of our study contribute not only to the literature on burnout but also tothe large body of research on POS and the emerging stream of research on political skill.The proponents of the burnout construct (Leiter & Maslach, 1988; Maslach, 1982; Maslach &Jackson, 1981) noted that emotional exhaustion (EE) should appear first as excessive demands drain anindividual’s emotional resources. As a coping strategy, individuals limit their involvement with othersand distance themselves psychologically (DP). Finally, individuals recognize the discrepancy betweentheir original optimistic attitude and their current attitude, and consequently, experience a feeling ofinadequacy in relating to people and performing their job (RPA). This implies that as emotionalexhaustion increases, DP and RPA are likely to increase. In this sample, role conflict was associatedemotional exhaustion and overall burnout, but not with either depersonalization or reduced personalaccomplishment. This may be due to the relatively low levels of perceived role conflict (3.67 on a 7-pointscale) and emotional exhaustion (3.16) and the relatively high level of perceived organizational support(4.51) in this organization.Results of our study have several theoretical implications. For instance, our results indicate thatperceptions of organizational support but not political skill could reduce feelings of emotional exhaustion.Although support is measured subjectively the source remains the organization, an external entity, andconsequently these results are consistent with Lazarus’ (1991) theory of stress with its emphasis onexternal support as a coping mechanism. In contrast, results indicate that political skill but not perceivedorganizational support could alleviate feelings of reduced personal accomplishment. Perceived control isthe key mechanism in Karasek’s (1979) JD-C model and political skill is a form of interpersonal controland influence. Thus, results for reduced personal accomplishment offer support to the key tenet of the JD-C model of stress.The results for political skill reinforce Ferris et al.’s (1999, 2005) conceptualization of politicalskill as an interpersonal construct. Politically skilled individuals are masters of social interactions whoenjoy a sense of personal security and self-confidence from prior experience in, and mastery over, theirwork environments and the individuals with whom they interact. These findings are consistent withPerrewé et al’s (2000) arguments and results of Perrewé et al. (2004, 2005) that higher levels of politicalskill reduce strain caused by role conflict and overload. Results for POS suggest that while perceptions ofsupport help cope with emotional exhaustion, such perceptions do nothing to prevent feelings ofinefficacy (reduced personal accomplishment), suggesting a potential limit for the benefits of support.75

Our results have one methodological implication. To our knowledge, two studies have reported arelationship between POS and burnout. Armstrong-Stassen (2004) reported that POS was negativelyrelated to burnout, but only the emotional exhaustion subscale of burnout was used in her study. In thisstudy, we related POS to the three dimensions of burnout. Cropanzano et al. (1997, Study 2) reportedPOS to be negatively related to an overall measure of burnout. Therefore, for comparison purposes, weexamined POS and overall burnout and found results similar to those of Cropanzano et al. However, atthe dimension level, results for overall burnout are misleading as POS related to lower emotionalexhaustion and less depersonalization but not to reduced personal accomplishment. Previous researchershave cautioned against collapsing the different dimensions of burnout to create an overall measure ofburnout as the different dimensions relate differently to other variables (see Boles et al., 2000). Thedifferent pattern of relationships exhibited by the different dimensions of burnout in our data reinforcesthe argument against collapsing burnout dimensions.Potential Limitations and Directions for Future ResearchOne potential limitation is our sample size. It is difficult to detect small effects in a sample of 120participants. Even though we found support for two two-way interactions, it is difficult to conclusivelyknow if lack of support for the other hypothesized interactions is because of a true lack of interactiverelationships or because of insufficient statistical power to detect small effects. This concern highlightsthe need to replicate results of this study. Second, the cross-sectional nature of our data preventsinferences of causality. It is quite possible employees experiencing burnout attributed it to role conflictand lack of organizational support. On the other hand, it is difficult to suggest that burnout couldencourage participants to report lower levels of political skill. While a longitudinal study would allow formore confidence in inferences of causality, difficult issues, such as the appropriate time interval betweenmeasurement of independent variables and burnout would need to be resolved.Results of our study suggest several exciting opportunities for research. For instance, in thisstudy, we examined the moderating roles of political skill and perceived organizational support on therelationship between role conflict and burnout. In addition to replicating results of our study, futureresearch should also investigate moderating effects of these two variables on the relationship betweenother role demands, such as role ambiguity and role overload, and burnout. Second, in this study, we onlyinvestigated the mitigating effects of perceived support from the organization itself. Future research couldexamine the relative importance of support from the organization, supervisor and one’s coworkers onburnout. It is quite possible that support from different sources might differentially moderate therelationship between different role demands and burnout dimensions.Third, more research is needed on the consequences of burnout. Examining how the differentdimensions of burnout influence outcomes, such as job satisfaction, job involvement, task performance,contextual performance, and dysfunctional behaviors, such as workplace aggression is likely to result intheoretical advancement of the burnout construct. For instance, one could expect emotional exhaustion tolead to reduced job involvement and organizational commitment whereas depersonalization is likely toresult in incivility and potentially escalate into workplace aggression.Practical ImplicationsResults of our study, taken together with previous research (e.g., Lee & Ashforth, 1996) indicatethat role conflict results in burnout. Organizations should reduce role conflict to the extent possible byclearly defining and communicating role-related expectations. Second, enhancing organizational support76

is another way to reduce feelings of emotional exhaustion and prevent depersonalization. Rhoades andEisenberger (2002) have offered a number of suggestions on ways to enhance organizational support.Third, enhancing political skill might be another avenue to reduce depersonalization and avoid feelings ofreduced personal accomplishment. To enhance political skill, proponents of the political skill construct(Ferris et al., 1999) suggest use of process-focused techniques, such as drama-based training,developmental simulations, and behavior modeling (Ferris, Anthony, Kolodinsky, Gilmore & Harvey,2002). In summary, results suggest that burnout can be reduced by decreasing role conflict, increasingpolitical skill, and enhancing organizational support of employees.ConclusionJob stress including burnout is on the rise (Cryer et al., 2003; Kahn & Langlieb, 2003). Becausejob stress is of major significance to both employees and organizations, it is important to continuesearching for mechanisms that can reduce the negative effects of job stress. Drawing on theoreticalframeworks of the stress-strain relationship (Karasek, 1979; Lazarus, 1991), we hypothesized and foundrole conflict to be positively related to, and perceptions of organizational support and political skill to benegatively related to burnout dimensions. The latter variables also moderated the relationship betweenrole conflict and different dimensions of burnout. In addition to discussing theoretical implications andfuture research directions, we suggest mechanisms for reducing employee burnout.77

ReferencesAiken, L.S., & West, S.G., Multiple Regression: Testing and Interpreting Interactions, Newbury Park,CA: Sage, 1991.Armstrong-Stassen, M., “The Influence of Prior Commitment on the Reactions of Layoff Survivors toOrganizational Downsizing.” Journal of Occupational Health Psychology, 9, (2004), 46-60.Baron, R.A., & Markman, G.D., “Beyond Social Capital: How Social Skills Can Enhance Entrepreneurs’Success.” Academy of Management Executive, 14, (2001), 106-116.Boles, J.S., Dean, D.H., Ricks, J.M., Short, J.C., & Wang, G., “The Dimensionality of the MaslachBurnout Inventory Across Small Business Owners and Educators.” Journal of VocationalBehavior, 56, (2001), 12-34.Capelli, P., Bassi, L., Katz, H., Knoke, D., Osterman, P., & Unseem, M., Change at Work. New York:Oxford University Press, 1997.Carlson, D.S., & Perrewé, P.L., The Role of Social Support in the Stressor-Strain Relationship: AnExamination of Work-Family Conflict. Journal of Management, 25, (1999), 513-540.Cascio, W.F., “Whither Industrial and Organizational Psychology in a Changing World of Work?”American Psychologist, 50, (1995), 928-939.Casper, W.J., Martin, J.A., Buffardi, L.C., & Erdwins, C.J., “Work-Family Conflict, PerceivedOrganizational Support, and Organizational Commitment of Employed Mothers,” Journal ofOccupational Health Psychology, 7, (2002), 99-108.Cohen, J., & Cohen, P., Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences,Hillsdale, NJ: Lawrence Erlbaum Associates, 1983.Cooper, C.L., Dewe, P.J., & O’Driscoll, M.P., Organizational Stress: A Review and Critique of Theory,Research and Applications, London: Sage, 2001.Cordes, C.L., & Doughtery, T.W., “A Review and an Integration of Research on Job Burnout,” Academyof Management Review, 18, (1993), 621-656.Cropanzano, R., Howes, J.C., Grandey, A.A., & Toth, P., “The Relationship of Organizational Politicsand Support to Work Behaviors, Attitudes and Stress,” Journal of Organizational Behavior, 18,(1997), 159-180.Cropanzano, R., Rupp, D.E., & Byrne, Z.S., “The Relationship of Emotional Exhaustion to WorkAttitudes, Job Performance and Organizational Citizenship Behaviors,” Journal of AppliedPsychology, 88, (2003), 160-169.Cryer, B., McCraty, R., & Childre, D., “Pull the Plug on Stress,” Harvard Business Review, 81, (2003),102-107.Eisenberger, R., Fasolo, P., & Davis-LaMastro, V., “Perceived Organizational Support and EmployeeDiligence, Commitment and Innovations,” Journal of Applied Psychology, 75, (1990), 51-59.Eisenberger, R., Huntington, R., Hutchison, S., & Sowa, D., “Perceived Organizational Support,” Journalof Applied Psychology, 71, (1986), 500-507.Ferris, G.R., Anthony, W.P., Kolodinsky, R.W. Gilmore, D.C., & Harvey, M.D., Development ofpolitical skill. In C. Wankel & R. DeFillippi (Eds.), Research in Management Education andDevelopment: Rethinking Management Education for the 21 st Century (Vol. 1, pp. 3-25).Greenwich, CT: Information Age Publishing, 2002.Ferris, G.R., Berkson, H.M., Kaplan, D.M., Gilmore, D.C., Buckley, M.R., Hochwarter, W.A., & Witt,L.A., Development and Initial Validation of the Political Skill Inventory, Paper Presented at theAnnual Meetings of the Academy of Management, Chicago, 1999.Ferris, G.R., Treadway, D.C., Kolodinsky, R.W., Hochwarter, W.A., Kacmar, C.J., Douglas, C., & Frink,D.D., “Development and Validation of the Political Skill Inventory,” Journal of Management, 31,(2005), 126-152.78

Fox, M.L., Dwyer, D.J., & Ganster, D.C., “Effects of Stressful Job Demands and Control onPhysiological and Attitudinal Outcomes in a Hospital Setting,” Academy of Management Journal,36, (1993), 289-318.Froggatt, K.L., & Cotton, J.L., “The Impact of Type A Behavior Pattern on Role Overload-Induced Stressand Performance Attributions,” Journal of Management, 13, (1987), 87-98.Ganster, D.C., Worker Control and Well-Being: A Review of Research in the Workplace, In S.L. Sauter,J.J. Hurrell Jr., & C.I. Cooper (Eds.), Job Control and Worker Health (pp. 235-280). London:Wiley, 1989.Ganster, D.C., Fusilier, M., & Mayes, B.T., “Role of Social Support in the Experience of Stress at Work,”Journal of Applied Psychology, 71, (1986), 102-110.Kahn, J., & Langlieb, A.M., Mental Health and Productivity in the Workplace: A Handbook forOrganizations and Clinicians, Hoboken, NJ: Jossey-Bass, 2003.Kahn, R.L., & Byosiere, P., Stress in Organizations. In M.D. Dunnette & L.M. Hough (Eds.), Handbookof Industrial and Organizational Psychology (2 nd ed.). Palo Alto, CA: Consulting PsychologistsPress, 1992.Karasek, R.A., “Job Demands, Job Decision Latitude and Mental Strain: Implications for Job Redesign,”Administrative Science Quarterly, 24, (1979), 285-308.La Rocco, J.M., House, J.S., & French, J.R.P., Jr., “Social Support, Occupational Stress and Health,”Journal of Health and Social Behavior, 21, (1980), 202-216.Lazarus, R.S., “Progress on a Cognitive-Motivational-Relational Theory of Emotions,” AmericanPsychologist, 46, (1991), 819-834.Lazarus, R.S., & Folkman, S., Stress, Appraisal and Coping, New York: Springer Publishing Company,1984.Lazarus, R.S., Stress and Emotion: A New Synthesis, London: Free Association Books, 1999.Leather, P., Lawrence, C., Beale, D., Cox, T., & Dickson, R., “Exposure to Occupational Violence andthe Buffering Effects of Intra-organizational Support,” Work and Stress, 12, (1998), 161-178.Lee, R.T., & Ashforth, B.E., “A Meta-Analytic Examination of the Correlates of the Three Dimensions ofJob Burnout,” Journal of Applied Psychology, 81, (1996), 123-133.Leiter, M.P., & Maslach, C., “The Impact of Interpersonal Environment on Burnout and OrganizationalCommitment,” Journal of Organizational Behavior, 9, (1988), 297-308.Maslach, C., Burnout: The Cost of Caring, New York: Prentice-Hall, 1982.Maslach, C., & Jackson, S.E., “The Measurement of Experienced Burnout,” Journal of OccupationalBehavior, 2, (1981), 99-113.Maslach, C. & Jackson, S.E., Maslach Burnout Inventory Manual (2 nd ed.), Palo Alto, CA: ConsultingPsychologists Press, 1986.Parasuraman, S., Greenhaus, J.H., & Granrose, C.S., “Role Stressors, Social Support, and Well-BeingAmong Two-Career Couples,” Journal of Organizational Behavior, 13, (1992), 339-356.Parker, S.K., & Sprigg, C.A., “Minimizing Strain and Maximizing Learning: The Role of Job Demands,Job Control and Proactive Personality,” Journal of Applied Psychology, 84, (1999), 025-939.Pearlin, L., The Social Contexts of Stress, In L.Goldberger & S. Breznitz (Eds.), Handbook of Stress (2 nded., pp. 303-315), New York: The Free Press, 1993.Peeters, M.C.W., Montgomery, A.J., Bakker, A.B., & Schaufeli, W.B., “Balancing Work and Home:How Job and Home Demands are Related to Burnout,” International Journal of StressManagement, 12, (2005), 43-61.Perrewé, P. L., Ferris, G.R., Frink, D.D., & Anthony, W.P., “Political Skill: An Antidote for WorkplaceStressors,” Academy of Management Executive, 14, (2000), 115-123.Perrewé, P. L., Zellars, K. L., Ferris, G. R., Rossi, A. M., Kacmar, C. J., & Ralston, D. A., “NeutralizingJob Stressors: Political Skill as an Antidote to the Dysfunctional Consequences of Role ConflictStressors,” Academy of Management Journal, 47, (2004), 141-152.79

Perrewé, P. L., Zellars, K. L., Rossi, A. M., Ferris, G.R., Kacmar, C. J., Liu, Y., Zinko, R., &Hotchwarter, W.A., “Political skill: An Antidote in the Role Overload-Strain Relationship,”Journal of Occupational Health Psychology, 10, (2005). 239-250.Rahim, M.A., & Psenicka, C., “A Structural Equations Model of Stress, Locus of Control, Social Support,Psychiatric Symptoms, and Propensity to Leave a Job,” Journal of Social Psychology, 136,(1996), 69-84.Rhoades, L., & Eisenberger, R., “Perceived Organizational Support: A Review of the Literature,” Journalof Applied Psychology, 87, (2002), 698-714.Rizzo, J., House, R.J., & Lirtzman, S.I., “Role Conflict and Role Ambiguity in Complex Organizations,”Administrative Science Quarterly, 15, (1970), 150-163.Schaubroeck, J., Jones, J.R., & Xie, J.L., “Individual Differences in Utilizing Control to Cope with JobDemands: Effects on Susceptibility to Infectious Disease,” Journal of Applied Psychology, 86,(2001), 265-278.Schaufeli, W.B., & Enzmann, D., The burnout companion to study and practice: A critical analysis,London: Taylor & Francis, 1998.Stamper, C. L., & Johlke, M. C., “The Impact of Perceived Organizational Support on the RelationshipBetween Boundary Spanner Role Stress and Work Outcomes,” Journal of Management, 29,(2003), 569-588.Smith, C.S., Tisak, J., Hahn, S.E., & Schmieder, R.A., “The Measurement of Job Control,” Journal ofOrganizational Behavior, 18, (1997), 225-237.Spector, P.E., Chen, P.Y., & O’Connell, B.J., “A Longitudinal Study of Relations between Job Stressorsand Job Strains While Controlling for Prior Negative Affectivity and Strains,” Journal of AppliedPsychology, 85, (2000), 211-218.Tubre, T.C., & Collins, J.M., “Jackson and Schuler (1985) Revisited: A Meta-Analysis of theRelationships between Role Ambiguity, Role Conflict and Job Performance,” Journal ofManagement, 26, (2000), 155-169.Viswesvaran, C., Sanchez, J.I., & Fisher, J., “The Role of Social Support in the Process of Work Stress: AMeta-Analysis,” Journal of Vocational Behavior, 54, (1999), 314-334.Wright, T.A., & Bonett, D.G., “The Contribution of Burnout to Work Performance,” Journal ofOrganizational Behavior, 18, (1997), 491-499.Wright, T.A., & Cropanzano, R., “Emotional Exhaustion as a Predictor of Job Performance andVoluntary Turnover,” Journal of Applied Psychology, 83, (1998), 486-493.Xie, J.L., & Schaubroeck, J., Bridging Approaches and Findings Across Diverse Disciplines to ImproveJob Stress Research, In P.L. Perrewé & D.C. Ganster (Eds.), Research in Occupational Stress andWell Being (vol. 1, pp. 1-53). Oxford, England: Elsevier Science, 2001.80

ASAC 2006Banff, AlbertaPeter Lee (student)York UniversityMULTIPLE IDENTIFICATIONS IN ORGANIZATIONSOrganizational identification research has often focused on single social referents asidentification targets for individuals. However, modern organizations offer multipletargets for identification and hence, organizational identities. Few organizational studieshave opted to focus on multiple social identifications by organizational members. Weattempt to address this gap by positing a model of multiple identifications fororganizational members.In America, it is natural for a [person] to identify [her/himself] with the businesscorporation [s/he] serves.(Burke, 1937, as quoted in Cheney, 1983a)While organizational identification (OI) has enjoyed a recent resurgence in organizationalresearch, Burke’s opening quote attests to the rich history that organizational identification theorizing andresearch has enjoyed. Albert and Whetten’s (1985) seminal article re-sparked interest in organizationalidentity and identification and provided the foundation for further research where it has become animportant area of organization research. The first detailed model of OI was proposed by March andSimon (1958) in their model of motivation, and in the ensuing 20 years, only a handful of studiesinvolving OI were published (e.g. Brown, 1969; Gouldner, 1957, 1958; Lee, 1969; Rotondi, 1975). In the1980s, organizational researchers re-discovered OI with the publication of Stuart Albert and DavidWhetten’s (1985) seminal article, “Organizational Identity”. Meanwhile, other social sciences such aspsychology (Tajfel & Turner, 1979/1986), sociology (Stryker, 1980), and communications (Cheney,1983a, 1983b) began publishing research on identity and identification at around the same time. Finally,OI has also caught the attention of critical theorists who view OI as another management tool for theexploitation of workers (e.g. Alvesson & Willmott, 2002).Work on OI grew out of the work on organizational identity, which drew heavily on socialidentity theory (Tajfel & Turner, 1979/1986) and later, self-categorization theory (Turner, 1987). Tajfeland Turner (1979/1986) define social identity as “those aspects of an individual’s self-image that derivefrom the social categories to which he perceives himself as belonging” (p. 16). The notion of in-groupsand out-groups plays a key role in the individual’s search for identity. Individuals wish to boost their selfesteemand will seek those social referents to do so. Thus, individuals will categorize themselves alongdimensions that provide a sense of oneness with a social referent. Ashforth and Mael (1989) extended thisidea into the realm of OI by claiming that OI is a specialized form of social identification: organizationalmembers will assess their organization’s categorizations by engaging in a comparison with their own selfcategorizations.Self-categorization theory (SCT) grew out of Tajfel and Turner’s initial work and represents anevolution of social identity theory. Hogg and Terry (2000) provide a definition of social categorization:“[s]ocial categorization of self and others into ingroup and outgroup accentuates the perceived similarityof the target to the relevant ingroup or outgroup prototype (cognitive representation of features thatdescribe and prescribe attributes of the group)” (p. 123). Thus, SCT involves the perceived integration ofindividual identity into the ingroup (i.e. social referent) prototype; individuals will tend to see themselves81

on similar, salient characteristics as the ingroup, such as race and sex (Chattopadhyay, Tluchowska, &George, 2004).Hogg and Terry (2000) view the concept of prototypes to be critical in differentiating SCT fromsocial identity theory. Prototypes clearly demarcate ingroup from outgroup attributes; indeed, establishingan identity involves disidentifying with other groups (Elsbach, 1999; Elsbach & Bhattacharya, 2001). Thepurpose of prototypes then is to highlight the outgroup differences and ingroup similarities (Ashforth &Mael, 1989). Prototypes provide bases for individuals to compare themselves; prototypes exhibitacceptable, salient characteristics of ingroups and can be represented by model group members orconveyed through more abstract means such as attitudes and beliefs (Hogg & Terry, 2000).Another key distinction of SCT involves the context sensitivity of prototypes. While the ingroupidentity may not change significantly, the composition of the outgroup can (Hogg & Terry, 2000), henceleading to prototype redefinition. Thus, over time, identification, through the definition of prototypes,changes leading to a dynamic view of social identity. Therefore, building on social identity theory, SCTposits that individuals maintain or enhance their positive self-concept through the salience of outgroupprototypes. Interestingly, salience also figures prominently in identity theory as posited by symbolicinteractionists (e.g. Stryker, 1980). Salience of prototypes then provides the conduit to which thepossibility of multiple organizational identifications can occur, a point to which I will return later.As organizations move toward an increased use of short-term transactional employment contractsat the expense of long-term relational ones (Matusik & Hill, 1998) and individuals engage inboundaryless careers (Arthur & Rousseau, 1996), organizational identification may seem to be a lessrelevant concept for today’s mobile workforce. However, it may be exactly because of such changes thatorganizational identification still remains, or perhaps is increasingly so, “a fundamental task oforganizations” (Pratt, 1998: 171). Furthermore, “an organization must reside in the heads and hearts of itsmembers” (Albert, Ashforth, & Dutton, 2000: 13). Lastly, Cheney (1983b) writes that “fosteringidentification is the ‘intent’ of many corporate policies, for with it comes greater assurance thatemployees will decide with organizational interests uppermost in mind” (p. 158).Researchers have defined organizational identification in several ways. Cheney (1983a) proposedthat “[i]dentification – with organizations or anything else – is an active process by which individuals linkthemselves to elements in the social scene” (p. 342), while Ashforth and Mael (1989) defined it as the“perception of oneness or belongingness to some human aggregate” (p. 21). Moreover, Dutton, Dukerich,and Harquail (1994) described organizational identification as “the degree to which a member defineshim- or herself by the same attributes that he or she believes define the organization” (p. 239), Pratt(1998) depicted organizational identification occurring “when an individual’s beliefs about his or herorganization become self-referential or self-defining” (p. 172); and finally, Elsbach (1999) suggested thatorganizational identification is a “self-perception based on (1) a sense of active connection between one’sidentity and the identity of an organization, and (2) a positive relational categorization of oneself and theorganization” (p. 179). While these definitions have slight variations, they agree that organizationalidentification involves an integration of the identity of a social referent into an individual’s self-concept.Furthermore, recent work on organizational identity and organizational identification has recognized that,while organizational identity may comprise core, distinctive, and enduring elements (Albert & Whetten,1985), organizational identification may be a more dynamic process (Brown, 2001; Elsbach, 1999; Pratt,1998). In particular, identity addresses the question, “Who am I?” while identification deals with thequestion, “How do I come to know myself in relation to the organization?” (Pratt, 1998: 171).However, recent OI research has involved individual organizational members and their relatedorganizations (e.g. Dukerich, Golden, & Shortell, 2002; Dutton, Dukerich, & Harquail, 1994; Kreiner &Ashforth, 2004; Pratt, 2000). In today’s increasingly dynamic and complex business environment,82

multiple organizational identities (and hence, multiple OIs) seem appropriate for study. Yet, as Pratt(1998) contends, research into the area of multiple identifications remains relatively unexplored. Feworganizational studies have opted to focus on multiple organizational identities (see Foreman & Whetten,2002; Golden-Biddle & Rao, 1997; Pratt & Rafaeli, 1997 for notable exceptions), with even fewerfocusing on multiple social identifications by organizational members, that is, having more than onesocial referent with which to identify (see Barker & Tompkins, 1994; George & Chattopadhyay, 2005 fornotable exceptions). We attempt to address this gap by positing a model of multiple identifications fororganizational members.Literature ReviewOrganizational IdentificationPratt (1998) defines organizational identification as “occur[ring] when an individual’s beliefsabout his or her organization become self-referential or self-defining” (p. 172, emphasis in original). Thisdefinition expands Ashforth and Mael’s (1989) characterization of organizational identification as “theperception of oneness or belongingness to some human aggregate” (p. 21) by suggesting two differentpaths to identification. In the first path, the organizational member recognizes the similarity in beliefsbetween the organization and him/herself. Thus, an individual looks to the organization for salientcharacteristics that are shared with one’s own beliefs, or what Pratt (1998) describes as “identificationthrough affinity, where… individuals seek similar organizations” (p. 174, emphasis in original). Here,there may be little individual change during the process of identification. For instance, while not anorganizational member, an individual may identify with their local symphony orchestra due to a mutuallove for live, classical music. In the same way, an organizational member working at a public broadcaster,such as PBS, BBC, or CBC, may identify with the public broadcaster due to its commitment to providinghigh-quality and innovative programming. This conceptualization of identification follows Ashforth andMael’s (1989) feeling of “oneness” with an organization.The second path sees the organization as driving a change in the organizational member. Here,the individual changes one’s self to become more similar to the organization; it is “identification throughemulation… when we incorporate organizational beliefs and values into our own identities” (Pratt, 1998:174, emphasis in original). Pratt (2000) describes how organizational newcomers to a network marketingorganization, Amway, began to change their identities as they spent more time with other organizationalmembers. Ibarra (1999) also found that newcomers look to and assimilate characteristics of successfulorganizational insiders (i.e., those who best represented the organizational values). This process ofidentification is similar to Tajfel and Turner’s (1986) work on social identity theory (and later, selfcategorizationtheory) whereby individuals engage in identification when elements of the organization areencapsulated into identities.Organizational identification differs from related constructs like person-organization (P-O) fit andorganizational commitment. Researchers investigating P-O fit focus on matching individuals andorganizations (for a review, see Kristof, 1996). Kristof (1996) defines P-O fit as “the compatibilitybetween people and organizations that occurs when (a) at least one entity provides what the other needs,or (b) they share similar fundamental characteristics, or (c) both” (pgs. 4-5). Her definition of P-O fitencompasses elements of identification, that is, her second condition, but also points to the broadercontext that P-O fit research addresses. Furthermore, Cable and Edwards (2004) and Kristof (1996)suggest that fit can either be complementary or supplementary. Complementary fit occurs when theindividual brings her/his own unique values or characteristics to the organization to “complete” theorganization. Supplementary fit involves individuals who have similar values and characteristics to theorganization. Identification is more closely aligned with supplementary rather than complementary fit,and so, can provide a proxy of the former.83

Similar to P-O fit, organizational commitment shares some characteristics with organizationalidentification, but also differs from organizational identification. Similar to organizational identification,organizational commitment looks at the relationship between an organization and individual.Furthermore, in one of the more popular conceptualizations of organizational commitment, theOrganizational Commitment Questionnaire (OCQ; Mowday, Porter, & Steers, 1982), there is conceptualoverlap between identification and commitment. Identification is a key part of the commitment process(Dutton, Dukerich, & Harquail, 1994; Meyer & Allen, 1997), but research has made a distinction betweenorganizational identification and organizational commitment (Ashforth & Mael, 1989; O’Reilly &Chatman, 1986; Pratt, 1998). In Meyer and Allen’s (1997) conceptualization of commitment,identification is one mechanism by which affective commitment develops; “that is, employees’ affectiveattachment to their organizations might be based on a desire to establish a rewarding relationship with anorganization” (p. 15, emphasis in original). Finally, Pratt (1998) distinguishes the two constructs byreviewing the underlying questions that each asks; namely, organizational identification asks, “How do Icome to know myself in relation to the organization?”, while organizational commitment asks, “Howsatisfied am I with the organization?” (p. 178). Thus, organizational identification is conceptualized asone of the antecedents of organizational commitment.Multiple Identities and Multiple IdentificationsOrganizations have been theorized as having multiple identities (Albert & Whetten, 1985; Pratt &Foreman, 2000): “organizations have multiple organizational identities when different conceptualizationsexist regarding what is central, distinctive, and enduring about the organization” (Pratt & Foreman, 2000:20). Furthermore, Albert & Whetten (1985) state that organizational identity is not simply classified in adichotomy (e.g. profit vs. non-profit, government vs. private), but that it is a hybrid, composed of multipletypes. They offer two types of multiplicity: ideographic and holographic. The former refers to the casewhen multiple identities exist within the same organization but are not shared by all; for example,different shifts within a rehabilitation hospital exhibit their own identities, as is the case with the nursesworking in day (“well” roles) versus night (“sick” roles) shifts in Pratt and Rafaeli’s (1997) study.Holographic multiple identities are shared by all; for example, in Golden-Biddle and Rao’s (1997) study,the non-profit organization known internally as a “family of friends” and externally as a communityorganization dedicated to protecting the public presented two discrete identities yet both identities wereshared by the board of directors and management team.Hence, multiple organizational identities can, and do, exist. These identities do not necessarilyconflict (e.g. Golden-Biddle & Rao, 1997) nor do they need to be shared by all (e.g. Pratt & Rafaeli,1997). The salience of the social referent, or prototype, determines the consequences of multipleidentifications. Organizational members will be clear as to what identity or identities they will subscribe;they will either adopt their unit identity (e.g. day-shift versus night-shift nurses in a rehabilitation unit) orshare the multiple organizational identities (e.g. management team and board of directors at a non-profitorganization). Furthermore, individuals’ OIs will be clear. The notions of ideographic and holographicmultiplicity suggest that individuals’ identification occurs at different levels of the organization; indeed,March and Simon (1958) and Gouldner (1957, 1958) stated that employees have more than one socialreferent for identification.Rousseau (1998) distinguishes these processes as situated identification and deep structureidentification. The former entails an “elemental form of identification [and] is a perception of a discretework setting, created by situational cues signaling shared interests” (Rousseau, 1998: 218). So, the nursesin Pratt and Rafaeli’s (1997) study fall into this category. The second type of identification strikes a morelasting chord; “[deep structure] identification occurs when the employment relationship alters the mentalmodel individuals have of themselves to incorporate the organization itself” (Rousseau, 1998: 218).84

Similarly, the management team and board of directors in Golden-Biddle and Rao’s (1997) study can bedescribed as having deep structure identification.Albert and Whetten’s (1985) and Rousseau’s (1998) conceptualizations of OI suggest theexistence of different levels of identification for individuals (cf. March & Simon, 1958). Furthermore, intheir formulation of multiple OIs, Ashforth and Johnson (2001) explicitly incorporate ordered identities;identities can exist at the job, workgroup, department, division, and organization levels. Using SCT andidentity theory arguments, they contend that the salience of identities produces nested or cross-cuttingidentities, and hence, identifications (Ashforth & Johnson, 2001). Nested identities are typically seen inhierarchical organization charts, while cross-cutting identities include formal (e.g. task forces) or informal(e.g. soccer team or knitting club) social categories. Ashforth and Johnson (2001) assert that the salienceof the identity (e.g. one that is more proximal, exclusive, and concrete) determines the ultimate memberidentification.A Model of Multiple Identifications in OrganizationsMultiple identities create multiple targets for identification. The salience of the identity targetdetermines how an organizational member will identify. Thus, expanding on the contributions fromAlbert and Whetten (1985), Rousseau (1998), and Ashforth and Johnson (2001), dimensions involvingthe salience of identities (i.e. higher versus lower order) can be intersected to generate an individual’sidentification responses, outlined in Table 1. This framework applies only in those situations whereformal identities are in question, thus Ashforth and Johnson’s (2001) inclusion of informal social rolessuch as the company book club do not apply. In addition, this model concerns organizational insiders(Pfeffer & Baron, 1988) and intraorganizational identification, as opposed to OI during organizationalchange, such as a spin-off (e.g. Corley & Gioia, 2004) or merger (e.g. van Knippenberg & van Leeuwen,2001).Table 1A model of multiple identifications in organizationsSalience of Higher Order Identity(e.g. organizational identity)Salience of Lower OrderIdentity(e.g. workgroup identity)LowHighLow 4. Identity Ambiguity 1. Deep Structure IdentificationHigh 3. Situated Identification 2. Identity Conflict /Identity MaintenanceWhile Ashforth and Johnson (2001) acknowledge that, due to the more proximal, exclusive, andconcrete properties of lower order identities, higher order identities are less salient. However, higher orderidentities (e.g. organizational identities) are also important factors in determining an individual’sidentification. Dutton and Dukerich (1991) showed how the external image of an organization caninfluence the organizational identity, affecting members’ OI. Furthermore, Dukerich, Golden, and85

Shortell (2002) found that physicians’ perceived attractiveness of an organizational identity led tostronger OI. Thus, the salience of lower order identities may not always prevail over the salience ofhigher order identities.Beginning with the off-diagonals (quadrants 1 and 3), deep structure identification occurs insituations such as that found in work conducted by Dukerich and colleagues (Dukerich, Golden, &Shortell, 2002; Dutton & Dukerich, 1991) as well as the study by Golden-Biddle and Rao (1997).Furthermore, deep structure identification develops more strongly over time (Rousseau, 1998). These“cosmopolitans” (Gouldner, 1957, 1958) are those individuals who proclaim to “bleed brown”, as is thecase with some employees at United Parcel Service, or “bleed blue”, as is the case with some employeesat IBM. Those with deep structure identification carry this identification across various social contexts,revealing the influence of the organization on the internal cognitive schema of an individual.Organizations can engage in substantive and symbolic management to highlight higher order identities(Ashforth & Johnson, 2001). For example, senior management can adopt a policy of ensuring thatorganizational size be kept to a minimum in order to foster higher order identification. Similarly, seniormanagement can focus on those organizational characteristics that are deemed to be central, enduring, anddistinctive by articulating a mission statement or underscoring organizational prototypes. Thus, insituations where the salience of higher order identities surpasses the salience of lower order identities,organizational members will engage in deep structure identification.Humans have a fundamental motivation for social identification; merely grouping a temporarycluster of strangers will engender feelings of togetherness (Brewer & Gardner, 1996). Hence, situatedidentification strikes at this core sense of collectivity. According to Ashforth and Johnson’s (2001)dimensions, situated identification occurs because of the greater exclusiveness, concreteness, andproximity of lower order identities. This contributes to greater salience of lower order identities. There aremany reasons suggested by van Knippenberg and van Schie (2000). First, following SCT, individuals willcategorize themselves into groups with those whom they feel share a similar identity (Turner, 1987).Therefore, the intragroup differences will be less at a subgroup level than at a higher group level.Furthermore, members will tend to be more homogeneous, contributing to the exclusiveness of thesubgroup. Second, higher order identities are less “real”; the identity as an equity investment banker hasmore overlap with an individual’s day-to-day organizational existence than the identity as an employee ofGoldman-Sachs. Finally, organizational members work with other insiders more often than outsiders,especially within the organization. Therefore, working alongside other insiders becomes an individual’sprimary ingroup, contributing to greater salience of lower order identities and hence, situatedidentification.While some researchers claim that individuals cannot undertake simultaneous multiple identities(e.g. Ashforth & Mael, 1989), others have empirically demonstrated the opposite (e.g. Barker &Tompkins, 1994; Rotondi, 1975; van Knippenberg & van Schie, 2000). When there is high salience ofboth higher and lower order identities, it is unclear exactly what identification response will ensue.Organizational members who identify strongly with both higher and lower order identities may eithermaintain the status quo, continuing high levels of identification, or encounter an identity conflict,resulting in a resolution of the conflict. For example, a brand manager at an advertising agency may passthe pronoun test by remarking that s/he works at a specific agency and “we organize the print andtelevision advertisements” for a particular client. However, there may be other instances where strongidentification may lead to ethical dilemmas, as in the case of whistleblowers (Near & Miceli, 1995); basedon one whistleblower case (Katz, 1999), he enjoyed working for the organization and the seniormanagement team, so when he realized what they were doing, he felt very conflicted about whistleblowing.He even initially went along with the fraud. This case illustrates Gouldner’s (1957, 1958)cosmopolitanism-localism tension, as he ultimately reverted to his professional identity as a charteredaccountant. Situational cues should tip OI one way or the other, as OI is posited to include evaluative and86

cognitive elements (Ashforth & Mael, 1989; Rousseau, 1998; Turner, 1987). For the brand manager,her/his identification will depend on whether s/he is talking to a fellow employee or meeting someone forthe first time. Similarly, a whistle-blower may at one time decide to identify with the corporatelysanctionedfraud, but at another, identify with his professional identity. Thus, depending on the context,high levels of both types of identity salience may lead to identity maintenance or identity conflict.Finally, when there are low levels of salience for both higher and lower order identities, it isproposed that identity ambiguity will result. Corley and Gioia (2004) follow a sensemaking (Weick,1995) perspective in their definition of identity ambiguity as referring “to a situation in which theorganization’s identity is vague, not because two or more competing identity claims exist but because noclaims exist or, more likely, because the meanings underlying those claims are in doubt or subject tomultiple plausible interpretations” (p. 201). They further claim that identity ambiguity differs fromidentity conflict (Albert & Whetten, 1985; Foreman & Whetten, 2002; Golden-Biddle & Rao, 1997),which stems from a holographic construction of a hybrid organizational identity (Golden-Biddle & Rao,1997). Instead, identity ambiguity arises from the lack of a clearly defined organizational identity, such asthat involving an organization in transition (e.g. Corley & Gioia, 2004; van Knippenberg & van Leeuwen,2001).However, this definition restricts itself to the organizational level of analysis; that is, identityambiguity concerns only organizational identity. This definition excludes the possibility ofintraorganizational identifications. Bartel and Dutton (2001) examine the boundaries between full-timeand contract workers or organizational insiders and outsiders. As the boundaries blurred, they discoveredambiguous memberships, defining them as “those situations in which memberships are experienced asvague, problematic, or unstable” (p. 116). Identity ambiguity then concerns an organizational memberwho neither identifies with her/his immediate social referents (e.g. job, workgroup, department) nor withher/his distal social referents (e.g. organization); uncertainty and vagueness plague the available identitysources and no salient prototype steps forward. While identity ambiguity could result from organizationaltransitions such as take-overs or mergers (e.g. Corley & Gioia, 2004; van Knippenberg & van Leeuwen,2001), low levels of identity salience within an organization can also result from individualdisidentification with the organization (Elsbach, 1999; Kreiner & Ashforth, 2004) or a decision to leavethe organization (O’Reilly & Chatman, 1986). Thus, with low salience of higher and lower orderidentities, an organizational member will encounter identity ambiguity and once again, depending on thecontext, adopt attitudes and behaviours that range from disidentification to sensemaking (e.g. Corley &Gioia, 2004).DiscussionMultiple identifications within organizations for individuals are possible. Unlike much of thework on multiple identity organizations (e.g. Albert & Whetten, 1985; Corley & Gioia, 2004; Foreman &Whetten, 2002; Golden-Biddle & Rao, 1997), this paper has conceptualized multiple identifications froman individual perspective; that is, the multiple identities available to an organizational member within anorganization (e.g. Ashforth & Johnson, 2001; Barker & Tompkins, 1994; Pratt & Rafaeli, 1997). Buildingon work by Turner (1987), Albert and Whetten (1985), Rousseau (1998), and Ashforth and Johnson(2001), two dimensions of identity salience were posited to interact with each other to produce fourdifferent identification strategies that an organizational member can undertake.This framework excludes individuals categorized as “non-standard” (Kalleberg, Reskin, &Hudson, 2000) but represent an interesting area for future research (see George & Chattopadhyay, 2005for an exception). Non-standard workers, such as consultants, contractors, and traditional contingentworkers, often spend much of their time at a client organization, rather than the employing organization.What implications does this type of employment relationship have for organizational identification?87

George and Chattopadhyay (2005) found that these workers, highly-skilled (computer programmers andsystem designers) and well-paid (average annual salary was $65,000), engaged in multiple identifications,identifying with their employing organization and client organization to the extent that they foundattractive attributes about each. This study reveals that more research is needed in this area, as boundaryconditions, antecedents, and outcomes have not been fully explored.ReferencesAlbert, Stuart & Whetten, David A., “Organizational Identity”, Research in Organizational Behavior, vol.7, L. L. Cummings & Barry M. Staw (eds.), 263-295, Greenwich, CT: JAI Press, 1985.Alvesson, Mats & Willmott, Hugh, “Identity regulation as organizational control producing theappropriate individual”, Journal of Management Studies, 39(5), (July 2002), 619-644.Ashforth, Blake & Johnson, Scott A., “Which hat to wear? The relative salience of multiple identities inorganizational contexts”, Social Identity Processes in Organizational Contexts, Michael A. Hogg& Deborah J. Terry (eds.), 31-48, Philadelphia: Psychology Press, 2001.Ashforth, Blake E. & Mael, Fred, “Social identity theory and the organization”, Academy of ManagementReview, 14(1), (January 1989), 20-39.Bartel, Caroline & Dutton, Jane E., “Ambiguous organizational memberships: Constructingorganizational identities in interactions with others”, Social Identity Processes in OrganizationalContexts, Michael A. Hogg & Deborah J. Terry (eds.), 115-130, Philadelphia: Psychology Press,2001.Brewer, Marilynn B. & Gardner, Wendi, “Who is This "We"? Levels of Collective Identity and SelfRepresentations”, Journal of Personality and Social Psychology, 71(1), (July 1996), 83-93.Brown, Andrew D., “Organization Studies and Identity: Towards a Research Agenda”, Human Relations,54(1), (January 2001), 113-121.Brown, M. E. “Identification and Some Conditions of Organizational Involvement”, AdministrativeScience Quarterly, 14, (1969), 346-355.Barker, James R. & Tompkins, Phillip K., “Identification in the Self-Managing Organization”, HumanCommunication Research, 21(2), (December 1994), 223-240.Cable, Daniel M. & Edwards, Jeffrey R., “Complementary and Supplementary Fit: A Theoretical andEmpirical Integration”. Journal of Applied Psychology, 89(1), (February 2004), 822-834.Chattopadhyay, Prithviraj, Tluchowska, Malgorzata & George, Elizabeth, “Identifying the Ingroup: ACloser Look at the Influence of Demographic Dissimilarity on Employee Social Identity”,Academy of Management Review, 29(2), (April 2002), 180-202.Cheney, George, “On the Various and Changing Meanings of Organizational Membership: A Field Studyof Organizational Identification”, Communication Monographs, 50(4), (December 1983a), 342-362.Cheney, George, “The Rhetoric of Identification and the Study of Organizational Communication”,Quarterly Journal of Speech, 69(2), (May 1983b), 143-158.Corley, Kevin G. & Gioia, Dennis A., “Identity Ambiguity and Change in the Wake of a Corporate Spin-Off”, Administrative Science Quarterly, 49(2), (June 2004), 173-208.88

Dukerich, Janet M., Golden, Brian R. & Shortell, Stephen M., “Beauty is in the Eye of the Beholder: TheImpact of Organizational Identification, Identity, and Image on the Cooperative Behaviors ofPhysicians”, Administrative Science Quarterly, 47(3), (September 2002), 507-533.Dutton, Jane E. & Dukerich, Janet M., “Keeping an Eye on the Mirror: Image and Identity inOrganizational Adaptation”, Academy of Management Journal, 34(3), (September 1991), 517-554.Dutton, Jane E., Dukerich, Janet M. & Harquail, Celia V., “Organizational Images and MemberIdentification”, Administrative Science Quarterly, 39(2), (June 1994), 239-263.Elsbach, Kim D., “An Expanded Model of Organizational Identification”, Research in OrganizationalBehavior, vol. 21, Barry M. Staw & Robert I. Sutton (eds.), 163-200, Greenwich, Connecticut:JAI Press, 1999.Elsbach, Kim D. & Bhattacharya, C. B., “Defining Who You Are By What You're Not: OrganizationalDisidentification and the National Rifle Association”, Organization Science, 12(4), (July/August2001), 393-413.Foreman, Peter & Whetten, David A., “Members' Identification with Multiple-Identity Organizations”,Organization Science, 13(6), (November/December 2002), 618-635.George, Elizabeth & Chattopadhyay, Prithviraj, “One Foot in Each Camp: The Dual Identification ofContract Workers”, Administrative Science Quarterly, 50(1), (March 2005), 68-99.Golden-Biddle, Karen & Rao, Hayagreeva, “Breaches in the Boardroom: Organizational Identity andConflicts of Commitment in the Nonprofit Organization”, Organization Science, 8(6),(November/December 1997), 593-611.Gouldner, Alvin W., “Cosmopolitans and Locals: Toward an Analysis of Latent Social Roles – I”,Administrative Science Quarterly, 2(3), (September 1957), 281-306.Gouldner, Alvin W, “Cosmopolitans and Locals: Toward an Analysis of Latent Social Roles – II”,Administrative Science Quarterly, 2(4), (December 1958), 444-480.Hogg, Michael A. & Terry, Deborah J., “Social Identity and Self-Categorization Processes inOrganizational Contexts”, Academy of Management Review, 25(1), (January 2000), 121-140.Ibarra, Herminia, “Provisional Selves: Experimenting with Image and Identity in ProfessionalAdaptation”, Administrative Science Quarterly, 44(4), (December 1999), 764-791.Katz, Jonathan G., “File No. 3-9807”, http://www.sec.gov/litigation/admin/34-40938.htm. [Accessed 28October 2005].Kalleberg, Arne L., Reskin, Barbara F. & Hudson, Ken, “Bad Jobs in America: Standard and NonstandardEmployment Relations and Job Quality in the United States”, American Sociological Review,65(2), (April 2000), 256-278.Kreiner, Glen E. & Ashforth, Blake E., “Evidence toward an Expanded Model of OrganizationalIdentification”, Journal of Organizational Behavior, 25(1), (February 2004), 1-27.Kristof, Amy L., “Person-Organization Fit: An Integrative Review of Its Conceptualizations,Measurement, and Implications”, Personnel Psychology, 49(1), (Spring 1996), 1-49.Lee, Sang M., “Organizational Identification of Scientists”, Academy of Management Journal, 12(3),(September 1969), 327-337.March, James G. & Simon, Herbert A., Organizations, New York: John Wiley & Sons, 1958.Meyer, John P. & Allen, Nancy J., Commitment in the Workplace, Thousand Oaks, CA: SagePublications, 1997.Mowday, Richard M., Porter, Lyman W. & Steers, Richard M., Employee-Organizational Linkages, NewYork: Academic Press, 1982.Near, Janet P. & Miceli, Marcia P., “Effective Whistle-Blowing”, Academy of Management Review,20(3), (July 1995), 679-708.O'Reilly, Charles & Chatman, Jennifer A., “Organizational Commitment and Psychological Attachment:The Effects of Compliance, Identification, and Internalization on Prosocial Behavior”, Journal ofApplied Psychology, 71(3), (August 1986), 492-499.89

Pfeffer, Jeffrey & Baron, James N., “Taking the Workers Back Out: Recent Trends in the Structuring ofEmployment”, Research in Organizational Behavior, vol. 10, Barry M. Staw & L. L. Cummings(eds.), 257-303, Greenwich, CT: JAI Press, 1988.Pratt, Michael G., “To Be or Not To Be? Central Questions in Organizational Identification”, Identity inOrganizations: Building Theory through Conversations, David A. Whetten & Paul C. Godfrey(eds.), 171-207, Thousand Oaks, CA: Sage Publications, 1998.Pratt, Michael G., “The Good, the Bad, and the Ambivalent: Managing Identification among AmwayDistributors”, Administrative Science Quarterly, 45(3), (September 2000), 456-493.Pratt, Michael G. & Foreman, Peter O., “Classifying Managerial Responses to Multiple OrganizationalIdentities”, Academy of Management Review, 25(1), (January 2000), 18-42.Pratt, Michael G., & Rafaeli, Anat, “Organizational Dress as a Symbol of Multilayered Social Identities”,Academy of Management Journal, 40(4), (August 1997), 862-899.Rotondi, Thomas “Organizational Identification and Group Involvement”, Academy of ManagementJournal, 18(4), (December 1975), 892-897.Rousseau, Denise M. “Why Workers Still Identify with Organizations”, Journal of OrganizationalBehavior, 19(3), (May 1998), 217-233.Stryker, Sheldon, Symbolic Interactionism, Don Mills ON: Benjamin/Cummings Publishing Co., 1980.Tajfel, Henri & Turner, John C. “The Social Identity Theory of Intergroup Behavior”, Psychology ofIntergroup Relations (2 nd ed.), Stephen Worchel & William G. Austin (eds.), 7-24, Chicago:Nelson-Hall Publishers, 1986.Turner, John C., “A Self-Categorization Theory”, Rediscovering the Social Group: A Self-CategorizationTheory, John C. Turner, Michael A. Hogg, Penny J. Oakes, Steven D. Reicher, & Margaret S.Wetherell (eds.), 42-67, Oxford, UK: Blackwell, 1987.van Knippenberg, Daan & van Leeuwen, Esther, “Organizational Identity After a Merger: Sense OfContinuity as the Key to Postmerger Identification”, Social Identity Processes in OrganizationalContexts, Michael A. Hogg & Deborah J. Terry (eds.), 249-264. Philadelphia: Psychology Press,2001.van Knippenberg, Daan & van Schie, Els C. M., “Foci and Correlates of Organizational Identification”,Journal of Occupational and Organizational Psychology, 73(2), (June 2000), 137-147.Weick, Karl E, Sensemaking in Organizations, Thousand Oaks, CA: Sage Publications, 1995.90

ASAC 2006Banff, AlbertaA.R. ElangovanUniversity of VictoriaJia Lin XieKevin Hill (student)University of TorontoTACIT KNOWLEDGE INTEGRATION IN ORGANIZATIONS:A KNOWLEDGE-ASSEMBLY PERSPECTIVEThis paper presents a process model of tacit knowledge integration using knowledgeassemblytheory (Hayes-Roth, 1977) as the underpinning. By associating explicitknowledge with knowledge components (cogits) and tacit knowledge with links amongthese components (unitized assemblies), the paper offers a three-step model – knowledgecontribution, recognition and valuation – of tacit knowledge integration and its keydrivers.IntroductionOrganizations function, compete and succeed by effectively utilizing and building on theknowledge their members possess. Both scholars and practitioners alike recognize that knowledge is themost important strategic resource for organizations and represents the key to building unique capabilitiesand sustainable competitive advantages in the global economy (Grant, 1996; Leiebeskind, 1996; Spender,1996). Not surprisingly, the effective management of knowledge in organizations has emerged as a keytopic of research interest over the last decade. Research in this area has addressed and continues toexplore a plethora of questions and perspectives, both interorganizational (e.g., learning-based alliances,Grant & Baden-Fuller, 1995) and interpersonal (e.g., status-based considerations, Thomas-Hunt, Ogden,& Neale, 2003).One stream within this field of inquiry has focused on the difference between explicit knowledgeand tacit knowledge and the implications for organizational learning and effectiveness. In a general sense,explicit knowledge can be viewed as the know-about and the tacit knowledge as the know-how inorganizations (Daft, 2001). Explicit knowledge is the formal, systematic knowledge that can be codified,documented, stored, and easily shared. Examples include databases, documents, policies, and procedures.Tacit knowledge, on the other hand, is based on personal experience, insights, intuition, and judgment,and is often difficult to codify. Not surprisingly, these two types of knowledge are differentiated primarilyby their potential for transferability and the mechanisms for transfer across individuals, space and time(Hansen, 1999). While the effective management of both types of knowledge is important inorganizations, up to 80% of an organization’s valuable knowledge can be tacit in nature (Grayson &O’Dell, 1998) thus limiting the ease with which it is captured, codified and transferred. Managing tacitknowledge, therefore, represents serious challenges as well as significant strategic opportunities formodern organizations.Managing tacit knowledge has proven to be challenging and arduous for numerous reasons. Notonly is it difficult to capture and transfer such knowledge, but is also extremely inefficient to do so (e.g.,two specialists with tacit knowledge can share or transfer such knowledge by teaching each other one-onone;Grant, 1996; Grant & Baden-Fuller, 1995). To circumvent these constraints, some researchers haverecommended focusing on tacit-knowledge integration rather than sharing or transfer. For example,91

Grant (1996) posited that successful integration of tacit knowledge offers a significant sustainablecompetitive advantage to the firm and argued that the primary role of the firm (and subunits within thefirm) should be to integrate the various sets of tacit knowledge.It is important to acknowledge here that the transfer of tacit knowledge is indeed possible in thelong run (e.g., apprenticeships) but the primary difficulties rest in effectively utilizing tacit knowledge inthe short and immediate term. Given that functional experts (possessors of tacit knowledge) inorganizations are typically involved in multiple groups at any given point in time (and, perhaps, withchanging membership), we cannot assume the group stability, continuity and longevity necessary for tacitknowledge transfer among its members. Thus, in advocating the importance of knowledge integration, werecognize that efficiency must be considered along with effectiveness in managing tacit knowledge.Knowledge integration allows the group to accomplish its goals without placing demands on individualgroup members to acquire expertise across domains.This paper has two primary objectives: (a) to understand the dynamics of tacit knowledgeintegration, and (b) to outline the drivers of the knowledge integration process at the group level. In doingthis, we will also clarify and define tacit knowledge and knowledge integration in a general sense. Interms of context, we are focusing especially on multi-functional groups that have been formed for, or arepresently active in, addressing complex issues, tasks or problems. In other words, we are equating themembers of these groups to experts possessing tacit knowledge in their respective functions.Conceptualizing Tacit KnowledgeA survey of the prior research in the area of knowledge management indicates conceptualizationof tacit knowledge has been fragmented at best and chaotic at worst. The net result has been a certaindegree of confusion regarding the concept and a subsequent paralysis in probing deeper into its nature,dynamics and management. For example, tacit knowledge (or sometimes knowledge in general) has beendiversely interpreted or conceptualized as communities of practice (Brown & Duguid, 1991), absorptivecapacity (Cohen & Levinthal, 1990), combinative capabilities (Kogut & Zander, 1992), thought worlds(Dougherty, 1992), organizational routines (Nelson & Winter, 1982), and knowledge conversion (Nonaka& Takeuchi, 1995). Although these myriad interpretations and perspectives on tacit knowledge haveoffered us a rich set of options for understanding of the concept, they have also stalled research progressin this area. The perspectives, although intuitively appealing, have not lent themselves easily to theorybuilding,extrapolation and adaptation. Part of the problem stems from the lack of theoretical groundingof these perspectives themselves. If the essence of what we have come to understand as tacit and explicitknowledge can be anchored in an underlying theoretical frame, it would make it possible for us toformulate ways to manage tacit knowledge. Knowledge-assembly theory, housed in the connectionistschool of thought, offers one such framework for positioning and understanding tacit and explicitknowledge.A Knowledge-Assembly Perspective on Tacit KnowledgeKnowledge-assembly theory (Hayes-Roth, 1977) focuses on the acquisition, representation, andprocessing of knowledge. Based on Hebb’s (1949) work on cell-assembly, knowledge-assembly theoryassumes that the representation and processing of knowledge change qualitatively as learning progresses.The theory posits that a knowledge-structure consists of cognitive components (called “cogits”) and links(called “associations”) among the components. As one’s familiarity with a component increases with itsaccess and use, the component is activated in an all-or-none fashion (i.e., the knowledge underlying thecomponent is accessed holistically or not at all) and acquires a certain “strength” in one’s cognition.Different components are assembled in configurations (called “assemblies”) using links or associations92

that signify relationships among the components and meaning. Further, a given component can be part ofmore than one assembly. If one component is activated in one’s cognition or memory, it activates thelinked components based on the strength of the associations among the components. That is, the entire setof knowledge underlying the linked components is accessed holistically. This knowledge-assemblyprogresses with an individual’s experience – it begins as a collection of individual components and endsas a single tightly integrated unit with strong associations among the components. In other words, theassociations among the related components in a configuration continue to be strengthened throughexperience to a point until the entire structure can be activated in an all-or-none fashion (called“unitization”). Thus, the processes of establishing, strengthening, and associating components and thenunitizing these assemblies or configurations are applied recursively as learning progresses.While knowledge-assembly theory and other related models within the connectionist perspectivehave garnered significant attention from cognitive psychologists and learning theorists, they have notbeen used to interpret the discourse on tacit and explicit knowledge that has pre-occupied managementscholars. Two core elements of knowledge-assembly theory – cognitive components and links – areespecially useful for highlighting the differences and implications of explicit and tacit dimensions ofknowledge. The cognitive components that serve as the fundamental building blocks of knowledgeassemblytheory are akin to explicit knowledge bits regarding any given topic or domain. Each knowledgebit represents a piece of information-structure that is cognitively or perceptually delineated from otherpieces. For example, when focusing on international trade as a domain, the explicit knowledge bits couldbe pieces of information about “value of currency”, “interest rates”, “tariffs”, “exports”, “trade balance”,“production costs”, “subsidies”, etc. As such, these explicit bits of knowledge can be acquired andtransferred to others since they can be easily codified (e.g., one can teach another what interest rates are).On the other hand, the associations or links that are formed among the cognitive components – theexplicit knowledge bits – constitutes tacit knowledge, especially in the state of unitization. The linksrepresenting relationships are assembled to render meaning and understanding of the structure as a whole.In the above example, the associations one forms among the different knowledge bits could be as follows:when interest rates go up, the value of a currency may increase, which would increase production costsfor domestic companies, which could lead to lower exports, and affect the trade balance between twocountries. With time and experience, the complexity, strength, and density of these associations wouldincrease to a point that the entire assembly is accessed quickly and on an all-or-none basis. In otherwords, the activation of one explicit knowledge bit, say interest rate hike, immediately activates the wholestructure with such speed and integration that it is even not possible for that individual to access singleassociations or parts of the assembly. Such unitization of knowledge within an individual’s cognitionmakes it difficult to dis-assemble the pieces, codify links and transfer what is known to anotherindividual. At this point the knowledge structure is uniquely whole to that individual and does not lenditself to easy articulation – in other words, this knowledge is tacit. See Figure 1 (in the Appendix).Conceptualizing the tacit dimension of knowledge as unitized knowledge grounds tacitknowledgein knowledge-assembly theory and fits well with the conventional ideas about the qualities oftacit knowledge. Further, it offers a common foundation for conceptualizing the various mechanisms ofmovement associated with explicit and tacit knowledge, i.e., transfer, sharing and integration.Understanding IntegrationBefore exploring tacit knowledge integration, it is important to address what we mean by“integration”. Dictionary definitions of the verb ‘integrate’ tend to emphasize how this action serves to“combine or be combined to form a whole” (The Concise Oxford English Dictionary, 2004). While theemphasis here is on combination, the tacitness of knowledge implies that knowledge cannot simply becombined when it resides primarily in diverse areas of the organization and its membership. Indeed,organizations are segmented into areas purposefully in order to obtain the benefits of specialization in93

developing distinctive competence. As such, the driving force behind the integration of knowledge insuch contexts is more than a mere combination.A second meaning of the word ‘integrate’, to “bring or come into equal participation in aninstitution or social group” can be used as part of a broader conceptualization (The Concise OxfordEnglish Dictionary, 2004). This meaning stems primarily from the context of the desegregation ofinstitutions. Yet, this meaning of “integrate” appears to be quite relevant due to the diversity ofknowledge bases that characterizes organizations and the groups (especially cross-functional) containedwithin them. Moreover, it applies to the extent that different knowledge bases can sometimes act as abarrier to understanding. As such, knowledge integration in teams can be viewed as the process of makingthe knowledge potential of the team available to its members. This aspect of the phenomenon makes itclear that a crossing of boundaries is often associated with integration.More comprehensive is the definition of integration as “a combination and coordination ofseparate and diverse elements or units into a more complete or harmonious whole” (Webster's Third NewInternational Dictionary of the English Language, 1961). Such an interpretation implies that integrationcan include a striving toward harmony among the parts without implying a singular state. It is moresensitive to the configural (rather than purely additive) nature of effective integration. Essentially, it is theway that parts are brought together, as opposed to their general aggregation, that distinguishes(knowledge) integration from other movements (e.g., pooling, transfer, sharing). In addition, as acoordinated activity, there is the implication that change is felt across the units of the group that achievesintegration, thus distinguishing it from assimilation (which refers more to the complete altering andsubsuming of one unit by another).A close examination of these meanings and reflection on the conventional uses of the term“integration” highlights four key aspects that collectively characterize the construct – crossing ofboundaries, change in the initial parts, interdependency, and seamlessness. Whether the emphasis is oncombination, participation or coordination, a common implication is that there is a “crossing ofboundaries” among the different parts (units, members) being integrated. In the case of knowledgeintegration, it implies the extending of one’s expertise or the mixing of a member or unit’s expertise withothers. A second facet of integration is a change in the initial parts (units, members). The notion ofparticipation of every part and the crossing of boundaries across parts implies that no part is leftunchanged through the process of integration – each part is influenced to varying degrees by one or moreof the other parts. Similarly, the emphasis on configuration and coordination of parts in integrationhighlights a pattern of interdependencies among the different parts in a seamless way (i.e., it is not a mereaddition but a meaningful assembly of parts that are intricately interwoven). Knowledge integration, aswith other forms of integration, would embody these four facets as well. Viewed through the lens ofknowledge-assembly theory, it can be defined as the meaningful combination of different unitizedknowledge assemblies into a larger knowledge assembly.A Multi-Level Model of Tacit Knowledge IntegrationIt is important to recognize that knowledge integration is as much a process as an outcome (muchlike groupthink or group polarization), and that it is context-specific, i.e., it does not happen in a vacuumbut requires the frame of a decision-situation to take place. We use a multi-functional team in anorganization tasked with specific responsibilities and making decisions (e.g., developing a business plan,initiating a major change, generating a strategy for internationalization, assessing an investmentopportunity) as our archetypal context as we outline a multi-stage, multi-level process model ofknowledge-integration.94

The process of tacit knowledge integration involves a progression through a series of steps forboth the individual and the group. These steps underpin the individual and group transformation that iscentral to integration, and can be interpreted through the lens of the knowledge-assembly theory. In theremainder of this section we discuss the steps in the tacit-knowledge integration process, the individualand group level effects during the process from a knowledge-assembly theory standpoint, and the keydrivers of this multi-step process.Steps in the Tacit Knowledge Integration ProcessTacit knowledge integration in a multi-functional team is the result of the group progressingthrough three key steps: knowledge contribution (KC), knowledge recognition (KR), and knowledgevaluation (KV).Knowledge Contribution. The integration of any knowledge possessed by the different membersof a team is wholly dependent on the willingness of each member to contribute his/her expertise to thetask at hand. By default, such contributions are likely to be explicit-knowledge related in the beginningbefore delving into tacit-knowledge as the group and problem-solving progresses.But the willingness to open up and share one’s expertise may be influenced by numerous factorsincluding the individual’s personal interest and engagement in the task, trust towards his/her teammembers, the nature of the reward systems and in the organization, and the dominant culture of theorganization. Expertise is a source of power in organizations (French & Raven, 1959) and employeesoften hoard information or closely guard knowledge about special techniques, skills and competencies forfear of losing their privileged status. While the tacit nature of this knowledge may make it close toimpossible for others to quickly copy, the perception of the threat of losing one’s privileged status couldlimit an expert’s contribution in a multi-functional team. This reluctance to contribute, or the need to bemore strategic in contributing, would also be influenced by the trust the team-members have towards eachother. Prior research on trust in organizations has noted that “stealing of ideas” and “taking credit forwork that’s not one’s own” are often rated as key triggers for the dissolution of interpersonal trust(Giacalone & Greenberg, 1997). In line with the folk saying “Once bitten, twice shy”, employees who donot trust their co-workers are often more likely to protect themselves by, among other approaches, closelyguarding their sources of power, including knowledge. On the other hand, if the expert perceives themulti-functional team setting as a vehicle for either solidifying or enhancing his/her standing, he/she maybe more open to contributing openly. The team member would also be more engaged and open tocontributing if he/she is personally interested in the task or problem that the team is tackling either forintrinsic or extrinsic reasons. The nature of reward system in the organization – individual vs. groupbased– and the transparency, credibility and perceived fairness of the assessment also play a role in amember’s willingness to contribute. If the team-member perceives systemic inadequacies or inaccuraciesin how performance (including contribution in a team setting) is assessed or does not see a proportionatelink between assessed performance and rewards (i.e., low instrumentality, Vroom, 1964), then he/she isless likely to contribute his/her expertise in the team setting. Lastly, the culture of the organization –especially the extent to which citizenship behaviours and achievement-orientation are emphasized –would play a role in a team-member’s willingness to contribute. Team-members in a work culturecharacterized by an achievement orientation (e.g. the performance culture, Cameron & Quinn, 1999)and/or citizenship behaviours (Organ, 1988) are more likely to be open to sharing their expertise andforthcoming with their contributions.Knowledge Recognition. The recognition of a team-member’s expertise by other members is thenext critical step in the process leading to tacit knowledge integration. Even if an individual member iswilling to and actively contributes his/her expertise towards addressing the task that the multi-functionalteam is faced with, it is only when that knowledge is accepted and absorbed by other team members that95

integration can take place. But even before acceptance comes recognition – team-members would need torecognize the expertise of the member making the contribution. Such recognition could be a function ofboth an ability to recognize as well as a willingness to recognize that expertise. Several researchers havestressed the importance of common knowledge – a basic level of knowledge broadly relating to theproblem at hand that is shared by all members – as necessary for an expert in one domain to sufficientlyunderstand, interact and utilize the knowledge of an expert from a different domain (Grant, 1996). Inother words, an expert financial analyst and an expert market researcher would need to have a sufficientlevel of common knowledge about the fundamentals of business (which would include aspects of financeand marketing) to be able to communicate and connect with each other, i.e., to “recognize” each other’sexpertise. In the absence of such common knowledge, there is no basis for a team-member to be able tounderstand or evaluate the contribution made by another member.While common knowledge would equip a team-member to be able to recognize the expertise ofhis/her fellow-members, it does not guarantee that he/she would do so. The motivation or willingness onthe part of a team-member to recognize other experts in the multi-functional team-setting would beinfluenced by the power dynamics within the group, the extent to which acknowledging that expertisecould lead to an outcome that is not preferred (or not lead to a preferred outcome), or could negativelyimpact his/her standing in the group and the organization. Interpersonal conflict and jostling for positionsand images of influence and social status could drive some team-members to ignore, belittle or refute theexpertise of rival team-members (Menon & Pfeffer, 2003). Regardless of the reasons, the process towardstacit-knowledge integration would be disrupted if team-members do not mutually recognize each other’sexpertise.Knowledge Valuation. The final step leading to tacit-knowledge integration in a multi-functionalteam involves an accurate valuation of each member’s expertise for the specific task or problem the groupis addressing. The mere recognition by a team member that his/her fellow team members are experts intheir domain does not guarantee that he/she deems their expertise relevant or necessary for the problem athand. Knowledge valuation builds on knowledge recognition and makes it more context specific. Anaccurate valuation is contingent both on the ability to perceive the relevance of that specific expertise aswell as an openness to its role in solving the problem at hand. The ability to perceive the relevance isintricately linked to the team-member’s sense of clarity regarding the problem but not necessarily thesolution. Clarity about the problem per se makes it possible for the team member to properly value therole the expertise of other team-members can play in solving it. For example, an expert financial analystin a team would be able to understand the relevance and value of the HR manager’s expertise if there is aclarity within the team that the poor performance of the company is due to low employee morale andpersonal productivity.While clarity about the problem facilitates knowledge valuation within a team, clarity regardingthe solution is not necessary. In fact, ambiguity or confusion about the solution serves to increase a teammember’sopenness to others’ expertise and contributions and make him/her less dogmatic about apreferred outcome. When a member (expert) is completely clear about the solution to a problem, which isoften viewed through the lens of his/her own expertise, he/she is more likely to dismiss or minimize theusefulness of the expertise of his/her team-mates (while recognizing them as experts in their domains)since the “answer is obvious”. Along with this vagueness regarding the solution, a realistic selfassessmentof one’s own expertise also fosters an openness to others’ expertise. When a member’s senseof his/her expertise and its role in addressing a problem is realistic, he/she is more likely to acknowledgethe need for other types of expertise. An over-blown or grandiose sense of one’s own expertise is likely tomake a member regard others’ contributions as redundant or unnecessary and de-value their expertise inthat situation. In sum, while knowledge recognition implies that team-members in a multi-functionalgroup acknowledge each other as experts in different domains, positive knowledge valuation ensures that96

their expertise is considered relevant and useful enough to allow tacit knowledge integration as it pertainsto the given problem.Successful progression through these three steps – knowledge contribution, recognition andvaluation - should result in the integration of explicit and tacit knowledge possessed by the members ofthe multifunctional group. In the next section, we use knowledge-assembly theory to explain theprogression of tacit knowledge integration in a multi-functional group. In the subsequent section, we thenprovide an overview of the likely drivers of this process, paying special attention to factors that wouldactivate individual motivation to proceed through these steps.Applying Knowledge-Assembly Theory to Tacit Knowledge IntegrationKnowledge-assembly theory offers a useful lens for interpreting both the individual-level andgroup-level dynamics during the different steps of the knowledge integration process.Individual-Level Knowledge IntegrationFigure 2 (see Appendix) highlights the transformation that takes place at the individual-level as agroup member progresses through the different steps, while Figure 3 (see Appendix) captures the paralleltransformation at the group level. Suppose there are three experts (possessing tacit knowledge) whocomprise a multi-functional team in an organization. Further, let’s assume that Member 1 possessesexpertise in Domain 1, Member 2 in Domain 2, and Member 3 in Domain 3. Figure 2 focuses on theprogression of Member 1 through the different steps of the KI process. From a knowledge-assemblystandpoint, in this case Member 1 possesses a unitized knowledge-assembly that comprises of cogits A,B, C, D, E & F with strong links among them. Similarly, Member 2 has a unitized knowledge-assembly inthe domain consisting of C, F, K, X, Y and V cogits, while Member 3 possesses a unitized knowledgeassemblywith cogits F, K. J, Q, H and M. So all three members have cogit F in their assemblies; inaddition, Member 1 and Member 2 have cogit C as a common cogit while Member 2 and 3 share cogit K.In this context, cogits F, C and K represent points of connection for the three members even though eachmember may have utilized them in unique ways as part of their respective assemblies.In the Knowledge Contribution step of the process Member 1 brings tacit knowledge in his/herdomain that includes explicit knowledge bits A to F and contributes it as a whole assembly to the group.Members 2 and 3 do the same with their unitized knowledge assemblies. But given that the contributionsof Members 2 and 3 are in a unitized form, all that Member 1 picks up at this stage is an awareness ofsome of the knowledge bits (cogits) that form part of their assemblies. In this case, Member 1 becomesaware of cogits J, Q, M from Member 3 and X, Y and V from Member 2. Such awareness would be basedon the common knowledge that the three members share about the issue, their work, organization, etc.In the Knowledge Recognition step, Member 1 starts to recognize the expertise of Members 2 and3 not only because of the common knowledge they all share, but also based on an awareness of thereputation of Members 2 and 3, prior experience in dealing with them, and the shared cogits of C and F.At this point, Member 1 starts to recognize how some of the cogits in the domains of Members 2 and 3might be relevant for his/understanding. Member 1 starts to establish weak links with these cogits – inthis case with cogits J, Q, X and Y – and represented by dotted arrows in Figure 2. At this stage, theseweak links connect back to Member 1’s unitized knowledge-assembly but have not being fullyincorporated into it. Cogits M and V continue to be more distant since Member A has not established anyrelationships or attached relevant meaning to them.97

In the Knowledge Valuation step, Member 1 gains even more clarity as he/she is able to assess orevaluate the value of the new cogits (representing some of the expertise of the other two members) for theproblem at hand. Such a valuation is driven by Member 1’s growing clarity about the context, theproblem the group is addressing, and the perceived relevance of the new cogits. This contributes to thestrengthening of some of these links that he/she has already formed – represented by solid arrows withcogits J and X. The links with cogits Q and Y continue to be weak as Member 1 does not perceive themas useful or relevant to addressing the specific problem or connecting to his/her own unitized knowledgeassembly. It is at this point, that cogits J and X start to penetrate Member 1’s knowledge-assembly.In the final step, Knowledge Integration, Member 1 has incorporated cogits J and X fully intohis/her knowledge assembly, i.e., cogits J and Q are now part of unitization. This is represented by thecreation of one or more strong links between these two new cogits and other cogits already in theknowledge-assembly. Some of these links are likely influenced or flavoured by the relationships orassociation that Members 2 and 3 had evoked in contributing their own unitized knowledge-assemblies –at least as gleaned by Member 1 during the previous three steps. In this case, strong links are builtbetween cogit J and cogits E and F, and cogit X and cogits F, C and B within Member 1’s knowledgeassembly. Further, cogits J and X are now linked strongly through cogit F. The links between the oldcogits in Member 1’s knowledge-assembly and the two new cogits are more nuanced, specific andsophisticated. Cogits J and X and the associated set of relationships are now part of the tacit knowledge ofMember 1. In essence, Member 1 has managed to integrate the tacit knowledge of Members 2 and 3 withhis/her own tacit knowledge with his/her unique imprint.Group-Level Knowledge IntegrationA similar set of dynamics occurs at the group level as Member 1, 2 and 3 progress through thedifferent steps. Please refer to Figure 3. In the Knowledge Contribution step, three unitized knowledgeassembliesconnected by the shared cogits F, C and K serving as lynch-pins represent the knowledge thatthe group collectively possesses. As the group progresses to the Knowledge Recognition step where themembers mutually recognize each other’s expertise, multiple weak links are formed between each of theknowledge-assemblies and some of the cogits housed in the other two knowledge-assemblies. These linkssignify the beginning of changes or modifications to the shared knowledge of the group hereto onlylimited to the shared cogits. There is a growing awareness of the multiple meanings and perspectives. Inthe Knowledge Valuation step, some of these weak links get strengthened as the group gauges therelevance and value of these cogits and appreciate their usefulness for solving the problem. Finally, in theKnowledge Integration step, these cogits are now fully incorporated within each of the knowledgeassemblies along with their detailed links. As a result, the overlap among the three knowledge-assemblies– representing the shared group expertise – increases well beyond the initial lynch-pin cogits of F, C andK. In other words, the group and the overlap among them increases. In other words, the group hasintegrated the three domains of tacit knowledge to create a new domain (knowledge assembly) thatconstitutes the group’s tacit knowledge for dealing with such situations in the future.The group’s newly unitized knowledge-assembly exhibits the four facets of integration outlinedin the previous section. The establishing of multiple links amongst the cogits housed in the initial threedifferent knowledge-assemblies reflects the crossing of boundaries of the multiple domains of expertise(knowledge-assemblies) brought to the group by the different members. As a result, each of the initialknowledge-assemblies is altered in a different way as each member incorporates cogits from the other twomembers’ knowledge-assemblies into their own unitized knowledge-assembly. The number, direction andpattern of the links formed as the new group knowledge-assembly takes shape highlights both aninterdependency of the cogits within that assembly in creating meaning as well as seamlessness as abigger assembly gets unitized.98

Key Drivers of the Knowledge Integration ProcessProgression through the three key steps of the knowledge-integration process – knowledgecontribution, recognition and valuation – depends on certain key factors. Some of these factors ensure thesuccessful enactment and completion of one step in the process, while others may drive progressionthrough more than one step. Collectively, however, they ought to be considered necessary conditions fortacit knowledge integration to occur. We posit that tacit-knowledge integration in a multi-functional teamoccurs when its members (experts) have collective clarity and commitment to solving the problem athand, mutual acceptance of each other’s expertise, and individual awareness of the limitations of ownexpertise.Clarity of Problem. Clarity regarding the problem is an essential condition for the knowledgevaluation step in the process. When team members are clear about the problem they are more likely tounderstand the need for and relevance of each other’s expertise to solving it as well their own. Problemclarity provides focus for one’s own contributions, helps highlight the boundaries of one’s own expertisein successfully addressing the problem, and promotes openness to other insights and contributions. Thisalso implies that the onus is on the leader of any multi-functional team aspiring to achieve a state ofintegration of tacit knowledge residing with individual group members to dispel any ambiguity about theproblem (while not encouraging any jumping to conclusions regarding the solution).Commitment to Solving Problem. A key success factor that has an impact on all four steps inthe process is commitment on the part of team members to solving the problem. Without the impetusstemming from a sense of commitment to addressing the problem, group members are likely to bereluctant or disinterested in contributing their expertise and less likely to take the trouble to even fullyunderstand the problem. Consequently, they are not likely to dedicate the time and attention necessary torecognize, assess and value their group members’ expertise and contributions. Further, a low level ofcommitment is also likely to dampen their enthusiasm for incorporating new knowledge or for tolerating aweakening or alteration of their expert assemblies (already unitized assemblies). Thus, ensuring andpromoting commitment to solving the problem becomes an imperative for any leader hoping to move amulti-expert group towards tacit knowledge integration.Realistic Evaluation of Own Knowledge. Openness to valuing and accepting other members’expertise is closely tied to a realistic evaluation of the scope of one’s own expertise. An overly subjectiveand inaccurate reading of self-knowledge is likely to negatively impact all three steps of the knowledgeintegration process. A lower than realistic assessment of self-knowledge will limit a member’scontribution since he/she may not fully realize what he or she could do to address the problem at hand.Similarly, a poor self-rating of one’s expertise or competence can result in unintended self-censorship dueto a lack of confidence in contributing. On the other hand, an exaggerated sense of one’s own expertiseand relevance for the problem is likely to make the member minimize the need or value of others’expertise. The erroneous conviction that he/she already has the “answer” may even prompt the member tonot adequately recognize the expertise of his/her fellow team members. But a realistic assessment of one’sknowledge and its adequacies fosters an openness to including and valuing others’ expertise and activelysearching for ways to complement one’s own.Trust. For different members in a multi-functional group to contribute their expertise toaddressing a problem and to be open to taking the risk of weakening or altering their unitized knowledgeassemblies, they need to trust others in the group. Group members need to believe in the inherentgoodwill and fairness of others in the group and the organizational as a whole to tap into their ownexpertise, contribute the best they can, and be open to others’ contributions. On the other hand, distrust islikely to breed mutual suspicion leading to hoarding of one one’s expertise and the rejection of others’.99

Further, high levels of trust are also important for group members to be open to being influencedsufficiently enough by others to alter their own unitized knowledge assemblies. A willingness to toleratethe weakening of links in one’s own knowledge assembly requires somewhat of a leap of faith since oneis doing so without becoming a full expert in the new domain being introduced. Such an openness iscontingent on the level of trust that exists within the group and towards organizational systems (e.g.,performance appraisal, rewards).Common Knowledge. A sufficient level of common knowledge is critical for the differentexperts in a multi-functional group to be able to communicate and understand each other, to recognize theinherent expertise that each member brings to the group, and to accurately value each member’s potentialcontribution to addressing the problem. It is important to recognize that while the intent is to tap into andintegrate the in-depth, tacit expertise that each member possesses, the integration is only possible if thereis a common foundation. Not only does the foundation offer a common point of reference and connectingframe for the different bits of tacit knowledge, it also aids significantly in recognizing the boundaries ofone’s own expertise. Multi-expert groups aspiring to tacit knowledge integration would need to ensurethat there is a common basis for relationships to form and communications to be effective.DiscussionBy focusing specifically on tacit knowledge integration, this paper has turned the researchspotlight on a much talked about, yet under-investigated, topic within the knowledge management area.Prior research in this area has been conducted at the level of the organization and/or has assumed thatknowledge is a commodity. Where the integration of new knowledge with existing knowledge structureshas been proposed (e.g., Cohen & Levinthal, 1990), theory has rarely specified how such integrationoccurs. Through the lens of knowledge-assembly theory, we have elaborated more precisely the nature ofthe knowledge integration process.It is important to note that the process and success of knowledge-integration poses someinteresting challenges for multi-functional groups. A close look at Figure 3 indicates that while there issignificant overlap among the knowledge-assemblies of the members of the group such that the grouplevelknowledge-assembly is enhanced, it is a step removed from a complete overlap. A total overlapwould mean that the knowledge-assemblies of all the members of the groups are identical. Such anextreme degree of similarity could result in a lack of diversity in viewpoints, analyses and decisionmakingresulting in groupthink (Janis, 1972). Similarly, while unitization of knowledge-assembliesrepresents the development of true tacit knowledge (collection of strong links that are accessed on an allor-nothingbasis), the tight and strong links can also block or slow down the absorption of new cogits andlinks into one’s unitized knowledge-assembly. But weaker links are more open to change and adaptationand can speed up the process of integration. This poses a dilemma – while strong links and unitization ofknowledge-assemblies is the ultimate goal and measure of success in tacit knowledge development, weaklinks are receptive to change and enhancement of knowledge-assemblies and favourable for theknowledge-integration process. While an individual may not be able to completely prevent unitization ofknowledge-assemblies in his/her domain, an awareness of the importance of being open to newknowledge and challenging or modifying one’s own expertise would be of benefit to knowledgeintegration.This paper constributes to the literature in several ways. First, we have highlighted the disparateset of meanings and interpretations that have both informed and stifled our understanding of tacitknowledge. Using knowledge-assembly theory as a theoretical underpinning we have offered aconceptualization of tacit knowledge as unitized knowledge-assemblies (with explicit knowledgerepresenting cogits). Second, we have put forward a definition of integration that goes beyond mere100

combination of parts to a meaningful reconfiguration of different parts into a new assembly with someessential features - crossing of boundaries, change in the initial parts, interdependency, and seamlessness.Third, we have outlined the three key steps – knowledge contribution, knowledge recognition andknowledge valuation – that constitute an integral part of the knowledge integration process. By viewingknowledge integration not just as a desired outcome but also as a process, we have drawn attention to theunderlying sequential dynamics and the importance of successfully passing through each stage. Fourth,we have described a process of tacit knowledge integration in a multi-functional team using knowledgeassemblytheory – both at the level of the individual and the group. By portraying how links to new cogitsare formed, subsumed into an existing unitized knowledge-assembly, and re-configured to create a newknowledge-assembly, we have offered a framework for understanding how tacit knowledge can beintegrated. Fifth, we have presented the key drivers necessary for the knowledge integration process –including progression through the three steps – to occur. Each of these drivers is a necessary but not asufficient condition for the knowledge integration process to occur and collectively they explain thesuccess or failure of knowledge-integration in multi-functional teams.This paper directly focuses on the integration of tacit knowledge in cross-functional teams. Thisis especially important in answering the call made by others to focus more on human systems than ITsystems as the key drivers of knowledge management (McDermott, 1999). Tacit knowledge, by its verynature is human-centered, and must be managed much like other human systems. Hence, our attention toimportant drivers like trust and commitment highlight the limitations of focusing too much on thecharacteristics of the knowledge itself divorced of its human characteristics.Many of the theoretical prescriptions for effective knowledge management offered to date havebeen built upon case-based research. In order to build confidence in the theory of tacit knowledgeintegration here, future research must complement case-based investigations with large-scale quantitativestudies. This paper also offers useful tips to managers and leaders in organizations who set up multifunctionalteams with the hope of optimizing decision making. To ensure that knowledge integrationtakes place successfully, leaders need to ensure that the key drivers necessary for the progression throughthe three steps – knowledge contribution, recognition and valuation – are in place. Proper attention tothese factors will foster effective integration and also serve as an alternative to more costly and timeconsumingapproaches of tacit knowledge-acquisition such as socialization and experience-basedlearning.101

AppendixFigure 1: Knowledge AssembliesADBEFCCogits with someweak linksADBEFCAssemblywith WeaklinksADBEFCAssembly withStrong Links -Unitized/tacitknowledge102

Figure 2: Individual-Level Transformations during the Tacit-Knowledge Integration ProcessADBECFXYJKCQM V ADBECFXYKRJQMVKVJADBEQCFAXDY JBEQCFXMYVKINote:KC = Knowledge Contribution, KR = Knowledge Recognition, KV = KnowledgeValuation,KI = Knowledge Integration.103

Figure 3: Group-Level Transformations during the Tacit-Knowledge Integration ProcessADJHBECFKQMXYVKCADJHBECFKQMXYVKRADJHBECFKQMXYVKIADJHBECFKQMXYVKVNote:KC = Knowledge Contribution, KR = Knowledge Recognition, KV = KnowledgeValuation,KI = Knowledge Integration.104

ReferencesBrown, John Seely, & Duguid, Paul, “Organizational learning and communities of practice: Toward aunified view of working, learning, and innovation,” Organization Science, 2(1), (February 1991),40-57.Cameron, K. S., & Quinn, R. E. Diagnosing and changing organizational culture. Reading: Addison-Wesley, 1999.Cohen, Wesley M., & Levinthal, Daniel A., “Absorptive capacity: A new perspective on learning andinnovation,” Administrative Science Quarterly, 35, (March 1990), 128-152.Daft, R., Organization theory and design., Cincinnati, OH: Southwestern, 2001.Dougherty, Deborah, “Interpretive barriers to successful product innovation in large firms,” OrganizationScience, 3(2), (May 1992), 179-202.French, John. R. P., Jr., & Raven, Bertram H. The bases of social power. In D. Cartwright (Ed.), Studiesin social power (150-167). Ann Arbor, MI: Institute for Social Research, 1959.Giacalone, R. & Greenberg, J. (Eds.), Antisocial behavior in organizations, Thousand Oaks, CA: Sage,1997.Grant, R.M. Toward a knowledge-based theory of the firm. Strategic Management Journal, 17 (Winter1996), 109-122.Grant, Robert M. & Baden-Fuller, Charles, "A Knowledge-Based Theory of Inter-Firm Collaboration,"Academy of Management Proceedings, (1995), 17-21.Grayson, C.J. & O’Dell, C.S., Mining your hidden resources, Across the Board, (April 1998), 23-28.Hansen, Morten, “The search-transfer problem: The role of weak ties in sharing knowledge acrossorganizational subunits,” Administrative Science Quarterly, 44(1), (March 1999), 82-111.Hayes-Roth, Barbara, “Evolution of Cognitive Structures and Processes”, Psychological Review, 84(3),(1977), 260-278.Hebb, D.O., Organization of Behavior, New York: Wiley, 1949.Janis, I., Groupthink. Boston: Houghton Mifflin, 1982.Kogut, Bruce, & Zander, Udo, “Knowledge of the firm, combinative capabilities, and the replication oftechnology,” Organization Science, 3(3), (August 1992), 383-397.Liebeskind, J.P., Knowledge, strategy and the theory of the firm. Strategic Management Journal, 17(Winter 1996), 109-122.Menon, Tanya & Pfeffer, Jeffrey, “Valuing internal vs. external knowledge: Explaining the preference foroutsiders,” Management Science, 49(4), (April 2003), 497–513.McDermott, Richard, “Why information technology inspired but cannot deliver knowledge management,”California Management Review, 41(4), (Summer 1999), 103-117.Nelson, Richard R. & Sidney G. Winter, An Evolutionary Theory of Economic Change. Cambridge,Mass.: Belknap Press of Harvard University Press, 1982.Nonaka, Ikujirao, & Takeuchi, Hirotaka. The knowledge-creating company: How Japanese companiescreate the dynamics of innovation. New York: Oxford University Press, 1995.Organ, D., Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: LexingtonBooks, 1988.Oxford Reference Online. Oxford University Press. University of Toronto Libraries. 17 January 2006.Thomas-Hunt, Melissa, Ogden, Tonya, & Neale, Margaret, “Who’s really sharing? Effects of social andexpert status on knowledge exchange within groups,” Management Science, 49(4), (April 2003),464-477.Spender, J.C., Making knowledge the basis of a dynamic theory of the firm, Strategic ManagementJournal, 17 (Winter 1996), 109-122.The Concise Oxford English Dictionary. Ed. Catherine Soanes and Angus Stevenson. Oxford UniversityPress, 2004.Vroom, V. H., Work and motivation, New York: Wiley, 1964.105

Webster's Third New International Dictionary of the English Language: Unabridged., Springfield, Mass.:G. & C. Merriam Co., 1961.106

ASAC 2006Banff, AlbertaMorris B. MendelsonJeff FroomanJ. Kevin Murphy (MBA student)University of New Brunswick Saint JohnPERCEPTIONS OF LEADERSHIP AS A DETERMINANT OFLEGITIMATE AND ILLEGITIMATE USE OF SICK LEAVEThis study examined the relationship between perceptions of one’s supervisor’sleadership style and employees’ self-reported use and abuse of sick time mediated by jobsatisfaction. Differential effects were found depending on whether respondents perceivedtheir leaders to be transformational or transactional. Implications, limitations, and futuredirections are discussed.In Canada an estimated 700,000 full-time employees, or about 7% of the total, were absent fromwork for all or part of any given week because of personal reasons, such as illness, disability, or personalor family responsibilities (Statistics Canada, 2002). Absenteeism in the United States is estimated to costits economy over 40 billion dollars per year (Dalton & Mesch, 1991). Further, approximately 83% ofcompanies surveyed by CCH, a Chicago tax and business law research and publishing firm, believe thatunscheduled absenteeism is likely to stay the same or get worse in upcoming years (Gale, 2003).Controlling employee absenteeism is critical to all employers attempting to survive in anincreasingly competitive marketplace. As argued by Gaudine and Saks (2001, p. 16), “[T]he ability of aprogramme or policy to decrease absenteeism by even a small amount may mean large financial savingsas well as important quality implications”. However, successfully developing controls for absenteeismrequires identifying the antecedents of absenteeism in order to develop, assess, and implement strategiesto control it. Broadly put, the research question we seek to answer in this paper, then, is: What are thedeterminants of absenteeism?We hypothesize in this paper that the leadership style of an employee’s immediate supervisor willbe one such determinant of absenteeism, and that job satisfaction will be an important mediator of thatrelationship. Specifically, we look at employees’ perceptions of their supervisor’s leadership style and theeffect these perceptions have on the legitimate and illegitimate use of sick days. We consider bothtransformational and transactional leadership styles. With an instrument designed to ascertain theemployees’ perceptions of their supervisors’ leadership styles, their level of job satisfaction, and theirattitudes towards use of sick days, two regression analyses were performed. The results of the firstregression showed that as the level of perceived transformational leadership increased, job satisfactionincreased; and as this increased, the illegitimate use of sick leave decreased. The results of the secondregression showed that as the level of perceived transactional leadership increased, job satisfactiondecreased; and as this decreased, the use of legitimate sick leave decreased but the illegitimate use of sickleave increased.107

Theory: Constructs and RelationshipsThe purpose of this paper is to test the relationship that may exist between employees’ perceptionof their immediate supervisor’s leadership style, and the employee’s self-reported use and abuse of sickleave, and whether job satisfaction mediates that relationship. In this section we discuss the three factorsand the relationships we predict to exist between them.Dependent Variable: AbsenteeismIt is generally agreed that absenteeism represents a significant cost to employers. However, fromthe earliest studies (e.g., Chadwick-Jones, Brown & Nicholson, 1973) little agreement has existed on auniform operational definition of “absenteeism.” For one thing, there are many reasons why employeesare absent from work. Some of these reasons are legitimate, such as staying home due to an illness,having a child or other dependent at home sick, or requiring time to attend important events such ashaving a dentist or doctor’s appointment that cannot be rescheduled. Compare these legitimate reasonswith those that are more difficult to defend such as skipping work to attend a ball game or to play a roundof golf.In the current research, we were particularly interested in making the distinction betweenlegitimate and illegitimate absences. The lack of consistency across studies however, in how thisconstruct has been measured, makes determining its antecedents exceedingly difficult. In other words,absenteeism is a complex issue that has resisted the development of simple strategies to overcome itsnegative impact on business. What we do know is that absenteeism (measured as an employee’s intentionto be absent from work) has been shown to be related to higher levels of employee stress, lower levels ofaffective commitment, and reduced job satisfaction (Lowe & Schellenberg, 2003; Oi-Ling, 2002). Morerecently, it has been demonstrated to be negatively predicted by transformational leadership and perceivedorganizational support (Gakovic, 2003). It should be noted, however, that Gakovic conducted his researchat the work unit level, not at the individual level, of analysis. Further, no other research to date hasinvestigated the psychological processes, such as job satisfaction, that help explain these relationships.Independent Variables: Perceived Leadership StylesThe aim of this paper is to determine the role of perceived leadership style and its effects onabsenteeism (as mediated by job satisfaction). Two styles of leadership will be contrasted--thetransformational style and the transactional style. While we contrast the two styles in this paper, we donot wish to imply the two to be opposites on a continuum or even mutually exclusive. Indeed, both Bass(1999) and Yukl (2002), for example, make much of the fact that transformational style and thetransactional style may appear side-by-side. This suggests there may even be circumstances in which acombination of the two leadership styles might be more effective than using just one style on its own.Because the two leadership styles are separate and distinct, each will be discussed independently of theother in the two subsections below. Furthermore, the leadership styles will be discussed primarily in termsof the behaviours that characterize them, since as Yukl observes the constructs to date are essentiallyunderstood in terms “of the component behaviours used to influence followers and the effects of theleader on the followers (2002, p. 253).”Transformational Leadership. Again according to Yukl, “the essence of transformationalleadership appears to be inspiring, developing, and empowering followers (2002, p. 261),” such that theyare motivated to engage in supernumerary behaviour--that is, they seek to achieve beyond expectationsheld of them. Bass and Avolio (1995) and Bass (1999) identified four types of behaviours of108

transformational leaders--idealized influence, inspirational motivation, individualized consideration, andintellectual stimulation.Idealized influence has to do with behaviours that arouse followers to feel a powerfulidentification and strong emotions toward the leader. Inspirational motivation involves behaviours thatmodel moral values as an example. It also includes communicating an inspiring vision and using powerfulsymbols to arouse greater effort and a feeling of belonging. Individualized consideration has to do withproviding coaching, support, and encouraging followers. Finally, intellectual stimulation involves thosebehaviours that influence followers to view problems from a fresh perspective and with an increasedawareness.A growing body of research has demonstrated important relationships between the use oftransformational leadership and individual and organizational outcomes. For example, transformationalleadership has been found to be associated with greater organizational commitment (e.g., Barling, Weber,& Kelloway, 1996; Koh, Steers, & Terborg, 1995; Piliai & Williams, 2004), more organizationalcitizenship behaviours (e.g., Koh et al., 1995; Piliai, Schriesheim, & Williams, 1999; Mackenzie,Podsakoff, & Rich, 2001), higher levels of job satisfaction (e.g., Bycio, Hackett & Allen, 1995),heightened job performance (e.g., Avolio, Howell, & Sosik, 1999; Howell & Hall-Merenda, 1999;Mackenzie et al., 2001; Piliai & Williams, 2004), and a reduced intent to turnover (Bycio et al., 1995).Several of these factors--reduced stress, increased employee commitment, and increased job satisfaction--have in turn have been found to be negatively related to absenteeism rates (Bassett 1994; Lowe &Schellenberg 2003, Oi-ling 2002).Given these results, we expect to find that perception of transformational leadership style in anemployee’s immediate supervisor will be negatively related to the use of sick time.H1a 1 :H2a:Perceived transformational leadership will be negatively related to the legitimateuse of sick leave.Perceived transformational leadership will be negatively related to theillegitimate use of sick leave.These predictions are in keeping with the general view of transformational leadership--thatemployees who feel trusted and empowered and have come to identify with their leader and their leader’svision, will be so committed to their jobs that they will desire to be at their workplaces more than mighteven be expected of them; in short, they will take almost no sick leave.Transactional Leadership. In contrast to the transformational perception of leadership,transactional leadership seeks to motivate followers by appealing to their self-interest (Burns, 1978). As aconstruct, transactional leadership actually encompasses four types of behavior: management bycontingent reward, laissez-faire management, passive management by exception, and active managementby exception (Bass and Avolio, 1995). These four types of behaviour run along a continuum which rangesfrom the use of positive methods to encourage employee performance to the use of negative methods toencourage employee performance.1 In numbering our hypotheses, an “a” is assigned to a hypothesis involving transformational leadership,while a “b” is assigned to a hypothesis involving transactional leadership.109

Management by contingent reward uses positive methods to encourage employee performance.Simply put, the leader states his expectations, and uses rewards and incentives to prompt employees toachieve those expectations. This is the most common type of behavior for a transactional leader.The remaining three types of behavior are less frequently used by transactional leaders. Laissezfairemanagement sits in the middle of the continuum. It involves a “hands-off” approach toward workersand their performance. To a large degree this type of behaviour is characterized by the leader ignoring theemployees, not responding to their problems and perhaps not even monitoring their performance. Passivemanagement by exception is the type of behaviour that involves the leader using punishment as aresponse to unacceptable performance after it has occurred. Finally, active management by exception sitsat the far end of the continuum. This type of behaviour relies on using measures to correct unacceptableperformance while it is occurring. This is probably the most negative method for encouraging employeeperformance, because it is akin to the leader watching over the shoulder of the employees and promptlycalling employees on each misstep they make.To generalize, though, the transactional leader can be said to motivate employees by reward andpunishment. Furthermore, for transactional leaders there is an inherent belief that social systems workbest with a clear corporate structure of command in which management is omnipotent, and where theexpectation is for everyone to faithfully follow managerial instruction. To facilitate this, there areusually clear rules and norms whereby it is understood what is required of subordinates, what rewardswill be received for following orders, and where formal systems of discipline are in place for thosewho fail to follow orders.A literature search yielded no studies relating transactional leadership and absenteeism. To ourknowledge these factors have not been studied in tandem before (and thus we believe our work istherefore the first to study this relationship). We hypothesize, though, that transactional leadership will bepositively related with both forms of absenteeism:H1b:H2b:Perceived transactional leadership will be positively related to the legitimateuse of sick leave.Perceived transactional leadership will be positively related to the illegitimateuse of sick leave.We make these predictions in view of how control oriented transactional leadership can be. Thetransactional leader tends to strip employees of their autonomy. The amount of direction the employeesreceive and the number of constraints put upon the employees by rules and norms tends to makeemployees feel as if they are not trusted and are powerless. Generally speaking, such employees will lackcommitment to their jobs and will show no great desire to be at their workplaces; in short, they will takefull advantage of their sick leave in both a legitimate and illegitimate manner.Mediator: Job SatisfactionThe focus of this study is addressing the question, “What are the determinants of absenteeism?” Inparticular, as indicated in the previous subsections, we are hypothesizing that leadership style maysignificantly affect absenteeism. In this subsection we also hypothesize that job satisfaction may playa role in determining levels of absenteeism.110

Satisfaction, in general, has been defined in many different ways, from need to equity to anexpectation of rewards for job performance (Jorgenson, Dunnette, & Pritchard, 1973). Some have definedsatisfaction as a derivative of need or valuement (Hackman & Lawler, 1971). Others such as Locke(1976), defined it as “. . . a pleasurable or positive emotional state resulting from the appraisal of one’sjob or job experiences” (p. 1304). Expectancy-valence models have conceptualized satisfaction as afunction of the individual's expectant rewards owing to her behaviours on the job (Pritchard & DeLeo,1973; Vroom, 1964). Job satisfaction, in particular, is said to include these dimensions: opportunity toparticipate in decision making, job enlargement, job enrichment, working conditions and internal-externalfeedback received for performance (Hackman & Lawler, 1971).Some studies have indicated that job satisfaction may be a determinant of absenteeism (Kohler &Mathieu, 1993; Lowe & Schellenberg, 2003; Oi-Ling, 2002; Scott & Taylor, 1985). Other studies haveshown that unhappy employees have higher turnover or are absent from work more often than satisfiedemployees. Job satisfaction has also been linked to other withdrawal behaviors, including lateness,grievances, drug abuse, and decisions to retire. Further, Steel and Rentsch (1995) found that jobsatisfaction and job involvement were negatively related to absenteeism. There have been however,conflicting findings. For example, job satisfaction and personal characteristics (such as age and familysize) have been found to be related to absenteeism in some studies, but not related in others (Scott &McLellan, 1990). In short, our own intuition would guide us to conclude that job satisfaction should benegatively related to absenteeism. After all, satisfied workers should be eager to come to work. And thisintuition is well supported by a majority of the studies cited. Thus, we propose our third hypothesis:H3: Job satisfaction will be negatively related to the use of sick leave (both legitimate andillegitimate).In terms of the relationship between job satisfaction and transformational and transactionalleadership style, the literature is thin. One study found that transformational leadership was positively related to jobsatisfaction Zaccaro (1991). We could find no studies linking transactional leadership and job satisfaction, however.Given the nature of these two types of leadership styles, though, we hypothesize the following:H4a:H4b:Perceived transformational leadership will be positively related to job satisfaction.Perceived transactional leadership will be negatively related to job satisfaction.Our thinking in regard to these two hypotheses is this: Satisfaction with one’s job seems to hingearound, to some significant degree, how one’s own supervisor makes one’s job a meaningful experienceby providing opportunities to participate, by providing autonomy and some authority, and by providinghelpful encouragement and mentoring. These characteristics constitute the very essence oftransformational leadership, and thus our supposition that transformational leadership leads to increasedjob satisfaction. Conversely, the transactional leader to a large degree discourages participation, restrictsautonomy, and provides feedback typically through the carrot and/or the stick, and thus our belief thattransactional leadership leads to reduced job satisfaction.Finally, given the limited amount of research done to date involving the interplay of our threevariables--leadership style, job satisfaction, and absenteeism--we feel it remains unclear what exactly therole is that job satisfaction might play. Therefore, we will check for its role as a possible mediatorbetween leadership style and absenteeism, as indicated in our fifth hypothesis:H5a: The relationship between perceived transformational leadership and the use of sick leave(both legitimate and illegitimate) will be mediated by job satisfaction.H5b: The relationship between perceived transactional leadership and the use of sick leave(both legitimate and illegitimate) will be mediated by job satisfaction.111

In other words, leadership style may directly affect absenteeism, or it may indirectly affectabsenteeism through job satisfaction (full mediation), or it may both directly and indirectly affectabsenteeism (partial mediation).Respondents and ProcedureMethodsOne hundred and twenty out of 165 unionized employees (response rate = 72.7%), working in sixoffices of a national mail delivery company in five New Brunswick cities, completed this study. Amajority of respondents (M age = 45.5 yrs., range = 20 to 59 years) were male (75%), worked as lettercarriers (73.3%) as opposed to inside workers (26.7%), and a significant majority worked the day shift(93.3%). A high proportion (95.7%) had at least completed a high school education. Union members weredistributed the questionnaire in a sealable envelope by union representatives as members entered theirwork area or the lunchroom. They were told that completion of the questionnaire was voluntary and thattheir responses would be anonymous and confidential. All respondents completed a consent form beforeparticipating.MeasuresThere were two scales adapted from previous research and three scales that were developed forthe purposes of this research.Job Satisfaction. The Job Descriptive Index (Warr, Cook, & Wall, 1979) is a 15 item scale usedto assess employees’ level of satisfaction with five aspects of their job (i.e., work itself, pay, promotionopportunities, supervision and coworkers) rated on a seven-point Likert-type scale (1 = extremelydissatisfied; 4 = neither satisfied nor dissatisfied; 7 = extremely satisfied). The scale showed stronginternal consistency (∀ = .87).Transformational and Transactional Leadership. Perceptions of one’s immediate supervisor’sleadership styles were measured using the Multifactor Leadership Questionnaire (Form 5X-Short)developed by Bass and Avolio (1995). The four transformational leadership scales (i.e., idealizedinfluence, inspirational motivation, intellectual stimulation, and individualized consideration) and the fourtransactional leadership scales (i.e., contingent reward, active management by exception, passivemanagement by exception, and laisez-faire management) had very good internal consistencies(transformational leadership, ∀ = .97; transactional leadership, ∀ = .82).Absenteeism. For the purposes of this research, we developed eight items (see Appendix A) todetermine employees’ agreement or disagreement with their own self-reported reasons for using sickleave. An exploratory factor analysis using varimax rotation 2 was computed resulting in a three factorsolution which accounted for 61%of the variance. The first factor, which we labelled ‘legitimate use ofsick time’ included four items (∀ = .69) reflecting participants’ agreement that they are absent only forlegitimate reasons (e.g., “Only under extreme cases do we use sick time”). The second factor labelled‘illegitimate use of sick time included two items (∀ = .57) to reflect participants’ agreement that theyabuse their sick time (e.g., “I abuse my sick time”). Both of these scales were rated on a seven-pointLikert type scale (1 = strongly disagree; 7 = strongly disagree). Although neither scale reached the .702 A factor analysis using direct oblimin rotation yielded similar results.112

eliability cut off suggested, the items were retained for subsequent analysis. Using the Spearman-BrownProphecy formula (DeVellis, 1991), we calculated what the reliability would be if the number of items forthese scales were increased. With four items, the reliability of the ‘illegitimate use of sick time” scaleincreases to ∀ =.73; with six items, it becomes ∀ = .80. This demonstrates that scales’ coefficient alphasare very sensitive to sample size. The first and second scales were rated on a seven-point Likert type scale(1 = strongly disagree; 7 = strongly agree). The final two-item factor was identified as the ‘number of sickdays used per year’ and was rated on a seven-point frequency scale (1 = 1 to 2 days; 7 = 13+ days) butshowed an unacceptable internal consistency (∀ = .20). Therefore it was dropped from furtherconsideration in this study.ResultsTo test the hypotheses, we used Baron and Kenny’s (1986) procedure for mediated regression.First, we first set out to satisfy the three conditions to demonstrate that job satisfaction would mediate theeffects of transformational and transactional leadership styles. To conduct such an analysis, separateregression analyses using SPSS 12 were computed to look at the separate effects of transformational andthen transactional leadership on the two outcome variables of interest (i.e., legitimate and illegitimate useof sick time). In the first condition for each analysis, the independent variables (transactional andtransformational leadership) must be shown to load onto the mediator (job satisfaction). In the secondcondition, the mediator must significantly predict the criterion variables (i.e., legitimate and illegitimateuse of sick time). In the third condition, when the mediator has been statistically controlled for, previouslysignificant relationships between the predictors and criterion variables will either be no longer significant(in the case of full mediation) or the strength of prediction will be reduced (partial mediation). The Fchange statistic was used to indicate if our hypothesized mediated effects would be supported.113

Table 1Regression analyses testing for direct and mediating effects of transformational leadership and jobsatisfaction on legitimate and illegitimate use of sick leave.Effects of Transformational LeadershipR 2 change F change βTransformational Leadership predictingJob Satisfaction .33 57.71*** 57***Job Satisfaction predictingLegitimate use of sick time .01 .90 .09Illegitimate use of sick time .03 3.42 a -.17 aTransformational Leadership predictingLegitimate use of sick time .00 .10 -.03Illegitimate use of sick time .05 6.75* -.23*Controlling for Job SatisfactionTransformational Leadership predictingLegitimate use of sick time .02 1.89 -.12Illegitimate use of sick time .01 23 -.20 aa p < .05 (one-tailed); * p < .05; *** p < .001The results of the analysis using transformational leadership (see Table 1) partially support ourhypothesis that job satisfaction would fully mediate the effect of transformational leadership onto thelegitimate and illegitimate use of sick time. More specifically, transformational leadership strongly loadedon the mediator variable (job satisfaction; β = .57) and the mediator significantly loaded only ontoillegitimate use of sick time (β = -.17) but not legitimate use of sick time (β = .09). After removing theinfluence of the mediator, the relationship between transformational leadership and legitimate use of sicktime was not significant (β = -.12) but it was the negatively predictive of illegitimate use of sick time (β =-.20) suggesting partial mediation.The results of the analysis using transactional leadership as an initial variable also provided somesupport for our hypotheses but with an unexpected finding as well (see Table 2). As expected, we foundthat perceptions of transactional leadership are negatively related to job satisfaction (β = -.50).Interestingly, we found that transactional leadership was negatively related to legitimate use of sick time(β = -.17) but was positively related to illegitimate use of sick time (β = .21). However, after controllingfor the mediator (job satisfaction), transactional leadership no longer demonstrated a significantrelationship to either the legitimate (β = -.17) nor the illegitimate use of sick time (β = .16), suggestingfull mediation.114

Table 2Regression analyses testing for direct and mediating effects of transactional leadership and jobsatisfaction on legitimate and illegitimate use of sick leave.Effects of Transformational LeadershipR 2 change F change βTransformational Leadership predictingJob Satisfaction .25 39.79*** -.50***Job Satisfaction predictingLegitimate use of sick time .01 3.45 .09Illegitimate use of sick time .03 3.42 a -.17 aTransformational Leadership predictingLegitimate use of sick time .03 3.54 a -.17 aIllegitimate use of sick time .04 5.24* -.21*Controlling for Job SatisfactionTransformational Leadership predictingLegitimate use of sick time .00 .00 -.17Illegitimate use of sick time .01 .69 -.16a p < .05 (one-tailed); * p < .05; *** p < .001DiscussionThe research question this study focused on was: What are the determinants of absenteeism? Weexamined if leadership style and job satisfaction might be two such determinants. Specifically, we lookedat the following. First, in regard to absenteeism, we considered both legitimate and illegitimate use of sicktime, to see if these might be affected differently. Second, in regard to leadership style, we examinedemployees’ perceptions of their immediate supervisor’s style, and whether they might perceive it to betransformational or transactional. Third, in regard to job satisfaction, we tested for its direct effects onabsenteeism, but also tested to see if it mediated the relationship between leadership style andabsenteeism. In general, we supposed that (1) transformational leadership would relate positively to jobsatisfaction and negatively to absenteeism, (2) that transactional leadership would relate negatively to jobsatisfaction and positively to absenteeism and (3) that job satisfaction would relate negatively toabsenteeism.The results of the current study provide support for the hypothesis that both transformational andtransactional leadership styles have significant relationships to job satisfaction. In the current research wefound that perceptions of transformational leadership positively predicted job satisfaction, whiletransactional leadership perceptions negatively predicted job satisfaction. More importantly, jobsatisfaction was found to partially mediate the relationship between transformational leadership andillegitimate use of sick leave, but that job satisfaction did not mediate the relationship betweentransformational leadership and legitimate use of sick leave. Interestingly, we also found that jobsatisfaction fully mediated the effect of transactional leadership on illegitimate use of sick leave. Thestriking difference however, was that the relationship between transactional and illegitimate sick leavewas positive (see Figure 1).115

Figure 1Observed relationships between predictors, mediators and outcomes.Transformational LeadershipLegitimate UseOf Sick Time.57***Job Satisfaction.09TransactionalLeadership-.50***-.17 aIllegitimate UseOf Sick Time-.20 aap < .05 (one-tailed); *** p < .001The results of this research highlight different causes for different types of absenteeism. Usingsick leave when one is legitimately ill or even for those instances when an individual requires time forappointments, is a reality in organizations. Using sick leave to engage in activities such as attending amovie, playing golf, or taking a trip to the countryside are illegitimate and potentially more avoidablefrom an organizational perspective. In our research, an important predictor of whether people are more orless likely to abuse their sick leave is respectively, how transactional or transformational they perceivetheir leaders.In this study, when immediate supervisors are perceived to be transformational, employees reportgreater satisfaction and less abuse of their sick leave. However, there was no relationship between thisleadership style and legitimate use of sick time, and this was contrary to our hypothesis. This surprisingfinding may have a number of plausible explanations. A transformational leader would not displaycharacteristics or take measures to unduly influence or harm an employee, who is legitimately sick,simply to achieve reduced absenteeism without regard for the employee’s health. Another plausibleexplanation is that a manager perceived to be transformational will be accepting of an employee’sdecision to use an appropriate amount of sick time to overcome an illness, and thus employees will feelencouraged to take legitimate sick time.The present study also suggests that perceiving one’s leader as transactional is more likely toresult in more illegitimate use of sick leave, but less use legitimate of sick leave. This too was surprisingand may be a very interesting and important finding. If a transactional leadership style motivatesemployees to come to work when they are ill but stay away from work when they are healthy, we are leftto wonder in what context the employee is present in the workplace.In short, employers have to have a clear idea of what it is they are trying to minimize (Kelly,1990). Kelly pointed out that employers should not expect to eliminate absenteeism, but should try tominimize it. If, as the case would appear, employees who perceive their leader as transactional return towork even though they are ill, employees could be exhibiting presenteeism behaviour. Presenteeism isbehaviour in which an employee attends work, but is not able to function at a normal level ofproductivity. The employee’s return to work, even though he or she is legitimately ill, can be viewed by116

some as dedicated behaviour. Others would perceive it as inappropriate risk-taking behaviour rather thana health-promoting one (Grinyer & Singleton, 2000). Appropriate use of sick-time would mean usingenough sick-time to return to work in a healthy state.We should seriously question the benefit of transactional leadership in that it may promotepresenteeism behaviour. In a recent study, Goetzel found that presenteeism accounted for an averageof 61% of an employee’s total medical and lost-productivity costs and that this behaviour could costemployers an average of $255 per worker per year. We can reasonably expect that employees whocome to work while sick will be operating at a reduced productivity level or capacity. As such,transactional leadership may compel employees to return to work prematurely even while usinglegitimate sick-leave. This obviously ignores the health of the affected employee and at the extreme,could result in the illness of the employee’s colleagues.Of equal interest in this research, we found that if employees see their leaders as transactional, notonly will they report being less job satisfied, they will be more likely to report abusing their sick leave ifat all possible. In other words, although an employee may avoid missing work for legitimate reasons dueto a fear of punishment or removal of reward, the same employee may be motivated to skip work whenthe opportunity presents itself. The behaviour is consistent with the goal of transactional leadership in thatit is set up as an exchange process aimed at motivation through appealing to one’s own self-interest. Itmust be noted that as the first study of its kind, subsequent research must be conducted on different andlarger samples before any definitive conclusions can be made. Nonetheless, these preliminary findinghave key theoretical and practical implications.We set out to demonstrate that leadership style is a critical antecedent to absenteeism. Given thatthere is a dearth of research on this important topic, the results of this study makes a critical contributionto the literature on unearthing some of the causes of this costly withdrawal behaviour. Although Bycio etal. (1995) showed a relationship between transformational leadership and turnover intentions, there hasbeen no studies that have investigated the relationship between transformational leadership andabsenteeism.LimitationsThere are a number of limitations of this research that must be addressed. The first limitation ofthis study is its generalizability. The current study was conducted in a unionized environment and wouldneed to be replicated in larger and more diverse samples from other populations. Additionally, the studywas conducted in a public service union and may not be applicable to private sector unions.Second, the absenteeism scale developed in this study had marginal internal consistencies eventhough three factors were identified. One plausible reason for the low reliability of the two retainedmeasures (i.e., legitimate and illegitimate use of sick leave) is that participants’ responses showed littlevariability. It may be preferable to sample in such a manner that the participants would vary moresignificantly with respect to the factors that are being estimated but little on other attributes (Fabrigar,Wegener, MacCallum, & Strahan, 1999). The third scale, which represented self-reported absenteeismover the last year, showed unacceptably low reliability and therefore, was dropped from furtherconsideration.It should also be noted that the accuracy of self-reported absences from work is unreliable overlong periods of time. This imprecision increases over longer time periods. Accuracy of self-reported datais as low as 57% for six months and 51% for periods of twelve months (see Severens, Mulder, Laheij, &Verbeek, 2002). Because of the imprecision related to the recall period, a recall period of two months117

might be more valuable in a self-reported absence questionnaire. For a two-month recall period, accuracyis at an 87% level (Severens et al., 2002). In this study, employees were reporting rates of absenteeism ofless than 8 days annually, whereas the employer was reporting 10.98 annually per internal worker and17.6 days annually per external worker. Some of the difference between the employee’s self-reported useof absence from work and the employer’s records may be reconciled by different interpretations ofabsenteeism. However, the difference between employee and employer reports of absenteeism ratesexposes the difficulty in making comparisons across studies and even within workgroups and areas.Accuracy of self-reported data and the definition of absenteeism are limitations to this report and may beto others attempting similar studies.Third, although we found mediated relationships, some of the tests were only marginallysignificant (i.e., p < .05, one-tailed). One possibility is that because of the poor reliability of the twooutcome measures, any real relationship was obfuscated. Follow up research must ensure that morerefined and strengthened measures be used to avoid not detecting real relationships when they do occur.Fourth, this study relied entirely on self-report data collected from one source, namely employees.This may have increased the likelihood of mono-method bias. Mono-method bias has been argued to leadto percept-percept inflation (i.e., inflation of the real relationships between variables, see Crampton &Wagner, 1994). Although we recognize a concern with mono-method bias, we believe its effects on theresults of this research are minimal. In fact, a number of authors (e.g., Crampton & Wagner, 1994; Wall,Michie, Patterson, Wood, Sheehan, Clegg et al., 2004; Wall & Wood, 2005) have more recently arguedthat mono-method bias does not occur as frequently as is often presumed. Crampton and Wagner (1994,p. 72) suggested that “…percept-percept inflation may be more the exception than the rule inmicroresearch on organizations.” Furthermore, it would be next to impossible to collect data regarding theillegitimate use of sick leave from a source other than employees as this data is not readily available fromother sources. Nonetheless, one method of ruling out mono-method bias as a threat would be to collectlongitudinal data and/or data from multiple sources (i.e., employees, supervisors).ConclusionBusiness, at a heightened level of urgency, is focused on reducing costs associated withabsenteeism. Difficulties with identifying how absenteeism costs can be reduced and what strategies arebest to achieve this reduction are at the heart of this study. Hence, it is vital for business to be proactive intheir choice of an appropriate leadership style, which may help them achieve the ultimate goal of reducingabsenteeism costs. Transactional leadership gives the impression that it is a viable option to help reduceabsenteeism. However, this strategy is aligned with the idea of presenteeism, which is actually shown tobe an absenteeism cost driver.Companies ought to focus on the value of their employees, and their employees coming to workin a healthy state. Therefore, the employees would be more likely to work for the good of theorganization. This is a greater value than simply getting an employee to work at any cost and for anybenefit, such as seen in the model of presenteeism.Transformational leadership is not shown to affect legitimate absenteeism, but it is shown to havea negative relationship with illegitimate absenteeism. Value in reducing abuse of absenteeism wouldaccrue to the company. A reduction in overall absenteeism based on reducing abusive absenteeism ispossibly the best option for most companies and is most likely to take place when a manager displays atransformational style of leadership.Future research should evaluate psychological factors at the point at which employees are drivento use sick-time for actual illness or for other reasons. New research must develop statistically reliable118

absenteeism scales to measure the intent of employees to use sick-time either legitimately orillegitimately.Finally, ongoing research should assess the value of leadership style, its impact on job satisfactionand ultimately absenteeism in the workplace. Future surveys of leadership style must also examine theeffects of group processes. Understanding how leadership behaviour impacts the reaction of a group, suchas in a unionized environment, group, or organizational process is critical to benefiting from the results ofthis paper.119

Appendix AATTITUDES REGARDING LEGITIMATE AND ILLEGIMATE USE OF SICK TIMESDNAND1 2 3 4 5 6strongly disagree somewhat neither somewhat agreedisagree disagree agree nor agreedisagreeSDNAND1 a I go to work even when I am sick 1 2 3 4 5 62 b Sick time can be used for personal 1 2 3 4 5 6obligations3 a I avoid using sick time even when I am very sick 1 2 3 4 5 64 b I abuse my sick time 1 2 3 4 5 65 a I use as much sick time as necessary 1 2 3 4 5 6to get completely over an illness6 a Only under extreme cases do I use sick time 1 2 3 4 5 67 Approximately, how many sick days do you 1 to 2 3 to 4 5 to 6 7 to 8 9 to 10 11 to 12normally use for illnesses annually?8 How many sick days do you use for other 1 to 2 3 to 4 5 to 6 7 to 8 9 to 10 11 to 12reasons annually?a = legitimate use; b = illegitimate use120

ReferencesAdams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in Experimental SocialPsychology, 2, pp. 267-299. New York: Academic Press.Avolio, B. J., Howell, J. M., & Sosik, J. J. (1999). A funny thing happened on the way to the bottom line:Humour as a moderator of leadership style effects. Academy of Management Journal, 42, 219-227.Barling, J., Weber, T., & Kelloway, E. K. (1996). Effects of transformational leadership training onattitudinal and financial outcomes: A field experiment. Journal of Applied Psychology, 81, 827-832.Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in socialpsychological research: Conceptual, strategic, and statistical considerations. Journal ofPersonality and Social Psychology, 51, 1173-1182.Bass, B. M. (1999). Two decades of research and development in transformational leadership. EuropeanJournal of Work and Organizational Psychology, 8, 9-32.Bass, B. M. & Avolio, B. (1995). MLQ Multifactor Leadership Questionnaire For Research. RedwoodCity, California: Mind Garden.Bycio, P., Hackett, R. D., & Allen, J. S. (1995). Further assessments of Bass's (1985) conceptualization oftransactional and transformational leadership. Journal of Applied Psychology, 80, 468-478.Chadwick-Jones, J. K., Brown, C. A., Nicholson, N. (1973). Absence from work: Its meaning,measurement and control. International Review of Applied Psychology, 22, 137-155.Chadwick-Jones, J. K., Brown, C. A., Nicholson, N. (1982). Social Psychology of Absenteeism. NewYork, New York: Praegar Publishers.Connelly, M. (2003). Chrysler group cracks down on absenteeism. Automotive News, 78, p. 29.Crampton, S. M., & Wagner III, J. A. (1994). Percept-percept inflation of microorganizational research:An investigation of prevalence and effect. Journal of Applied Psychology, 79, 67-76.Dalton, D. R., & Mesch, D. J. (1991). On the extent and reduction of avoidable absenteeism: anassessment of absence policy provisions. Journal of Applied Psychology 76, 810-817.D’Aveni, R. A. (1994). Hypercompetition. New York: Free Press.DeMeuse, K. P. (1985). A compendium of frequently used measures in industrial/organizationalpsychology. The Industrial-Organizational Psychologist, 23, 53-59.Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use ofexploratory factor analysis in psychological research. Psychological Methods, 3, 272-299.Gakovic, A. (2003). Social exchange, transformational leadership, and employee behaviour: The role oforganizational identification and career commitment. Dissertation Abstracts International, 63, p.4404.Gale, S. F. (2003). Sickened by the cost of absenteeism, companies look for solutions. WorkforceManagement, 82, 72-73.Gaudine, A. P., & Saks, A. M. (2001). Effects of an absenteeism feedback intervention on employeeabsence behavior. Journal of Organizational Behavior, 22, 15-29.Grinyer, A., & Singleton, V. (2000). Sickness absence as a risk-taking behaviour: A study oforganisational and cultural factors in the public sector. Health, Risk & Society, 2,7-21.Hackman, J. R. & Lawler, E. E. (1971). Employee reactions to job characteristics. Journal of AppliedPsychology, 55, 259-286.Hemingway, M. A. & Smith, C. S. (1999). Organizational climate and occupational stressors as predictorsof withdrawal behaviours and injuries in nurses. Journal of Occupational and OrganizationalPsychology 72, 285-299Hinrichs, J. R. (1986). A replicated study of job satisfaction dimensions. Personnel Psychology, 479-503.121

Howell, J. M., & Hall-Merenda, K. E. (1999). The ties that bind: The impact of leader-member exchange,transformational and transactional leadership, and distance on predicting follower performance.Journal of Applied Psychology, 84, 680-694.Jorgensen, D. O., Dunnette, M. D., & Pritchard, R. D. (1973). Effects of the manipulation of aperformance-reward contingency on behavior in a simulated work setting. Journal of AppliedPsychology, 57, 271-280.Kelly, L. (1990). Understanding Absenteeism. Worklife Report, 7, 7-10.Koh, W. L., Steers, R. M., & Terborg, J. R. (1995). The effects of transformational leadership on teacherattitudes and student performance in Singapore. Journal of Organizational Behavior, 16, 319-333.Kohler, S. S., & Mathieu, J. E. (1993). An examination of the relationship between affective reactions,work perceptions, individual resource characteristics, and multiple absence criteria. Journal ofOrganizational Behavior, 14, 515–530Locke, E. A. (1976). Nature and causes of job satisfaction. In M.D. Dunnctte (Ed.). Handbook ofIndustrial and Organizational Psychology, pp. 1297-1349. Chicago: Rand McNally.Lofquist, L. G., & Dawis, R. V. (1969). Adjustment to Work. New York: Appleton-Century-Crofts.Lowe, G., & Schellenburg, G. (2003). Correlates of employees' perceptions of a healthy workenvironment. American Journal of Health Promotion, 17, 390-399.Mackenzie, S. B., Podsakoff, P. M., & Rich, G. A. (2001). Transformational and transactional leadershipand salesperson performance. Journal of the Academy of Marketing Sciences, 29, 115-135.McShane, S. L. (1984). Job satisfaction and absenteeism: a meta-analysis re-examination. CanadianJournal of Administrative Science, 1, 61-77.Nicholson, N. J. G. (1982). The absence culture and the psychological contract: Who's in control ofabsence? Academy of Management Review, 10(3), 397-407.Oi-ling, S. (2002). Predictors of job satisfaction and absenteeism in two samples of Hong Kong nurses.Journal of Advanced Nursing, 40, 218-229.Onyett, S., Pillinger, T., & Muijen, M. (1997). Job Satisfaction and Burnout Among Members ofCommunity Mental Health Teams. Journal of Mental Health, 6, 55-66.Piliai, R., & Williams, E. A. (2004). Transformational leadership, self-efficacy, group cohesiveness,commitment, and performance. Journal of Organizational Change Management, 17, 144-160.Piliai, R., Schriesheim, C. A., & Williams, E. S. (1999). Fairness perceptions and trust as mediators fortransformational and transactional leadership: A two-sample study. Journal of Management, 25,397-444.Pritchard, R. D., & DeLeo, P. J. (1973). Experimental valence-instrumentality relationship in jobperformance. Journal of Applied Psychology, 57, 264-270.Scott, K. D., & McClellan, E. (1990). Gender differences in absenteeism. Personnel Management,19,229-253.Scott, K. D., & Taylor, G. S. (1985). An examination of conflicting findings on the relationship betweenjob satisfaction and absenteeism: a meta analysis. Academy of Management Journal, ?? 599-612Severens, J. L., Mulder, J., Laheij, R. J. F., & Verbeek, A. L. M. (2002). Precision and accuracy inmeasuring absence from work as a basis for calculating productivity costs in The Netherlands.Social Science & Medicine, 51, 243-249.Steel, R., & Rentsch, J. (1995). Influence of cumulation strategies on the long-range prediction ofabsenteeism. Academy of Management Journal, 38(6), 1616-1634.Vroom, V, H. (1964), Work and motivation, New York, Wiley.Wall, T. D., Michie, J., Patterson, M., Wood, S. J., Sheehan, M., Clegg, C. W., & West, M. A. (2004). Onthe validity of subjective measures of company financial performance. Personnel Psychology, 57,95-118.122

Wall, T. D., & Wood, S. J. (2005, March). The romance of HRM and business performance: The case forbig science. (Institute of Work Psychology Working Paper No. 354). Sheffield, UK: University ofSheffield.Warr, P., Cook, J., & Wall, T. (1979). Scales for the measurement of some work attitudes and aspects ofpsychological well-being. Journal of Occupational Psychology, 52, 129-148.Xie, J. L. (2000). Interactive effects of absence culture salience and group cohesivness: A mult-level andcross-level analysis of work absenteeism in the Chinese context. Journal of Occupational &Organizational Psychology, 73(1), 31-52.Yukl, G. A. (2002). Leadership in Organization (5 th edition). Upper Saddle River, New Jersey: Prentice-Hall 2002.Zaccaro, S. J. (1991). Prior absenteeism, supervisory style, job satisfaction, and personal characteristics: An investigationof some mediated and moderated linkages to work absenteeism. Organizational Behavior & Human DecisionProcesses, 50, 24-41.Zedeck, S. (1987). Satisfaction in union members and their spouses. Paper presented at the JobSatisfaction: Advances in Research and Practice Conference, Bowling Green, Ohio.123

ASAC 2006Banff, AlbertaChristopher Fredette (student)Joseph Krasman (student)Schulich School of Business, York UniversitySTRAIGHT FROM THE SOURCE’S MOUTH:EXPLORING THE IMPLICATIONS OF FEEDBACK SEEKING ON THE SOURCEThis paper explores the effect feedback-seeking has on the person from whom feedbackis sought – the “source”. An exploratory semi-structured qualitative methodology wasapplied, sampling 49 professionals, yielding a thematic framework integrating 27emergent themes into four conceptual domains. Findings indicate that both positive andnegative responses may be categorized along two conceptual axes – asking versus givingand functional versus relational.IntroductionFeedback-seeking behaviour (Ashford, Blatt & VandeWalle, 2003; Ashford & Cummings, 1983)refers to the proactive search by individuals for evaluative information about their performance. Theconcept was introduced by Ashford and Cummings (1983) over 20 years ago to refute the traditional viewof feedback as information that individuals only passively receive at the behest of the organization, forexample, during the performance review (Ilgen, Fisher & Taylor, 1979). Since that time, research onfeedback-seeking has proliferated (for reviews see Ashford et al., 2003; Madzar, 1995; Morrison, 2002).Researchers have identified numerous factors that influence people’s feedback-seeking behaviour. Thesehave included characteristics of the individual, characteristics of the source or feedback provider, andcharacteristics of the context in which the feedback-seeking exchange takes place. In addition, researchershave demonstrated the positive consequences associated with feedback-seeking. These have includedincreased job satisfaction and job performance as well as decreased turnover intentions (Morrison, 1993;Morrison & Weldon, 1990).Taking a step back and looking at the phenomenon from a structural perspective reveals thatfeedback-seeking is an interpersonal exchange. One party serves as the seeker and another party serves asthe source. Research has shown that feedback sources traditionally include one’s supervisor, coworkers,and subordinates (Ashford, 1993; Greller & Herold, 1975). The seeker asks the source for feedback andthe source chooses the mode of response, responding with evaluative performance information or otherdesirable materials. To date, all of the research energies have focused on the seeker side of the exchange,where the seeker asks the source for feedback and the source responds. Critically, there has not been anyresearch on what feedback sources experience when being asked for, and responding to, feedback-seekingrequests.We believe that answering this question is crucial for two key reasons. First, if feedback-seekingis to be promoted in the workplace, it is imperative to understand how sources are affected by feedbackseekingbehaviour. Second, the experience sources have in the feedback-seeking exchange may be aphenomenon worthy of investigation in its own right. For example, as will be discussed, in the mentorshipliterature, researchers have begun to examine how mentors – rather than protégées – experiencementorship interactions. Similarly, in the helping literature, researchers have started to investigate howhelp givers – instead of help-seekers – experience helping encounters.124

Existing and Related LiteratureWhile the effects of feedback-seeking on the source have not been examined, similar questionshave been addressed in other, related fields. One prominent example is the mentorship literature.Originally, mentorship was considered a phenomenon that applied strictly to the protégée. Over time,however, scholars began to consider its effects on the mentor as well; first conceptually and thenempirically. For example, some proposed that mentorship enables people who have reached a plateau intheir mid-life transition, to experience satisfaction from making productive use of their accumulated skillsand wisdom (Levinson, Darrow, Klein, Levinson & McKee, 1978). Halatin (1981) argued that mentorsshould experience feelings of self-importance and satisfaction from helping protégées progress towardtheir professional goals. He also cautioned that mentors may feel exploited if over-consulted by protégées.Halatin and Knotts (1982) noted that mentorship is resource intensive, demanding a lot of time andenergy from the mentor. It was not until the 1990s, however, that empirical work on the experience ofmentors began to appear. Interestingly, Raggins and Cotton (1993) examined factors that influencepeople’s willingness to serve as a mentor such as their gender, rank, and age. Raggins and Scandura(1994) identified 17 costs and 24 benefits of being a mentor, whereas Burke, McKeen and McKenna,(1994) examined how being a mentor affects one’s job satisfaction and performance. Finally, Allen(1997, 2003) identified 19 micro-motives and three macro-motives for a person deciding to become amentor.Similarly, the helping literature offers another example of this form of theoretical development.While the majority of research in this area has focused on the help seeker, a few works have examined theeffects help-seeking has on the help-giver. Rudolph, Roesch, Greitemeyer and Weiner (2004) showed thathelp-givers can experience positive or negative emotions from the act of being asked for help. The type ofemotion help-givers experience was shown to depend on the causality that he or she attributes to themisfortune of the help-seeker. As a result, when the help-giver attributes the cause to be within the controlof the help-seeker (an internal attribution), he or she experiences negative emotions such as anger.Alternatively, when the help-giver attributes the cause to lie beyond the control of the help-seeker (anexternal attribution), he or she experiences positive emotions such as sympathy.The theme of this year’s conference called for “considering the well worn paths around ourdisciplines” and “seeking higher ground”. A well worn path in the feedback-seeking literature has beenthe exclusive examination of the topic entirely from the seeker’s perspective. Inspired by this year’stheme, we carried out an exploratory study to examine the effect feedback-seeking has on the source and,in particular, what sources like and dislike about being asked for feedback. In the section that follows, wedescribe our methodology and analysis. In further sections of this paper we explore the thematiccategories of ‘likes’ and ‘dislikes’ that emerged from our data, in addition to explaining the dimensionsthat underlying these categories. We conclude our paper with a discussion of future research directions.125

Research Tradition and MethodologyQualitative data for this study was collected as part of a larger field study conducted using aquantitative self-report survey instrument. The larger data collection instrument focused on collectinginformation about perceptions of the costs and benefits of seeking feedback from the seeker’s perspectiveusing a series of multiple item measures composed of individual items measured with Likert-style scales.In contrast, our qualitative data was collected using a pair of open-ended questions, which respondentswere prompted to answer after they had completed the quantitative materials. The questions wereformatted such that respondents were presented with the first question followed by a series of five distinctthought-spaces in which to freely respond, the second question was then presented in the same fashion.Taking the perspective of the feedback source, participants were encouraged to provide their thoughts andopinions in as much detail as possible.This data collection methodology is consistent with the central tenets of semi-structuredinterviewing, as it provides the opportunity for respondents to interpret and respond to each questionindependently and in their own words. Moreover, whereas interviewing techniques require interviewerintervieweeinteraction that may introduce answer-satisficing based on social desirability or a desire toanswer quickly, our methodology afforded participants with additional independence and anonymity. Inour opinion, increasing interviewee independence allows respondents to consider each question in length,reflecting on their opinions from multiple dimensions and provides the opportunity to offer multipleperspectives without having to integrate or coordinate multiple divergent thoughts into a single metaframework.Similarly, improving anonymity offers participants an added measure of privacy, allowing forthe free expression of alternative views and unorthodox thoughts without challenge or need forjustification.SampleOur sample consisted of 49 respondents selected based on their participation in a larger studyfocused on feedback seeking behaviour and the perception of personal costs associated with seekingfeedback from others. The average demographic profile of our participants is characterized as 41%female, averaging 35.4 years old, with 7.1 years of organization tenure and 4.5 years of job tenure. Selfreportedcurrent employment ranged from entry level to senior management, with the vast majority 66.7%indicating positions of middle management or beyond. Current positions involving subordinateinteraction indicate ample real-world experience providing both formal and informal feedback to others,thereby demonstrating a relevant familiarity, a qualification necessary to participate in our investigation.MethodIn selecting an appropriate research methodology to implement in our investigation, we sought amethod which would satisfy our requirements for an independent participant-driven approach to thecollection of free expression in a new and emerging area of study. From this perspective, applying aCritical Incident Technique (CIT) approach was an appropriate choice for the collection of vivid thoughtsand expressions which could later generate important themes used in the development of theory (Gremler,2004). The flexibility inherently associated with Critical Incident Techniques allowed us to capitalize onthe distinct advantages of this method in a number of ways (Flanagan, 1954). First, CIT offers researchersa wider lens through which to observe real participant-driven experiences, rather than restrictingobservation to a limited set of variables or concepts (Walker & Truly, 1992). Second, CIT is an inductiveapproach that focuses on generating concepts or theories from patterns which emerge directly fromrespondents, rather than focusing on hypothesis testing or validation of a preconceived model (Olsen &Thomasson, 1992). Third, using a CIT approach limits much of the researcher’s influence on participantresponses by inviting respondents to share their perceptions on an issue or topic, rather than rating126

participant perceptions of a series of researcher-initiated questions (de Ruyter, Perkins & Wetzels, 1995).Finally, undertaking data analysis using a grounded theory-based content analysis strategy is altogetherconsistent with the past applications of Critical Incident Techniques as documented across a variety offields and functions (Gremler, 2004).Functionally, our approach to the collection of participant data was both practical and pragmatic,involving a series of distinctly focused activities. Working with our sample, selected on the basis ofconvenience, we posed the following two questions: first, ‘What do you like about giving feedback?’ andsecond, ‘What do you dislike about giving feedback?’ Consistent with the previously discussed format,respondents were provided a series of five distinct thought-spaces in which to freely respond to each openended question. Once collected, each thought-space was transcribed verbatim, reflecting eachparticipant’s statement in both content and intent. Upon completing the transcription process, eachthought-space was individually separated resulting in a total of 341 statements (184 ‘likes’ and 157‘dislikes’) generated from 49 participants, averaging a total of 6.95 statements per respondent (3.75‘likes’ and 3.20 ‘dislikes’ per respondent). We then separated the ‘like’ statements from the ‘dislikes’, andbegan the process of content analysis. Consistent with the principles of grounded theory, each authorapproached the data with no preconceived theoretical frameworks or hypotheses, instead allowing themesto emerge from participant writings (Strauss & Corbin, 1998). Working individually, both authorsinitially focused our energies on open coding the work-bench level content of each statement (Strauss &Corbin, 1998:101), dividing the materials into practical yet consistent categories. Later, axial coding ledus to assess the comparability of each work-bench category’s similarity and difference (Strauss & Corbin,1998:123), which organized our practical themes into emergent emic concepts – the basis for theoreticalintegration. Pragmatic adherence to this method allowed us to approach the analysis of our datasystematically, while simultaneously affording us the opportunity to surface and challenge our ownassumptions, thus grounding us in the expressed experience of our participants.Likes of Being Asked for FeedbackAlthough all study participants received indistinguishable research materials and were presentedwith identical questions, participant responses quickly deviated from the strict interpretation of “What doyou like about being asked for feedback?” and moved toward a broader understanding of their experiencewith feedback processes. To this end, our presentation of the findings associated with the “likes”represents the full spectrum of participant thought-spaces, illustrating four unique perspectives whichemerged during our iterative analysis. These four perspectives are based on the emergence of twocategorical fault lines which are reflected in the passages contained in each respondent thought-space.The first is based on the distinction between asking and giving, in which respondents spoke either interms of being asked for feedback or alternatively in the act of giving feedback. In contrast, the seconddistinction is derived from the participant’s conception of the object of conversation. In this regard, eachrespondent-expression illustrated either a relational interaction or a functional interaction, demonstratingthe use of one to the exclusion of the other. Although seemingly simplistic, the emergence of thiscategorical framework established a fundamental foundation upon which we will build throughout thecourse of this paper.127

Table 1: Likes and Dislikes of Being Sought for FeedbackRelationalBeing Asked for FeedbackLikes (total=44) Feeling valued for myopinion. (24) Feeling valued for myexpertise. (10) Feeling respected for myabilities. (10)Dislikes (total=32) Coping with remaininghonest. (11) Insufficient feedback. (11) Perpetuating unhealthydependence. (10)Giving FeedbackLikes (total=45) Being helpful by offeringassistance. (20) Feeling like part of theteam. (16) Encouraging and praisingothers. (9)Dislikes (total=51) Fear of giving negativefeedback. (26) Not adhering to feedback.(13) Fear of being evaluated.(12)FunctionalLikes (total=41) Monitoring task-basedinformation. (17) Exploiting opportunities tobe heard. (14) Communicating to shareideas and information. (10)Dislikes (total=38) Ulterior motivation. (20) Abrupt or unexpectedasking. (9) Bad timing. (9)Likes (total=54) Testing the value ofpersonal knowledge. (12) Improve knowledge basesof seekers. (17) Contribute to a solution.(12) Leverage opportunities toexercise influence. (13)Dislikes (total=36) Takes too much time. (18) Attracts unwantedaccountability. (10) Being objective. (8)Being Asked for Feedback – Relational ResponsesIntuitively, it makes sense that people would reflect positively on being asked to provide theirideas, input and opinions. Consistent with this notion, our respondents communicated a strong sense ofpersonal value or worth associated with being asked for feedback by others. In sum, 44 of 184 responsesreflected feelings of positive self-value in which the source indicated feeling good about themselves as aresult of others’ request for feedback. Respondents offered a variety of statements ranging in focus fromthe very broad “Makes me feel that my input is valued.” to the very focused “I like that the person askingme for feedback thinks of me in a superior way. Considers me knowledgeable and experienced.” While itis apparent that both statements are self-reflecting, it is equally evident that all statements of this genre arenot homogeneous in nature. In this regard, thematic analysis revealed the emergence of three similar yetdistinct categories associated the notion of feeling good about being asked for feedback. These includedfeeling valued for my opinion, feeling valued for my expertise, and feeling respected for my abilities.128

Feeling Valued for my Opinion. By far the most prominent sub-dimension in this area, 24 of 44respondent expressions indicated a positive feeling of value associated with others seeking feedbackinformation from them. Whereas some respondents simply stated, “Flattering”, others expressedthemselves in more depth, “That they value my opinion” or “It makes me feel important”, still others gavemore elaborate responses that offered insight into the underlying reasons for their positive response to therole of source. In this regard, particularly insightful contributions were offered by more articulateparticipants, “Person who is asking for feedback is looking and interesting in my personal opinion – andthis is important to me, because my thoughts are counted.” Similarly, statements such as “Gives you afeeling of self worth and importance.” or “It shows that my opinion is valued and that I am appreciatedfor my feedback.” illustrate the nature and scope of the this dimension. In our pool of respondents,feedback sources clearly communicated a sense of personal value and perhaps validation associated withthe seeking activities of others.Feeling Valued for my Expertise. Perhaps less prominent than the previous category,expressions of a personal sense of value resultant from recognition of the source’s knowledge, skills andabilities was no less clear. A number of respondent reflections (10 of 44) clearly provided an impressionof value derived with being asked to provide feedback of a technical or task-specific nature. Exemplars ofthis type of statement include “I like the fact that a coworker would look at me to asked about feedbackbecause it reflects my knowledge.”, “It is nice to be used as a resource. Seen as knowledgeable.” and “Ilike the way I feel when people solicit my advice. Again, is a motivation for me to improve myknowledge and be able to answer their questions.” In contrast to the previous sub-dimension, statementscategorized in this theme demonstrated feeling a sense of personal value associated with others’perception of their knowledge. In this regard, one participant expression is particularly striking, “Beingasked validates you and makes you feel valued for your knowledge and expertise.”Feeling Respected for my Abilities. Although represented by fewer participant reflections thanthe first category, this theme is composed of ten (10 of 44) statements clustered together based on thenotion of others’ respect without any indication of affect. Rather than indicating a positive sense of selfvalue,respondent expressions in this category were oriented toward capturing the deference of feedbackseeking others. In this regard, typical participant statements were relatively neutral in orientation,revealing a preference for respect as illustrating in the following accounts: “That individual has respectfor my opinion”, “Giving my opinion when someone respects my experience and is requesting help.” and“Person respects your opinion.” In contrast to more neutral statements, one participant thought-spaceincluded a more extreme illustration of the importance of deference in the feedback source-seekerrelationship, “I feel superior to the person or group of people being evaluated.” Regardless of the degreeto which respondent statements emphasized seeker-deference to the source, there appears to be significantdistinction between feeling-valued responses and those focusing on respect. This point is particularlyevident in another respondent statement “That the person believed that I would be a good source offeedback – this person obviously respects my opinion.”The Act of Giving Feedback – Relational ApproachesBased on our request to indicate what respondents liked about being asked for feedback, manyparticipants introduced statements which focused on the relational benefits of giving feedback to thosewho seek it, in addition to reflections about the value of being asked their ideas, input and opinions.Indicative of this distinct response orientation, participant thought-spaces included statements such as “Togive positive feedback reinforces a lot of assurance in the person that what he/she is doing is on the righttrack this in fact may cause they to strive to achieve more.” and “Get to know your colleagues andwhat they’re going through – i.e. builds rapport – recognizes strengths and weaknesses” which clearlyillustrate an alternative outlook. In total, 45 of 184 responses illustrated a thematic concept characterized129

oth as feedback-giving and relationally focused. Within this perspective, thematic analysis revealed theemergence of three similar yet distinct categories associated the notion of feeling good about givingothers feedback. These included being helpful by offering assistance to those who seek it, feeling like partof the team, and giving positive feedback to encourage and praise others.Being Helpful by Offering Assistance. Analysis of participant statements, both in terms ofnumber (20 or 45) and in terms of content, demonstrated that individuals enjoyed giving feedback toothers when it was helpful or offered assistance to the seeker. Straight forward statements like “It helpssomeone improve so I feel better for having helped” illustrate that sources derive value, and therebybenefiting from providing seekers with feedback. Similarly, statements such as “Knowing that I may helpsomeone work through a difficult situation.” or alternatively “It gives me a sense of accomplishment thatI can help someone else succeed” further reinforce the notion of inherent benefits resulting from givinghelpful feedback to seekers. In response to our question regarding feedback ‘likes’, one thought-spacesuccinctly read, “Being helpful”. The critical factor that differentiates being helpful (relational helping)from other dimensions presented in our investigation, emerged during a process of iterative reflection inwhich the authors recognized that relational helping lacked an instrumental undercurrent. The findingresonated strongly with both authors and will become clear as our discussion progresses.Feeling like Part of the Team. Participant thought-spaces illustrating this thematic dimensiontotalled 16 of the 45 expressions included in this category. Respondent expressions ranged in scope fromsimple statements like “Opportunity to deepen work relationship” to more complex thoughts such as“That the feedback was used and had positive outcomes and created linkages.” Although somewhatvaried in description, respondent remarks provide insight into the relational nature of seeker-sourcefeedback exchanges, illustrating a feeling of inclusiveness associated with act of giving and receivingfeedback. This perspective is consistent with the notion presented in ‘feeling valued for my opinion’,however it is applied from a different point of view and with a differing focus, as illustrated in thefollowing respondent passages “When being asked for feedback, I’m part of the team.”, “The relationwith team members. The cooperation and information sharing between team members.” and “That the endresult would be beneficial to the parties involved”. In our pool of respondents, feedback sources clearlycommunicated a sense that giving others the feedback they sought was a valued indicator of associativestrength and team inclusion.Encouraging and Praising Others. Represented by only a limited number (9 of 45) of thoughtspaceexpressions, this dimension is expectedly one of the most homogeneous in our investigation. Tothis end, participant statements centred on using positive reinforcement to make a positive difference inthe seekers’ self-conception, situational understanding or performance expectations. Predominantexamples of the use of praise and encouragement reported by respondents include “Being able to givepositive encouragement to my peers”, “Providing someone with the recognition they deserve for a jobwell done” as well as “To give positive feedback reinforces a lot of assurance in the person that whathe/she is doing is on the right track.” In addition, one exemplar captures the essence of this dimensionwith particular concision, “It allows me to make the other person feel good by giving good praises onexcellent performance.” Clearly, this illustration emphasizes the way in which the act of giving feedbackis an integral contributor to the strength of established seeker-source relationships.Being Asked for Feedback – Functional ResponsesDifferentiating the participant responses on the basis of orientation – function versus relational –is reflective of the way in which respondents classified and organized their replies. This form ofdifferentiation indicated to us that respondents recognized that the seeker-source feedback process is amulti-dimensional phenomenon, in which similar means may be used to reach differing ends. From thisperspective, the feedback process takes on an instrumental flavour where participants express the ways in130

which seeker-source linkages may be used to facilitate task or performance based endeavours. Consistentwith this notion, our respondents provided 41 (of 184) incidents that demonstrate functional responses torequests in which sources were asked for feedback by others. These responses are based on participantstatements that illustrate both a functional approach as well as an ‘asked for feedback’ orientation.Respondents offered a variety of statements ranging in focus from the very broad “An opportunity to giveimportant input” to the very focused “Get to know what’s going on within the company in terms ofassignments and projects outside of your area.” While it is apparent that both statements reflect functionalpurposes, it is equally evident that the statements of this genre are not homogeneous in nature. In thisregard, thematic analysis revealed the emergence of three similar yet distinct categories associated withfunctional responses to being asked for feedback. These included monitoring and collecting task-basedinformation, exploiting opportunities to be heard, and communicating to share ideas and information.Monitoring Task-Based Information. Consisting of 17 (of 41) participant statements, the focusof this category centres on the need to collect and decipher information within an organizational context.Participant comments revealed functional responses to requests for feedback included “Provides informalstatus update of where projects stand.” In addition, other respondents noted the ability to monitor “Formof monitoring ongoing projects – management of projects within department”, gather facts “It providesme with current new that might be happening” and to establish objectives “set up goals and follow up” asamong the things they appreciated about being asked for feedback. One participant provided aneloquently written exemplar of this dimension, “In the new organizational structures they are often veryflat and ‘reporting’ lines are fuzzy often working with others on projects appear to be asking for favours –therefore without being asked for feedback there are not opportunities to get the best work done unlessparticipants voluntarily ask for feedback.”Exploiting Opportunities to be Heard. “A chance to voice out my opinion” or “The opportunityto provide my opinion on whatever issue that is being asked” best characterize this emergent theme. Inthis instance, this dimension is represented by a total of 14 participant expressions, indicating that those inthe position of feedback source believe that they are afforded the opportunity to contribute their opinionsor thoughts when seekers request feedback. The statements which clustered to support this theme all sharetwo important points: first, the notion of chance or opportunity provided by the feedback request; and,second the ability to command the attention of others. Further exemplars include “Being able to raiseimportant points that would otherwise not occur”, “A chance to be honest if changes need to be done” and“It gives me a chance to express myself.” This is a particularly interesting finding as it suggests that somesources may be willing to use feedback channels as a means of forwarding other issues or agendas.Communicating to Share Ideas and Information. Although not an overly novel finding toemerge from our thematic analysis, no less than 10 of 41 respondent comments emphasized using theseeker-source feedback exchange as a means of communicating ideas and information. From thisperspective, participant compositions focused on feedback as a mode of communication for learning“May be able to learn something new” or for reflection “Gives you a chance to temporarily think aboutanother aspect of the company’s functions”. In addition, others illustrated the utility afforded by seekersourcediscourse, suggesting that they liked being asked for feedback as it provided an “opportunity toshare experience” as well as an “invitation to start communication that can lead to better mutualunderstanding of what should be done and what way”. The notion that increases in communication iscommensurate with new idea creation was underscored as a further functional opportunity resulting fromthe feedback seeker-source relationship.The Act of Giving Feedback – Functional ApproachesBased on our request to indicate what respondents liked about being asked for feedback, manyparticipants produced statements which focused on the functional benefits of giving feedback to those131

who seek it. Consistent with the previous functional approach, respondents’ demonstrated recognition ofthe seeker-source feedback dynamic as a multi-dimensional phenomenon. Indicative of this distinctresponse orientation, participant thought-spaces included statements such as “It allows me to learn morejust by looking over someone else’s work – maybe I will find errors that I could have made myself” and“Assisting individuals with certain tasks that increase their understanding or assist in accomplishing afirm wide goal” which clearly illustrate an alternative outlook. In total, 54 of 184 responses illustrated athematic concept characterized both as feedback-giving and functionally focused. Thematic analysisrevealed the emergence of four distinct categories associated with functional approaches to giving othersfeedback. These included using feedback to test the value of personal knowledge, improve theknowledge-bases of seekers, contribute to a solution and take advantage of opportunities to exerciseinfluence.Testing the Value of Personal Knowledge. Emergent from our data analysis was the notion ofusing channels of feedback as a means to test and challenge the source’s knowledge, expertise andunderstanding of the environment in with he or she operates. This category was comprised of 12respondent statements from a total of 54 expressions. Comments such as “I find that I have an opportunityto exercise skills in assessment that I do not regularly use at work”, “Giving feedback gives me a changeto improve my interpersonal skills and professional tactfulness on how to provide constructive criticism”and “Allows me to correct and improve my work performance” are consistent with the concept of usingfeedback processes as an instrument against which the source may gauge comparative knowledge, skilland ability sets.Improve Knowledge-bases of Seekers. On the basis of performance-oriented improvement,participants generated 17 reflections which centred on feedback as a means of improving the performanceresults of seekers. Statements consistent with this perspective include, “I like providing constructivecriticism as it makes (hopefully) for better workers” or “I like to see people asking for feedback as itshows they care about doing things right and are willing to improve and learn from it”. Exemplarsrepresentative of this dimension focus on seeker-source feedback dynamics as a means of transmittingperformance enhancing information, as reflected in “Being able to communicate what was done right andwrong, so that expectations are clearly known for next project” and “The attitude of the individualtowards new things. The readiness to take on new technologies”.Contribute to a Solution. By focusing on the concept of being helpful from an instrumentalperspective, respondent statements such as “Increases efficiency of ‘getting the job done’” or “I likeproviding feedback in my area of expertise as it ensures things are done correctly” illustrate an air ofintervention within the framework of feedback giving activities. Support for this perspective wascomprised of 12 respondent generated statements, consistent with the means-ends approaches outlinedabove. In this regard, the following reflection is particularly insightful with respect to the notion ofintervention, “Important to give my feedback especially if a product or service is important to mybusiness”.Leverage Opportunities to Exercise Influence. Perhaps most reflective of ways in whichrespondents understood seeker-source feedback relationships as having multiple dimensions are thestatements which expressed the potential to make change by influencing others. As expected, none of the13 participant statements contained any reference to use of feedback as a means of manipulating theopinions or activities of seekers. However, statements such as “A chance to make change, if keep silent,nothing that I want to see will happen”, “The opportunity to influence the direction of a project with yourideas” and “Opportunity to influence another person” clearly reflect an assumption that feedback is anappropriate avenue through which to influence others and engage in change activities. This is clearlycontradictory to many of the ‘likable’ facets of feedback processes discussed in the relational, and yetboth dimensions appear to be able to co-exist simultaneously.132

Dislikes of Being Asked for FeedbackConsistent with the approach outlined in the discussion of ‘likes’ findings, all study participantsreceived indistinguishable research materials and were presented with identical questions. In our analysis,we found that participant responses quickly deviated from the strict interpretation of “What do you dislikeabout being asked for feedback?” and moved toward a broader understanding of their experience withfeedback processes. Similarly, our explication of the findings derived from the ‘dislikes’ data collectionrepresent the collective range of participant thought-spaces, and illustrate counter-points to the fourperspectives which emerged during our previous analysis. Again, these four perspectives are based on theemergence of two categorical fault lines which are reflected in the passages contained in each respondentthought-space. The first is based on the distinction between asking and giving, in which respondentsspoke either in terms of being asked for feedback or alternatively in the act of giving feedback; whereas,the second distinction results from respondent-expressions illustrating either a relational or a functionalnature, thereby using one frame to the exclusion of the other. Throughout our analysis, the emergence ofthis analytic framework established a fundamental foundation upon which we continue to buildthroughout the course of this paper.Being Asked for Feedback – Relational ResponsesWhile, it makes sense that people would feel positively about being asked to provide their ideas,input and opinions, when asked about their dislikes, our respondents communicated a sense ofapprehension associated with being asked for feedback by others. In sum, 32 of 157 responses reflectedfeelings of personal concern in which the source indicated feeling uneasy about themselves as a result ofothers’ request for feedback. Respondents offered a variety of statements ranging in focus from the verybroad “I dislike when I am being asked questions about my managers and I feel that I can not becompletely honest” to the very focused “The ‘responsibility’ of feeling that, when asked for a solution toa problem, one must come up with an answer. A feeling of helplessness or letting them down when onecan’t”. While both statements are self-reflecting and deal with relational rather than functional concerns,the differences between them are representative of the heterogeneity among the statements of this genreilk revealed in our analysis. In this regard, thematic analysis revealed the emergence of three similar yetdistinct categories associated the notion of feeling uneasy about being asked for feedback. These includedcoping with remaining honest in the face of undesirable questions, fearing that he or she will not be ableto provide sufficient feedback and perpetuating unhealthy dependence.Coping with Remaining Honest. “I dislike when I am being asked questions about my managersand I feel that I not be completely honest” is a respondent statement that encapsulates much of the contentcontained in the 11 thought-spaces upon which this theme based. Many of the respondent statementsillustrated their concern with being asked questions which could pressure them to compromise theirintegrity, such as “May be hard to be honest. Do not want to hurt the feelings of others.” In addition,some respondent statements went so far as to mention strategies they utilize in order to cope withchallenging questions. One strategy mentioned by a participant was manufacturing positive comments,“… say something good even if half-true …”Insufficient Feedback. An important theme to emerge from our data is the source’s concern withnot providing seekers with sufficient feedback to assist them. As one respondent stated “May not be ableto provide any valuable feedback which can help the person asking”. Similarly, another expressed“Employee may be wondering if info provided will satisfy the person asking for feedback”. In total, thistheme was derived from 11 participant statements, all indicating concern that their contribution would failto satisfy the seekers’ needs or expectations.133

Perpetuating Unhealthy Dependence. A further theme that emerged from the data, composedof 10 respondent statements, was the source’s concern with seekers becoming overly reliant on them forfeedback. For example, respondents made comments such as “You do not want to become someone’scrutch”, “May be propping up another employee”, “The individual repeatedly comes back with similarrequests since they were able to get some feedback in the past” and “May lead to the person repeatedasking you for feedback in the future”. Not only were sources concerned with perpetuating a state ofover-reliance, but they were also concerned with seekers losing their ability to resolve their performancerelatedissues on their own. For example, respondents made comments such as “…inhibits free thinking”and “…you want others to think independently”.The Act of Giving Feedback – Relational ApproachesBased on our request to indicate what respondents disliked about being asked for feedback, manyparticipants introduced statements which focused on the relational consequences of giving feedback tothose who seek it, rather than focusing on being asked. Indicative of this distinct response orientation,participant thought-spaces included statements such as “Not everyone takes feedback positively. Somepeople asked for it but never talk to you again.” and “I dislike when I spend my time answering questions,but the person asking doesn’t pay attention or does not make good use of the information.” which clearlyillustrate this alternative outlook. In total, 51 of 157 responses illustrated a thematic concept characterizedboth as feedback-giving and relationally focused. Within this perspective, thematic analysis revealed theemergence of three dominant categories illustrating concerns relating to giving others feedback, each issimilar in orientation yet distinct in focus. These concerns included a fear of giving negative feedback,that feedback would not being followed, and a fear of being judged based on feedback given. In addition,a limited number of respondent comments demonstrated concerns that the feedback provide would not begiven credit in the future.Fear of Giving Negative Feedback. Perhaps the most dominant theme to emerge from this dataset was the sources’ dislike with having to give seekers negative feedback. This category is representedby 26 participant statements, however, the greater consequences and implications associated withproviding seekers with negative feedback can be seen throughout the discussion of the findings in the‘dislikes’ section of this paper. Whereas some respondent statements provided a general response“Having to give negative feedback”, others elaborated on the reasons underlying their aversion.Participant statements illustrated discomfort with delivering negative feedback, stemming from notwanting to cause emotional distress to seekers. Comments such “May hurt the person’s feelings”, “Theymay get offended”, “May seem disrespectful of the job they’ve been doing”, and “Person may feeldiscouraged” illustrate this view. In addition, other respondent comments outlined an aversion to tellingseekers they have performed poorly stemming from not wanting to damage the preexisting seeker-sourcerelationship. For example, one respondent wrote “Fear of losing rapport with the person being evaluated”while another mentioned “It can potentially disrupt my otherwise good relationships”.Not Adhering to Feedback. A second theme to emerge from within this perspective stems fromthe sources’ dislike with seekers not adhering to their feedback. Many participant statements (13)illustrated that sources felt that if the seekers were going to ask them for feedback they should follow theiradvice. Comments such as “When my feedback does not get incorporated”, “Nothing done about it”,“Will my suggestions be implemented?”, “Risk of being ignored”, and “Listen but no step taken”reflected this concern. One reason why sources did not like their feedback being ignored was that they felttheir time had been wasted. For example, respondents mentioned “Feedback may be discarded, resultingin a waste of time”, “The feedback may not be accepted so I have used my time without value”, and “Idislike when I spend my time answering questions, but the person asking doesn’t pay attention or doesnot make good use of the information”. Another reason why sources did not like having their feedback134

overlooked was that it made them feel like the seeker had not asked for their feedback in earnest. Forexample, respondents mentioned “I don’t like not being taken seriously” and “Not taken seriously”.Fear of Being Evaluated. Respondent statements such as “You are on a measuring scale whenyou are giving feedback to your seniors or managers or even team leads”, “Direct feedback requires whatwill the person think after you give them the feedback (consequences)” and “To have to give negativefeedback to someone my seem disrespectful of the job that they’ve been doing” illustrate a theme thatemerged from our data analysis demonstrating a source’s concern with having their feedback evaluated byothers and attributions made about them as a result. Similarly, respondents made comments such as“…opening myself up to being criticized for my feedback” further demonstrating concern about thejudgments that feedback seekers will make. Researchers of feedback-seeking have long noted theimpression-management concerns that seekers take into account in deciding whether or not to ask forfeedback (Ashford & Cummings, 1983; Morrison & Bies, 1991). For example, seekers do not want toappear anxious, nervous, or incapable to seekers. This data, composed of 12 open-ended participantstatements, suggests that sources have similar concerns with regard to avoiding the appearance ofincompetence or ineptitude in eyes of the seekers, based on the feedback that they provide.Although not supported by a substantial number of participant statements (5), an emergent notionderived from our data comes from an expression of aversion for situations in which seekers do notacknowledge them for feedback when it leads to positive work outcomes. Participant comments indicatefeeling short changed when seekers use their feedback, benefit as a direct result, receive praise andrecognition from others for a job well done, but do not acknowledge the role played by the source’sfeedback. In this regard, respondent statements included, “Occasionally my feedback has beenunrecognized and not appreciated appropriately – other takes credit for new idea”, “Copy idea as theirown”, and “Someone else taking the glory for a job well done because of my input”.Being Asked for Feedback – Functional ResponsesConsistent with the differentiation strategies argued in our discussion of the feedback ‘likes’,participants appeared to classify, categorize and organize their thoughts on the basis of orientation –function versus relational – as reflected in the way in which respondents classified and organized theirthought-statements. This form of differentiation reinforced our belief that respondents recognized that theseeker-source feedback process is a multi-dimensional phenomenon, in which similar means may be usedto reach differing ends. From this perspective, the feedback process takes on an instrumental flavourwhere participants expressed displeasure with many of the ways in which seeker-source linkages could beused to inhibit or obscure task or performance based endeavours. Consistent with this notion, ourrespondents provided 38 (of 157) incidents that illustrate a functional pattern of responses to requests inwhich sources were asked for feedback by others. These responses to ‘dislikes’ are based on participantstatements that illustrate both a functional approach as well as an ‘asked for feedback’ orientation.Respondents offered a variety of statements ranging in focus from the very broad “May feel like you arebeing used” to the very focused “If I’m very busy and asked not to be interrupted, someone asking forfeedback can really side track you.” Although consistently instrumental in orientation; these statementsillustrate the variation in participant statements making clear that a number of thematic categoriescoexisted within the functional domain. In this regard, thematic analysis exposed three similar yet distinctcategories associated with functional responses to being asked for feedback. These included: a suspicionof ulterior motivations; abrupt or unexpected asking; and simply exercising bad timing.Ulterior Motivation. Emergent from our analysis are the source’s expressions of displeasurewith being asked for feedback for reasons other than a seeker’s genuine attempts at performanceimprovement. No less than 20 respondent statements illustrated a concern with seeker motivations,providing comments such as “If the request is not genuine; not open for ideas”, “If the person is not135

sincere about the feedback process…”, and “Sometimes not sure about the intention of that someone”.Other respondents gave more explicit characterizations of the nature of these ulterior motives, providingdescriptions in which seekers were attempting to test the source’s knowledge. This was evident fromcomments such as “Sometimes it can appear that I am being tested”, “I dislike when people ask mequestions to test my knowledge rather than to learn”, and “When I know I’m right about something andthe other person questions me because they’ve been with the job longer”. Another motive was seekersasking for feedback just to follow formal procedure. For example, respondents said “…they are justasking to go through the motions”, “…it’s just a formal procedure”, and “Sometimes getting feedback isjust part of the job.” A further motive was seekers attempting to appease the source such as “…offering itas a hollow gesture to appear cooperative”. Another suggested motive was a seeker attempting to absolvehimself or herself of decision-making responsibility as illustrated by the comment “Sometimes used totransfer the tough decisions or some workload to others”.Abrupt or Unexpected Asking. Emerging from the thematic analysis, participant statementsindicated concern with respect to responding to unanticipated or abrupt questioning. For example, severalrespondents described their discomfort with being “put on the spot”. One reason why sources dislikedbeing asked for feedback abruptly is that they felt it limited their ability to give a thoughtful and properlyarticulated reply to the seeker as illustrated, “If being asked abruptly, ‘words’ may not come out right –need time to think it over and organize thoughts”. Comprised of 9 participant thought-statements, thiscategory illustrates an awareness of the importance of context in feedback process as represented by thefollowing statements “If asked in a formal meeting”, “Based on how it is asked at times it puts me on thespot” and “It can put you on the spot or uncomfortable position depending on the situation”. In addition,one respondent commented, “Sometimes would rather give a measured and thought-out response whenfeedback is instead demanded on the fly. Have to come back with changes much later.” From thisperspective, participant statements suggest that, when and where the seeker chooses to ask for feedback,play a role in the source’s pleasure or displeasure with the feedback process.Bad Timing. Building on the previous theme, a similar theme emerged from participantstatements expressing an aversion to being asked for feedback at times when they are preoccupied withtheir own work. In this regard, many respondent comments exemplified this theme stating, “Beingbothered when I am busy”, “If I’m very busy and asked not to be interrupted, someone can really sidetrackyou”, and “Sometimes I dislike being asked questions when I am very busy and I can not bebothered with other issues”. The emergence of this category is based on 9 respondent reports whichsuggest that feedback sources find addressing feedback-seeking requests to be a task that is adjunct totheir core job responsibilities. In addition, the busy times at which sources find it inconvenient to be askedfor feedback include both specific times during the day such as “dinnertime” or more extended lengths oftime such as “quarter-end and year-end reporting periods”.The Act of Giving Feedback – Functional ApproachesBased on our request to indicate what respondents disliked about being asked for feedback, manyparticipants produced statements which focused on the functional consequences of giving feedback tothose who seek it. Consistent with previous descriptions of functional approaches, respondents’demonstrated recognition of the seeker-source feedback dynamic as a multi-dimensional phenomenon.Indicative of this distinct response orientation, participant thought-spaces included statements such as“Written feedback requires time. Time is often a constraint” and “No direct benefit if you give advice ona project that is outside of your area of responsibility” which clearly illustrate an alternative outlook tothose presented in the three previous domains. In total, 36 of 157 responses illustrated a thematic conceptcharacterized both as feedback-giving and functionally focused. Thematic analysis revealed theemergence of three distinct categories associated with functional approaches to giving others feedback.136

These included the role of time as a constraint on the source’s ability to provide feedback, avoidingunwanted accountability resulting from providing input, and remaining objective in the feedback process.Takes Too Much Time. Prominent among the topics which participants discussed was theimpact of time as a constraint on the feedback process. Consistently, our thematic analysis revealed adislike with the amount of time it takes to address feedback-seeking requests. Moreover, this themeaccounted for 18 of the 36 participant statements relating to the functional dislikes of giving feedback.Ranging from simple straight forward statements such as “Takes time”, “It may be time consuming”, and“Time involved” to more articulate expressions like “Sometimes, time is an issue and dealing with otherpeoples problems distracts from your own” or “Time consuming takes me away from other work that Imust complete” it is evident that many participants shared this functional concern. In addition, somerespondents highlighted the specific facets of the feedback-seeking exchange that they found most timeintensive,including reviewing and readings documents, meeting with the seeker, researching answers,and writing feedback out where required.Attracts Unwanted Accountability. Avoiding unwanted accountability was a further theme toemerge from our data. Characterized by respondent comments as a source’s concern for being heldaccountable for the effectiveness of the feedback that they provide, this theme was captured in by a totalof 10 respondent statements. Sources felt that giving feedback was not merely a conveyance of theirexpert opinion. Rather, it caused them to become responsible for the outcome of the task to which theirfeedback was targeted. For example, respondents provided comments such as “Places some unwantedresponsibility on me”, “Sometimes it mean you are responsible for whatever you said”, and “Sometimesafter you give the feedback and you own the problem…” indicating a potential transfer of responsibilityinferred in the seeker-source feedback encounter.Being Objective. A final theme revealed in the analysis of our data relates to a source’s concernwith not being objective in their feedback replies. In particular, 8 respondents alluded to the difficulties ofseparating their opinions of the seeker’s performance from their opinions of the seeker as a person. Forexample, respondents made comments such as “May not like the person, therefore conflicted in oursuggestions”, “If I’m not pleased with a product/service it’s difficult to be constructive in my criticism”,and “Having to separate personal likes/dislikes when providing feedback on certain individuals”. Thispurposeful reflection on the need for objectivity indicated to us that participants were aware of theinfluence and responsibility associated with their position as sources in the feedback process. As a result,it occurred to us that ‘being objective’ acted as an internal check and balance mechanism in the feedbackprocess which may impede the source’s urge to respond feedback seekers with instrumental manipulation.Discussion, Implications and ConclusionThe purpose of this investigation centred on exploring the nature and implications of feedback ofseeking on the source. To this end, the results of our paper illustrate the finding that sources have acomplex understanding of feedback as a multiple dimensional phenomenon. Respondent statementsdemonstrated a holistic understanding of feedback a process composed of many individual facets andprocesses as reflected in the multitude of themes which emerged from our data analysis. In addition,participants made equally clear the message that individual feedback-seeking activities are neitherconsidered, nor responded to, as a homogeneous cluster. Further, participants demonstrated an ability tofluidly segregate and reintegrate their feedback experiences among a number of divergent dimensions. Inthis regard, participants demonstrated a keen awareness of the ability to fracture and fragment theirexperiences dealing with feedback situations into multiple dimensions without loss of coherence in theaggregate. Relying on insight derived from immersing ourselves in the data, later challenging ourassumptions through iterative reflection, we have explored an understudied facet of the organizationalinteraction, resulting in a conceptual framework to capture and better understand this phenomenon in137

future investigations. Following an integrative approach to theory development, the categoricalframework illustrated herein capitalizes on both the differentiation between situations which areconsidered relational versus those considered functional in orientation, as well as the distinctions betweenthose which focused on the feedback-asking role versus those focused on the feedback-giving role.Although all study participants received indistinguishable research materials and were presentedwith identical questions, participant responses quickly deviated from the strict interpretation of “What doyou like (dislike) about being asked for feedback?” and moved toward a broader understanding of theirexperience with feedback processes. To this end, our findings integrate both the ‘likes’ and the ‘dislikes’,representing the full spectrum of participant thought-spaces, illustrating four consistent yet independentlyunique perspectives emergent from our iterative analysis. These four perspectives are based on participantdifferentiation of two categorical domains, which are reflective of the passages contained in eachrespondent thought-space. The first is based on the distinction between asking and giving, in whichrespondents spoke either in terms of being asked for feedback or alternatively in the act of givingfeedback. This categorical differentiation is particularly telling in that it suggests that participantsconsider their role in the interactions associated with providing feedback in two ways: first as an activityoccurring at an intra-personal level; and second, as an inter-personal activity occurring within a dyadicrelationship. Consistent across each domain of the relational category presented in the thematic analysis,participants demonstrated that being asked for feedback prompts a response within the individual.Illustrations of introspection abound in themes such as the Feeling Valued for my Opinion, Feeling likePart of the Team, Coping to Remain Honest, and Fear of Giving Negative Feedback. In contrast,participant generated statements which focused on the act of giving feedback departed from the intrapersonallevel of consideration, instead focusing on their role in the dyadic interaction between seeker andsource. This distinction between intra-personal reflection and inter-personal process – captured at thesource – goes beyond the traditional focus of the feedback-seeking literature, by exploring source-beliefsand attributions, thereby setting the stage for future investigations.The second distinction, that which distinguishes relational activities from those with a functionalorientation, is derived from the participant’s perception of the nature of the conversation. In this regard,each respondent-expression illustrated either a relational interaction or a functional interaction,demonstrating the use of one to the exclusion of the other. Although seemingly simplistic, the emergenceof this categorical framework illuminates a fundamental divergence in the way that sources’ perceive andframe feedback encounters. Perhaps most interestingly, our findings suggest that participants are fluidlytransitioning between ask-give and relational-functional domains with little forethought, reflection orconscious effort. If this proposition could be tested and verified it may yield great benefit across severalbodies of existing literature beyond our own.Future research should proceed in two key directions. First, scales should be developed for eachof the categories and themes that emerged from the data. These should then be tested on a wider audiencein order to assess the generalizability of this study’s findings. Second, researchers should considerpossible individual difference factors as well as contextual factors that may moderate the positive andnegative implications sources experience from being asked for feedback.In summation, given that the concept of feedback-seeking was introduced (Ashford andCummings, 1983) over 20 years ago, and to date the vast majority of all research energies have focusedon the seeker side of the exchange. There is clearly a critical need to explore the dynamics that feedbacksources experience when being asked for, and responding to, feedback-seeking requests and to this endwe have begun to do so. In this paper we have undertaken an exploratory investigation into the domain ofsource, a critical and as yet under-explored area of the feedback-seeking literature. From this perspective,it is our belief that the arguments presented in this paper make a number of significant contributions to anunderdeveloped area of study. Primary among these contributions are the 27 emergent themes generated138

from the collection and analysis of qualitative data derived from a sample of 49 practising managers andprofessionals. In addition, the categorical framework developed in this investigation offers an importantavenue for future exploration, specifically in the terms of the development and validation of futuremeasurement instruments. Finally, if the arguments presented in this paper do nothing more than sparkthe interest and imagination of others searching for answers about the source’s role in the feedbackprocess, our research community will be well served.139

ReferencesAllen, T., “The mentor’s perspective: A qualitative inquiry and agenda for future research,” Journal ofVocational Behaviour, 51. (1997), 70-89.Allen, T., “Mentoring others: A dispositional and motivational approach,” Journal of VocationalBehavior, 62. (2003), 134-154.Ammons, R., “Effects of knowledge of performance: A survey and tentative theoretical formulation,” TheJournal of General Psychology, 84. (1956), 279-299.Ashford, S., “The feedback environment: An exploratory study of cue use,” Journal of OrganizationalBehaviour, 14(3). (1993), 201-224.Ashford, S., Blatt, R., & VandeWalle, D., “Reflections through the looking glass: A review of research onfeedback-seeking behaviour in organizations,” Journal of Management, 29(6). (2003), 773-779.Ashford, S. & Cummings, L., “Feedback as an individual resource: Personal strategies of creatinginformation,” Organizational Behaviour and Human Performance, 32. (1983), 370-398.Burke, R.; McKeen, C.; McKenna, C., “Benefits of mentoring in organizations: The mentor’sperspective,” Journal of Managerial Psychology, 9(3). (1994), 23-32.de Ruyter, K., Perkins, D.S. & Wetzels, M., “Consumer-defined service expectations and post purchasedissatisfaction in moderately-priced restaurants,” Journal of Consumer Satisfaction,Dissatisfaction, and Complaining Behavior, 8. (1995), 67-80.Flanagan, J.C., “The critical incident technique,” Psychological Bulletin, 51 (July, 1954), 327-358.Greller, M. & Herold, D., “Sources of feedback: A preliminary investigation,” Organizational Behaviourand Human Performance, 13. (1975), 244-256.Gremler, D.D., “The critical incident technique in service research,” Journal of Service Research, 7(1).(2004), 65-89.Halatin, T. & Knotts, R., “Becoming a mentor: Are the risks worth the rewards?” SupervisoryManagement, 27(2). (1982), 27-29.Ilgen, D., Fisher, C. & Taylor, M., “Consequences of individual feedback on behavior in organizations,”Journal of Applied Psychology. 64. (1979), 359-371.Kluger, A. & DeNisi, A., “The effects of feedback interventions on performance: A historical review, ameta-analysis, and a preliminary feedback intervention theory,” Psychological Bulletin, 119(2).(1996), 254-284.Levinson, D., Darrow, D., Klein, E., Levinson, H. & McKee, B., The Seasons of a Man’s Life, New York,NY: Knopf. 1978.Madzar, S., “Feedback seeking behaviour: A review of the literature and implications for HRDpractitioners,” Human Resource Development Quarterly, 6(4). (1995), 337-350.Morrison, E., “Newcomer information seeking: Exploring types, modes, sources, and outcomes,”Academy of Management Journal, 36. (1993), 557-589.Morrison, E., “Information seeking in organizations,” Human Communication Research, 28. (2002), 229-242.Morrison, E. & Weldon, E., “The impact of an assigned performance goal on feedback seeking behavior,”Human Performance, 3. (1990), 37-50.Olsen, M.J.S & Thomasson, B., “Studies in service quality with the aid of critical incidents andphenomenology,” in QUIS 3: Quality in Services Conference, Eberhard E Scheuing, BoEdvardsson, David Lascelles, and Charles H. Little, eds. Jamaica, NY: International ServiceQuality Association, 481-505.Raggins, B. & Scandura, T., “Gender differences in expected outcomes of mentoring relationships,”Academy of Management Journal, 37(4). (1994), 957-971.Raggins, B. & Scandura, T., “Burden or blessing? Expected costs and benefits of being a mentor,”Journal of Organizational Behavior, 20. (1999), 493-509.140

Rudolph, U., Roesch, S., Greitemeyer, T. & Weiner, B., “A meta-analytic review of help giving andaggression from an attributional perspective: Contributions to a general theory of motivation,”Cognition and Emotion, 18(6). (2004), 815-848.Strauss, A. & Corbin, J., Basics of Qualitative Research: Techniques and procedures for developinggrounded theory, 2 nd ed, Thousand Oaks CA: Sage Publications, 1998.Walker, S. & Truly, E., “The critical incident technique: Philosophical foundations and methodologicalimplications,” in AMA Winter Educators’ Conference Proceedings: Marketing Theory andApplications, Vol. 3, Chris T. Allen and Thomas J. Madden, eds. Chicago: American MarketingAssociation, 270-275.141

ASAC 2006Banff, AlbertaIvy Kyei-Poku (Doctoral Student)Richard Ivey Business SchoolUniversity of Western OntarioORGANIZATIONAL STRUCTURE AND ORGANIZATIONAL COMMITMENT: ACONTINGENCY APPROACHStudies that have endorsed the contingency approach to organizational analysis haveconcentrated on the organization structure-environment interface to the detriment of theinterface between these and the individual level characteristics. This conceptual paperattempts to show that contingency theory remains useful and may be extended toindividual level phenomena, such as, the study of organizational commitment. This paperargues that different organizational commitment forms will be differentially importantunder varying organizational structures.Organizational commitment research has flourished during the past few decades. The largeinterest in this area appears to be a result of the link found between organizational commitment andseveral significant employee behaviors; however, the boundary conditions (i.e., context) in which theserelationships are fostered have remained relatively unexamined. Organizational context is important in thedetermination of job-related attitudes and behaviors and germane to all phases of organization research(Clinebell & Shadwich, 2005). However, strikingly little research attention has been given to these twobodies of research. One plausible explanation for the relative decline of interest and subsequentindependence of the two bodies of work might be the evolving difference in objectives.The theories of organizational commitment have gained prominence in micro-organizationalbehavior literature as a predictor of turnover intentions and actual turnover (Allen & Meyer, 1996; Meyer& Allen, 1997). More recently organizational commitment has also been linked to factors such asemployee well-being (Wasti, 2005). Organizational social structure, a domain of organizational theoryand macro-organizational behavior, was first conceptualized by Max Weber in the early 1900s, whopublished his theory of bureaucracy (Hatch, 1997). In contrast to organizational commitment,organizational structure has largely been applied to explaining the variables in promoting organizationaleffectiveness and performance (Ambrose & Schminke, 2003; Donaldson, 2001).A second plausible explanation may be due to inadequate theoretical conceptualizations of thelinkage between structure and behavior (Jones, 1984). This is somehow of a paradox given that one majortheme of organizational theory is the design of organizational structures to enhance individual andorganizational effectiveness.It is within the social context or structure that decisions are made and these factors may have aneffect on the levels of employee commitment. Stated differently, not much of the literature onorganizational commitment considers macro level variables, such as organization social structure giventhe importance of these variables in promoting organizational effectiveness and performance (Ambrose &Schminke, 2003). For example, Morris and Steers (1980) examined several aspects of organizationalstructure with attitudinal organizational commitment and collectively they accounted for 20% (p

organizational size (i.e., the total number of organizational members) may increase promotionalopportunities or side bets, thereby increasing calculative commitment.Nonetheless, these studies were conducted in an era when researchers were making a distinctionbetween the two forms of commitment; attitudinal and calculative commitment (Meyer, Stanley,Herscovitch & Topolnytsky 2002). Two limitations become evident; first, there is now a consensus thatorganizational commitment is multidimensional with varying antecedents, correlates and consequences(Meyer et al. 2002) and there is, therefore, the need to concurrently examine all dimensions oforganizational commitment if researchers are to better understand an employee’s relationship with anorganization (Jaros, 1997). Second, calculative commitment is a behavioral view of organizationalcommitment and is viewed as being committed to a particular course of action rather than to an entity(i.e., the organization).The purpose of this paper is to comprehensively examine the linkage between organizationalsocial structure and the Meyer and Allen's three-component model of organizational commitment (Meyer& Allen, 1991; 1997). I employ this model because among the various multidimensionalconceptualizations of organizational commitment, this model has gained substantial popularity (Watsi,2005). Second, I propose that an employee’s level in the organization (i.e., organizational level) maymoderate the structure-commitment relationship. The explanation for positing a moderating effect is thatthe overall lack of consistent findings between structure and commitment may be because relations arenot direct but also moderated by other variables (Mathieu & Zajac, 1990). Third, using contingencytheory I attempt to explain why the different forms of organizational commitment would be more salientto individuals under different organizational structures.The dominant approach to explaining organizational structures during the past three decades hasbeen contingency theory (Donaldson, 2001). The central theme of this theory is that the performance ofan organization is contingent on the fit between internal structures and the external environment(Donaldson, 2001). An implication of the previous statement is that contingencies of organizationalstructure include some that are within the organization and some that are outside of it (Donaldson, 2001).However, organizational theorists have focused more directly on the external fit between the organizationand its environment (Hollenbeck, Moon, Ellis, West, Ilgen, Sheppard, Porter, & Wagner III, 2005; Laird& Lee, 1979). It is, therefore, an inherent weakness if the theory fails to integrate both the macro (i.e.,environment-structure) and micro (i.e., structure-employee attitude or behavior) dimensions (Laird & Lee,1979).The fundamental premise here is that, contextual factors such as, organizational size and theenvironment present constraints to which the organization must adjust by modifying its organizationalstructure and procedures (Lin & Germain, 2003). The varying organizational structures create differentpersonnel or human resource requirements (Hollenbeck et al., 2005) as well as reinforce certain values,attitudes, perceptions and behaviors more strongly than others (Tosi & Slocum, 1984). With properadjustment optimum performance may be achieved (Lin & Germain, 2003). Contingency theory arguesthat there is no one best way to organize; rather performance depends on an appropriate fit. Further, theconcentration of a fit or a match between the organizational structure and employees may be based onexchanges made between employees and their employing organization and expectations of reciprocity canplay an important role in the employee’s adjustment to the organization’s environment (Kristof, 1996).Organizational CommitmentOrganizational commitment or the psychological bond that characterizes the employee’srelationship with the organization seems to hold the interest of so many researchers due to the powerfulimplications it has on the organization and its members (Becker, 1992; Meyer & Allen, 1997), and the143

perceived relationship to many organizational outcomes. For example, from an organizationalperspective, lower levels of turnover, absenteeism and higher productivity are manifestation of strongemployee commitment. From an individual perspective, lower turnover intentions (Allen & Meyer, 1996;Meyer & Allen, 1997) and psychological well-being (Watsi, 2005) are indications of organizationalcommitment.In commitment research, there are two common perspectives: the behavioral perspective and theattitudinal perspective. The behavioral approach has focused mainly on identifying conditions underwhich a behavior, once exhibited tends to be repeated, as well as on the effects of such behavior onattitude change (Meyer & Allen, 1991). The attitudinal approach, on the other hand, has focused largelyon identification of the antecedent conditions that contribute to the development of commitment and atthe behavioral outcomes of commitment (Meyer & Allen, 1991). The distinction between the attitudinaland behavioral approaches is clearly reflected in the research traditions that have become associated witheach.A major contributor to the behavioral view was Salancik (1977), who draws extensively fromthe dissonance literature. He proposed that organizational commitment results from the process in whichthe employees become committed to the implications of their own actions. The focus of the behavioralcommitment approach was that individuals were committed to a particular course of action rather than toan entity, in this case their organization. The goal, therefore, has been to identify the conditions underwhich an act, once taken, would be likely to continue (Meyer & Allen, 1997). Several conditions or“binding variables” have been posited including the irrevocability of the initial act, its publicness, and thevolition associated with it (Meyer & Allen, 1997). For example, if an employee perceived that anindividual of significant status (e.g., a manager) was aware of his or her action, commitment increasedbecause of its publicness.The attitudinal commitment approach focuses on the process by which employees come to regardtheir relationship with the organization. The attitudinal approach is the most studied type oforganizational commitment. Some of the earliest work was done by Porter, Steers, Mowday, and Boulian(1974). They examined commitment as though it were uni-dimensional, hence, developing measures thatwould produce one composite score representing commitment. According to their study, commitment wasdefined as “the strength of an individual’s identification with, and involvement in, a particularorganization” (p. 604). Furthermore, committed employees were described as: (a) having the desire toremain in the organization, (b) willing to exert effort on the organization’s behalf, and (c) believing in andaccepting the values of the organization. This implied that an individual who demonstrated thesecharacteristics was considered committed to the organization. This led to the development of a 15-itemscale, known as the Organizational Commitment Questionnaire (OCQ). Since the creation of the OCQ, ithas become one of the most widely used measurements of commitment (Becker, 1992).One of the problems with this approach is that different researchers defined this composite viewof commitment in very different ways. This indicated that commitment was more complex than originallythought. For example, Angle and Perry (1981) employed factor analysis on the OCQ and found twodistinct factors, which the authors labeled value commitment and commitment to stay. Evidently,commitment as assessed by OCQ is not uni-dimensional but has at least two dimensions. The twodifferent conceptualizations were eventually termed affective commitment (i.e., Porter’s view) andcontinuance commitment (i.e., Becker’s view) by Meyer and Allen (1997) and Meyer, Paunonen,Gellatly, Goffin, and Jackson (1989). Meyer and Allen (1984) were among the first researchers todescribe these two views of commitments. Furthermore, Meyer and Allen (1991) defined anotherdimension: normative commitment (i.e., Wiener’s view, discussed below).144

Accordingly, the diversity in the conceptualization and measurement of organizationalcommitment has made it difficult to interpret the results of an accumulating body of research (Meyer &Allen, 1991). Meyer and Allen (1991) noted that common to the various definitions of organizationalcommitment is the view that organizational commitment is a “psychological state that first characterizesthe employees’ relationship with his/her organization, and secondly, has implications for the decision tocontinue membership in the organization” (p. 67). Meyer and Allen (1991) therefore go beyond theexisting distinction between attitudinal and behavioral commitment and argue that commitment, as apsychological state, is not restricted to value and congruence of organizational goals (i.e., feelings orbeliefs or both concerning the employee’s relationship to the organization). Commitment can also be areflection of a desire, a need, or an obligation to maintain membership with the organization. Thisassertion, then, falls out of the traditional social psychological definition of an attitude (Meyer & Allen,1991; 1997). A three-component conceptualization was then articulated leading to the development of theThree-Component Model. The model postulates that an employee simultaneously experiencescommitments to the organization based on emotional attachment (affective, i.e. Porter et al., 1974), afeeling of obligation (normative, Meyer et al., 1989), and perceptions that the social and economic costsof leaving the organization are prohibitively high (continuance, Becker’s [1960] Side-Bet Theory).Meyer and Allen (1997) argued that it is more appropriate to consider the views as components ofcommitment rather than types because an employee’s relationship with an organization might reflectvarying degrees of all three. Although there are many varied definitions of organizational commitment,they appear to reflect at least three general themes - affective, normative, and continuance.Affective commitment refers to employees’ feelings of belonging and attachment to theorganization (Meyer & Allen, 1991; 1997). Affective commitment stems from the work of Kanter (1968)who defined commitment as the “willingness of social actors to give energy and loyalty to theorganization” (p. 499). Several writers including Porter et al. (1974) directed attention to a sense ofbelonging and the experience of loyalty. Recently, researchers such as O’Reilly and Chatman (1986)referred to commitment as employees' feelings of pride toward and desire for affiliation with anorganization and Allen and Meyer (1990) defined affective commitment as an “emotional attachment to,identification with, and involvement in the organization” (p. 11).Continuance commitment, unlike affective commitment, relates to perceived costs of leaving theorganization, both financial and non-financial (e.g., accrued pensions) and perceived lack of alternatives(e.g., high unemployment rates) (Becker, 1960; Meyer & Allen, 1991; 1997). Continuance commitment isbased on the costs and benefits of continued membership, and has its roots in Becker’s (1960) Side-BetTheory of Commitment. This theory posits that employees make certain investments or side-bets (e.g.,tenure toward pension, promotions, and work relationships) that become sunk costs that diminish theattractiveness of external employment opportunities.In the commitment literature, continuance commitment has been shown to be related toemployees’ perceptions about skill transferability (Allen & Meyer, 1990), such that employees whoassumed their training skills were less transferable to other similar organizations expressed strongercontinuance commitment (Meyer & Allen, 1997). Whitener and Walz (1993) also found that investmentssuch as job security, status, and retirement funds were positively correlated with continuancecommitment. Age, tenure, education, position, career satisfaction, and pay have been the commonapproach for testing side-bets. Findings obtained from these studies, however, have been mixed (Meyer &Allen, 1997) because, for some employees, the perceived cost of leaving the organization increases asthey grow older and their organizational tenure increases. For others, the costs might decrease, becausetheir experience and skills increase, which may be of value to other employers (Meyer & Allen, 1997).The conceptualization of continuance commitment describes an individual-organizational interaction, andalteration of investments over time (Mathieu & Zajac, 1990)145

The third and least researched component of commitment is normative (Meyer & Allen, 1997).Normative commitment is concerned with the moral obligation employees feel about remaining with theirorganization (Meyer & Allen, 1991; 1997; Wiener, 1982). This is based on the earlier work of Weiner(1982). He proposed that individuals internalize norms and behave as they do because they feel it ismoral. Weiner (1982) stated that normative commitment is formed on the basis of a collection ofpressures that individuals feel from family and culture during their early socialization and during theirsocialization as newcomers to the organization. The presumed process here is one of internalizationbecause socialization processes are extremely rich and varied; they carry messages of what is right andmoral, and of particular attitudes and behaviors, such as rewards and punishments or modeling andimitation of others (Meyer & Allen, 1997). It has also been suggested that normative commitmentdevelops on the basis of certain kinds of investments that seem difficult for an employee to reciprocate.This may include nepotism-hiring policies that favor the employees’ family members (Meyer & Allen,1997). This can create a sense of indebtedness and employees may be obliged to stay with theorganization. Finally, it has been hypothesized that normative commitment develops on the basis of a“psychological contract” between the employee and the organization (Meyer & Allen, 1997).Psychological contract is the mutual obligation between the employee and employer; this relation can beexplicit (e.g., pay) or implicit (e.g., favors) in nature (Anderson & Schalk, 1998). The relationship ischaracterized by the employer managing the career development and employee well-being. In return forthis sense of security and well-being, the employee continues to work hard and remain loyal to theemployer (Anderson & Schalk, 1998).Organizational StructureThe organization’s structure is the social architecture that describes; (a) how large numbers ofpersons are clustered into smaller groups and differentiated and (b) how the independent actions of thesedifferentiated groups are coordinated to produce some holistic product or service (Hollenbeck et al., 2005)and it is one of the most ubiquitous aspects of organizations (Ambrose & Schminke, 2003; Clegg &Hardy, 1996). Donaldson (2001) noted that structure includes — but is not limited to, power andreporting relationships such as those identified in organization charts, behaviors required of organizationmembers by organizational rules, and patterns of decision making (e.g., decentralization) andcommunication among organization members. Moreover, it encompasses both formal and informalaspects of relationships between members.Research has demonstrated that organizational structure interacts with a variety of factors toinfluence organizational performance. These factors include environmental change (e.g., Lawrence &Lorsch, 1967), organizational size (e.g., Pugh, Hickson, Hinings, & Turner, 1969), organizationalproduction technology (e.g., Perrow, 1967; Woodward, 1965), and organizational strategy (Chandler,1962). Donaldson (2001) argues that there are three main contingencies that affect structure: taskuncertainty, task interdependence and size. These aforementioned contingencies determine the structurean organization would adopt.The most prevalent distinction for describing fundamental differences in organizational structureis that of mechanistic and organic structural forms (Burns & Stalker, 1961; Slevin & Covin, 1997;Stopford & Baden-Fuller, 1994). Mechanistic structures are characterized as rigid, tight, and traditionalbureaucracies. In mechanistic settings, power is centralized; communications follow is rigid andhierarchical. Managerial styles and job descriptions are uniform and formal rules and regulationspredominate decision-making. By contrast, organic organizations are characterized by flexible, loose,decentralized structures. Formal lines of authority are less clear, power is decentralized, communicationchannels are open and more flexible, and formal rules and regulations take a back seat to adaptability inhelping employees accomplish goals (Burns & Stalker, 1961; Lawrence & Lorsch, 1967). It is important146

to note that these two structural forms represent ends of a continuum, not a dichotomy. No organization isperfectly organic or mechanistic; most display some characteristics of both, and intermediate stages existbetween the two archetypes.Research has demonstrated that organizational structure affects organizational performance(Covin & Slevin, 1989; Jennings & Seaman, 1990). Structure has been shown to influence individual-andteam-level outcomes as well. For example, organic structures have typically been associated withincreased job satisfaction (Rahman & Zanzi, 1995), affective commitment (Meyer & Allen, 1997) andlearning (Slevin & Covin, 1997). In this present paper I consider the effect of structure on organizationalcommitment. I explore structure from what Brass (1984) calls a “traditional view”, that is I examinedcentralization, formalization and organization size as focal structural dimensionsThe application of structural dimensions involves a comparison of mechanistic and organic formsof organization. For instance, mechanistic forms are characterized by high levels of centralization andformalization and also levels of complexity (i.e., vertical). This paper takes the view that, albeit verticalcomplexity represents an important structural dimension of organizational structure, it has been shown tobe positively correlated with organizational size (Schminke, Cropanzano, & Rupp, 2002). In someresearch it has been used as a surrogate for size (Wally & Baum, 1994).Organizational Commitment and Organizational StructureThe conceptual and empirical groundwork laid by Morris & Steers (1980) suggests thatorganizational structure may influence organizational commitment and the social exchange relationshipsworkers form with their organizations. In this section I discuss the role of organization structure and thethree forms of organizational commitment.Centralization reflects the concentration of power and decision making authority in anorganization. Hage and Aiken (1967) identified two subcomponents of centralization: participation indecision making and hierarchy of authority. Participation in decision making refers to the relativeconcentration of decision-making power with regard to policy making. It speaks to how involvedindividuals are in issues such as hiring, promotion, and setting departmental policies. The secondsubcomponent of centralization, authority hierarchy, refers to the relative concentration of decisionmakingpower with regard to performing tasks. It concentrates on the autonomy an individual has inperforming his or her day to day job. Low levels of participation in decision making and high levels ofhierarchy of authority each reflect a centralized structure.Employee involvement and participation (decentralization) has long been seen as an importantpart of organizational life and a key to achieving increased organizational effectiveness and positiveemployee perceptions (Shadur, Kienzle, & Rodwell, 1999). The assumption is that if employees areadequately informed about matters that concern them and they are allowed to make decisions relevant totheir task there will be benefits for both the organization and the individual. It is also a mindful andintended effort by management to provide visible extra role or role-expanding opportunities forindividuals at a lower level of in the organization to have greater voice in more areas of organizationalperformance (Glew, O Leary-Kelly, Griffin, & Van Fleet, 1995).Centralization may be linked to attitudes of organizational commitment, especially the affectiveform of commitment. Individuals with strong affective commitment want to remain with the organizationbecause they desire to. The mechanism presumably involved in creating the desire for these individualsvary across different conceptualizations (Herscovitch & Meyer, 2001), however; include involvement inthe organization and identification with the organization’s mission and values (Meyer & Allen, 1991).147

Consequently any situational variable, such as decentralization, that contributes to the likelihood that anindividual will become involved, or derive his or her identity from the association with the organizationor from working towards an objective that is congruent with personal values would contribute toward thedevelopment of affective commitment (Herscovitch & Meyer, 2001). To expound on the abovestatements, for instance, to the extent that individuals have input into policy decisions as well asautonomy over decisions about their tasks (decentralization), that is, employees have more discretion,control and other resources to execute their tasks should lead to increased perceptions of identificationwith and involvement in the organization (affective commitment). Further, drawing on empowermentliterature it has been reported that employees who consider themselves empowered-due to theredistribution of power by the organization-experience higher levels of organization loyalty because theyfeel more involved in the achievement of the organizational goals.In regards to normative commitment; I argue that decision-making processes that encouragemember participation may allow the employees to perceive the organization as caring about their valuesand contributions to the organization. This logic is consistent with organizational support theory (OST)that suggests that if an employee perceives that the organization cares about him or her, the employeemay feel a sense of obligation to contribute to the organization's well-being and help the organizationreach its goals (normative commitment) by returning the treatment (Allen, Shore & Griffeth, 2003). Thenotion of organizational support is accepted as being important in enhancing organizational effectiveness(Allen, et al., 2003) and increases performance (Glew, et al., 1995).For the link between (de)centralization and continuance commitment, higher levels ofparticipation as well as having autonomy (authority hierarchy) over one’s own task should have a positiveimpact or no impact on the mind-set of perceived cost. That is, decentralization which is associated withhow an employee is involved in and identifies with his or her organization through participation indecision-making, sharing of ideas to create value for the organization and having autonomy over theirown tasks should increase employee’s perceived cost of leaving the organization due to lack ofalternatives to make up investments. I predict a main effect of decentralization on organizationalcommitment. Specifically,Proposition 1a: The relationship between levels of participation in decision making will be positivelyassociated with (a) affective commitment (b) normative commitment and (c) continuance commitment.However the link will be strongest with affective commitment than normative and continuancecommitmentProposition 1b: The relationship between lower levels of authority hierarchy will be positively associatedwith (a) affective commitment (b) normative commitment and (c) continuance commitment. However thelink will be strongest with affective commitment than normative and continuance commitmentOrganizational size is the second structural variable that may affect perceptions of organizationalcommitment. Size has been examined as a structural component in its own right (Meyer, 1972); adimension of an organization's context (Pugh et al., 1968). Kimberly (1976) described four aspects oforganizational size: physical capacity, available personnel, inputs or outputs, and discretionary resources.My focus here is with the extent to which size affects attitudes and consequently behavior of organizationmembers, the second aspect, available personnel and discretionary resources, are therefore, appropriate.Given the nature of organizational size the organizational commitment form, continuancecommitment may be most germane to predicting organizational effectiveness and performance. Stevens etal. (1978) reported an association between organizational size and calculative commitment positing thatlarger organizations may increase the employees’ chances of promotions and other forms of side bets,148

thus increasing commitment levels. Continuance commitment extends the definition of calculativecommitment; first, continuance commitment considers how easy it is to leave one organization foranother job (Cooper-Hakim & Viswesvaran, 2005), second, is emphasizes the role of perceived ratherthan actual benefits.The antecedents for continuance commitment are perceived sacrifice and perceived lack ofalternatives. However, in a recent meta-analysis, Meyer et al. (2002) reported that the twosubcomponents, perceived sacrifice and lack of alternatives, are more highly correlated than initiallyreported. Even more important, perhaps, is the fact that the sacrifice subcomponent has a strongerpredictive power of attitudes and behaviors than the alternatives subcomponent. In light of these findings,they suggest that it would be appropriate to operationalize continuance commitment as a perceivedsacrifice. However, studies continue to report the two antecedents as the bases for the same commitmentmind-set. To be consistent with the other studies as well as for the sake of parsimony we will theorizeusing the two antecedents but recognize that this is an issue that warrants additional research.Drawing on the work of Stevens et al. (1978), who posit that larger organizations may increasethe chances of promotions and other forms of side bets, I extend this reasoning and argue that asorganizations grow employee perceptions that there are internal opportunities and alternatives availablewould increase. This process would also decrease employee recognition or perceptions of cost relatedfactors. Labor market and employee retention literatures also assert that organizational size enhances thedevelopment of an internal labor market, that is, promotion opportunities (Baron, 1984; Baron, Davis-Blake, & Bielby, 1986; Kirschenbaum & Mano-Negrin, 1999).Furthermore, organizational size has an impact on turnover through career progression paths(Kirschenbaum & Mano-Negrin, 1999). Stated differently, organization size and developed internalorganizational labor markets produce lower departure or turnover rates since promotion opportunitieshave a strong negative influence on departures for career-related reasons Thus, larger organizationspromote and maintain internal labor markets by creating job opportunities as primary mechanisms toinduce the organizations' retention potential (Kirschenbaum & Mano-Negrin, 1999). Further,Organizational size is salient and known to affect employment conditions and career outcomes (Baron,Davis-Blake & Bielby, 1986). For example, when an employee perceives that their growing organizationaffords them the opportunity to develop their career that same mechanism may also reduce the cost basedmind-set. Thus I believe that larger organizations will evoke the perception of more opportunities foradvancement and lessen the threatened loss of investments. Consequently, perceptions of external (i.e.,outside the organization) alternatives is diminished.In keeping with the same line of reasoning as that of the continuance commitment, organizationalsize may be linked to normative commitment in the sense that large organizations are more concernedthan are small organizations with legitimating their personnel practices and compensation policies byconforming to accepted principles of human resource management. Thus, institutional forces may inducelarge organizations (at least in the United States) to pay high wages and provide better fringe benefits andopportunities for promotion (Kalleberg & Van Buren, 1996). Several writers have also argued that largeorganizations pay high wages and provide more fringe benefits in order to elicit sufficient applicants forjob openings (Kalleberg & Van Buren, 1996). When an individual is the recipient of these benefits itcreates a need to reciprocate (normative commitment). Thus organizational size may influence normativecommitment positively.Organizational size may influence affective commitment in two ways. First, as organizationsgrow, it may be increasingly difficult for organizations to demonstrate to its members how valuable theyare to the organization as a whole. Given that as organizations grow larger they become morebureaucratic (Robbins, 1990). In brief, with regards to organizational size, it will seem intuitive that larger149

organizations are seen as less personable and harder to identify with (Mathieu & Zajac, 1990). Theresulting propositions are:Proposition 2a: Larger organizations will be positively associated with higher levels of perceivedalternatives for employees within their organizations as oppose to outside the organization, thusdecreasing continuance commitment..Proposition 2b: Larger organizations will be positively associated with lower levels of perceived cost forthe employees within their organizationsProposition 2c: Larger organizations will be positively associated with higher levels of normativecommitment.Proposition 2d: Larger organizations will be negatively associated with lower levels of affectivecommitment.The commitment component relevant to highly formalized organizational structures is normativecommitment. Formalization describes the extent to which rules, procedures, instructions, andcommunications are written down (Pugh, Hickson, Hinings, & Turner, 1968). In highly formalizedsystems, little flexibility exists in determining how decisions are made or what outcomes are due in agiven situation; procedures and rewards are also dictated by the rules. Some authors have suggested thatincreased formalization may influence commitment to some extent by facilitating both job and role clarity(Morris & Steers, 1980).Role clarity is defined as an individual's understanding of job expectations, and leads to jobclarity, and the process for fulfilling these expectations, and the consequences of one's role performance(Hartenian, Hadaway & Badovick, 1994). In the absence of role clarity, outcomes for the individual andthe organization include dissatisfaction and tension, anxiety, and lower performance (Hartenian, et al.,1994). To the extent that the employees perceives the formal rules as easing the their day to day workactivities may contribute to the process by which an individual acquires the social knowledge and skillsnecessary to assume an organizational role (i.e., employee socialization) thus also comes to appreciate thevalues, abilities, expected behaviors, and social knowledge essential for assuming an organizational roleand for participating as an organizational member. Herscovitch and Meyer (2001) reports that normativecommitment is characterized by the mind-set that one remains with the organization based on internalizedset of norms concerning appropriate conduct through socialization.Formalization may, therefore, allow individuals to make more objective decisions about whetherto internalize the organization’s goals as their own. Moreover, when organizational members perceiveorganizational rewards as appropriate this cognition may contribute to an individual's confidence that heor she is being compensated appropriately for the work performed. The process may reflect recognition ofemployee’s obligation within the psychological contract (Rousseau, 1995) perceived to be in effect withthe organization. Therefore, we predict a main effect of formalization on normative commitment.Formalized organizations are likely to be larger organizations and thus more complex andcharacterized by specialization. The interdependence that accompanies such specialization limits anindividual employee's autonomy because his or her actions must coordinate to a substantial degree withactions of other employees. This process will have a negative impact on affective commitment.With regards to organizational size, larger organizations may be associated with moreopportunities for job mobility and career advancement because there are more slots within theorganization; hence, increased extrinsic and accumulated interests. Continuance commitment is where150

individuals are viewed as remaining in an organization for extrinsic rewards (Hrebiniak & Alutto, 1972).We can, therefore, assume that the perception of rewards should positively influence continuancecommitment.Proposition 3: The relationship between levels of formalization will be positively associated with higherlevels of (a) normative commitment (b) high levels of continuance commitment but lower levels of (c)affective commitment. The link will be stronger with normative commitment than continuancecommitmentThe Moderating effect of Organizational LevelIn addition to considering both the organizational and individual level relationship, I argue that anemployee’s hierarchical position or organizational level may play a role in the extent to which the threeforms of organizational commitment are related to performance. A meta-analytic review by Mathieu andZajac (1990) indicates that individuals higher in the organizational hierarchy tend to have greater levels ofcommitment than employees at lower levels. This is because, for example, higher ranking individuals arelikely to be more involved in their work than are their low ranking counter parts (Brown, 1996).Additionally, those with seniority within the organization often receive treatment they perceive asadvantageous (i.e., better benefits) and also tend to have more influence on organizational policies(schminke et al., 2002). Hence, relative to lower level employees, those who are higher in theorganizational hierarchy tend to experience higher levels of affective, normative and continuancecommitment.For example, from the stand point of continuance commitment, employers offer a variety ofinducements (side-bets) that encourages the upper echelons to stay on with the organization. This furthercomplicates their desire to turnover because employees often want to avoid the high social and economiccosts of leaving (e.g., relocation costs, disrupted social networks) (Sinclair, Tucker Cullen & Wright2005). Thus continuance commitment strengthens as these side bets accumulate, rendering employeesmore likely to stay with the organizationHypothesis 4: Organizational level will moderate the relationship between the structural dimensions ofthe organization—(a) participation, (b) authority hierarchy, (c) formalization, and (d) organizationalsize—and organizational commitmentImplications for Theory and PracticeThe most immediate contributions of assessing organizational structure and the Meyer and AllenThree-Component Model of Organizational Commitment is the propositions that provide direction forfuture research. Should these propositions be supported, there are additional implications fororganizational structure and organizational commitment in general. Second, this paper offers a first step incross level theory that is I described the relationship between independent and dependent variables atdifferent levels.There are also implications for management practice. A central goal of mechanistic organizationsis to eliminate the human element from decision making (Weber, 1947) by relying more on formal rulesand procedures. Therefore, in mechanistic organizations, formal procedures will be seen as the standardand normative and continuance commitments may be the components that are prevalent. Practitionersmay choose to enhance these commitments to achieve a better fit and consequently a more productiveorganization.151

In contrast, the central goal of an organic organization differs from that of its mechanisticcounterpart. Organic organizations are based on interpersonal transactions. They rely more oninterpersonal factors like face-to-face communication (Lengel & Daft, 1988), discussion and elaboration(Courtright, Fairhurst, & Rogers, 1989), and informal control systems (Ouchi, 1980). The significance ofthese interpersonal influences in organic organizations should increase affective commitment. As aresult, I suggest that practitioners will concentrate more on fostering the affective component oforganizational commitment in organic organizations than in mechanistic organizations.ConclusionMy aim has been to strengthen both the organization social structure and organizational commitmentliteratures by utilizing contingency theory to explain the matching of people to structures. Thisincorporation of literatures suggests a number of directions through which the understanding of humanbehavior in the workplace can be advanced. First, these dimensions can be used to understand how andwhy certain types of commitments are variably suited to different types of structures.Second, by assessing some factors of social structure and utilizing the multidimensional view oforganizational commitment, organizational theorists can now more thoroughly study how structuralfactors may influence the nature of work and affect behavior.152

ReferencesAllen, N. J., & Meyer, J.P. (1990). The Measurement and Antecedents of Affective, Continuance andNormative Commitment to the Organization. Journal of Occupational Psychology, 63, 1-18.Allen, N.J., & Meyer J.P. (1996). Affective, Continuance and Normative Commitment to theOrganization: An Examination of Construct Validity. Journal of Vocational Behavior, 49, 252-276.Allen, D.G., Shore, L.M., & Griffeth, R.M. (2003). The Role of Perceived Organizational Support andSupportive Human Resource Practices in the Turnover Process. Journal of Management, 29, 99-118.Anderson, N., & Schalk, R. (1998). The psychological Contract in Retrospect and Prospect. Journal ofOrganizational Behavior, 19, 637-647.Ambrose, M.L., & Schminke, M. (2003). Organizational Structure as a Moderator of the Relationshipbetween Procedural Justice, Interact ional Justice, Perceived Organizational Support andSupervisory Trust. Journal of Applied Psychology, 88, 295-305.Angle, H.L., & Perry, J.L. (1981). An Empirical Assessment of Organizational Commitment andOrganizational Effectiveness. Administrative Science Quarterly, 26, 1-15.Baron, J. N., Davis-Blake, A., &, Bielby, W. T. (1986). The Structure of Opportunity: How PromotionLadders Vary within and among Organizations. Administrative Science Quarterly 31, 248-273.Becker, H. S. (1960). Notes on the Concept of Commitment. American Journal of Sociology, 66, 32-42.Becker, T.E. (1992). Foci and Bases of Commitment: Are they distinctions worth making? Academy ofManagement Journal, 35, 232-244.Burns, T., & Stalker, G. M. (1961). The Management of Innovation. London: Tavistock.Brown, S. P. (1996). A Meta-Analysis and Review of Organizational Research on Job Involvement.Psychological Bulletin, 120, 235–255.Chandler, A. D., Jr. (1962). Strategy and Structure: Chapters in the History of the American IndustrialEnterprise. Cambridge, MA: MIT Press.Clegg, S. R., & Hardy, C. (1996). Organizations, Organization, and Organizing. In S. R. Clegg, C. Hardy,& W. R. Nord (Eds.), Handbook of organization studies (pp. 1-28). Thousand Oaks, CA: Sage.Clinebell, S., & Shadwich, G. (2005). The Importance of Organizational context on Employees’Attitudes: An Examinations of Working in Main Offices Versus Branch Offices. Journal ofLeadership Organizational Studies, 11, 89-100.Cooper-Hakim, A., & Viswesvaran, C. (2005). The Construct of Work Commitment Testing anIntegrative Framework. Psychological Bulletin, 131, 241-259.Courtright, J. A., Fairhurst, G. T., & Rogers, I. E. (1989). Interaction Patterns in Organic and MechanisticSystems. Academy of Management Journal, 32, 773-802.Covin, J. G., & Slevin, D. P. (1989). Strategic Management of Small Firms in Hostile and BenignEnvironments. Strategic Management Journal, 10, 75-87.Donaldson, R. (2001). The contingency theory of Organizations. Thousand Oaks: Sage.Glew, D.J., O Leary-Kelly, A.M., Griffin, R.W., & Van Fleet, D.D. (1995). Participation inOrganizations: a Preview of the Issues and Proposed Framework for Future Analysis. Journal ofManagement, 21, 395-421.Hage, J., & Aiken, M. (1969). Routine technology, social structure, and organizational goals.Administrative Science Quarterly, 14, 366-376.Hatch, M. J. (1997). Organization Theory: Modern, Symbolic, and Postmodern Perspectives. New York:Oxford University Press.Hartenian, L.S, Hadaway, F. J, & Badovick, G. J. (1994). Antecedents and consequences of RolePerceptions: A path analytic approach. Journal of Applied Business Research, 10, 40-52.Herscovitch, L., & Meyer, J.P. (2002). Commitment to organizational Change: Extension of a Three-Component Model. Journal of Applied Psychology, 87, 474-487.153

Hrebiniak, L.G.; & Alutto, J.A. (1972) Personal and Role-Related Factors in the Development ofOrganizational Commitment. Administrative Science Quarterly, 17, 555-573.Hollenbeck, J. R, Moon, H., Ellis, A.P.J., West, B.J., Ilgen, D.R.., Sheppard, L., Porter C.O.L.H., &Wagner III, J. A. (2005). Structural Contingency theory and Individual Differences: Examinationof External and Internal Person-Team Fit. Journal of Applied Psychology, 87, 599-606.Jaros, S.J. (1997). An Assessment of Meyer and Allen's (1991) Three-Component Model ofOrganizational Commitment and Turnover Intentions. Journal of Vocational Behavior, 51, 319-337.Jennings, D. F., & Seaman, S. L. (1990). Aggressiveness of Response to New Business Opportunitiesfollowing Deregulation: An Empirical Study of Established Financial Firms. Journal of BusinessVenturing, 5, 177-189.Jones, G. (1984). Task Visibility, Free Riding, Shirking: Explaining the Effect of Structure andTechnology on Employee Behavior. Academy of Management Review, 9,684-695.Kanter, R.M. (1968). Commitment and social organizations: A study of commitment Mechanisms inUtopian Communities. American Sociological Review, 33, 499-517.Kalleberg, A. L, & Van Buren, M. E (1996). Is bigger better? Explaining the Relationship betweenOrganization Size and Job Rewards. American Sociological Review, 16, 47-68.Kirschenbaum, A., & Mano-Negrin, R. (1999). Underlying Labor Market Dimensions of Opportunities:The Case of Employee Turnover Human Relations.52, 1233-1255.Kristof, A. L. (1996). Person-organization fit: An integrative review of its conceptualizations,measurement, and implications. Personnel Psychology, 49, 1-49.Laird W M., & Lee, D. (1979). An Alternative to Macro-Micro Contingency Theories: An IntegrativeModel. Academy of Management Review, 4, 333-345.Lawrence, P. R., & Lorsch, J. W. (1967). Organization and environment. Homewood, IL: Irwin.Lengel, R. H., & Daft, R. L. (1988). The selection of communication media as an executive skill.Academy of Management Executive, 2, 225-232.Lin. X., & Germain, R. (2003). Organizational Structure, Context, Customer Orientation andPerformance: Lessons from Chinese State-Owned Enterprise. Strategic Management Journal, 24,1131-1151Mathieu, J.E., & Zajac, D.M. (1990). A Review and Meta-Analysis of the Antecedents, Correlates, andConsequences of Organizational Commitment. Psychological Bulletin 108, 171-194.Meyer, J. P., & Allen, N. J. (1984). Testing the “Side-bet Theory” of Organizational Commitment: SomeMethodological Considerations. Journal of Applied Psychology, 69, 372-378.Meyer, J., & Allen, N. (1991). A Three-Component Conceptualization of Organizational Commitment.Human Resource Management Review, 1, 69-89.Meyer, J.P., & Allen, N.J. (1997). Commitment in the Work Place: Theory Research and Application.London, UK: Sage.Meyer, J. P., Stanley, D.J, Herscovitch L, & Topolnytsky L. (2002). Affective, continuance, andnormative commitment to the organization: A meta-analysis of antecedents, correlates, andconsequences. Journal of Vocational Behavior, 61, 20-52.Meyer, J.P., Paunonen, S.V., Gellatly, I.H., Goffin, R.D., & Jackson, D.N. (1989). OrganizationalCommitment and Job Performance: It's the Nature of the commitment that Counts. Journal ofApplied Psychology, 74, 152-56.Morris, J. H. & Steers, R. M. (1980). Structural Influences on Organizational Commitment. Journal ofVocational Behavior, 17, 50-57.O'Reilly, C.A., & Chatman, J. (1986). Organizational Commitment and Psychological Attachment: Theeffects of Compliance, Identification, and Internalization on Prosocial Behavior. Journal ofApplied Psychology, 71, 492-499.Ouchi, W. G. (1980). Markets, Bureaucracies, and Clans. Administrative Science Quarterly, 25, 129-141.154

Perrow, C. (1967). A framework for the comparative analysis of organizations. American SociologicalReview, 32, 194-208.Porter, L.W., Steers, R. M., Mowday, R.T., & Boulian, P.V. (1974). Organizational Commitment, Jobsatisfaction,and Turnover among Psychiatric Technicians. Journal of Applied Psychology, 59,603-609.Pugh, D.S., Hickson, D.J., Hinings, C.R., & Turner, C. (1969). Dimensions of organizational structure.Administrative Science Quarterly, 13, 65-105.Rahman, M., & Zanzi, A. (1995). A comparison of organizational structure, job stress, and satisfaction inaudit and management advisory services (MAS) in CPA firms. Journal of Managerial Issues, 7,290-305.Rousseau, D.M. (1995). Psychological Contracts in Organizations: Understanding Written andUnwritten Agreements. London: Sage.Salancik, G.R. (1977). Commitment and the Control of Organizational Behavior and Belief. In NewDirections in Organizational Behavior, edited by B.M. Staw and G.R. Salancik. Chicago: St.Clair Press.Schminke, M, Cropanzano, R., & Rupp D.E. (2002). Organizational Structure and Fairness Perceptions:The Moderating Effects of Organizational Level. Organizational Behavior and Human DecisionScience processes, 89, 881-905.Shadur, M.A., Kienzle, R., & Rodwel, .J.J. (1999). The Relationship between Organizational Climate andEmployee Perceptions of Involvement .Group & Organization Management, 24, 479- 503.Sinclair, R.R., Tucker, J.S., Cullen, J.C., & Wright, C. (2005). Performance Differences among FourOrganizational Commitment Profiles. Journal of Applied Behavior, 90, 1280-1287.Stevens, J.M., Beyer, J.M. & Trice, H.M. (1978). Assessing Personal, Role, and Organizational Predictorsof Managerial Commitment. Academy of Management Journal, 21, 380-396.Tosi, J.R. & Slocum J.W. Jr, (1984). Contingency Theory: Some Suggested Directions, Journal ofManagement, 10, 1984, pp. 9-26.Wally, S., & Baum, J.R. (1994). Personal and Structural Determinants of the Pace of Strategic DecisionMaking. Academy of Management Journal, 37, 932-956.Wasti, S.A. (2005). Commitment profiles: Combinations of Organizational Commitment Forms and JobOutcome. Journal of Vocational Behavior, 67, 290-308.Weber, M. (1947). The Theory of Social and Economic Organization. New York: Free Press.Wiener,Y. (1982).Commitment in organizations: A normative view. Academic Management Review, 7,418-428.Whitener, E.M., & Walz, P.M. (1993). Exchange Theory Determinants of Affective and ContinuanceCommitment and Turnover. Journal of Vocational Behavior, 42, 265-281.Woodward, J. (1965). Industrial organization: Theory and practice. London: Oxford University Press.155

ASAC 2006Banff, AlbertaZeying Wan(student)Richard Ivey School of BusinessUniversity of Western OntarioCOMMUNITIES OF PRACTICE AND INDIVIDUAL PERFORMANCEThe requirement of learning organizations calls for a dynamic and knowledge-basedenvironment – one provided by “Communities of practice” (CoPs). CoPs refer togroups of people informally bound together by sharing knowledge and practice for acommon enterprise. This paper illustrates how CoPs enhance the individualperformance by building a conceptual model and test this model by some pilot data.IntroductionThe challenges facing today’s business reality of distribution and globalization draw muchattention from decision makers regarding the development of core competencies through learning. In thisnew setting firms cannot relay too much on established practices, they have to learn new knowledge andinvent new practices, which in turns enhance the individual level and organization level performance. Therequirement of learning calls for a dynamic and knowledge-based environment – one provided by“Communities of practice” (CoPs). In general, CoP refers to a group of people informally bound togetherby sharing knowledge and practice and expertise for a common enterprise (Wenger 1998; Wenger andSnyder 2000). CoPs used to be marginal and outside of the mainstream relationships, but now theybecome popular especially within some large complex organizations because of their capability of selfmanagementand knowledge sharing. They can reduce the administrative load on formal structures andredistribute responsibility on informal structures. Many corporations have benefited from identifying andcultivating CoPs (Allee 2000; Wenger and Snyder 2000; Rouser and Dorsey 2003), albeit under varyingnames: “learning communities” at Hewlett-Packard Company, “family groups” at Xerox Corporation,“thematic groups” at the World Bank, “peer groups” at British Petroleum and “knowledge networks” atIBM Global Services (Gongla and Rizzuto, 2001).Because of its value, CoP is currently a popular topic in the fields of knowledge management(KM), organizational learning, and strategy. The former studies show that after identifying or cultivatingCoPs, companies can improve their ability in knowledge management and gain competitive advantage byenhancing the effectiveness of KM (Smits and Moor 2004), facilitating the knowledge diffusion (Storckand Hill 2000), facilitating knowledge transfer (Rouser and Dorsey 2003), tailoring IT support (Agresti2003), promote the spread of best practices (Wenger 2000), shaping the knowledge strategy (Wenger2004). As Allee (2000) pointed out, CoPs are powerful vehicles both for sharing knowledge andachieving business results. However, most studies related CoPs were conducted at the group ororganizational level and ignore the individual level which is more fundamental to understand themechanism of CoPs. A notable exception is Lesser and Storck’s work (2001). They looked at the CoP asan engine of the development of social capital that leads to individual behavioral changes and in turnpositively influences business performance, e.g., enhance organizational effectiveness and profitability. Ithighlighted the importance of investigating the individual behavior during the learning process. Anotherlimitation of this field is all the empirical studies on this topic are based on qualitative data, e.g.,interviews, case studies. Therefore, the purpose of this paper is to illustrate how CoPs enhance theindividual performance by building a conceptual model and test this model by using quantitative researchmethod.156

The paper is organized as follows. First, the literatures on CoPs are reviewed, and based on thereview a conceptual model and a refined definition of CoPs is provided. Second, the linkage betweenCoPs and individual performance is built and hypotheses are proposed. Third, hypotheses are tested bysome pilot data and its results are presented. Finally, the paper is concluded by discussing theimplications of future research.Communities of PracticeWhat is Communities of Practice? Most studies have attempted to define the CoPs by comparingit with other similar groups, and have argued that CoPs are not teams (Wenger 1998), organizational unitsor communities of expertise (Agresti 2003), or actor-network (Fox 2000), not merely knowledgenetworks or virtual communities (LaContora and Mendonca 2003). They are not merely ba which inJapanese means a shared space for emerging relationships (Nonaka and Konno 1998).Among all these groups, ba is seemingly most similar to the concept of CoP, because both ba andCoP can transcend time, space, and organization boundaries to create and transfer knowledge. But thereare some important differences between CoP and ba (Nonaka and Toyama 2003). CoP has an identity andits boundary is firmly set by the task, culture, and history of the community, while the boundary of ba isfluid and can be changed quickly. The membership of a CoP is fairly stable, and it takes time for a newparticipant to learn about the community to become a full participant, while the membership of ba is notfixed. CoP is created and disappears according to common purpose and interest, while ba is created,functions and disappears according to needs (short time purpose). Compared to CoP, ba has a ‘here andnow’ quality, a more flexible form and a shorter life cycle. A CoP can be viewed as an organicconfiguration of various ba, where people learn and practice together in order to employ storedknowledge and to create new knowledge. The main differences between CoPs and some types of groupsare summarized in table 1. (Based on the comparison provided by Wenger and Snyder 2000; LaContoraand Mendonca 2003; Jones and Esnault 2004)The definitions of CoPs have changed over time (Lave and Wenger 1991, Brown and Duguid1991, Manville and Foote 1996, Wenger 1998, Lesser and Storck 2001, Gongla and Rizzuto 2001,LaContora and Mendonca 2003, Ward 2003). According to Hilderth et al (1998), at least fifteencharacteristics have been highlighted in the literature by that time. Adding the newest definitions, thispaper classifies all the important characteristics of CoPs into three dimensions: identity dimension,practical dimension, and structural dimension (Table 2).157

Table 1The comparison of CoPs with other groupsOrganizational Project Virtual Network CoPs baunit team communitiesDefined by functions project interest relationships Knowledge needBound by identity project Not fixed Not fixed identity Not fixedDriven by Jobproject information Mutual Value, purpose needrequirements goalneedsLife cycle long short Medium Medium Medium shortLife cycledetermined byUntil the nextreorganizationProjectDeadlineEnd ofneedAs long aspeople areinterestedAs long aspeople havea reason toconnectUntil thecommonvalue, purposeare lostFormal/informal formal formal informal informal informal informalKey practice routine project Shareinformationinteraction SituatedlearningShareknowledgeLearning no no no no yes yesenvironmentmembership stable unstable Less stable Less stable Less Stable unstableTable 2Three dimensions of CoPsIdentity dimension Practice dimension Structural dimensionLave and Wenger, 1991 Common language Share practice, situated A group of peoplelearningWenger, 1998 Joint enterprise Share repertoire and A group of peopleresources, commonpracticeWenger and Snyder, Joint enterprise Share expertise and A group of people2000knowledgeManville and Foote, A store of knowledge Exposure to common A group of professionals1996problems and a commonpursuit of solutionsLesser and Storck 2001 Common interest Sharing and learning A groupGongla and Rizzuto Domains of knowledge N/A Institutionalized, informal2001Davenport and Hall,2001LaContora andMendonca, 2003Ward, 2003N/AN/ACommon history, sharedvision of the futureaccomplishmentsSituated learning,situated action,distributed cognitionShare information,knowledge, insight,experience and toolsShare knowledge,common practices andactionsnetwork of professionalsSocial infrastructureA groupA communityThe identity dimension reflects the sense of shared value, purposes or interests behindcommunity’s activities. These common purposes, interests, or values distinguish them from other kind ofgroups and other CoPs; they also define who the members are (structure dimension) and what they aredoing (practice dimension). In a sea of information, identity also helps us sort out what we pay attention158

to, what we participate in, and what we stay away from. It is the premise of the existence of CoPs anddetermines the CoP’s life cycle. The life cycle was defined by Wenger et al (2002) as five stages:potential, coalescing, launching, maturing, and dispersing.The practice dimension is a platform for learning and sharing knowledge. It reflects members’ways of doing things and their knowledge and experience. The key knowledge practice can be defined inCoPs as situated learning, which means learning as a practice is situated in its context. Rather thanviewing learning as the acquisition of certain forms of knowledge, Lave and Wenger have tried to place itin social relationships, situations of co-participation (Smith 2003). The learning process is tied to ongoingactivities and practices and is done by community members through social interaction rather than byisolated individuals (Fox 2000). It is the context within which work activities are situated that shapes theway individuals acquire knowledge and enhance capabilities to get their work done. Situated learning thusdiffers from formal training, in which knowledge is codified and transferred to learners in special trainingsessions. The relationship between members also differs from the relationship between teachers andstudents. In CoPs, members not only learn from others but also make a contribution to others. Situatedlearning occurs as members adjust to each other’s need. It also underpins a number of pedagogicalapproaches with different names (deep learning, learning by doing, action learning) that attempt to offerrich simulations of the world within a secure and secluded learning environment (Davenport and Hall2001).Furthermore, knowing what is not enough. Knowing-doing gap in knowledge managementprovides a view says that differences in performance come less from what people know, but more fromtheir ability to translate knowledge into actions (Pfeffer and Sutton 1999). Situated learning itself is aknowledge practice, which enables people to transform understanding into appropriate actions thateventually delivers appropriate performance (Leijen and Baets 2003).The structural dimension reflects the relationship among community members. The interactionsbetween members build a social network that supports the members’ knowledge practices and knowledgetransfer. This social network as the structure of CoPs links individuals and specifies the strength of theirrelationship.Based on this review, a conceptual model of CoPs is built (Figure 1). People come together fortheir common purposes, interests, or values; they function as a community through relationships ofmutual engagement that bind members together into a social network; they also build capability of sharedknowledge and experience by situated learning. Although these three dimensions are emphasized indifferent studies, there is no single definition covered all. Hence, a community of practice is defined inthis paper as a group of people informally bound by their common purposes, interests, or values, learningand practicing together for sharing knowledge and experience, with all these practices supported by thesocial network formed by the members’ interactions.159

Figure 1the conceptual model of Community of PracticeIdentitySocialnetworkSituatedlearningCommunity of practiceCommunities of Practice and Individual PerformanceIn the group level and organizational level studies of CoPs, researchers found it is hard tomeasure the value and outcomes of CoPs. One reason is that the effects of community activities are oftendelayed; another one is that the outcomes of CoPs generally appear in the work of teams and businessunits, not in the communities themselves. According to Wenger and Snyder (2000), the best way to assessthe value of a CoP is by listening to members’ stories, which can clarify the complex relationships amongactivities, knowledge, and performance. This difficulty and complexity also explains why most studies onCoPs use qualitative data supporting their arguments.However, on individual level the benefits of CoPs in terms of developing people’s capabilitiesand enhancing people’s performance are more direct and easier to be located. At the same time, theorydevelopment also provides more insights of the mechanism of CoPs. By looking at the three dimensionsof CoPs, this paper argues that it is the knowledge practices-situated learning that enhances the individualperformance. Situated learning provides a solution of education, training and performance support; theyalso provide anytime, anywhere virtual forums for learning and knowledge sharing (LaContora andMendonca 2003). In general terms, knowledge is interpreted information and is anchored in the beliefsand commitment of individuals (Nonaka et al 2001). Knowledge and the capability to create and utilizeknowledge are considered to be the most important source of individual’s performance, which eventuallyprovides a synergistic advantage of organizations (Argote 1999).Situated learning is an effective process of acquiring knowledge. Along with the increasing ofknowledge, the individual’s capabilities increase. It also enhances the individual’s learning efficiency,which reduce the cost and time of training and reduce the information overload of the employees. In mostorganizations, it is more practical of knowing how to find and apply relevant information efficiently thantrying to master vast amounts of knowledge. In CoPs, rather than engaging in an extensive search throughan organization’s repository of knowledge, the members turn first to other members to learn where to find160

elevant knowledge (Cross and Baird 2000). For example, before accessing to a database, they seekinformation from other experienced members firstly to get the direction of a specific point in the database.Obviously it saves the time of problem solving and makes it more efficient.Situated learning also enhances the individual’s ability to respond to environmental changes. Anovice learns by observation, by shadowing, by asking on the spot. A major component of situatedlearning is learning how to behave in a given group and understanding what resources used for whichpurpose. It can decrease the learning curve of new employees and enable them get into the organizationalmemory and make it easier to learn their new roles and responsibilities (Lesser and Storck 2001). Basedon these analyses, the hypothesis was drawn as follows:H1: situated learning could enhance the individual performance.We already identified the key knowledge practice which is the core of communities of practice:situated learning, by which people gain the knowledge of how to be a good member of a community(Davenport and Hall 2001). In CoPs, all these practices are supported by the social network built by theinteraction of members. The social network integrates knowledge and allows it to diffuse amongmembers.Social networks as the structural dimension of CoP link individuals and specify the strength oftheir relationship. According to Wenger’s latest argument (2002), “the heart of a community is the web ofrelationships among community members, and much occurs in one-on-one exchanges”. In his view, acommunity of practice need not focus too much on public events. The relationships between memberswill be strengthened by every phone call, e-mail or conversation. When the individual relationshipsamong community members are strong, the events will be much richer and varied. Since participantsknow each other better, they usually join community events with multiple reasons, e.g., finding an expertfor problem-solving, sharing experience with others, discussing some hot topics. These events provideplenty of time and space for people to enhance the social network informally and increase the chance ofknowledge transfer between members.In the psychology literature, researchers found that there are four types of social networks whichprovide different kinds of supports: emotional aid, material aid, information, and companionship(Wasserman and Galaskiewicz 1994, P56). By this classification, the social networks within CoPs mainlyprovide the support of information and knowledge sharing. ‘Who you know significantly affects what youeventually know’ (Cross and Baird 2000). Accordingly, the second hypothesis is:H2: the social network provides the support of individual’s situated learning.In one particular CoP, some characteristics of its social network such as size (the number ofnetwork members) and frequency of interaction also should be considered. According to Wasserman andGalaskiewicz (1994, P62), the network with bigger size is generally more supportive, and the networkwith higher frequency of interaction is more supportive.H2a: the member who belongs to a social network with bigger size will have more chance ofsituated learning.H2b: the member who belongs to a social network with higher frequency of interaction will havemore chance of situated learning.The identity dimension is also important. As mentioned above, it defines who the members areand what they are doing. CoP develops around things that matter to its members. The practices also161

eflect the member’s own understanding of what is important. From this perspective, the members withclear understanding of the identity will be more engaged in the community and more involved in thesocial activity and knowledge practice within the CoP.H3a: the member with more clear understanding of community’s identity will have higherfrequency of interaction with other members.H3b: the member with more clear understanding of community’s identity will know more membersin the community.H3c: the member with more clear understanding of community’s identity will have more chance ofsituated learning.The research model is illustrated in Figure 2.Figure 2 Research modelH3aFrequencyH2bClarityofIdentityH3cSituatedLearningH1PerformanceH3bNetworkSizeH2aMethodologySample and Data Collection MethodsThe overall approach taken to empirically test the relationships implied by the hypotheses was afield study using a survey methodology. Data were collected from student subjects enrolled in the PhDprogram of a Business School in Canada. According to the definition of CoPs, there is a particular162

community of practice within the PhD students. People come together with the common purpose ofpreparing for their academic career and with common interest in management and business relatedstudies; they work together and share knowledge with other members in the social network formed by theinteraction of students.A web based survey was designed using Microsoft Office Front Page, and the link along with acover letter was send to all the PhD students by Email. After the follow-up email, a total of 54 completedsurveys were returned. The respondent rate is 73%. Sample characteristics are shown in Table 3.Table 3sample informationGenderMale:Female:GradeFirst year:Second year:Third year:Fourth year:GroupFinance-Economics:General Management:Information systems:Management Science:Operations Management:Accounting:Organization Behavior:Marketing:361816141595141074095Measurement of Research VariablesThree variables were measured using multi-item scales (situated learning and clarity ofcommunity identity use 7-point scales, performance use 8-point scale). The items have been generatedbased on relative literature and have been discussed with other PhD students. Performance data are selfreportedand could potentially induce certain biases. Network size and the frequency of interactionsbetween members are measured by asking the respondents to write down the number of people whohe/she knows in the community and ask them to recall the frequency of interactions in the past threemonths (90 days) using different approaches (face-to-face meeting, telephone conversations, instantmessages, Email, Written messages). Network research typically relies on single-item sociometricquestions, which by themselves do not provide information regarding measurement reliability.ResultsTo test the hypotheses, PLS Graph was used, which is a latent structural equations modelingtechnique that utilizes a component-based approach for estimation. The measurement model of constructswas assessed firstly, and the structural relationships were examined next.163

Measurement ModelThe scales are assessed in terms of item loadings, internal consistency, and discriminant validity.Item loadings and internal consistencies greater than .70 are considered acceptable (Fomell and Larcker1981). Based on these criteria, one item in ‘performance’ construct, one item in ‘situated learning’construct, and three items in ‘frequency of interaction’ construct with low loading are dropped. As can beseen from the loading in table 4 and internal composite reliability scores in table 5, the rest items used inthis study meet these guidelines after rerunning the PLS. All items exhibit high loadings (>.70) on theirrespective constructs. Furthermore, all constructs in the model exhibit good internal consistency asevidenced by their composite reliability scores. The scores are more than adequate (>.70), ranging from.875 to .949.Table 4Results of PLS-loadingConstruct and itemsPLS Outer Model LoadingFrequency of interaction:Meeting.9021Email.8616Size 1.000Situated learning:S1.8782S2.9396S3.9035S4.8286S5.8885Clarity of IdentityId1Id2Id3Performance:P1P2P3P4.8361.8510.8820.8487.9300.8692.8361164

ICR(# of items)Table 5Inter-construct correlationsFrequencyNetworksizeClarityofidentitySituatedlearningPerformanceFrequency 0.875 (2) 0.88Network size 1 (1) -0.047 1Clarity of0.892(3) 0.197 0.240 0.86identitySituated learning 0.949 (5) 0.389 0.222 0.354 0.89Performance 0.927 (4) -0.099 0.109 0.083 -0.156 0.87*ICR: Internal composite Reliability. The bolded numbers on the leading diagonal are the square root of thevariance shared between the constructs and their measures. Off diagonal elements are the correlations amongconstructs. For the discriminant validity, diagonal elements should be larger than off-diagonal elements.There are two ways to assess discriminant validity: (1) indicators should load more strongly ontheir corresponding construct than on other constructs in the model and (2) the square root of the averagevariance extracted (AVE) should be larger than the inter-construct correlations. As can be seen in theTable 6, all indicators load more highly on their own construct than on other constructs. Furthermore, asshown by comparing the inter-construct correlations and AVE (shaded leading diagonal) in table 5, allconstructs share more variance with their indicators than with other constructs. Thus, these results showthe adequate discriminant validity of these constructs.Table 6Correlation between items and factor scoresClarity of SituatedNetwork Size Frequency Identity Learning PerformanceId1 .157 .101 .836 .258 -.001Id2 .219 .128 .851 .357 .055Id3 .228 .260 .882 .285 .139SL1 .111 .253 .307 .878 -.082SL2 .210 .359 .351 .940 -.209SL3 .147 .332 .267 .903 -.229SL5 .310 .380 .414 .888 -.101SL4 .161 .369 .190 .829 -.048Size 1.000 -.047 .240 .222 .109F-meeting -.055 .902 .173 .373 -.063F-Email -.025 .862 .176 .303 -.119P1 .106 -.091 .047 -.066 .849P2 .054 -.076 .061 -.187 .930P3 .189 -.095 .123 -.120 .869P4 .064 -.100 .057 -.116 .836*The bolded numbers are item loadings on their own construct, others are cross loading. For the discriminantvalidity, the loadings should be higher than cross-loadings.165

The structural ModelThe path coefficients and explained variances for the model are shown in Figure 3. Clarity ofIdentity explains 3.9% of the variance of frequency of interaction and 5.8% of variance of network size.Frequency of interaction, clarity of identity, and network size together explain 26% of the variance insituated learning, while situated learning explain 2.4% of the variance in performance. As shown inFigure 2, PLS results provide support for hypotheses 2b, 3a, and 3b. However, the hypothesis 1, 2a, and3c are not supported.Figure 3PLS Results0.197**Frequency0.347***ClarityofIdentity0.0390.242SituatedLearning-0.156Performance0.24*NetworkSize0.1810.26 0.0240.058*** Significant at .01** Significant at .05* Significant at .1DiscussionThe results show that community members who frequently interact with others have more chanceshare knowledge with other members in their daily work; and the frequency of interaction is moreimportant than how many people they know in the community. The results are correspondent to thetheoretical argument that in CoPs, the structural dimension provides support to the practical dimension.The results also show that members with more clear understanding of community’s identity willbe more active in the community; they not only know more members in the community but also havemore frequent interactions with other members.To explain why the hypothesis 1 is not supported, we have to consider two factors: one is samplesize, another is the respondents and the setting in which the study took place. The sample size is 54,which is very small. The setting for the study was an educational institution and respondents werestudents in a business school, which is quite different with the organizational setting. For the PhDstudents, learning is their full time job. The knowledge they learned in class and the knowledge theyobtained in reading by themselves contribute more to their performance. In comparison, situated learningonly provides marginal contribution for a full-time student. However, in an organizational setting, the roleof situated learning and the interaction with other people are more important for acquiring knowledge. For166

the people who do not have too much chance of training and formal learning as a student, CoPs willprovide the learning environment they need and will contribute more for their performance.Another difference between this setting and organizational setting is that senior students who ingeneral have better performance than junior students tend to work at home and have less interaction withother students. Because of the required course attendance, junior students have to spend more time inschool and in turns have more chance interact with other students. Although they learned a lot from thecommunity, it doesn’t reflect in their performance in short term. Maybe for this reason the pathcoefficient between situated learning and performance is not only not significant but also negative. Hence,it is necessary to lunch a survey in an organizational setting for investigating the influence of CoPs onindividual performance in future research.Additionally, the scales used for measuring the constructs are newly developed based on somequalitative research and need to be refined. The self-reported performance also potentially induces certainbiases and leads to the inaccuracy of measurement.ConclusionAs the new frontier, communities of practice are emerging in organizations. For researchers, thefirst step is to understand what they are and how they work. The goal of this study is to enrich ourunderstanding of CoPs by building a theoretical model and empirically testing some hypotheses. Theresults show that CoPs have three dimensions with different functions. It is the practice dimension-theself-motivated situated learning that directly enhances individual performance. At the same time, thesituated learning is supported by the structural dimension-social network and guild by the identitydimension.Not limited in improving individual performance, other benefits of CoPs, such as shaping theknowledge strategy and facilitating the knowledge diffusion, also need to be further investigated. I hopethis study will stimulate others to extend this research topic further.This study has some value for theory development as well as for practice. For the managers, it isimportant to realize that CoPs are the “hidden fountainhead” of knowledge and are the key to thechallenge of the knowledge economy (Wenger and Snyder 2000). It requires specific managerial effortsto identify and cultivate CoPs and further integrate them into organization’s structure so that theirpotential can be leveraged. They can support CoPs by encouraging the member’s participation, providingresources they need and acknowledge the value of CoPs. They can also provide a healthy environment forthe growth of CoPs by adjusting the reward systems, work processes, and some relative policies.Although CoPs are self-organized and do not need heavy institutional infrastructures, they will makemore contributions when they get adequate support and their knowledge practices fit with theorganizational environment.167

ReferencesAgresti, W., “Tailoring IT support to communities of practice,” IT Pro (published by the IEEE ComputerSociety), (November/December 2003), 24-28.Allee, V., “Knowledge networks and communities of practice,” Journal of the Organization DevelopmentNetwork, 32 (4), (2000).Argote, L., Organizational learning: creating, retaining and transferring knowledge. Norwell, MA,Kluwer. 1999.Brown, J. & Duguid, P., “Organizational learning and communities of practice: toward a unified view ofworking, learning, and innovation,” Organization Science 2 (1), (1991), 40-57.Cross, R. & Baird, L., “Technology is not enough: improving performance by building organizationalmemory,” Sloan Management Review, 41 (3), (2000).Davenport, E. & Hall, H., “New knowledge and micro-level online organization: ‘Communities ofPractice’ as a development framework”, Proceedings of the 34 th Hawaii international conferenceon system sciences, 2001.Fomell, C., & Larcker, D.F., “Evaluating structural equations models with unobservable variables andmeasurement error,” Journal of Marketing Research, (1981), 381-391.Fox, S., “Communities of practice, Foucault and actor-network theory”, Journal of Management Studies,37 (6), (2000), 853-867.Gongla, P. & Rizzuto, C.R., “Evolving communities of practice: IBM global services experience,” IBMSystems Journal, 40 (4), (2001).Hildreth, P., Kimble, C. & Wright, P., “Computer mediated communications and internationalcommunities of practice,” Proceedings of Ethicomp’98, Erasmus University, the Netherlands,1998, 275-286.Hildreth, P., Kimble, C. & Wright, P., “Communities of practice in the distributed internationalenvironment,” Journal of Knowledge Management, 4 (1), (2000), 27-37.Jones, C. & Esnault. L., “The metaphor of networks in learning: communities, collaboration andpractice”. Proceeding in Networked learning conference 2004.Kimble, C., Hildreth, P. & Wright, P. Knowledge management and business model innovation, IdeaGroup Publishing, 2001.LaContora, J. & Mendonca, D., “Communities of practice as learning and performance support systems,”IEEE, (2003), 395-398.Lave, J. & Wenger, E. Situated learning: legitimate peripheral participation, Cambridge, CambridgeUniversity Press, 1987.Leijen, H.V. & Baets, W. R.J., “A cognitive framework for reengineering knowledge-intensiveprocesses,” Proceedings of the 36th Hawaii international conference on system sciences, 2003.Lesser, E.L. & Storck, J., “Communities of practice and organizational performance”, IBM SystemsJournal, 40 (4), (2001), 831-841.Manville, B. & Foot, N., “Harvest your workers’ knowledge,” Datamation, 42, (1996), 78-83.Nonaka, I. & Konno, N. “The concept of ‘Ba’: building a foundation for knowledge creation”, CaliforniaManagement Review, 40(3), (1998).Nonaka, I., Toyama, R. & Konno, N., “SECI, BA and leadership: a unified model of dynamic knowledgecreation,” in Nonaka, I. and Teece, D. (Eds), Managing Industrial Knowledge. Creation, Transferand Utilization, Sega Publications, London, 2001.Nonaka, I. & Toyama, R., “The knowledge-creating theory revisited: knowledge creation as asynthesizing process”, Knowledge Management Research & Practice 1, (2003), 2-10.Pfeffer, J. & Sutton, R.I., “Knowing what to do is not enough: turning knowledge into action,” CaliforniaManagement Review 42, (1999), 83-108.Rouser, K. & Dorsey, A., “Strong threads of learning – knowledge transfer within communities ofpractice,” IEEE, (2003), 141-145.168

Smits, M. & Moor, A., “Measuring knowledge management effectiveness in communities of practice,”Proceedings of the 37th Hawaii international conference on system sciences, 2004.Smith, M.K., “‘Communities of practice’, the encyclopedia of informal education”,http://www.infed.org/biblio/communities_of _practice.htm. [Accessed 14 February 2004].Ward, A. (2003). Http://www.workfrontiers.com/what _is_community.html, [Accessed April 2004].Wasserman, S. & Galaskiewicz, J., Advances in social network analysis: Research in the social andbehavioral sciences. SAGE Publications, Inc, 1994.Wenger, E., Communities of practice: learning, meaning, and identity, Cambridge University Press,1998.Wenger, E., “Communities of practice, learning as a social system”, Systems Thinker, June, 1998.Wenger, E., McDermott, R. & Snyder, W., Cultivating communities of practice, a guide to managingknowledge. Harvard Business School Press, 2002.Wenger, E., “Knowledge management as a doughnut: shaping your knowledge strategy throughcommunities of practice,” Ivey Business Journal, (2004).169

ASAC 2006Banff, AlbertaManuscript 42 OBD. Lance FerrisUniversity of WaterlooJeanetta VenaExport Development CanadaJohn L. MichelaUniversity of WaterlooUNCERTAINTY, THREAT, AND WORK ATTITUDESIN ANTICIPATION OF A MERGERPrevious studies on mergers (Schweiger & DeNisi, 1991) show perceived uncertaintynegatively impacts work attitudes. The current study suggests uncertainty alone does notimpact work attitudes. Uncertainty interacted with organizational dependence such thatuncertainty only impacts attitudes when dependence is high. Uncertainty was also relatedto objective circumstances (branch proximity).Change can have a powerful impact on individuals, especially if it threatens something of value,such as one’s livelihood. Organizational change has been thought to have the potential to create feelingsof uncertainty, loss of identity, threats to individuals' self-esteem, and threats to job security (Ackerman,1982; Callan, Terry, & Schweitzer, 1994; Nadler, 1982; Schweiger & Ivancevich, 1985). Corporatemergers — in which jobs and roles are frequently reorganized, reclassified, or eliminated — are a salientkind of organizational change that can be stressful and threatening (Buono & Bowditch, 1989; Matteson& Ivancevich, 1990).Although past research on mergers documents reactions such as distress and uncertainty in thecontext of mergers, less is known about the psychological processes that intervene between the mergercontext and outcome variables. For instance, while uncertainty has been identified as a key element feltby employees in the merger process (Schweiger & DeNisi, 1991), it is not entirely clear how uncertaintyoperates or combines with other variables to impact on organizational outcomes. The present study soughtto clarify the role and operation of uncertainty by examining its contextual antecedents and how the effectof uncertainty can be moderated by another variable, dependence on the organization.Mergers, Uncertainty, and ThreatOrganizational mergers are viewed as stressful events for employees, who are thrust into a periodof uncertainty surrounding employment stability, management, and responsibilities (Buono & Bowditch,1989; Scheck & Kinicki, 2000; Schweiger & DeNisi, 1991). Research has shown that, both during andfollowing mergers, employee job satisfaction and organizational commitment may decline, and turnoverintentions and stress levels may increase (Fried, Tiegs, Naughton, & Ashforth, 1996; Terry, Carey, &Callan, 2001; Schweiger & DeNisi, 1991).Uncertainty has been identified as one of the concomitants of mergers which may impactnegatively upon work attitudes (Davy, Kinicki, Kilroy, & Scheck, 1988; Scheck & Kinicki, 2000).Schweiger and DeNisi (1991) demonstrated that perceived uncertainty increases after the announcementof a merger, and that it is accompanied by decreases in organizational commitment and job satisfaction.However, the explanation for uncertainty's association with decreases in favourable work attitudes seemsincomplete. Is uncertainty alone enough to impact work attitudes?170

We suggest that uncertainty is a necessary but not sufficient condition. A more operative variable,threat, combines uncertainty along with dependence on the organization. Our logic parallels the logic ofexpectancy-value models of motivation (Mitchell, 1974). In its simplest form, uncertainty, likeexpectancy, is a cognitive variable or state involving beliefs about future outcomes; dependence, on theother hand, implies a valued outcome. Dependence on the organization exists when people value what theorganization provides (income, pride, social connections, etc.) and they perceive their presentorganizational membership to be particularly favourable for obtaining those provisions (cf., Thibaut &Kelley, 1959, on CL alt ). Thus uncertainty can interact with dependence to yield threat for peopleparticularly concerned about potential loss or other harm deriving from the merger.The concept of threat can explain why employees’ anticipation of a merger could underminework attitudes (e.g., affective commitment and job satisfaction). Affective commitment has been thoughtto develop partially through social exchange processes (Liden, Wayne, Kraimer, & Sparrowe, 2003). Insocial exchange theory (Blau, 1964), two entities exchange valued resources. In the case of employeeorganizationexchange, employees provide organizations with their work efforts in exchange for pay,rewarding social interactions, and other provisions. Employees evaluate the ensuing exchange process andform affective commitment to the organization consistent with their satisfaction with this exchange.Satisfaction, in turn, is a function of the amount of the provisions obtained, their value to the individual,and how these provisions compare with expectations based on factors such as past experience and beliefsabout what alternative employers would provide (Lawler, 1981).Uncertainty introduces the possibility of a breach of this exchange relationship. When uncertaintycombines with dependence, this potential breach matters, and thus the commitment and satisfactionnormally accruing from the exchange may decline from this threat. Accordingly, we expected to find thatwork attitudes such as affective commitment and satisfaction with the organization as employer would benegatively impacted by threat in a merger context, but not by uncertainty per se. Hence, we submit:Hypothesis 1: Perceived uncertainty and dependence on the organization will interact to producea detrimental impact on work attitudes.Continuance Commitment as DependenceMeyer and Allen (1991) proposed that organizational commitment consists of threecomponents — affective, continuance, and normative commitment — which bind an employee to theorganization, but which otherwise differ in terms of their antecedents, consequences, and mind-sets(Meyer & Herscovitch, 2001; Meyer, Stanley, Herscovitch, & Topolnytsky, 2002). For example,affective, continuance, and normative commitment respectively reflect the mind-sets of one who desiresto remain with the organization; of one who remains with an organization due to what they stand to loseby leaving (or their dependence on the job); or of one who feels obliged to remain with the organization.Affective commitment has received the most attention among the commitment components whenit comes to mergers; the present study examined continuance commitment as well. Continuancecommitment can be thought of as the extent to which one is dependent on the employer, or the extent towhich one would be adversely affected by the loss of one’s job. Thus, by measuring dependence on thejob (i.e., continuance commitment), and examining its interaction with a measure of uncertainty, wewould have an assessment of threat per se.Situational Antecedents of Uncertainty171

Schweiger and DeNisi’s (1991) research on uncertainty in mergers showed that levels ofperceived uncertainty can be modified. By introducing a communication plan providing informationabout the merger, they were able to lower perceived uncertainty levels. While organizationalcommunications concerning mergers represent one source of information which impacts uncertainty, thecurrent study looked for an effect of a rather different source of information — an aspect of the objectivesituation that employees were in. Specifically, we looked for differences in perceived uncertainty as afunction of the distance between branches of merging organizations. Employees of branches in closeproximity to branches of the other organization involved with the merger were expected to experiencehigher levels of uncertainty than employees of branches that are not in close proximity. This predictionreflects the reality that uncertainty surrounding branch closures should be greater when one is providing aredundant service with another branch. On this basis, we predict:Hypothesis 2: Perceived uncertainty will be higher in branches that are in closer proximity tobranches of the organization with which the merger is planned.It should be noted that in this hypothesis there appears a component of threat, perceiveduncertainty, but not threat itself. This is because the other component of threat, dependence, would beexpected to vary among individuals within bank branches, because it is tied to individual circumstances.METHODParticipants and ProcedureThis study was conducted in a large financial services institution that had recently announced anintended merger, but had not yet received regulatory approval. Participants were 62 employees from fourarbitrarily selected branches of the consumer banking division. Participants completed the questionnairesand returned them directly to a research assistant. Employees were assured that their responses would notbe divulged to the organization, and they could decline to participate without consequences. The responserate of those invited to participate was 100%.MeasuresPerceived uncertainty. Schweiger and DeNisi’s (1991) 21-item scale of Perceived Uncertaintywas modified to a 16-item measure by eliminating items that our organizational contact deemed irrelevantor otherwise inappropriate (items 2, 6, 7, 14, and 18 from the original survey). A 7-point Likert scale (1 =never a source of uncertainty; 7 = always a source of uncertainty) was used. Example items include"Whether you will be laid off" and "Whether you will be forced to take a demotion" (α = .94).Affective commitment. A modified version of Meyer and Allen’s (1997) 6-item measure ofaffective commitment was used to measure affective commitment, using a 7-point Likert scale (1 =strongly disagree, 7 = strongly agree). Our organizational contact requested the question "I feelemotionally attached to my organization” be replaced with the following item: "I feel loyal to myorganization." (α = .86)Dependence. Two items from Meyer and Allen’s (1997) scale of continuance commitment, "Toomuch of my life would be disrupted if I decided I wanted to leave my organization now" and "I feel asthough I have too few options to consider leaving this organization," assessed employees’ dependence.The remaining items were omitted to reduce overall questionnaire length. A 7-point scale (1 = stronglydisagree, 7 = strongly agree) was used. The two items produced an alpha of .52. Although this lies below172

the cutoff of .60 often recommended, the alpha is understandable from having only two items, and we willsee that the measure did contribute to prediction of responses to the anticipated merger and thus wasadequate.Satisfaction. A two-item measure (“Generally speaking, I am very satisfied with (organizationname) as my employer,” “I would recommend (organization name) as a good place to work”) was used tomeasure job satisfaction. A 7-point scale (1 = Not at all true, 7 = Very true) was used. (α = .93)Work Attitudes and ThreatResultsThe means, standard deviations, and correlations are presented in Table 1. We first examinedwhether affective commitment and satisfaction were lower among employees experiencing greater threat,using moderated multiple regression. To capture threat, we examined the interaction of uncertainty anddependence, analyzing continuance commitment (the indicator of dependence) as a moderator of theeffect of uncertainty on affective commitment and satisfaction.Table 1Means, standard deviations, and intercorrelations between measuresVariableMean(SD)1. Uncertainty 3.63(1.20)2. Dependence 4.17(1.45)3. Satisfaction 5.57(1.41)4. Affective Commitment 4.91(1.06)Note. All N = 62. * p .10.173

Figure 1Interaction of dependence and uncertainty predicting affective commitmentAffective Commitment65.554.543.53LowHighLowDependenceHighDependenceOverall UncertaintyA similar pattern was obtained in the prediction of satisfaction (Figure 2). Uncertainty wasassociated with lower satisfaction when dependence was relatively high. In a regression equationpredicting satisfaction, the interaction of uncertainty and dependence was significant, t(58) = −2.56, p .10). The full model was significant, F(3, 58)= 9.51, p < .001, and accounted for an overall R 2 of .33. Upon examining the simple slopes of the highand low (+/− 1 S.D.) dependence regression lines, the slope of the high dependence regression line wassignificant, t(58) = −4.37, p < .001, while the slope of the low dependence line was not, t(58) = 0.31, p >.10.174

Figure 2Interaction of dependence and uncertainty predicting satisfactionJob Satisfaction7.56.55.54.53.5LowDependenceHighDependence2.5LowHighOverall UncertaintyUncertainty in Relation to Branch Proximity to Merging OrganizationThe four branches surveyed were divided into two groups, "far" and "near" branches. Eachbranch classified as "far" was more than 1 kilometer (.62 miles) in distance from a branch of the potentialmerging organization. We examined mean differences on uncertainty between these groups of branches.The near-far distinction was coded as an a priori contrast in one-way analysis of variance with 4 levels ofthe branch factor. Along with the significant omnibus effect in the ANOVA, F(3, 58) = 8.46, p < .001, thea priori contrast between the near and far branches also yielded statistically significant results (t(58) =3.93, p < .001). As expected, employees in the near branches expressed more uncertainty (M = 4.17, SD =1.06; far branches M = 3.07, SD = 1.09)), supporting Hypothesis 2. This near-far difference is depicted inFigure 3. No significant differences between near-far branches in continuance commitment (dependence)were found, consistent with continuance commitment being an individual difference variable.175

Figure 3Mean differences of uncertainty between branches in near or far proximityOverall Uncertainty4.543.532.521.51FarProximityNearDiscussionWe predicted that the employees who were most threatened by potential consequences of themerger would show the lowest levels of affective commitment and satisfaction with their employer. Thefindings supported our predictions. The form that these findings took, and the finding that branchproximity was associated with employees' levels of uncertainty, illuminated the nature and operation ofthreat in a merger context.The statistical interaction of dependence and uncertainty (Figures 1 and 2) suggests thatdependence on one's organization acts as a boundary condition for merger-related uncertainty to have animpact on work attitudes. Employment uncertainty becomes truly threatening, and thus impactful, whenone is at risk of losing something valued. Because ratings of uncertainty were associated with proximityof one's bank branch to a branch of the other organization in the merger (Figure 3), the experience ofthreat was connected with objective circumstances shared with others in the same branch. At the sametime, threat was connected to idiosyncratic aspects of one's capacities and options, as captured by themeasure of dependence. This discussion will consider the theoretical implications of these findings,followed by practical implications.Theorized MechanismsThe observed decline in affective commitment in a merger context is consistent with otherresearchers' findings of decreases after mergers (Schweiger & DeNisi, 1991; Terry et al., 2001). But howcan we explain these decreases before a merger had occurred? A number of theoretical mechanisms arepossible.176

Social exchange. Our predictions concerning threat were suggested by our application of socialexchange theory to this situation—specifically the relevance of whether the employee places special valueon what is obtained from the employer in the exchange. However, under the simplest interpretations ofsocial exchange theory, the benefits exchanged between employee and employer had not changed as longas the merger had not occurred. Or had the exchange changed? One possibility is that in anticipation of amerger, an inherently aversive and negative element in the form of threat itself was introduced, at least forpeople who stood to lose something substantial. Thus the "ledger" of rewards/costs, and resultingsatisfaction and attitudes, should have become less favourable for some people, consistent with ourempirical findings.The preceding explanation can be labelled as cognitive for its emphasis on perception, evaluation,and integration or other processing of information to arrive at an attitude toward the organization. Otherexplanations draw more on motivational concepts, having to do with implications of the impendingmerger for emotion, self-esteem, and identity.Because a threat to employment also threatens one's status in the family and community as wellas at work, it threatens one's self-esteem and identity (see Abrams and Hogg, 1988; Ashforth & Mael,1989; Callan et al., 1994; Dutton, Dukerich, & Harquail, 1994; Mael & Ashforth, 1992; Schweiger &Ivancevich, 1985). Devaluation of, and/or dis-identification from the organization are two ways ofbolstering self-esteem and maintaining a positive and stable self-identity in a merger context.Devaluation. If employees cope with merger threat by devaluing their employing organization,this response can blunt the psychological effect of the organization's possible rejection of the employee.The motivation here is psychologically defensive and ego-protective: when the employee places lessvalue on the organization and employment with it, he or she has less to lose in the merger. Part of theimplicit logic is that it would not be so bad for the organization to reject me, because my organization isnot so great. This interpretation squares with the present results, in that both of this study's outcomevariables are "work attitudes" and thus concern evaluation of the organization and the job. Lower ratingsobserved on these attitude measures — observed for the most threatened employees — are quiteconsistent with devaluation.Dis-identification. An alternate response, dis-identification, has been described as an activedisconnection and distancing of oneself and one’s values from some social entity (Elsbach, 1998;Dukerich, Kramer, & Parks, 1998). Identification with the organization has been identified as one of theantecedents of affective commitment (Meyer & Herscovitch, 2001). Consequently, the results concerningaffective commitment might reflect influences of threat on identification.Various studies have pointed to dis-identification or social distancing as a response to threatssimilar to those faced by participants in our study (Steele, 1997; Major, Spencer, Schmader, Wolfe, &Crocker, 1998). For example, in Astrachan's (1990) study of relationships among work group members,participants reported engaging in distancing and withdrawal when the end of the group's time forexistence approached. For example, one participant said "I don’t go out of my way to see these peoplelike I used to." In addition, Astrachan’s research implies that the concept of devaluation can be separatedfrom the concept of dis-identification through distancing, because forms of withdrawal were seen that didnot imply devaluation or denigration of group members.Inevitably we must ask, which is it, among the preceding mechanisms that explains anticipatoryreductions in work attitudes among the most threatened employees? We suggest that no single answerwill emerge in future research, because different process could come to the fore depending on specifics ofthe context and characteristics of the work force. For example, devaluation may become more likely whenthe merger announcement isn't communicated with sensitivity; distancing and dis-identification may be177

less likely among workers in a smaller professional community who may end up back together in somenew configuration after merger-induced layoffs. Thus one purpose of detailing various theories andprocesses in this Discussion has been to identify a larger range of variables to measure and a larger rangeof questions to address in future research.Geographic Antecedents of UncertaintyAlthough mergers have been shown previously to increase uncertainty overall (Schweiger &DeNisi, 1991), it was of interest to learn that employees in different parts of an organization canexperience different levels of uncertainty. Specifically, employees in branches with closer proximity to abranch of the merger partner reported significantly more uncertainty. Evidently, employees perceivedthere to be an increased likelihood of layoffs, transfers, and similar consequences of the merger whentheir services seemed more likely to be seen by management as redundant within a neighbourhood.Because uncertainty is a component of threat in our analysis, greater threat thus can be expected amongemployees, on average, in parts of the organization with greater uncertainty.Implications for Research and ManagementResults of the current study suggest that measuring uncertainty alone is not sufficient to captureemployees' full experience of merger situations. A more extensive, indirect assessment of organizationalchange-related threat could be developed, by combining objective indicators of the situation (e.g., in thecurrent study, branch proximity), subjective assessments as with the uncertainty measure, and measures ofemployees' individual circumstances (corresponding to the concept of dependence). Past research hasdemonstrated the effectiveness of indirect, interaction-based measurement of anticipatory coping for otherkinds of threat (Gladis, Michela, Walter, & Vaughan, 1992).The interactive nature of threat indicates that line managers and human resource managementspecialists should address both of the contributors to threat, uncertainty and dependence. The associationof perceived uncertainty with branch location indicates that different groups of employees may needdifferent kinds or amounts of management action, so these differential needs should be identified andaddressed. The uncertainty component of threat traditionally has been addressed by communicating factsof the situation to employees. Schweiger & DeNisi (1991) found that employees in organizations with“realistic merger preview” communications reported significantly less uncertainty than employees inorganizations which lacked such a program. Given our evidence for impacts of threat before the mergerwas actually approved, communication plans should begin as early as possible in the merger process,including in the pre-merger phase.178

ReferencesAbrams, Dominic & Hogg, Michael A., “Comments on the Motivational Status of Self-Esteem in SocialIdentity and Intergroup Discrimination,” European Journal of Social Psychology, 18, (1988),317-334.Ackerman, Linda S, “Transition Management: An In-Depth Look at Managing Complex Change,”Organizational Dynamics, 11, (1982), 46-66.Ashforth, Blake E. & Mael, Fred, “Social Identity Theory and the Organization,” Academy ofManagement Review, 14, (1989), 20-39.Astrachan, Joseph H., Mergers, Acquisitions, and Employee Anxiety: A Study of Separation Anxiety in aCorporate Context, New York: Praeger Publishers, 1990.Blau, Peter M., Exchange and power in social life, New York: Wiley, 1964.Buono, Anthony F. & Bowditch, James L., The Human Side of Mergers and Acquisitions, San Francisco:Jossey-Bass, 1989.Callan, Victor J., Terry, Deborah J. & Schweitzer, Robert, “Coping Resources, Coping Strategies andAdjustment to Organizational Change: Direct or Buffering Effects?,” Work & Stress, 8, (1994),372-383.Davy, Jeanette A., Kinicki, Angelo J., Kilroy, John & Scheck, Christine, “After the Merger: Dealing withPeople’s Uncertainty,” Training and Development Journal, 42, (1988), 56-61.Dukerich, Janet M., Kramer, Roderick &,McLean Parks, Judi, “The Dark Side of OrganizationalIdentification,” In David A. Whetten & Paul C. Godfrey (Eds.), Identity in Organizations:Building Theory Through Conversations, Thousand Oaks, CA: Sage Publications, Inc., 1998.Dutton, Jane E., Dukerich, Janet M. & Harquail, Celia V., “Organizational Images and MemberIdentification,” Administrative Science Quarterly, 39, (1994), 239-263.Elsbach, Kimberly D., “The Process of Social Identification: With What do we Identify?,” In David A.Whetten & Paul C. Godfrey (Eds.), Identity in Organizations. Building Theory ThroughConversation, Thousand Oaks, CA: Sage Publications, Inc., 1998.Fried, Yitzhak, Tiegs, Robert B., Naughton, Thomas J. & Ashforth, Blake E., “Managers’ Reactions to aCorporate Acquisition: A Test of an Integrative Model,” Journal of Organizational Behavior, 17,(1996), 401-427.Gladis, Madeline M., Michela, John L., Walter, Heather J. & Vaughan, Roger D., “High School Students'Perceptions of AIDS Risk: Realistic Appraisal or Motivated Denial?,” Health Psychology, 11,(1992), 307-316.Kelley, Harold H., “Attribution Theory in Social Psychology,” Nebraska Symposium on Motivation: Vol.15, Lincoln, NE: University of Nebraska Press, 1967.Lawler, Edward E., Pay and Organization Development, Reading, MA: Addison-Wesley, 1981.Liden, Robert C., Wayne, Sandy J., Kraimer, Maria L. & Sparrowe, Raymond T., “The DualCommitments of Contingent Workers: An Examination of Contingents’ Commitment to theAgency and the Organization,” Journal of Organizational Behavior, 24, (2003), 609-625.Mael, Fred & Ashforth, Blake E, “Alumni and Their Alma Mater: A Partial Test of the ReformulatedModel of Organizational Identification,” Journal of Organizational Behavior, 13, (1992), 103-123.Major, Brenda, Spencer, Steven, Schmader, Toni, Wolfe, Connie & Crocker, Jennifer, “Coping withNegative Stereotypes about Intellectual Performance: The Role of PsychologicalDisengagement,” Personality and Social Psychology Bulletin, 24, (1998), 34-50.Matteson, Michael T. & Ivancevich, John M., “Merger and Acquisition Stress: Fear and Uncertainty atMid-Career,” Prevention in Human Services, 8, (1990), 139-158.Meyer, John P. & Allen, Natalie J., “A Three-Component Conceptualization of OrganizationalCommitment,” Human Resource Management Review, 1, (1991), 61-89.Meyer, John P. & Allen, Natalie J., Commitment in the Workplace: Theory, Research, and Application,Newbury Park, CA: Sage Publications, 1997.179

Meyer, John P.,& Herscovitch, Lynne, “Commitment in the Workplace: Toward a General Model,”Human Resource Management Review, 11, (2001), 299-326.Meyer, John P., Stanley, David J., Herscovitch, Lynne & Topolnytsky, Laryssa, “Affective, Continuance,and Normative Commitment to the Organization: A Meta-Analysis of Antecedents, Correlates,and Consequences,” Journal of Vocational Behavior, 61, (2002), 20-52.Mitchell, Terrence R., “Expectancy Models of Job Satisfaction, Occupational Preference and Effort: ATheoretical, Methodological, and Empirical Appraisal,” Psychological Bulletin, 81, (1974), 1053-1077.Nadler, David A., “Managing Transitions to Uncertain Future States,” Organizational Dynamics, 11,(1982), 37-45.Scheck, Christine L. & Kinicki, Angelo J., “Identifying Antecedents of Coping with an OrganizationalAcquisition: A Structural Assessment,” Journal of Organizational Behavior, 21, (2000), 627-648.Schweiger, David M. & DeNisi, Angelo S., “Communication with Employees Following a Merger: ALongitudinal Field Experiment,” Academy of Management Journal, 34, (1991), 110-135.Schweiger, David L. & Ivancevich, John M., “Human Resources: The Forgotten Factor in Mergers andAcquisitions,” Personnel Administrator, (1985), 30, 47-56.Terry, Deborah J., Carey, Craig J. & Callan, Victor J., “Employee Adjustment to an OrganizationalMerger: An Intergroup Perspective,” Personality and Social Psychology Bulletin, 27, (2001),267-280.Thibaut, John W. & Kelley, Harold H., The Social Psychology of Groups, Oxford, England: Wiley, 1959.180

ASAC 2006 Jacques Barrette, Ph.D. 1Banff, Alberta Louise Lemyre, Ph.D. 2,3Wayne Corneil, Sc.D. 2Nancy Beauregard, M.Sc. 3Louise Legault, Ph.D. 3École de gestion 1École de Psychologie, Faculté des Sciences Sociales 2Gap-Santé, Institut de Santé des populations 3University of OttawaCONTEXTE DE TRAVAIL ET ENGAGEMENT ENVERS L’ORGANISATION :LE ROLE MEDIATEUR DU SOUTIEN DU SUPERIEUR HIERARCHIQUE 1RÉSUMÉL’objectif de cette recherche est de vérifier dans quelle mesure le soutien du superviseuragit comme une variable médiatrice entre les caractéristiques intrinsèques etextrinsèques d’emploi et l’engagement affectif. Les données ont été recueillies auprès de1664 cadres de la fonction publique fédérale du Canada. Les caractéristiquesintrinsèques mesurées sont l’autonomie au travail, le degré d’utilisation des habiletés, leconflit de rôles, l’ambigüité de rôles, la charge de travail, le conflit intra et intergroupes, la justice distributive, procédurale et interpersonnelle. La modélisation paréquations structurelles (EQS) montre que l’hypothèse de la médiation est corroborée. Lesrésultats indiquent que la justice procédurale, l’utilisation des habiletés, l’autonomie,l’ambigüité de rôles, les conflits intra et inter groupes partagent leur varianceprédictrice entre le soutien du superviseur et l’engagement affectif envers l’organisation,alors que la charge de travail, la justice interpersonnelle et le conflis de rôles sont liésdirectement avec le soutien du superviseur et indirectement avec l=engagement. Cesrésultats indiquent que les employés attribuent au superviseur un grande partie de laresponsabilité de la création du milieu de travail et des décisions concernant leurstâches, et ce même si certaines de ces caractéristiques intrinsèques et extrinsèques sontindépendantes de son pouvoir. En effet, plusieurs de ces caractéristiques relèvent souventde décisions prises par l’organisation ou de situations hors de son contrôle.L’importance théorique et pratique des résultats est discutée.1Cette recherche a été rendue possible grâce à une subvention de recherche accordée par le Conseil derecherches en sciences humaines du Canada aux Docteur. Les auteurs aimeraient remercier Dr. DanielCoulombe, professeur titulaire à la faculté de psychologie de l’université d’Ottawa pour ses conseils éclairéslors des analyses statistiques. Les auteurs aimeraient également remercier leur partenaire communautairel’Association des Professionnels Exécutifs du Canada et surtout les exécutifs participants à cette étude.181

IntroductionLe thème de l’engagement organisationnel attire, depuis plus de 20 ans, une attentionconsidérable dans les milieux de recherche en psychologie industrielle et organisationnelle (Monday,Steers et Porter, 1979; Mottaz, 1988; Allen et Meyer, 1990; Riketta, 2002). Cet intérêt soutenu s’expliquepar le fait que l’engagement est considéré comme un déterminant majeur de multiples comportementsorganisationnels tels que l’absentéisme, le roulement et le rendement au travail (Scholl, 1981; Spector,1997; Steers, 1977). De toutes les formes d’engagement, l’engagement affectif est celui qui est le plussouvent cité dans les revues de la littérature (Mathieu et Zajac, 1990; Meyer et Allen, 1997; Mottaz,1998; Riketta, 2002), et ce en raison de son influence particulière sur la plupart des comportementsbénéfiques pour l’organisation. L’engagement affectif renvoie à l’attachement émotionnel d’un employéenvers son organisation, les employés les plus fortement engagés s’identifiant et s’impliquant activementau sein de leur organisation, tout en appréciant en être membre (Allen et Meyer, 1990). Cet engagementest d’autant plus important lorsqu’il s’adresse aux gestionnaires qui dirigent une organisation puisque leurattitude envers leur propre organisation influence les comportements de leurs subordonnés. Dans un telcontexte, il n’est donc pas étonnant de constater l’intérêt marqué d’un nombre croissant d’organisationspour la création et l’implantation d’environnements de travail favorables à l’engagement affectif.Plusieurs antécédents à l’engagement ont été identifiés 2 mais peu d’études ce sont attardées au rôle quepeut jouer le superviseur en ce qui concerne le soutien qu’il accorde à ses subordonnés dansl’accomplissement de leur travail. La présente étude vise donc à identifier les caractéristiques du milieude travail qui influencent la perception qu’a l’employé du soutien que lui donne son superviseur, et devoir dans quelle mesure ce soutien perçu agit comme médiateur de son engagement affectif enversl’organisation.Le rôle du superviseurLe superviseur est le principal intervenant et représentant de l’organisation, ses décisions et sonstyle de gestion sont des éléments déterminants qui tendent à moduler les comportements d’engagementdes subordonnés (Chiok Foong Loke, 2001; McNeese-Smith, 1997). Plus particulièrement, les réponsesaux efforts individuels et de groupe ainsi qu’à leurs besoins et problèmes sont souvent interprétés commedes actions reflétant la culture entière de l'organisation (Jones, Flynn, et Kelloway, 1995). SelonStinglhamber et Vandenberghe (2003), la perception chez les subordonnés d’un traitement positifémanant du superviseur traduirait une action favorable de l’organisation à leur égard. La compréhensionde l’association entre le soutien du superviseur et ses conséquences sur les comportements d’engagementdu subordonné s’appuie sur la théorie de l’échange social (exchange theory) de Levison (1965). Cettethéorie avance que les subordonnés voient les actions prises par les agents de l'organisation comme unereprésentation des actions de l'organisation elle-même; un processus décrit comme la personnification del'organisation. Levison (1965) précise que les actions prises par les agents organisationnels sont vuescomme une indication des intentions de l'organisation plutôt que d'être attribuées seulement au motifpersonnel des agents (Cette phrase est lourde et pas claire. A simplifier). Cette personnification estencouragée par les responsabilités légales, morales et financières de l'organisation pour les actions desuperviseurs et par le pouvoir que ceux-ci exercent sur les subordonnés. Puisque les superviseurs agissentcomme les agents principaux de l’organisation, la perception de traitement favorable de la part dusuperviseur contribue à la perception du soutien organisationnel (Rhodes et Einsenberg, 2000) Ainsi, de lamême façon qu'un subordonné se forme une perception globale concernant la valeur que lui accorde sonorganisation, il développe une perception générale concernant le degré avec lequel le superviseur valorisesa contribution et se préoccupe de son bien-être (Kottke et Sharafinski, 1988). Selon McNeese-Smith(1997) l'engagement affectif envers l’organisation serait influencé lorsque le superviseur exprime son2Voir les revues de littérature de Mathieu et Zajac (1990), de Mottaz (1998) et de DeCotiis et Summers (1987)pour le répertoire des nombreux antécédents à l’engagement.182

appréciation, de la confiance et une communication ouverte. Le soutien du superviseur apparaît donccomme une variable prédictrice de l’engagement affectif envers l’organisation.Antécédents de l’engagement et du soutien du superviseurDans leur méta-analyse, Mathieu et Zajac (1990) concluent que les caractéristiques de l’emploitelles que l’autonomie (p.ex. contrôle, latitude décisionnelle), le défi (p.ex. responsabilités, habiletésvariées utilisées), les variables de rôles (p.ex. conflit de rôles) ont une importance significative dans ledéveloppement de l’engagement affectif. Les travaux d’Allen et Meyer (1990) rapportent des résultatssimilaires pour d’autres dimensions de l’expérience intrinsèque du travail (p. ex. défi dans l’emploi, statutde rôles, cohésion d’équipe, participation, habiletés). Ces dimensions s’avèrent aussi modérément etsignificativement associées avec l’engagement affectif avec des corrélations de Pearson observées del’ordre de 0,46 à 0,87. Dans une moindre mesure, l’engagement affectif a été positivement etsignificativement associé à des dimensions extrinsèques du travail, c’est-à-dire des dimensions qui nesont pas reliées directement à la réalisation de la tâches. Parmi celles-ci on retrouve en particulier lesprocessus de gestion des ressources humaines (DeCostiis et Summers, 1987), la perception de justice (p.ex. récompenses organisationnelles) (March et Simon, 1958), l’âge, l’éducation, le niveau hiérarchique etle salaire (Mathieu et Zajac, 1990).Parallèlement aux études sur les antécédents de l’engagement, plusieurs chercheurs ont tentéd’identifier les antécédents du soutien organisationnel perçu par le subordonné. Ces chercheurs ont concluque le soutien organisationnel et l’engagement affectif sont des processus distincts qui partagent lesmêmes antécédents. Ces antécédents sont définies comme un ensemble de conditions intrinsèques etextrinsèques d’emploi qui influencent le soutien organisationnel tout autant que l’engagement(Eisenberger, Fasolo, Davis-LaMastro, 1990; Guzzo, Noonan, Elron, 1994; Hackett, Bycio, Hausdorf,1994; Settoon, Bennett, Liden, 1996; Wayne, Shore, Liden, 1997). Des études antérieures ont soulignéque la perception de soutien du superviseur est un mécanisme sous-jacent à de l’engagement enversl’organisation et une composante importante, si non la plus importante du soutien envers l’organisation(Kottke et Sharafinski, 1988; Stinglhamber et Vandenberghe, 2003) avec des corrélations de l’ordre de0,60 (Eisenberger et al., 2002. En conséquence, il est permis de penser que la perception du soutien dusuperviseur et le soutien de l’organisation partagent des mécanismes explicatifs communs (dimensionsintrinsèques et extrinsèques du travail). Autrement dit, les variables qui prédisent l’engagement devraientaussi prédire le soutien du superviseur. D’autre part, le superviseur joue un rôle important dans ledéveloppement du sentiment d’engagement. Sur ce plan, la recherche de Stinglhamber et Vandenberghe(2003) indique que les subordonnés manifestent plus d’engagement affectif envers leur organisation dansla mesure où ils reçoivent du soutien de leur superviseur. Compte tenu du rôle important que joue laperception du soutien du superviseur dans la prédiction de l’engagement, nous formulons l’argument àl’effet que cette variable agira en tant que variable médiatrice entre les antécédents contextuelsintrinsèques et extrinsèques de travail et l’engagement envers l’organisation.Le rôle du soutien organisationnel comme variable médiatrice a déjà reçu l’attention de plusieurschercheurs. En effet, les études de Rhoades, Eisenberger et Armeli (2001) et de Eisenberger, Fasolo etDavis-LaMastro (1990) indiquent que la perception du soutien organisationnel est un antécédent àl’engagement et agit également comme variable médiatrice entre certaines caractéristiques de travail(p.ex. justice procédurale) et l'engagement affectif envers l’organisation. Considérant le rôle prépondérantdu soutien du superviseur sur le soutien organisationnel, il est attendu que les antécédents qui ont unimpact sur le soutien organisationnel aient aussi par concomitance un effet sur le soutien du superviseur.Sur la base de ces considérations, l’objectif de la présente recherche est donc d’explorer d’unepart l’importance des caractéristiques intrinsèques et extrinsèques du travail dans la prédiction du soutiendu superviseur, et d’autre part, de vérifier dans quelle mesure le soutien du superviseur agit comme183

médiateur entre ces caractéristiques et l’engagement envers l’organisation. Les paragraphes suivantsprésentent une synthèse des recherches liant différentes conditions intrinsèques et extrinsèques àl’engagement envers l’organisation et au soutien du superviseur.Conditions intrinsèques : soutien du superviseur et engagementLes conditions intrinsèques réfèrent au contenu de l’emploi ou au contexte de l’environnementimmédiat de travail (p.ex. utilisation de ses habiletés, responsabilités personnelles, etc.). Plusieurschercheurs affirment que ce sont prioritairement les conditions intrinsèques qui ont trait à la nature destâches qui déterminent l'attitude des employés envers leur organisation (Mottaz, 1988). S’appuyantsurtout sur le modèle d’Hackman et Oldham (1976), plusieurs caractéristiques d’emploi ont été mises enrelation avec le soutien et l’engagement organisationnel. Deux caractéristiques ont retenu l’attention deschercheurs, soit l’autonomie et la variété des habiletés utilisées dans le poste.Autonomie et habiletés. L'autonomie est définie comme la perception qu’a le subordonné ducontrôle qu’il détient sur la planification, l’exécution et la réalisation de son travail. L’autonomie dans leposte de travail est considérée comme un déterminant de l'engagement organisationnel (Mottaz, 1998). Laréduction de l’autonomie (p.ex. peu de contrôle, supervision trop étroite) entraînerait une diminution del’engagement (Etzioni, 1975; Salancik et Pfeffer, 1978; De Cotiis et Summers, 1987). Une autredimension intrinsèque liée à l’implication au travail d’importance a trait à la variété des habiletés utiliséesdans le poste (Hackman et Oldham, 1976). Un emploi riche en habiletés tout en représentant un défi pourle subordonné, lui offre des opportunités de développement ainsi que la possibilité de rencontrer sesbesoins d’accomplissement. Plusieurs études ont montré des liens positifs entre cette dimension etl’engagement affectif envers l’organisation (Durham et Grube, & Castaneda, 1994; Stone et Porter, 1975;Hall et Schneider, 1970).Conflits inter/intra groupes. Le conflit est un processus qui débute lorsqu’un individu ou ungroupe d’individus perçoit un écart ou de l’opposition relativement à un autre individu ou un groupe auniveau des intérêts, croyances, valeurs ou buts qui lui sont importants (De Dreu et al., 1999). De manièregénérale, les conflits intra groupe sont fréquents au travail. Sur ce plan, les résultats de Jehn et Mannix(2001) montrent que tous les types de conflits intra groupe (relationnels, tâches, processus) sont corrélésnégativement avec la performance du groupe. D’autre part, avec l’augmentation du travail d’équipe et lacréation d’équipes inter fonctions, les conflits inter groupes sont devenus eux aussi inévitables (Whettenet Cameron, 2004). En effet, la réalisation de projets organisationnels nécessitant la contribution demultiples équipes est un défi compte tenu des barrières structurelles et psychologiques qui se posent entreles groupes (Van Knippenberg, 2003). Les conflits intra ou inter groupes ont un impact sur plusieursdimensions organisationnelles tels que la performance au travail, la satisfaction, la rétention du personnel,l’efficacité, l’innovation et l’engagement au travail (De Dreu & Weingart, 2003a,b; Jehn, 1995 ;Murnigham et Colon, 1991; Jehn, Northcraft et Noale, 1999 ; Richter, Scully et West, 2005).Bien que les membres d’une équipe aient un rôle à jouer dans la gestion de leurs propres conflits,il n’en demeure pas moins que cette responsabilité relève aussi du superviseur. Ce dernier est en grandepartie responsable du climat de travail qui prévaut dans une unité de travail. La négligence d’intervenirlors d’un conflit mal géré affectera l’efficacité de l’unité et de l’organisation. La littérature sur ce sujetindique qu’une approche de résolution de problèmes est une des stratégies les plus efficaces pourfavoriser la résolution des conflits intra et inter groupes (Blake, Sheppard & Mouton, 1966; De Dreu,Harinek & Van Vianen, 1999; Nauta & Landers, 2000; Thomas, 1992). Toutefois, indépendamment de lastratégie utilisée pour gérer efficacement ces situations, la présence de conflits impliquentpresqu’inévitablement l’intervention et le soutien du superviseur (Allport, 1954; Brewer et Brown, 1998;Hartley, 1996; Lawrence et Lorsch, 1967; Richter, Van Deck & West, 2004). Cette position théoriquevoulant que le soutien du superviseur agisse comme facteur médiateur entre la présence de conflits et184

l’engagement affectif envers l’organisation reste, à notre connaissance, à établir empiriquement. Étantdonné la nature et la centralité des implications sur le plan de l’intervention organisationnelle impliquées,la compréhension de cette question mérite que l’on s’y attarde de plus près.Rôles stresseurs. Les rôles stresseurs illustrent des situations intrinsèques de travail et réfèrentaux demandes liées à l’emploi avec lesquelles l’individu a de la difficulté à transiger (Lazarus etFolkman, 1984). Notamment, trois types de rôles retiennent l’attention des chercheurs soit (a) la charge detravail qui implique une demande de travail au-delà de ce que le subordonné peut raisonnablementaccomplir dans une période de temps, (b) l’ambigüité de rôles qui implique un manque de clartéconcernant les attentes associées à leur rôle, la façon de le remplir ou les conséquences de leurperformance, et (c) le conflit de rôles qui implique une incompatibilité mutuelle dans les responsabilités(Cooper et Marshall, 1976). Mathieu et Zajac (1990) concluent dans leur revue de la littérature que cestrois rôles ont un impact significatif sur le développement de l’engagement. Les recherches montrent quel’ambiguïté de rôles décroît l’engagement en réduisant le lien perçu entre les rôles du subordonné etl’atteinte d’objectifs organisationnels tandis que le conflit de rôles entraîne une diminution du sens desresponsabilités (Salancik, 1977a). Par ailleurs, la revue de littérature de Rhodes et Eisenberg (2002)identifie ces trois rôles comme des conditions ayant un impact négatif sur le soutien organisationnel, unconstruit fortement corrélé avec le soutien du superviseur (Eisenberger et al., 2001). Pour leur part,Bedeian et Armenakis (1981) ont montré que l’ambiguïté de rôles et le conflit de rôles sont reliésnégativement et significativement au soutien du superviseur.Variables médiatrices. Selon Hutchison et Garstka (1996) et Guzzo et ses collaborateurs (1994)la relation entre plusieurs caractéristiques intrinsèques et l’engagement affectif envers l’organisationserait indirecte. Des variables médiatrices perceptuelles agiraient pour en affecter l’importance(Stinglhamber et Vandenberghe, 2003). Il est raisonnable de penser que le soutien du superviseur puisseêtre l’une de ces variables. Par exemple, l’autonomie et l’utilisation d’habiletés sont grandementtributaires du style de gestion et de la confiance accordée par le superviseur (Stumpf and Hartman, 1984 ;Salancik 1977b). D’autre part, nous savons que le soutien du superviseur agit comme tampon entre lesconditions de travail stressantes et plusieurs manifestations de santé et comportements organisationnelssous-optimaux (p.ex. la consommation excessive d’alcool, satisfaction, absentéisme) (Karasek &Theorell, 1990). Enfin, le superviseur s’impose comme un acteur de première ligne pour agir lors deconflits intra ou inter groupes, faute de quoi il risque de voir la performance de son unité affectée(Whetten & Cameron 2004). Ainsi, on peut avancer que le soutien du superviseur produira un étatpsychologique interne (médiateur) pouvant modifier l’influence de ces caractéristiques intrinsèques surl’engagement affectif du subordonné envers l’organisation.Conditions extrinsèques : soutien du superviseur et engagementLes conditions extrinsèques sont des facteurs contextuels venant de l’environnement de travailqui comportent une influence sur les comportements organisationnels des membres (Wernimont, 1996).Ces conditions sont souvent associées aux politiques organisationnelles touchant entre autres lesrécompenses reçues (p.ex. rémunération) (Hackman et Oldham, 1976). Parmi les caractéristiquesextrinsèques liées à la fois à l’engagement et au soutien qui retiennent l’attention des chercheurs ontretrouve au premier plan la justice distributive, procédurale et interpersonnelle.Justice distributive, procédurale et interpersonnelle. La justice distributive évalue jusqu’à quelpoint le subordonné perçoit équitable les récompenses reçues, et ce en comparaison à ce que reçoivent lesautres. La justice procédurale assure que l’on respecte dans tous les cas où elles s’appliquent, les règles etprocédures inscrites dans les politiques de l’organisation (Shermerhorn, Hunt et Osborn, 2000). Pour sapart, la justice interactionnelle renvoie à la composante sociale des pratiques organisationnelles etimplique la qualité du traitement interpersonnel dans l’allocation des ressources (Bies, 1987). Les185

echerches indiquent que la perception de justice est un déterminant important de l’engagementorganisationnel (Mottaz, 1988). Plus particulièrement, des études empiriques montrent une relationpositive entre le soutien organisationnel et la justice procédurale et interactionnelle (Moorman et al.,1998; Rhoades et Eisenberger, 2002; Shore et Shore, 1995). Cette perception du soutien organisationnelagit comme médiateur entre les sentiments de justice et l’engagement affectif. (Allen, Shore, et Griffeth,1999; Moorman, Blakely, et Niehoff, 1998). Comme le superviseur est le principal agent del’administration de cette justice, on peut s’attendre à un phénomène similaire en ce qui concerne lesoutien du superviseur. Dans les cas où le superviseur a un contrôle limité sur les caractéristiquesextrinsèques de justice (p.ex. politiques de rémunération), sa façon de les gérer risque tout de mêmed’affecter grandement la perception qu’auront ses subordonnés de son soutien. Les recherches deKonovsky et Pugh (1994) convergent dans ce sens. Selon ces auteurs, la confiance envers le superviseur(un concept s’apparentant au soutien du superviseur) agit comme médiateur entre la justiceorganisationnelle et les comportements et attitudes du subordonné. Tenant compte de ce rôle prépondérantdu superviseur dans divers processus (p.ex. attribution de récompenses, répartition des responsabilités,allocation des ressources) et de l’influence qu’ont ces processus sur les attitudes des subordonnés(Hutchison et Garstka, 1996), on devrait s’attendre à ce qu’une évaluation négative ou positive de cesformes de justice affecte la perception qu’a le subordonné de son superviseur, tout comme ces évaluationsaffectent le soutien organisationnel perçu (Wayne, et al., 1997; Rhoades et Eisenberger, 2002) etl’engagement organisationnel (Mathieu et Zajac, 1990; Price et Mueller, 1986; Putti, Aryee et Liang,1989). De plus, même si le superviseur peut avoir un contrôle limité sur les caractéristiques extrinsèquesde justice (p.ex. politiques de rémunération), sa façon de les gérer peut moduler grandement la perceptionqu’auront ses subordonnés de son soutien. Il appert donc que le superviseur peut agir comme médiateurentre la justice procédurale, distributive et interactionnelle, et l’engagement envers l’organisation.Objectif de la recherchePour récapituler, l’objectif de cette recherche est de vérifier dans quelle mesure le soutien dusuperviseur agit comme variable médiatrice dans la prédiction du lien entre les conditions de travailintrinsèques et extrinsèques et l’engagement affectif envers l'organisation. En s’appuyant sur les résultatsd’études antérieurs, nous anticipons que le soutien du superviseur agira comme une variable médiatriceentre (a) les caractéristiques intrinsèques d’emploi (autonomie, habiletés utilisées, conflit de rôles,ambigüité de rôles, charge de travail, conflit intra/inter groupes) et l’engagement organisationnel, et (b)entre les caractéristiques extrinsèques de justice (distributive, procédurale, interpersonnelle) etl’engagement organisationnel. Compte tenu qu’aucun modèle théorique ni recherche empirique ne nouspermet d’avancer d’hypothèses précises sur le partage de la variance des facteurs dans l’explication dusoutien du superviseur et de l’engagement affectif, nous posons à priori l’hypothèse d’une médiationtotale dans la prédiction du lien entre les caractéristiques intrinsèques et extrinsèques de travail etl’engagement affectif.MÉTHODOLOGIEParticipants et mode de recrutementLes données proviennent de l’enquête nationale de l’APEX (Association professionnelle descadres de la fonction publique fédérale). En février 2002, un questionnaire a été envoyé sous la directionde l’équipe de chercheurs par le Service des ressources humaines de l’ensemble des ministères et agencesdu gouvernement du Canada. Au total, 3670 participants de direction (cadres) de la fonction publiquefédérale ont été ciblés. Les trousses bilingues comprenaient un questionnaire, une lettre de recrutement etde consentement éclairé, une enveloppe de retour préaffranchie et préadressée à une boîte postale horsgouvernement. Les réponses sont demeurées anonymes et les résultats ont été compilés par les186

chercheurs. Près de 40 % des cadres ont répondu. Le tableau 1 présente la composition en pourcentage del’échantillon comparativement à la composition de la population des cadres de la fonction publique.L’analyse comparative (Treasury Board, 2002) indique que l’échantillon reflète étroitement lacomposition de la population générale des cadres en terme de sexe, du niveau hiérarchique (EX1 à EX5)et de l’âge. En tout, 2.1 % des fonctionnaires détiennent un diplôme de niveau secondaire, 10.4 % deniveau collégial et le reste de niveau universitaire. L’étude a reçu l’approbation du comité de déontologiede l’Université d’Ottawa.Variables indépendantesLes échelles pour mesurer l’autonomie, l’utilisation des habiletés, la charge de travail, le conflitde rôles, l’ambigüité de rôles et les conflits inter et intra groupes proviennent du Generic Job StressQuestionnaire.Table 1Analyse comparative de la composition de l’échantillon decadres de la fonction publique fédéraleSexe Niveaux hiérarchiques a Âge bHommes FemmesEX1 EX2 EX3 EX4 EX5 1 2 3 4 5 6Population 68.1 % 31.9 % 52 % 25 % 16 % 5 % 2 % 6% 13% 24% 35% 13% 4%Échantillon 66.8 % 33.2 % 51 % 26 % 15 % 6 % 2 % 5% 13% 23% 35% 19% 4%a) cinq niveaux hiérarchique (EX1 à EX5) distinguent les répondants, le niveau EX5 étant le plus élevé.b) 1 = âge de moins de 40 ans, 2 = de 40 à 44 ans, 3 = de 45 à 49 ans, 4 = de 50 à 54 ans, 5 = de 55 à 59 ans, 6 = plus de 60 ans.(GJSQ; Hurrell & Mclaney, 1988) développées par les chercheurs du National Institute forOccupational Safety and Health (NIOSH, U.S.).Autonomie. L’autonomie mesure la liberté d’action détenue par le répondant sur divers aspects deson travail. Les répondants évaluent en 10 énoncés sur une échelle Likert de 5 points (très peu àénormément) en répondant aux consignes suivantes: indiquez le degré d’influence que vous exercezactuellement dans chacun des différents secteurs ci-dessous. Par "influence" nous entendons le degré decontrôle que vous exercez sur ce qui est fait par les autres et la liberté de choix que vous avez dans votrepropre travail. Quelle influence avez-vous: 1) lorsqu’il s’agit d’obtenir le matériel et l’équipementnécessaires pour faire votre travail?; 2) etc. La fidélité interne (Cronbach alpha) de l’échelle est de 0,84.Utilisation des habiletés. Cette dimension mesure jusqu’à quel point le répondant perçoit utiliserses habiletés. (p.ex. dans quelle mesure vous donne t-ont la chance de faire ce dans quoi vous excellez?).Les répondants devaient évaluer 3 énoncés sur une échelle en 5 points (rarement à très souvent), Lafidélité interne (alpha) de l’échelle est de 0,84.Charge de travail. La charge de travail a été mesurée avec 5 énoncés reflétant la quantité detravail effectuée et la contrainte de temps (p.ex. dans quelle mesure votre emploi vous oblige t-il àtravailler très fort). Les répondants devaient évaluer chaque énoncé sur une échelle en 5 points (rarementà très souvent). La fidélité interne (alpha) de l’échelle est de 0,84.187

Ambiguïté et conflit de rôles. L’ambigüité de rôles a été mesurée au moyen de 4 énoncés sur uneéchelle en 7 points (totalement faux à totalement vrai) (p.ex. :je connais mes responsabilités - scoreinversé). De la même façon, le conflit de rôles fut mesuré au moyen de 5 énoncés (p.ex. je reçois desdemandes contradictoires de deux personnes ou plus). La fidélité interne (alpha) est de 0,84 et de 0,72pour chaque échelle.Conflits intra/inter groupes. Les mesures concernant les conflits intra groupe ont été obtenues endemandant aux répondants d’évaluer, sur une échelle en 5 points (absolument faux à absolument vrai),différentes conditions de travail (p.ex. mon groupe travaille dans l’harmonie). La même procédure a étéappliquée pour la mesure des conflits inter groupes (p.ex mon groupe et les autres groupes travaillent defaçon harmonieuse pour atteindre les objectifs généraux de l’organisation. La fidélité interne (alpha) dechaque échelle est respectivement de 0,83 et 0,80.Justice distributive, procédurale, interactionnelle. Les mesures de la justice distributive (4énoncés), procédurale (7 énoncés) et interactionnelle (4 énoncés) proviennent de l’échelle de Colquitt(2001). Les répondants devaient évaluer chaque énoncé sur une échelle en 5 points (très peu àénormément). La justice distributive réfère au processus organisationnel appliqué pour établir larémunération au rendement (p.ex. votre prime au rendement correspond-elle aux efforts que vous avezfaits dans votre travail?). La justice procédurale reflète la démarche utilisée par leur superviseur lors duprocessus d’évaluation du rendement (p.ex. avez-vous pu exprimer votre point de vue et vos sentimentspendant ce processus? La justice interactionnelle concerne l’évaluation par le répondant des dimensionsqualitative de ses interactions avec son superviseur lors du processus d’attribution de la prime aurendement (p.ex. lors du processus d’attribution de la prime au rendement votre superviseur vous a t-iltraité avec respect). La fidélité interne (alpha) de chaque échelle est respectivement de 0,90, 0,80 et 0,94.Variable médiatriceLa mesure du soutien perçu du superviseur provient de l’échelle développée par Barh (1995) etDuxbury et Higgins (2001). Cette échelle a été spécifiquement développée pour mesurer le soutien dusuperviseur auprès d’employés de la fonction publique fédérale du Canada. Les répondants devaientévaluer 16 énoncés sur une échelle en 5 points (fortement en désaccord à fortement d’accord) enrépondant à l’instruction suivante: mon directeur/ superviseur: 1) me complimente lorsque je fais bienmon travail; 2) etc. La fidélité interne de l’échelle est de 0,90.Variable dépendanteLa mesure de l’engagement affectif (6 énoncés) provient de l’échelle de Meyer, Allen et Smith(1993). Les répondants devaient indiquer leur degré d’accord sur une échelle en 6 points (fortement endésaccord à fortement en accord) avec une série d’énoncés (p.ex. : je serais heureux de compléter macarrière dans cette organisation). La fidélité interne (alpha) de l’échelle est de 0,86.Traitement des variablesEn nous appuyant sur une procédure similaire à celle utilisée par Bagozzi et Heatherton (1994) etBrooke, Russell et Price (1988), nous avons réduit le nombre d’énoncés en créant 3 ou 4 indicateurs parconstruit. Pour chacun des construits, l’énoncé ayant la plus haute pondération (reliability) et celui ayantla plus basse pondération ont été combinés pour créer un indicateur. Cette façon de procéder a été suiviejusqu’à ce que tous les énoncés d’un même construit aient été utilisés.188

Analyses préliminairesValidité de construit. La première étape avant d’évaluer un modèle structurel est de vérifier lavalidité de construit des mesures utilisées (Byrne, 1994). Dans un premier temps, une analyse factorielleen composante principale avec rotation varimax a été effectuée sur l’ensemble des énoncés sans imposer àpriori un nombre de facteurs. Onze facteurs ont émergé clairement expliquant 75 % de la variance et tousles énoncés associés aux facteurs obtiennent des poids pondérés (factors loading) supérieurs à 0,73. Dansun deuxième temps, une analyse factorielle confirmatoire ML (maximum likelihood estimation) a étéutilisée afin de vérifier la structure factorielle. Les résultats indiquent que le modèle à 11 facteurs décritadéquatement les données. Tous les indicateurs d’adéquation des données rencontrent les exigences debase confirmant la structure factorielle, soit un indice d’erreur d’approximation RMSEA = 0,03, inférieurau critère limite recommandé de 0,05 (Brown & Cudeck, 1993). De plus, les indices de qualitéd’ajustement (goodness of fit) sont tous supérieurs à la limite exigée de 0,90 (CFI = 0,96, CGI = 0,92,NNFI = 0,96) (Bentler & Bonett, 1980). En conséquence, les scores factoriels sont utilisés dans lesanalyses de modélisation subséquentes.Rôle modérateur du soutien du superviseur. Nous avons d’abord vérifié la pertinence du choixd’un modèle médiateur plutôt que d’un modèle modérateur en ce qui a trait au rôle du soutien dusuperviseur dans la prédiction statistique de l’engagement affectif. Un modèle modérateur serait appropriési la relation causale entre le soutien du superviseur et l’engagement affectif changeait en fonction desautres variables (Baron & Kenny, 1986). Pour ce faire, nous avons conduit 9 régressions hiérarchiquesmodératrices soit une pour chaque facteur, en contrôlant l’âge, le genre, le niveau d’éducation et leniveau hiérarchique. Aucun effet d’interaction ne s’est avéré significatif, permettant de conclure qu’unmodèle modérateur est inapproprié.Statistiques descriptives et corrélations. Nous avons par la suite évalué les conditions de base(Baron et Kenny, 1986) sous-jacentes à l'effet de médiation du soutien du superviseur. Les scores ont étécalculés en prenant la moyenne des énoncés pour chaque échelle. Les moyennes, écart-types, lescoefficients de fidélité interne (coefficient alpha) et les corrélations entre les variables sont présentés à latable 2. Les corrélations entre les variables de conditions de travail (intrinsèques, extrinsèques) et laperception du superviseur et l=engagement affectif, et celles entre la perception de soutien etl’engagement sont toutes significatives (p ≤ 0,000). Ces résultats remplissent la première condition demédiation et montrent qu’il est approprié de poursuivre les analyses d'effets médiateurs.Analyses par équations structurellesRôle médiateur du soutien perçu du superviseur. L’approche de modélisation par équationsstructurelles (structural equation modeling) à l’aide du logiciel AMOS 5 (Arbuckel, 1997) a été utiliséepour évaluer les autres conditions pour la médiation. Dans un premier temps, nous avons vérifié le modèlede médiation totale du soutien du superviseur, c'est-à-dire le modèle liant directement tous les facteursintrinsèques et extrinsèques au soutien du superviseur, et par la suite en liant ce dernier facteurdirectement à l’engagement. Les indices de qualité des ajustements (goodness of fit) recommandés parplusieurs chercheurs (Byrne, 1994; Long, 1983) (table 3) s’avèrent adéquat (CFI = 0,96, GFI = 0,97, TLI= 0,95, RMSEA = 0,04). Tous les paramètres spécifiés du modèle sont significatifs (p ≤ .001). Dans cemodèle, 59 % de la variance dans le soutien du superviseur est expliquée par les facteurs intrinsèques etextrinsèques de travail, et 18% de la variance de l’engagement affectif est expliquée par le soutien dusuperviseur. Sur la base des indicateurs d’ajustements LM (Lagrange Multiplier test), un modèle demédiation partielle a été ensuite testé (figure 1). La différence significative dans le chi-carré (Π 2 diff(129,14, dl = 5, N = 1664) = ≤ 0,05) montre que le modèle de médiation partielle représente mieux lesdonnées brutes comparé au modèle de médiation totale. De plus, les indices de qualité des ajustements189

(goodness of fit) sont tous supérieurs dans le modèle de médiation partielle (CFI = 0,98, GFI = 0,98, TLI= 0,97), et la variance expliquée augmente de 7 % (R 2 = 0,25). Comme évaluation supplémentaire dumodèle de médiation, nous avons comparé le modèle de médiation partielle avec le modèle alternatif quilie tous les facteurs à la fois au soutien du superviseur et à l’engagement affectif. Dans ce modèlealternatif, les cheminements du modèle de médiation partielle demeurent significatifs alors que 2 des 4cheminements nouveaux ne sont pas significatifs et les deux autres ont un niveau de significationinférieur au modèle précédent. Ces résultats permettent de conclure que le modèle de médiation partielleprésenté en figure 1 est celui qui reflète le mieux les données.Tableau 3Résultats des analyses d=équations structurellesModèles Π 2 Df CFI GFI TLI RMSEA R 21. Modèle médiation totale 421,89 103 0,96 0,97 0,95 0,04 0,182. Modèle médiation partielle 292,75 98 0,98 0,98 0,97 0,03 0,25Dans le modèle retenu (modèle de médiation partielle), tous les paramètres estimés sontsignificatifs à p # .001. Les résultats montrent que 59 % de la variance du soutien du superviseur estprédite par les conditions intrinsèques et extrinsèques de travail vécues par le subordonné et que le soutienprédit par la suite 25 % de la variance dans l’engagement envers l’organisation. À l’exception de lajustice distributive, laquelle variable est non significative dans le modèle, tous les estimés Bêta vontdans le sens attendu et sont significatif à p ≤ 0,001. Sur le plan des liens négatifs, les estimés montrentque la charge de travail, le conflit de rôles, l’ambigüité de rôles, les conflits intra groupe et inter groupessont tous associés négativement avec la perception du soutien du superviseur. L’ambigüité de rôles et leconflit inter groupes sont aussi associés négativement avec l’engagement envers l’organisation. Sur leplan des bêta positifs, les résultats montrent que de tous les facteurs, le sentiment de justiceinterpersonnelle est celui le plus fortement relié au soutien du superviseur. Le soutien du superviseur agitcomme médiateur total entre ce facteur et l’engagement. De plus, les résultats montrent que la justiceprocédurale, l’autonomie et l’utilisation d’habiletés sont associées positivement à la fois au soutien dusuperviseur et à l’engagement envers l’organisation. Ces résultats suggèrent que plus la justiceprocédurale, la justice interpersonnelle, l’utilisation d’habiletés et l’autonomie sont élevées, plus lesubordonné a une perception positive de son superviseur et parallèlement, plus il se sent engagéaffectivement envers son organisation. De tous les facteurs, ce sont les sentiments de justiceinterpersonnelle et procédurale qui sont les meilleurs prédicteurs statistiques du soutien du superviseur.En résumé, les conflits intra groupe, le conflit de rôles, la charge de travail et la justiceinterpersonnelle ont un lien direct avec le soutien du superviseur mais aucun directement avecl’engagement affectif. Ces derniers résultats indiquent que le soutien du superviseur à un effet médiateurentre ces 4 facteurs et l’engagement affectif. Les 5 autres facteurs soit la justice procédurale, l’utilisationdes habiletés, l’autonomie, l’ambigüité de rôles et les conflits inter groupes partagent leur varianceprédictrice entre le soutien du superviseur et l’engagement envers l’organisation. Ces résultats suggèrentun effet médiateur partiel du soutien du superviseur entre ces 5 facteurs et l’engagement affectif.190

Conflit intragroupeConflit de rôle-.13-.24Charge travail-.06.59JusticeJustice procéduraleHabiletés.50.29.18.18Soutiendusuperviseur.25AutonomieAmbiguïté des rôlesConflit intergroupe-.27-.17.12.14.12-.14-.18.25EngagementaffectifFigure 1: Modèle théorique de la relation médiatrice partielle du soutien dusuperviseurDiscussionCette étude a comme objectif d’explorer l’importance du rôle du soutien du superviseur commemédiateur entre diverses situations intrinsèques et extrinsèques de travail et l’engagement affectif dusubordonné envers l’organisation. Comme la littérature sur ce plan est encore relativement absente, laprésente étude vise une meilleure compréhension des diverses conditions d’emploi intrinsèques etextrinsèques qui influencent la perception qu’ont les subordonnés de leur superviseur et de la façon dontcette perception influence à son tour l’engagement envers l’organisation.Nos résultats mettent en évidence le lien négatif des rôles stresseurs (charge de travail, conflits derôle, ambigüités de rôles), et des conflits intra/inter groupes avec la perception qu’ont les subordonnés deleurs superviseurs. Des analyses supplémentaires montrent que l’ensemble de ces facteurs explique 18 %de la variance dans la perception du soutien du superviseur. Ainsi, le subordonné qui perçoit plus de rôlesstresseurs dans son emploi et plus de conflits intra et inter groupes se sent moins supporté par sonsupérieur immédiat. À son tour, ce sentiment négatif envers son superviseur serait associé à unediminution de l’engagement envers l’organisation. Deux des rôles stresseurs, soit la charge de travail et leconflit de rôles, sont particulièrement intéressants car ils apportent un éclairage nouveau sur le planthéorique. En effet, nous savons suite aux travaux de Salancik (1977a) et Mowday et al. (1982) que les191

ôles stresseurs sont négativement associés à l’engagement organisationnel. Nos résultats suggèrent queces deux facteurs sont d’avantage des antécédents du soutien du superviseur lequel agit ensuite commemédiateur sur l’engagement. Ainsi, tout porte à croire que la détérioration du contexte de travail en ce quia trait à ces deux aspects est attribuée en premier lieu au superviseur.Parmi l’ensemble des facteurs étudiés, le sentiment de justice interpersonnelle est celui le plusfortement associé à une perception positive envers le superviseur. Ce résultat converge vers lesconclusions des recherches récentes sur la justice interactionnelle (Materson, et al., 2000). Nos résultatsmontrent que plus le subordonné se voit traité avec justice par son superviseur, plus sa perception positiveenvers ce dernier s’accroît. Le subordonné apparaît très sensible au traitement de respect et de dignitéqu’il reçoit de son superviseur et en retour ce sentiment envers son superviseur conditionne une partie deson engagement envers l’organisation. En ce qui a trait au sentiment de justice procédurale, nos résultatsconvergent vers ceux de Konovsky et Pugh (1994). La façon d’appliquer les procédures signalel’engagement de l'organisation envers le subordonné et place une valeur sur sa contribution à l’atteintedes buts organisationnels (DeCostiis et Summers, 1987). Ainsi, la perception d’équité dans l’applicationdes politiques (justice procédurale) favorise à la fois une perception positive du superviseur tout autantqu’une augmentation de l’engagement.Nos résultats montrent également une médiation partielle entre 5 facteurs (justice procédurale,autonomie, utilisation d’habiletés, ambigüité de rôle, conflits inter groupes) et l’engagement. Nos résultatsportant sur l’autonomie rejoignent ceux de Hutchison (1997) qui a noté que le niveau de décentralisationdes décisions était fortement lié au soutien du superviseur. Ainsi, même si l’organisation du travail (p.ex.description de tâches) et la latitude décisionnelle attribuée au poste dépend énormément de décisionsprises par l'organisation, les subordonnés attribuent un part de cette liberté d’action à leur superviseur. Enfait, le subordonné semble posséder une profonde conviction que le supérieur immédiat détient unpouvoir discrétionnaire important pour décider ou du moins influencer l’autonomie de son travail.Sur le plan de la justice distributive aucun résultat significatif n’a été obtenu. De toutes les formesde justice, c’est celle qui s’éloigne le plus du pouvoir d’intervention du superviseur puisqu’elle concernestrictement les politiques de l’organisation sur lesquelles le superviseur n’a pratiquement aucun contrôle.Ce résultat est intéressant en soi car il suggère que le subordonné a la capacité cognitive de distinguerentre des situations contextuelles attribuables au superviseur et celles relevant uniquement del’organisation.Le conflit intra groupe, lequel réfère aux conflits qui surviennent avec les collègues immédiats,est directement associé à la perception de soutien du superviseur. Il en est de même pour les conflits intergroupes. Nous savons qu’une situation de conflit trop élevé réduit la capacité cognitive de percevoir,traiter et d’évaluer l’information, et interfère avec le rendement (Walton et Dutton, 1969). Le succès deprojets organisationnels n’est possible qu’en présence d’un climat de collaboration à l’intérieur del’équipe et entre les membres de diverses équipes. Bien que les conflits puissent être de naturepersonnelle, ils ont le plus souvent comme source des facteurs hors du contrôle du subordonné (p.ex.compétition pour des ressources, conflits d’objectifs, contraintes de temps) (Whetten et Cameron, 2004).Nos résultats suggèrent que le subordonné face à ces deux types de conflit s’attend à ce que sonsuperviseur s’implique. Autrement, la confiance envers ce dernier en est affectée tout autant que sonengagement envers l’organisation. Il est plausible d’avancer qu’un niveau élevé de conflits retombeinévitablement sur la responsabilité du superviseur et que la négligence de s’en occuper affectera sacrédibilité face à ses subordonnés et réciproquement, sur l’engagement de ces derniers.En conclusion, la nature transversale du devis de recherche, la possibilité d’échantillonnage nonaléatoire et l’utilisation de données auto-rapportées limitent notre capacité d’inférer des liens causals entreles variables à l’étude. Toutefois, la plus grande contribution de notre étude se situe, sans aucun doute au192

niveau théorique. Nos résultats confirment le rôle médiateur de la perception de soutien du superviseurdans la relation entre l’engagement et les caractéristiques de l’emploi. Ils apportent un appui empiriquejusqu’ici absent, aux idées exprimées par plusieurs chercheurs en ce qui concerne le rôle critique que jouele superviseur dans l'échange entre l'organisation et le subordonné (Butz, Dietz. Konovsky, 2001;Clugston, Howell, Dorfman, 2000). Nos résultats rejoignent aussi le point de vue exprimé par Allen etMeyer (1990) et Mottaz (1998) qui affirmaient que l’engagement serait très influencé par lacommunication entre le superviseur et son subordonné et par une relation de soutien.Ces résultats ont plusieurs retombés pratiques. Les individus faisant face à des conditions detravail difficiles ont besoin du soutien de leur superviseur pour maintenir élevé leur engagement enversl’organisation. Ce soutien est d’autant plus important que nos résultats suggèrent que les subordonnésattribuent au superviseur un grande partie de la responsabilité de la création du milieu de travail et desdécisions concernant leurs tâches et le contexte de travail, et ce même si, en principe, certaines de cescaractéristiques intrinsèques et extrinsèques sont en partie indépendantes de son pouvoir. Tant qu’auxorganisations aux prises avec un faible niveau d’engagement des subordonnés, elles auraient tout intérêt àprendre des actions pour développer le rôle de soutien des superviseurs. Ces rôles peuvent bien sûrs êtredéveloppés par l’entremise d’une formation adéquate mais il serait aussi important de fournit auxsuperviseurs le soutien et les outils nécessaire pour développer cet aptitude. D’autre part, compte tenu del’importance de ce rôle, il conviendrait de sélectionner des leaders qui ont déjà une propension à exprimerdes comportements de soutien tels que l’écoute, l’encouragement et la rétroaction constructive.193

RéférencesArbuckle. J.L. & Wothke, W. (2005). Amos 5.0 User’s Guide. Smallwaters Corporation. Chicago.Allen, N.J. & Meyer, J.P. 1990. The measurement and antecedents of affective, continuance andnormative commitment to the organization. Journal of Occupational Psychology, 63: 1-18.Allen, Shore, et Griffeth (1999). A model of perceived organizational support. Unpublished manuscript.University of Memphis and Georgia State University. In Rhodes, L. and Eisenberg, R. (2002)Percieved Organizational Support: A review of the Litterature. Journal of Applied Psychology, 87(4) 698-714.Allport, G.W. (1954). The nature of prejudice, Reading, MA: Addison-Wesley.Bagozzi, R. & Heatherton, T. (1994). A general approach to representing multifaceted personalityconstructs : Application to self-esteem. Structural Equation Modeling, 1, 35-76.Bhar, J. (1995). Developing a measure of supervisor support. Unpublished master dissertation. School ofBusiness, Carleton Unversity, Ottawa, Ontario.Baron, R. M. & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychologicalresearch: Conceptual, strategic, and statistical considerations. Journal of Personnality and SocialPsychology, 51(6), 1173-1182.Bedeian, A.G. & Armenakis, A.A. (1981). A path-analytic study of the consequences of role conflict androle ambiguity. Academy of Management Journal, 24, 417-424.Bies, R.J. (1987). The predicament of injustice: The management of moral outrage. Research inOrganizational Behavior, 9, 289-319.Blake, R.R., Sheppard, H.A., & Mouton, J.S. (1966). Managing intergroup conflict in industry. Houston,TX: Gulf Publishing Company.Brewer, M.B. & Brown, R.J. (1988). Intergroup relations. In D.T. Gilbert & S.T. Fiske (Eds.), Thehandbook of social psychology (4 th ed., Vol. 2, pp. 554-594). New York: McGraw-Hill.Brooke, PP., Russell, D.W., & Price, J.L. (1988). Discriminant validation of measures of job satisfaction,job involvement, and organizational commitment. Journal of Applied Psychology, 73, 139-145.Butz, R.M., Dietz, J., & Konovsky, M.A. (2000). Top management and immediate supervisors as distincttargets of trust. Paper presented at the 16 th Annual Conference of the Society for Industrial andOrganizational Psychology, San Diego, CA, April, 2001.Byrne, B.M. (1994). Structural equation modeling with EQS and EQS/Window: Basic concepts,applications and programming. Sage Publications. London.Chiok F.L.J. (2001). Leadership behaviours : Effects on job satisfaction, productivity and organizationalcommitment. Journal of Nursing Management, 9, 191-204.Clugston, M., Howell, J.P., & Dorfman, P.W. (2000). Does cultural socialization predict multiple basesand foci of commitment? Journal of Management, 26, 5-30.Colquitt, J.A. (2001). On the dimensionality of organizational justice: A construct validation of ameasure. Journal of Applied Psychology, 86(3), 386-400.Cooper, D.L. & Marshall, J. (1976). Occupational sources of stress: A review of the literature relating tocoronary heart disease and mental ill health. Journal of Occupational Psychology, 49, 11-28.De Dreu C.K.W. & Weingart L.R. (2003a). A contingency theory of task conflict and performance ingroups and organizational teams, In M.A. West, D. Tjosvold, & K.G. Smith (Eds.), Internationalhandbook of organizational teamwork and cooperative working (pp. 151-166) Chichester, UK:Wiley.De Dreu C.K.W. & Weingart L.R. (2003b). Task versus relationship conflict: A meta-analysis, Journal ofApplied Psychology, 88, 741-749.De Dreu, C.K.W., Harinck, F. & Van Vianen, A.E.M. (1999). Conflict and Performance in Groups andOrganizations, In C.L. Cooper & I.T. Robertson (Eds.), International Review of Industrial andOrganizational Psychology, Vol. 14, pp. 376-405 Chichester, UK: Wiley.DeCostiis, T.A. & Summers, T.P. (1987). A path analysis of a model of the antecedents andconsequences of organizational commitment. Human Relations, 40, 455-470.194

Dunham, R.B., Grube, J.A., & Castaneda, M.B. (1994). Organizational commitment. The utility of anintegrative definition. Journal of Applied Psychology. 79, 370-380.Duxbury, L., & Higgins, C., (2001). The 2001 National Work-Life Conflict Study: Report One, 2001,Ottawa: Health Canada.Eisenberger, R., Ameli, S., Rexwinkel, B., Lynch, P.D. & Rhoades, L. (2001). Reciprocation ofperceived organizational support. Journal of Applied Pscychology, 86, 42-51.Eisenberger, R., Fasolo, P., & Davis-LeMastro, V. (1990). Perceived organizational support andemployee diligence, commitment, and innovation. Journal of Applied Psychology, 75, 51-59.Etzioni, A. (1975). A comparative analysis of complex organizations. New York: Free press.Guzzo, R.A., Nooman, K.A., & Elron, E. (1994). Expatriate managers and the psychological contract.Journal of Applied Psychology, 79, 617-626.Hackett, R., Bycio, P., & Hausdorf, P. (1994). Further assessments of Meyer and Allen's (1991) threecomponent model of commitment. Journal of Applied Psychology, 79: 15-23.Hackman, J.R., & Oldham, G.R. (1976). Motivation through the design of work: Test of a theory.Organizational Behavior and Human Performance, 16, 250-279.Hall, D.T. & Schneider, B. (1970). Correlates of organizational identification as a function of careerpattern and organizational type. Administrative Science Quarterly, 17, 340-350.Hartley, J.F. (1996). Intergroup relations in organizations. In M.A. West (Ed.), Handbook of work grouppsychology. Chichester, UK: Wiley.Hurrell, J.J. Jr. & McLaney, M.A. (1988). Exposure to job stress: A new psychometric instrument.Scandinavian Journal of Work, Environment, & Health, 14, 27-28.Hutchison, S. & Garstka, M. (1996). Sources of perceived organizational support: Goal setting andfeedback. Journal of Applied Social Psychology, 26, 1351-1366.Hutchison, S. (1997). A path model of perceived organizational support. Journal of Social Behavior andPersonality, 12, 159-174.Jehn, K. (1995) A multimethod examination of the benefits and detriments of intragroup conflict.Administrative Science Quarterly, 40, 256-282.Jehn, K., & Mannix, E. (2001). The dynamic nature of conflict: A longitudinal study of intragroupconflict and group performance. Academy of Management Journal, 44(2), 238-251.Jones, B., Flynn, D.M. & Kelloway, E,K. (1995). Perception of support from organization in relation towork stress, satisfaction, and commitment, in Steven, L. Sauter and Lawrence R. Murphy,Organization Risk Factors for Job Stress, American Psychological Association, Washington,D.C.Jehn, K.A, Northcraft, G.B, & Neale, M.A. (1999). Why differences make a difference: A field study ofdiversity, conflict, and performance in workgroups, Administrative Science Quarterly, 44, 741-763.Karasek, R. & Theorell T. (1990). Healthy work. Stress, productivity, and the reconstruction of workinglife. New York: Basic Books.Konovsky, M.A., & Pugh, S.D. (1994). Citizen behavior and social exchange. Academy of ManagementJournal, 37, 656-669.Kottke, J.L., & Sharafinski, C.E. (1988). Measuring perceived supervisory and organizational support.Educational and Psychological Measurement, 48, 1075-1079.Lawrence, P.R. & Lorsch, J.W. (1967b). Organization and environment: Managing differentiation andintegration. Boston: Harvard Graduate School of Business Administration.Lazarus, R.S. & Folkman, S. (1984). Sress appraisal and coping. New York: Springer.Levison, H. (1965). Reciprocation: the relationship between man and organization. AdministrativeScience Quarterly, 9, 370-390.March, J.G. & Simon, H.A. (1958). Organization. New York: Wiley.Mathieu, J. & Zajac, D. (1990). A review and meta-analysis of the antecedents, correlates andconsequences of organizational commitment. Psychological Bulletin, 108, 171-194.195

Masterson, S.S., Lewis, K., Goldman, B.B., & Taylor, M.S. (2000). Integrating justice and socialexchange: The differing effects of fair procedures and treatment on work relationship. Academyof Management journal, 43, 738-748.McNeese-Smith, D.K. (1997). The influence of manager behavior on nurse’s job satisfaction, productivityand commitment. Journal of Nursing Administration, 27, 47-55.Meyer, J. Allen, N., & Smith, C. (1993). Commitment to organizations and occupations: Extension of athree component conceptualization. Journal of Applied Psychology, 78: 538-551.Moorman, R.H., Blakely, G.L., & Niehoff. B.P. (1998). Does perceived organizational support mediatethe relationship between procedural justice and organizational citizenship behavior ? Adacemy ofManagement Journal, 41, 351- 357.Mottaz, C.J. (1988). Determinants of organizational commitment. Human Relations, 41(6), 467-482.Mowday, R.T., Steers, R.M., & et Porter, L.W. (1979). The measurement of organizational commitment.Journal of Vocational Behavior, 14,224-247.Mowday, R.T., Porter, L.W. & Steers, R.M. (1982). Employee-organizational commitment linkages. NewYork: Adacemy Press.Murnighan, J.K. & Conlon, D.E. (1991) The dynamics of intense work groups: A study of British stringquartets, Administrative Science Quarterly, 36, 165-186.Nauta, R. & Sanders, K. (2000). Interdepartmental negotiation behavior in manufacturing organizations.International Journal of Conflict Management, 11(2), 135-161.Price, J.L. & Mueller, C.W. (1986). Absenteeism and turnover of hospital employees. Grenwich, CT: JAIPutti, J.M., Aryee, S. & Laing, T.K. (1989). Work values and organizational commitment: A strudy in theAsian context. Human Relations, 42, 275-288.Rhoades, L. Eisenberger, R., & Armeli, S. (2001). Affective commitment to the organization: Thecontribution of perceived organizational support. Journal of Applied Psychology, 86(5), 825-836.Rhodes, L. & Eisenberger, R. (2002) Perceived Organizational Support: A review of the Litterature.Journal of Applied Psychology, 87 (4) 698-714.Richter, A., Van Dick, R., & West, M.A. (2005). The relationship between group and organizationalidentification and effective intergroup relations. Best Paper Proceedings of the Annual Meeting ofthe Academy of Management, New Orleans, USA.Riketta, M. (2002). Attitudinal organizational commitment and job performance: A meta-analysis.Journal of Organizational Behavior, 23, 257-266.Salancik, G.R. (1977a). Commitment and the control of organizational behavior and belief. In B. M.Straw & G.R. Salencik (Eds), New direction in Organizational Behavior. Chicago: St Clair PressSalancik, G.R. (1977b). Commitment is too easy. Organizational Dynamics, Summer, 62-80.Salancik, G.R. & Pfeffer, J. (1977) An examination of need satisfaction models of job satisfaction,Administrative Science Quarterly, 22, 427-456.Settoon, R.P., Bennett, N., & Liden, R.C. (1996). Social exchange in organizations: Perceivedorganizational support, leader-member exchange, and employee reciprocity. Journal of AppliedPsychology, 81, 219-227.Shermerhorn, J.R., Hunt, J.G. & Osborn, R.N. (2000). Comportement humain et organisation, 2 ièmeédition, Erpi : Québec, St-Laurent.Shore, L.M. & Shore, T.H. (1995). Perceived organizational support and organizational justice. In. R.S.Cropanzano & K.M. Kacmar (Eds.) Organizational politics, justice, and support: Managing thesocial climate in the workplace (pp. 149-164). Westport, CT: Quorum Books.Stinglhamber, F., & Vandenberghe, C. (2003). Organizations and supervisors as sources of support andtargets of commitment: A longitudinal study. Journal of Organizational Behavior, 24, 251-270.Stone, E.F. & Porter, L.W. (1975). Job characteristics and job attitude: A multivariate study. Journal ofApplied Psychology, 60, 57-64.Stumpf, S.F. & Hartman, K. (1984). Individual exploration to organizational commitment or withdrawal.Academy of Mnagement Journal, 27, 308-329.196

Thomas, KW (1992) Conflict and negotiation processes in organizations, In M.D. Dunnette & L.M.Hough (Eds.), Handbook of industrial and organizational psychology (2 nd ed pp. 651-717 PaloAlto, CA: Consulting Psychologists Press.Treasury Board (2002). Population Affiliation Report, January. Ottawa: Author.Van Knippenberg, D. (2003). Intergroup relations in organizations. In M.A. West, D. Tjosvold, & K.G.Smith (Eds.), International handbook of organizational teamwork and cooperative working (pp.381-400). Chichester, UK: Wiley.Walton, R. E., & Dutton, J.M. (1969). The management of interdepartment conflict: A model and review.Administrative Science Quarterly, 14, 73-84.Wayne, S.J., Shore, L.M., & Liden, R.C. (1997). Perceived organizational support and leader-memberexchange: A social exchange perspective. Academy of Management Journal, 40, 82-111.Wernimont, PF. (1996). Intrinsic and Extrinsic Factors in Job Satisfaction. Journal of AppliedPsychology, 50, 41-50.Whetten, D.A., Cameron, K.S. Developing Management Skills, (4 eds.), Addison-Wexley, 2004.197

ASAC 2006Banff, AlbertaEmily Block (Student)University of Illinois at Urbana-ChampaignLaura Erskine (Student)Marshall School of BusinessUniversity of Southern CaliforniaINDIVIDUAL DETERMINANTS OF THRESHOLD MODELS OFCOLLECTIVE ACTION: A LABORATORY SIMULATIONThrough the use of a laboratory simulation of collective action, we demonstrate thatdifferent constellations of individuals have different collective action structures, and thatpersonality and cultural orientation play a significant role in predicting the order in whichan individual will participate in a collective act.IntroductionCollective action is defined as situated social action (Brown &Goldin, 1973; Marx & Wood,1975). It occurs where a group of individuals act toward a common end and any individual’sparticipation is contingent on the participation of others. These studies have explored a variety of contextssuch as civil disturbances (eg., Quarantelli &Dynes, 1970), strikes (eg., Biggs, 2002), disaster response(eg., Aguirre, Wenger & Vigo, 1998), and crowds (eg., Aveni, 1977; Blumer &Lee, 1946; Le Bon,1895/1995; Park, 1972; Zimbardo, 1969) in order to understand their causes and consequences.The majority of research on collective action has primarily looked at emergent collective acts inpolitical or adversarial contexts (eg., Biggs, 2002; Seidler, Meyer & Macgillivray, 1976). However, ascollective action is defined as any action where individuals make decisions whether to participate basedon the decisions of others; there are many types of organizational activities that fall under the label ofcollective action. These collective acts can be associated with both positive and negative consequencesfor firms. For example, collective action can inform perspectives on organizational change, by explaininghow individuals initiate or accept new ideas or structures. These may include both product and processinnovations such as the acceptance and use of new technologies as well as developing buy in for newinnovations or ideas. Additionally, negative collective acts, such as protests, strikes and boycotts can alsoinfluence organizational outcomes and goals. These may be external to organizations, influencingorganizational reputation and profits, as well is more insidious acts within the organization such as thediffusion of the acceptance of ethical violations or social loafing. Given the potential for collective actionin and around organizations, it becomes important to have a greater understanding of these processes,their antecedents and dynamics. Of particular interest is which individuals and under what conditionsthose individuals will decide to join in a collective act.The importance of understanding organizational collective action is amplified as our theoriessuggest that organization members do not think, feel, or behave in isolation (Mowday &Sutton, 1993;Pfeffer, 1981). Groups are a fundamental unit of social life; they shape and impact individual perceptions,decision-making and action and are ubiquitous in organizations. Collective action is situated in groupsand therefore may be an important element in the consideration of group level activity.However, theories on groups and teams have failed to capitalize on work in sociology andeconomics, where theories of collective action have provided several insights that aid in the predictionand management of group level action. These theories generally agree that an individual’s decision toengage in a particular group level act involves some sort of rational cost benefit assessment (Oliver,198

1993), and that collective acts occur when a proportion of individual action creates reduced costs ofjoining for previously non-participating individuals. However, these theories differ in the degree to whichthese decisions are based on individual or social characteristics. In addition, these theories have primarilybeen tested by exploring single incidents of collective action post hoc, such as looking at the politicalconditions leading to strikes and protests (eg., Andrews &Biggs, 2004; Biggs, 2005; Seidler, Meyer &Macgillivray, 1976) and are thus unable to discern how different individuals and groups of individualswith different characteristics react to similar conditions.The purpose of this paper is to better understand the elements of collective acts which areespecially relevant to organizations; the factors influencing an individual’s choice to engage in such acts.Deriving from Granovetter’s (1978) notion that individuals vary in their propensity to engage in collectiveaction, we aim to predict the individual antecedents to participation an emergent collective act. Weaddress challenges regarding the dearth of cultural research in organizational research and theoversocialized nature of sociological theories of collective action (eg., Schein, 1996) by looking at acrosslevels in order to join individual and organizational phenomena. Through replication of the conditions inwhich the action occurs, we also aim to demonstrate that different constellations of individuals result invaried structures of collective acts. By doing so we contribute to multi-level theory development (Klein,Tosi & Cannella, 1999), as “understanding the psychological dynamics of a macro theory can also makethat theory more useful” (Staw, 1991 p. 810).We also aim to build theories of collective action in two key ways. First, as mentionedpreviously, we test our hypotheses using a controlled setting, allowing replication of the same situation inorder to observe a variety of different individuals’ reactions to the same event. Second, we test this theoryin the context of a collective act different from those which are usually the subject of observation insociological studies, where the emphasis is on acts that are aimed at changing some fundamental structureor belief system (Gould, 1993; Khawaja, 1993; Kim &Bearman, 1997; Marx &Wood, 1975; McAdam,1986). However, many collective acts in organizations occur without any larger social or political goal.By looking at a collective act which is more routine and less organized and value based than those intraditional social movement analysis, we are able to isolate the characteristics of individuals in a valuefree environment, more likely to be generalizable to organizational collective acts such as the diffusion ofinnovation or discontent.Theories of Collective ActionDefining Collective Action.Collective action has been defined quite disparately in economic and sociological literatures.Generally, these definitions can vary on one or more of three broad elements; who takes action, what theaction is directed toward, and how the action is organized. With respect to the first element, the action isviewed as being taken either by groups themselves (Marx &Wood, 1975; Weller &Quarantelli, 1973) orby individuals in the context of a group level activity (Brown &Goldin, 1973; Marx &Wood, 1975).Group level perspectives are interested in the actions of the group as a whole, while individual levelperspectives are interested in an individual’s participation in group level activities.There is also variation in the definition of collective action depending on the entity to which theaction is directed. While some researchers do not specifically cite the larger action to which the collectiveact is directed (Brown &Goldin, 1973; Granovetter, 1978), most scholars of collective action explicitlydefine action in terms of the entity toward which the action is focused. For example, Gusfield (1968)suggests that these are directed movements with a specific ideology. Oliver and colleagues (1985) gofurther with this definition to suggest that collective action is the “provision of a public good.”199

Finally, some definitions of collective action go a step further in their definitions to includespecifics regarding the manner in which the group is organized and managed. For example, Oliver andcolleagues (1985) suggest that collective action requires the commission of resources toward achieving apublic good. Weller & Quarantelli (1973) suggest that norms, either established or emerging, are requiredto direct the action of the group. Gusfield (1968) lists formal leadership and specific agendas asdefinitional elements of collective action.In this paper, we define collective action in a broad sense applicable to many forms oforganizational action, following Granovetter (1978), as those in which both the decision to participate aswell as the cost of individual participation may be contingent upon the decisions of others. This broaderdefinition is less restrictive and more widely applicable in the organizational context. In ourconceptualization, collective interests and ends are not assumed, and no public or social good is necessaryfor individuals to act in concert. Actions where individual interests may be served by a collective actionwith no greater social benefit fall under this definition. Finally, we define collective action asencompassing both directed and undirected acts with elements such as leadership, resources andorganization as potential but unnecessary elements of collective acts.Theories of Collective ActionThe study of collective action has permeated research in sociology and economics and hasproduced several common insights. Research in game theory (eg., Olson, 1965; Rasmussen, 1989),critical mass theory (eg., Oliver, Marwell & Teixeira, 1985), social movement theory (Gould, 1993;Khawaja, 1993; Kim &Bearman, 1997), and threshold models (Granovetter, 1978) have addressed issuessuch as the form and content of activities ranging from market behaviors to political protests. Thesetheories generally agree that an individual’s decision to act involves some sort of rational cost benefitassessment (Oliver, 1993). However, these theories differ in the degree to which these decisions are basedon individual or social characteristics. Simple game theoretical models (eg., Olson, 1965) are focused onthe role of individual preferences in the promotion of individual participation, whereas sociologicalcollectivity models such as critical mass theory (Oliver, Marwell & Teixeira, 1985) are focused onlooking at the interaction of individuals and their environments. While sociological and economic insightsare derived from these theories, they have several drawbacks that limit their applicability toorganizational collective acts, and will be addressed as each theory is explored. Below we briefly discusseach of these perspectives on collective action and argue how a more individual level perspective caninform and build upon these inherently multilevel theories.Individual Models. The first theory of collective action was described by Olson (1965). Beforethis time, any form of inaction in a social arena was considered the result a dearth of common interest(Oliver, 1993). However, Olson suggested that in order for individuals to participate in a collective act,they consider objectively the costs and benefits of participating and must believe that their participationwill make a difference. When there are disincentives for individuals to join in an action that benefits thewhole, issues of “free riding” are raised (Olson, 1965). Before Olson’s conceptualization of the free riderproblem, the group was considered unproblematic and collective action was viewed as a product ofcommon interests. Taking an individualistic perspective, he suggested that participation in group activitieswas a social dilemma where everyone wants to benefit from the participation of others while evading thecost of participation (Olson, 1965; Rasmussen, 1989). This social dilemma can be resolved by negotiationand the creation of incentives for individual participation (Heckathorn, 1996).There are some flaws in the application of this individualistic theory to collective acts. First, thenotion that each individual engages in utility maximization is problematic due to the difficulties of200

assigning weights to individual utilities (Oliver, Marwell & Teixeira, 1985). In addition, “only under themost extreme condition of consensus does action spring spontaneously from the actors’ individual goals,”(Coleman, 1966 p. 616). Thus, looking only at individual goals in isolation of the collectivity makes theunderstanding of collective action problematic.Collectivity Models. Sociological studies took the challenge of understanding the role of thecollectivity as a whole. These models view human interaction as “the furnace of collective action” (Kim&Bearman, 1997 p. 72). They are complex models which focus less on the motivations of the individualand more on, the distribution of traits across individuals within a collectivity (Oliver, 1993). Thesesociological models compare the characteristics of the individual with the characteristics of the socialaggregate; assume that an individual’s choices and goals are contingent on the choices and goals ofothers, and that any collective behavior requires analysis of the system of interaction between individualsand their environments (Schelling, 1978). In plain language, in order to understand people’s behavior, wehave to understand how the actions of others affect what people do. Sociological models such as criticalmass theory, social movement perspectives, and threshold perspectives recognize interdependencebetween individuals and their environments (Coleman, 1990; Fireman & Gamson, 1979; Granovetter,1978) and take into account both structural and individual factors in predicting group behavior (Aveni,1977; Smelser, 1963; Turner &Killian, 1972; Turner &Killian, 1987).Critical mass theory suggests that some threshold of participants in a group have to act before anyaction “explodes into being,” (Oliver, 1993). The argument is that individuals take into account howmuch others have already contributed when they make their own decision about participating and thatstart up costs for individual participation are high until a critical mass of individuals participate(Heckathorn, 1996; Marwell &Oliver, 1993; Oliver, Marwell & Teixeira, 1985; Oliver, 1993). Thisperspective, however, does not address which individuals are likely to act first in order to create thiscritical mass. Another theory that addresses collective action from a sociological perspective is socialmovement theory. This perspective shares common assumptions with critical mass theory that collectiveactions occur where supply is jointly held so that the marginal costs decrease with each participant(Heckathorn, 1996) and that there is a cost benefit calculation that occurs on the part of individuals thatdetermines action (Khawaja, 1993; Tilly, 1978). However, this perspective has been primarily applied tostudies of collective action aimed at changing the status quo, based on common interests and thepromotion of a common good.Studies of social movements demonstrate that social movements are successful when promotersof a collective action increase the value of the stated goal by building solidarity and common interests(Feree, 1992; Fireman &Gamson, 1979; Klandermans, 1988). The existence of this increasing goalsalience decreases the cost of participation (Heckathorn, 1996) and, as such, social movement literaturehas largely focused on consciousness raising and consensus mobilizing (eg., Kim &Bearman, 1997;McAdam, 1982; McAdam, 1986; Snow et al., 1986).Although social movement theory has been effective in studying a certain class of collectiveaction, it is focused on organized value-based acts and does not account for spontaneous collective actions(Nagel, 1981). A more broad treatment of collective action has been forwarded by Granovetter (1978)who described threshold models of collective action. This perspective assumes varying predispositionsagainst action and that these predispositions manifest in a threshold, which is a “number or proportion ofothers who must make a decision before a given actor does so; this is the point where net benefits begin toexceed net costs for that particular actor” (Granovetter, 1978 p. 1420). Threshold models are aimed atexploring binary decisions between two distinct and mutually exclusive alternatives. Individualdifferences are an important part of these models as the costs and benefits of engaging in any action varydepending on the individual (Granovetter, 1978).201

The intent of Granovetter’s initial paper on threshold models was to “predict, from the initialdistribution of thresholds, the ultimate number, or proportion making each of the two decisions”(Granovetter, 1978 p. 1424). He suggests that the distribution of individual thresholds will vary in a givencollectivity and that seemingly similar distributions of individuals have radically different outcomes. Heprovides the compelling example (see Granovetter, 1978, p. 1424-1425) of 100 people milling around in asquare as a potential riot situation where individual thresholds are distributed from zero to 99. In thissituation, the rioting continues, similar to a domino effect for individuals with thresholds zero to 99 untileveryone is involved in a riot. He then describes replacing the person with the threshold one with a personwith a threshold two and how the outcome would differ. After the first person riots, there would be no onewith the threshold one. As a result, no additional people are activated and the “riot” ends at that point withonly one participant.This perspective extends and complements simple models and earlier collectivity models inseveral ways. First, it assumes that decisions to participate are not simply based on an isolated cost benefitanalysis of fixed individual preferences, but contingent on the behavior of others. Additionally, asindividual thresholds vary, it explains why individuals are more likely to be early movers, thus moreclosely examining the dynamics forwarded in critical mass theory. Finally, this perspective on collectiveaction assumes that interests can be varied and that collective action does not have to result from a desirefor some common good or need be the result of organized action. “These models can be applied toprocesses not usually called collective behavior such as voting, residential segregation, diffusion ofinnovations, educational attainment, strikes, migration and markets as well as the more typical processesof crowd behavior and social movements” (Granovetter, 1978 p. 1421).202

Predicting Individual ThresholdsIndividual differences have a great deal of explanatory power. These stable traits, attitudes andbeliefs that influence human behavior have been shown to play a major role in the behavior of individualsin both organizational and private settings (Alderfer, 1977; Arvey &Bouchard, 1994; Barrick &Mount,1991; House, Shane & Herold, 1996; Maslow, 1954). These characteristics can both influence theperception of situations and play a role in individual interpretations of the organizational environment(Hackman &Oldham, 1976; House, Shane & Herold, 1996). As a result of the power of individualdifferences, it is likely that an individual’s propensity for collective action is partially a function of theseindividual traits.The study of individual differences is particularly important in the study of collective action.Research has demonstrated that personality is generally more predictive of behavior in ambiguoussituations than in strong situations where behavioral norms are well known and socially enforced(Monson, Hesley & Chernick, 1982). Under conditions which are likely to foster emergent collectiveaction, these behavioral norms are likely to be ambiguous and therefore the behavior of individuals islikely to be determined significantly by preexisting individual characteristics. We consider two keyindividual differences: personality and cultural orientation. We examine the five most studied personalitytraits and four cultural orientations in order to predict which of these will contribute to individualthresholds for collective action. These two types of important functions of individual behavior aregrowing in importance in organizational research (Earley &Mosakowski, 2000).PersonalityThe first set of individual differences that we explore in predicting individuals’ thresholds forcollective action is personality. Two important studies confirmed the existence of a five-factor model ofpersonality (Barrick &Mount, 1991; Borgatta, 1964; Norman, 1963). During the 1980s severalresearchers provided evidence for the robustness of the five-factor model (Conley, 1985; Costa &McCrae,1988; Digman &Inouye, 1986; Goldberg, 1981; McCrae &Costa, 1985; McCrae &Costa, 1987; McCrae&Costa, 1989). We explore each of the five factors: extraversion, neuroticism, agreeableness,conscientiousness, and openness to experience.Extraversion. The first individual personality difference that we predict will directly relate toone’s propensity for collective action is extraversion. It is one of the most central factors in the study ofpersonality and describes individuals who are generally sociable, gregarious, assertive, and active(McCrae &John, 1992; Phillips &Bedeian, 1994). The assertive dimension of extraverts has been linkedto the seeking of high status and rewards (Cattell, 1981). We hypothesize that extraversion will beassociated with lower thresholds for collective action. As extraverts are generally more talkative andassertive, they will be more likely to express discontent. The active component of extraversion alsosuggests that these individuals will be more likely to act in response to this discontent. Thus,Hypothesis 1a: Extraversion will be negatively related to an individual’s threshold for CA.Neuroticism. A second personality variable that we predict will be related to one’s threshold forcollective action is neuroticism. Individuals who score high on the neuroticism dimension of personalitywould be likely to exhibit anxiety, depression, anger, or worry (Costa &McCrae, 1988). Neurotics arelimited in social skills and avoid situations that demand taking control (Judge, Locke & Durham, 1997).They tend to seek out short term relationships and situations in which trust in others high social skills arenot necessary (Raja, Johns & Ntalianis, 2004). We hypothesize that neuroticism will be associated with203

lower thresholds for collective action. As those with neurotic personalities generally avoid taking control,they will be likely to avoid the act of participating in collective action. Thus,Hypothesis 1b: Neuroticism will be positively related to an individual’s threshold for CA.Agreeableness. Individuals who score high on the Agreeableness factor are generally courteous,flexible, trusting, good-natured, cooperative, forgiving, soft-hearted, and tolerant (Costa &McCrae, 1988).This tolerance is likely to manifest in tendencies to stay patient in uncertain situations and trust those inauthority positions. We hypothesize that agreeableness will be associated with lower thresholds forcollective action. As those with agreeable personalities are generally courteous, cooperative, and flexiblethey are less likely to be among the first individuals to act to change an existing situation. Theseindividuals are less likely to get anxious under conditions of uncertainty and are more tolerant ofdivergent perspectives making them less likely to act against authority. Thus,Hypothesis 1c: Agreeableness will be positively related to an individual’s threshold for CA.Conscientiousness. A fourth personality characteristic that we hypothesize will predict collectiveaction is conscientiousness. Those individuals who display high levels on this factor are hard working,dependable, achievement-oriented, and exhibit perseverance (Costa & McCrae, 1988). Conscientiousindividuals are likely to be risk adverse (Goldman, 1990) and have high performance and job satisfaction(Barrick &Mount, 1991; Judge et al., 2002; Judge et al., 1999). Highly conscientious individuals are alsomore likely to be concerned with accomplishing tasks than with tangible rewards (Stewart, 1996). Wepredict that conscientiousness will be positively related to higher thresholds for collective action. Asthese individuals are likely to be focused on the task and exhibit high levels of perseverance, they will beslow to abandon a task in progress. In addition, they are risk adverse and therefore less likely to joinemergent action until it is proven to be acceptable. Thus,Hypothesis 1d: Conscientiousness will be positively related to an individual’s threshold for CA.Openness to experience. The final personality characteristic that we consider in relation to theprediction of individuals’ thresholds for collective action is openness to experience. Individuals who areopen to experiences are imaginative, curious, broad-minded, intelligent, and artistically sensitive (Costa&McCrae, 1988). These attributes are all associated with positive attitudes toward potential learningsituations (Barrick &Mount, 1991). We believe that individuals with high scores in openness toexperience will have lower thresholds for those collective acts that result from discomfort with theuncertainty of a situation. These individuals will likely have a greater tolerance for ambiguity and willexhibit a greater ability to operate for longer periods of time under conditions of uncertainty. Thus,Hypothesis 1e: Openness will be positively related to an individual’s threshold for CA.204

CultureAn additional set of individual factors that are likely to influence one’s threshold for collectiveaction is one’s cultural orientation. In the 1980’s Hofstede examined differences among national culturesand developed a model that identifies four primary dimensions to assist in differentiating culturalorientation: Power Distance, Individualism, Masculinity, and Uncertainty Avoidance (Hofstede, 2001;Hofstede et al., 1990). Studies deriving from this research have demonstrated that an individual’s culturalorientation is predictive in causal attributions and negotiations (Valenzuela, Srivastava & Lee, 2005), ideageneration (Jung &Avolio, 1999), and leadership emergence (Pillai &Meindl, 1998).Power Distance. Power distance focuses on the degree of equality, or inequality, between peoplein the country's society (Hofstede et al., 1990). A high power distance ranking indicates acceptance ofinequalities in power and wealth. These individuals are more likely to subscribe to a caste system thatdoes not allow significant upward mobility. A low power distance ranking indicates the differencesbetween individuals are de-emphasized; equality and opportunity for everyone is stressed. Theimplications of power distance for individual thresholds for collective action are likely to depend on thecontext of the collective act. When collective acts are being undertaken to support the existing authorityor power structure behavior will differ from when collective acts can be viewed as going againsttraditional authority structures. Under conditions where the authority structure is supported, thoseindividuals with a high power ranking will be more ready to join. However, in those situations where theauthority structure is challenged, individuals with a high power ranking will have lower thresholds forcollective action. Thus,Hypothesis 4a: Under conditions in which authority is being challenged, those exhibiting a highpower distance ranking will have lower thresholds for CA..Individualism-Collectivism. Individualism focuses on the degree that society (or anorganization) reinforces individual or collective achievement and interpersonal relationships (Hofstede,2001; Hofstede et al., 1990). A high ranking indicates that individuality and individual rights areparamount (Earley &Mosakowski, 2000; Hofstede et al., 1990). A low ranking typifies societies (ororganizations) that promote close ties between individuals. These individuals will take responsibility forfellow group members (Hofstede, 2001), subordinate personal interests to the goals of the larger group(Finkelstein &Hambrick, 1996), and defer to team based decisions (Kirkman &Shapiro, 2001). Theimplications of individualism and collectivism will also be largely dependent on the context of thecollective act. As individualists are more likely to prioritize their needs over those of the group, they willbehave differently in collective acts whose primary benefit is to benefit the collectivity than those with alarge individual benefit. Under conditions where the primary benefit is to the collectivity, individualistswill be slow to join in a collective action, however, in those situations where the individual benefit ishigh, such as social loafing, those individuals will have lower thresholds for collective action. Thus,Hypothesis 4b: Under conditions in which individual benefit from joining a collective act is high,those exhibiting individualist cultural orientations will have lower thresholds for CA.Uncertainty Avoidance. Uncertainty Avoidance focuses on the level of tolerance for uncertainand ambiguous situations (Agarwal et al., 2004; Hofstede et al., 1990). A high score indicates a lowtolerance for uncertainty and ambiguity. This often results in the creation of laws, rules, regulations, andcontrols in order to reduce the amount of uncertainty (Hofstede, 2001). A low score indicates less concernabout ambiguity and uncertainty and more tolerance for a variety of opinions. These individuals are alsomore likely to be entrepreneurial and innovative. Collective action situations are almost always those thatlack certainty, rules, or controls; at least until a sufficient number of participants have joined and thenature of the act becomes much more concrete. Implications of uncertainty avoidance for individual205

participation in collective action are that those with high levels of uncertainty avoidance will feeluncomfortable in these uncertain situations and will have lower thresholds for participation. Thus,Hypothesis 4c: Those exhibiting high uncertainty avoidance will have lower thresholds for CA.Masculinity. Masculinity versus its opposite, femininity refers to the distribution of rolesbetween the genders. Those individuals with higher scores on masculinity are likely to follow thetraditional masculine work role model of achievement, control, and power (Hofstede, 2001). They arealso likely to believe in the importance of gender differentiation for roles within organizations (Hofstede,2001). Those individuals with lower scores on masculinity (or higher scores on femininity) support a lowlevel of differentiation between genders. In addition to lacking certainty, collective acts are also situationswhere there is an authority vacuum. Those in power have either deliberately or otherwise, withhelddirection for those in “subservient” positions. In these situations it is also unclear where the power lies.Implications of masculinity for individual participation in collective action suggest that those with highlevels of masculinity will be uncomfortable in situation that lacks an authority structure and no one seemsto be in control and will have lower thresholds. Thus,Hypothesis 4d: Those exhibiting high masculinity will have lower thresholds for CA.MethodsOur research design and the use of a laboratory setting in order to study individual thresholds forcollective action were chosen for several reasons. In discussing the difficulties of studying collectiveaction, Marx and Wood suggested; “The unpredictability of collective behavior, and practical and ethicaldifficulties of doing research in the middle of it, suggest the need for an experimental approach, … shortof this hypothesis testing must wait upon when and what the natural world chooses to yield” (Marx&Wood, 1975 p. 373). In addition, testing of Granovetter’s (1978) theory suggesting that differentdistributions of individuals in the same situation will vary requires the replication enabled throughsituational research. Where several researchers have attempted situational research in the field through thecreation of disaster scenarios (Aguirre, Wenger & Vigo, 1998), the random, emergent nature of theparticipants precludes the possibility of collecting comprehensive individual difference profiles for eachparticipant. Therefore, the use of the laboratory to simulate and replicate conditions for collective actionis a novel and appropriate choice for hypothesis testing.However, as with all methodological choices, the use of a laboratory situation has drawbacks andchallenges. Most prominent in this particular setting is the difficulty of creating genuine conditions forcollective action involving the “spontaneity, indignation or fear that are often involved in “real” collectivebehavior” (Marx &Wood, 1975 p. 373). Our research design addresses this by creating a realistic situationcommon to undergraduates that does not require the evocation of the complex emotions associated withvalue based collective acts. Most undergraduates have experienced classroom situations in which aninstructor fails to arrive for class. We replicated this common situation by having the experimenter notshow up for the simulation in order to see how the subjects collectively respond.Sample. Our sample consists of N=204 undergraduate students at two universities. The firstuniversity is a satellite campus of a large state system on the east coast. The second university is amedium-sized, private research university on the west coast. Students were provided extra credit in theirintroductory management classes for their participation in the experiment.206

Procedures. Groups of 20-30 students were told that they were participating in an experimentstudying small group interaction. Students were given stickers with large, identifying numbers and seatedaround tables in a room equipped with cameras and recording equipment. They were instructed that theexperiment would take no more than 1 hour and 15 minutes. They were given initial instructions andsome surveys assessing personality variables were administered. In the second sample, they wereinstructed to complete a number of scales assessing individual differences online prior to their arrival atthe experiment. In the first sample, their first task upon arrival was to fill out the same surveys. Initialinstructions included information about voluntary withdrawal and confidentiality.At this point, the experimenter instructed the students that she was going to drop off the surveysand pick up some equipment, and that she would return shortly to commence the interactive portion of theexperiment. No further instructions were given to the participants and the experimenters then observedthrough one-way mirrors. As the aim was to motivate collective action, we sought to determine whichstudents would tire of waiting for the experiment to apparently commence and suggest to otherparticipants that they should stop waiting and leave, or leave without discussion. As the students whodecided to withdraw from the experiment exited the room, they were ushered into a new room to bedebriefed by one or both of the authors and fill out reflection scales.After the results of an initial pilot study, after which no students left in response to being seated inthe room for up to one and a half hours, we determined that there was a need for the use of confederatesin this study. According to Granovetter’s (1978) theory, if no person exists with a threshold of zero, therewill be no person to activate the distribution and no collective action will take place. By usingconfederates, we were able to activate the collective action by ensuring the presence of one individualwith a threshold of zero. In our second and third pilot study, the use of one confederate proved to beineffective in reliably activating the collective act. As a result, we determined that a number ofconfederates should be used in order to artificially push forward the distribution at points where it stalled.Confederates were chosen by the instructors of the introductory management classes asindividuals who were motivated and capable of understanding the theoretical foundation of the study.Confederates were trained by the authors. This training included instruction on the nature of thresholdmodels and behavioral recommendations to interject (by leaving the room) at the point where theyobserved that the distribution to be stalled.Measures. Individual thresholds for collective action were measured by the number of subjectsto leave the room before a particular subject followed suit. In the case that more than one subjectsimultaneously indicated that they were leaving, we coded their thresholds to be equal.We measured individual differences using a number of different scales. Personality was measuredusing the NEO five factor personality scale from the International Personality Item Pool (IPIP) wheresubjects were asked to respond to fifty items according to a five point Likert-type scale as to the varyingaccuracy of the statements. Five scales measuring the five factors were (a) neuroticism (sample item: “Ioften feel blue.”), (b) extraversion (sample item: “I feel comfortable around people.”), (c) agreeableness(sample item: “I believe that others have good intentions.”), (d) conscientiousness (sample item: “I makeplans and stick to them.”), and (e) openness to experience (sample item: “I enjoy hearing new ideas.”).The internal reliability for these five scales are α = 0.86, 0.86, 0.77, 0.81, and 0.82 respectively.We measured cultural dimensions following Hofstede (1994). Participants were asked to respondto 20 items in terms of their image of an ideal job and their own personal life on a 5 point Likert-typescale. Items included “How important is it for you to have sufficient time for your personal or familylife?” and “To what extent do you agree or disagree with the statement ‘most people can’t be trusted’?”The responses allowed index scores to be calculated for the four dimensions of national value systems.207

The internal reliability for these scales were α = 0.64 (power distance), α =0.73 (individualism), α =0.69(uncertainty avoidance), and α =0.62 (masculinity).We also controlled for a number of factors. Granovetter (1978) suggests that factors such associal class, education, and occupation are likely to have a profound impact on the nature of individualthresholds. By limiting our study to business undergraduates, we control directly for education and socialposition through the likelihood of future earnings. However, we controlled for school in order to captureany potential differences between the two sample populations. We also controlled for gender and groupnumber. Finally, we controlled for group size as it has been demonstrated to influence group level activityin the context of collective action (Aguirre, Wenger & Vigo, 1998; Oliver &Marwell, 1988).ResultsIn order to test Granovetter’s (1978) theoretical proposition that different groups of individualswill exhibit vastly different behaviors based on the varying distribution of thresholds, we plotted thedistributions of the threshold of each of the groups in our sample. Figure 1 represents the distributions ofsix of those groups 1 . The varying functional forms of these demonstrations support Granovetter’sproposition.1Only six groups are presented in order to promote legibility. These six groups were randomlyselected.208

Figure 1Distributions of Thresholds by GroupDistributions of ThresholdsThresholds3025201510501 3 5 7 9 11 13 15 17 19 21 23Number of IndividualsGroup 1Group 2Group 3Group 4Group 5Group 6Table 1 reports descriptive statistics and correlations among all variables. The results ofregressions predicting individuals’ thresholds for collective action are located in Table 2. Poissonregression analysis was used to estimate our models. We hypothesized that personality and culturalvariables will predict individual thresholds for collective action. Results provide general support for ourarguments.209

TABLE 1Means, Standard Deviations, and Correlations for All VariablesVariable N Mean S.D. 1 2 3 4 5 6 7 8 9 10 11 12 131 Threshold 310 14.86 4.14 -2 Group 311 17.86 5.60 0.17 -3 Gender 310 20.58 5.04 0.08 0.08 -4 Group Size 304 10.84 3.70 0.15 0.80 0.03 -5 Sample 305 6.13 1.97 -.16 -.89 -.08 -.68 -6 Power Distance 291 14.00 3.71 -.02 -.30 0.03 -.19 0.36 -7 Individualism 311 21.75 5.44 -.11 -.12 -.05 -.08 0.15 -.01 -8 Masculinity 156 1.11 0.59 -.05 0.40 0.05 -0.12 -.43 -.07 -.03 -9 Uncertainty Avoidance 229 1.60 0.52 -.00 0.17 0.10 0.33 -.18 0.06 0.02 0.27 -10 Neuroticism 314 0.10 0.29 0.11 -.00 0.09 0.14 0.03 0.02 -.20 0.00 0.34 -11 Extraversion 314 0.12 0.32 -.01 0.09 0.18 0.04 -.11 -.09 0.10 -.01 -.12 -.41 -12 Openness to Experience 313 0.27 0.45 0.16 0.11 0.14 -.02 -.16 -.08 0.01 -.10 -.06 -.12 0.53 -13 Agreeableness 309 0.22 0.43 0.16 0.15 0.12 0.01 -.18 -.13 0.01 -.11 -.09 -.29 0.32 0.46 -14 Conscientiousness 314 0.82 0.39 0.12 0.05 -0.10 -.02 -.07 -.07 0.13 -.05 0.08 -.233 0.27 0.38 0.42210

Table 2Results of Regression Analysis for Organization-Related OutcomesModel 1Model 2Control Variables HypothesisTestingGroup 0.01 0.01(0.01) (0.01)Gender 0.09 0.05(0.04) (0.05)Group Size 0.10 0.14(0.11) (0.01)Sample -.07 -.12(0.11) (0.12)Power Distance 0.00(0.00)Individualism -.00 †(0.00)Masculinity -.00 **(0.00)Uncertainty Avoidance -.00(0.00)Neuroticism 0.14 **(0.04)Extraversion -.07 †(0.04)Openness to Experience -.09 *(0.04)Agreeableness 0.10 *(0.05)Conscientiousness 0.12 **(0.04)Adjusted R Square 0.01 0.06Number of Observations 204 204ChiSqu(df) 30.16(4) *** 98.11(13) ***† p < 0.10 * p < 0.05 ** p < 0.01 *** p < 0.001 All significance tests are two-tailed.211

We found strong support for our hypothesis that personality is predictive of individual thresholds.Hypothesis 1a was moderately supported. Individuals who scored high on the extraversion dimension ofpersonality were more likely to have lower thresholds and remove themselves from the study early.However, the coefficient was only marginally significant at 0.89. Hypothesis 1b was strongly supportedas those individuals who were more highly neurotic were less likely to have high thresholds. Hypothesis1c predicted that agreeableness would be positively related to the number of individuals engaging in acollective act before one chooses to join. This hypothesis was supported. We also found support for theremaining two personality variables, conscientiousness and openness to experience. Both thesecharacteristics were positively related to an individual’s threshold for collective action suggesting thatindividuals who are conscientious and open to new experiences are likely to participate later in emergentcollective acts.We found partial support for our arguments relating culture to individual thresholds for collectiveaction. Hypothesis 2b predicted that individualistic cultural orientations would be negatively related to thenumber of others who choose to join a collective act when there are strong individual benefits derivingfrom participation. We found marginal support for this hypothesis which was significant at a 0.068 level.We also found support for hypothesis 2d, which predicted that those with masculine cultural orientationswould be more apt to join early in a collective action. Our hypotheses regarding uncertainty avoidanceand power distance were directionally consistent but not significant.DiscussionThe purpose of this paper was to develop a better understanding of the psychological processesunderpinning collective action. We found support for Granovetter’s (1978) arguments regarding thedistribution of thresholds impacting the eventual outcome of collective acts. The varying functional formof the twelve groups of subjects demonstrates that similar conditions will result in a variety of differentoutcomes depending on the distribution of individual thresholds among participants. This provides strongsupport for threshold models of collective action, which has been elusive due to the difficulties in thecreation of replicable conditions to study multiple groups of individuals. This also provides justificationfor the importance of determining the factors that influence individual thresholds.We argue that sociological theories of collective action would benefit from a better understandingof which individuals are more likely to engage in collective acts. We also theorize that the individualdifferences of personality and cultural orientation are likely to profoundly impact individual’s thresholdsfor collective action. Our results provide general support for the importance of these factors in predictingindividual behavior in collective action situations.We found strong support for our arguments linking personality and collective action. Individualswho were highly extraverted were likely to have lower thresholds for collective action, where those whowere high on neuroticism, agreeableness, conscientiousness and openness to experience were likely tohave higher thresholds for collective action. The assertive dimensions of extroverted individuals makesthem likely to take control of situations and therefore more likely to emerge as leaders and instigators ofcollective acts. Neurotic individuals are less likely to be instigators as they tend to avoid situations thatdemand taking control (Judge, Locke & Durham, 1997).We also found that agreeable individuals were less likely to be instigators. We hypothesized thatthis tendency was due to the fact that these individuals are less likely to get anxious under conditions ofuncertainty and more tolerate of divergent perspectives making them more tolerant situations and lesslikely to act against authority. Support for this hypothesis could also be related to the jovial socialenvironment that emerged among participants in the experimental setting. Agreeable individuals would be212

more likely to be satisfied with the informal social interaction taking place in the room, and therefore lesslikely to leave.It is not surprising the conscientious individuals and those open to experiences also have highthresholds for collective action. Given the nature of the situation, the promise of an authority figurereturning for an important task, the reason that conscientious individuals remained in the room is likelybecause of their greater desire to successfully complete the task. Characteristics such as persistence andresponsibility are important attributes in the perception of “completing” the experiment. The situation,while not unfamiliar to students who have experienced instructors being late to class, was probablysomewhat surprising. The return of the experimenter would likely have led to a potential learningsituation and this was a likely contributor to those with high levels of openness to experience remaining inthe room.We also found some support for our arguments related to cultural factors. Although ourpredictions for power distance and uncertainty avoidance were not significant, they were directionallycorrect. This may result from a limited range of responses in the given population or it is possible that thestrength of the situation overrode these particular characteristics. It is likely that those participants whowere individualists put their own interests at the forefront and left quickly because they did not perceive abenefit to staying in the room. In contrast, collectivists likely remained together in the room, takingresponsibility for completing the experiment as a group. Our hypotheses posited some reasons for therelationship between masculinity and low thresholds such as an absence of authority structure in theroom. It is also possible that the sex of the experimenter was influential. In all cases the experimenter wasfemale. Those with less respect for women in authority positions may have had less respect for theinstructions to wait until the experimenter returned.Theoretical ImplicationsThis research shows the importance of studying collective action as it has both micro and macroimplications. Investigating individual differences shows that they play a key role in individual thresholdsfor collective action. This contributes on a micro level to understandings of the individual’s role in groupprocesses. While each individual has stable personality characteristics that determine their behavior, weshow that behavior is not determined in isolation. The variety of distributions that result from a randomassignment of individuals shows that individual determinants of thresholds play a critical role indetermining the activities and behaviors of groups.On a macro level, insights from the study of incidents of collective action are particularly relevantto the study of organizational change. Although many perspectives of organizational change suggest thatchange occurs exogenously (eg., DiMaggio &Powell, 1983; Sherer &Lee, 2002; Tushman &Romanelli,1983), recent treatments have increasingly looked inside organizations for the sources of change. Theseperspectives on endogenous change such as theories of innovation (Amabile, 1988; Kanter, 1988) andinstitutional entrepreneurship (Sherer &Lee, 2002) suggest that new ideas and practices can be generated,accepted, and promoted within established organizations and fields. Often this type of change causesindividuals to look beyond established norms, values, traditions and structures. Thus, one element ofunderstanding organizational change requires the understanding of how individuals can show support forelements different than their established expectations. “Innovative firms may be exactly those entities thathave best learned to channel these psychological tendencies into appropriate organization-level actions”(Staw, 1991 p. 817). For this reason, the study of collective action is important in understanding themicroprocesses of organizational change.Our findings also highlight the importance of the consideration of context when exploring grouplevel phenomena. Differing hypotheses for variables like individualism depend on the varying incentives.213

In this case, individuals saw personal benefit from joining in the action but in other situations, such asthose in which the cost to individuals is high and the benefit is primarily collective, those individualswould be less likely to join in. This demonstrates the need for a more nuanced approach to hypothesizingand a context based pragmatic approach to applying our theories by recognizing the perceived “valence”of the change. If a change is viewed as being negative and against organizational norms (eg., leaving aroom after being asked to stay) some predictor variables will have the opposite effects than they wouldhave if the change is viewed positively.We make a final contribution to organizational research by combining the emphasis ofpsychologists and sociologists. This multi-level examination of collective action looks at individualcharacteristics and how they can impact the behavior of groups. The use of both personality and culturalvalues bridges the level of analysis between groups and individuals. Much of organizational behavior canbe viewed as a collection of actions by essentially independent actors (Staw, 1991). Staw (1991) usesorganizational demography as an example of the value of studying individual distributions onorganizational outcomes. The same can be said for individual thresholds being used to predict collectiveaction.Practical ImplicationsCollective action can inform perspectives on organizational change by explaining howindividuals initiate or accept new ideas or structures. This can manifest in organizational settings in awide variety of situations. Imagine someone starting a new company and inviting current colleagues toleave with him or her. The number of individuals that join is an example of collective action. The speedand/or willingness to accept and utilize new technology, new documentation techniques, new sales tools,or new collaborative tools are also organizational examples where we could see thresholds for collectiveaction impacting the success of failure of these elements. Finally, whether consumers will join in theusage of a new product (eg., iPods, hybrid cars, or organic produce) or employees will choose toparticipate in new ways of working (eg., telecommuting, empowered teams, or virtual workarrangements) are impacted by both individual characteristics of the choice as well as individualthresholds for collective action.By recognizing the individual characteristics underpinning instigators, organizations may bebetter positioned to encourage or stymie such action creating structures to respond to situations likely tobe characterized by collective action (Nagel, 1981 p. 87), raising the costs of participation for individuals(Gurr, 1968; Oberschall, 1973) or breeding support from key individuals (Gamson, 1975).214

ReferencesAgarwal, R., R. Echambadi, A. M. Franco, and M.B. Sarkar, "Knowledge transfer through inheritance:Spinout generation, development, and survival," Academy of Management Journal, 47 (2004),501-522.Aguirre, B.E., D. Wenger, and G. Vigo, "Test of the emergent norm theory of collective behavior,"Sociological Forum, 13 (1998), 301-320.Alderfer, C. P., "A critique of Salancik and Pfeffer's examination of need-satisfaction theories,"Administrative Science Quarterly, 22 (1977), 658-669.Amabile, T. M., "A model of creativity and innovation in organizations," In Research in OrganizationalBehavior,10, B. M. Staw and L. L. Cummings, eds., Greenwich, CT: JAI Press (1988).Andrews, K., and M. Biggs, "The dynamics of protest diffusion: The 1960 sit-in movement in theAmerican South," Working Paper, (2004).Arvey, R. D., and T. J. Bouchard, Jr., "Genetics, twins, and organizational behavior," In Research inOrganizational Behavior,16, B. M. Staw and L. L. Cummings, eds., Oxford, UK: JAI Press(1994).Aveni, A. F., "Not-So-Lonely Crowd - Friendship Groups in Collective Behavior," Sociometry, 40(1977), 96-99.Barrick, M. A., and M. K. Mount, "The Big Five Personality Dimensions and Job Performance: A Meta-Analysis," Personnel Psychology, 44 (1991), 1-26.Biggs, M., "Strikes as sequences of interaction - The American strike wave of 1886," Social ScienceHistory, 26 (2002), 583-617.—, "Strikes as forest fires: Chicago and Paris in the late nineteenth century," American Journal ofSociology, 110 (2005), 1684-1714.Blumer, H., and A. M. Lee, New Outlines of The Principles of Sociology, New York, NY: Barnes &Noble, (1946).Borgatta, E. F., "The structure of personality characteristics," Behavioral Science, 12 (1964), 8-17.Brown, M., and A. Goldin, Collective Behavior, Pacific Palisades: Goodyear, (1973).Cattell, R. B., "Where next in human motivation research? Some possible crucial experiments," InDimensions of personality: Papers in honor of H.J. Eysenck, R. Lynn, ed., Oxford: Pergamon(1981).Coleman, J. "Equality of Educational Opportunity." Washington, DC: U.S. Government Printing Office(1966).—, Foundations of Social Theory, Cambridge, MA: Harvard University Press, (1990).Conley, J. J., "Longitudinal Stability of Personality-Traits - a Multitrait-Multimethod-MultioccasionAnalysis," Journal of Personality and Social Psychology, 49 (1985), 1266-1282.Costa, P. T., and R. R. McCrae, "Personality in Adulthood - a 6-Year Longitudinal-Study of Self-Reportsand Spouse Ratings on the Neo Personality-Inventory," Journal of Personality and SocialPsychology, 54 (1988), 853-863.Digman, J. M., and J. Inouye, "Further Specification of the 5 Robust Factors of Personality," Journal ofPersonality and Social Psychology, 50 (1986), 116-123.DiMaggio, P. J., and W. W. Powell, "The iron cage revisited: Institutional isomorphism and collectiverationality in organizational fields," American Sociological Review, 48 (1983), 147-160.Earley, P. C., and E. Mosakowski, "Creating hybrid team cultures: An empirical test of transnationalteam functioning," Academy of Management Journal, 43 (2000), 26.Feree, M. M., "The Political Context of Rationality: Rational Choice Theory and Resource Mobilization,"In Frontiers in Social Movement Theory, A. D. Morris and C. M. Mueller, eds., New Haven, CT:Yale University Press (1992).Finkelstein, S., and D. C. Hambrick, Strategic leadership: Top executives and their effects onorganizations, New York, NY: West Publishing Company, (1996).215

Fireman, B., and W. A. Gamson, "Utilitarian logic in the resource mobilization perspective," In TheDynamics of Social Movements, M. Zald and J. McCarthy, eds.: Winthrop Press (1979).Gamson, W. A., The strategy of social protest, Homewood, IL: Dorsey, (1975).Goldberg, L. R., "Basic Dimensions for a General Psychological Theory - Implications for Personality,Social, Environmental, and Developmental Studies," Journal of Personality Assessment, 45(1981), 657-658.Goldman, M., "Experience-dependent neuropsychological recovery and the treatment of chronicalcoholism," Neuropsychology Review, 1 (1990), 75-101.Gould, R. V., "Collective Action and Network Structure," American Sociological Review, 58 (1993), 182-196.Granovetter, M., "Threshold models of collective behavior," The American Journal of Sociology, 83(1978), 1420-1443.Gurr, T. R., "A causal model of civil strife: A comparative analysis using new indices," AmericanPolitical Science Review, 62 (1968), 1104-1124.Gusfield, J. R., "The study of social movements," In Encyclopedia of the Social Sciences,14, New York,NY: Macmillan (1968).Hackman, J. R., and G. R. Oldham, "Motivation through the design of work - Test of a theory,"Organizational Behavior and Human Performance, 16 (1976), 250-279.Heckathorn, D. D., "The dynamics and dilemmas of collective action," American Sociological Review, 61(1996), 250-277.Hofstede, G. H. "Values Survey Module 1994: English Language Questionnaire." hofstede@bart.nl(1994).—, Culture's consequences: Comparing values, behaviors, institutions, and organizations across nations,Thousand Oaks, CA: Sage Publications, (2001).Hofstede, G. H., B. Neuijen, D. D. Ohayv, and G. Sanders, "Measuring organizational cultures: Aqualitative and quantitative study across twenty cases," Administrative Science Quarterly, 35(1990), 286.House, R. J., S. A. Shane, and D. M. Herold, "Rumors of the death of dispositional research are vastlyexaggerated," Academy of Management Review, 21 (1996), 203.Judge, T. A., A. Erez, J. E. Bono, and C. J. Thoresen, "Are measures of self-esteem, neuroticism, locus ofcontrol, and generalized self-efficacy indicators of a common core construct?," Journal ofPersonality and Social Psychology, 83 (2002), 693-710.Judge, T. A., C. A. Higgins, C. J. Thoresen, and M. R. Barrick, "The Big Five personality traits, generalmental ability, and career success across the life span," Personnel Psychology, 52 (1999), 621-652.Judge, T. A., E. A. Locke, and C. C. Durham, "The dispositional causes of job satisfaction: A coreevaluations approach," In Research in Organizational Behavior,19, B. M. Staw and L. L.Cummings, eds., Greenwich, CT: JAI Press (1997).Jung, D. I., and B. J. Avolio, "Effects of leadership style and followers' cultural orientation onperformance in group and individual task conditions," Academy of Management Journal, 42(1999), 208-218.Kanter, R. M., "When a thousand flowers bloom: Structural, collective, and social conditions forinnovation in organization," In Research in Organizational Behavior,10, B. M. Staw and L. L.Cummings, eds., Greenwich, CT: JAI Press (1988).Khawaja, M., "Repression and Popular Collective Action - Evidence from the West Bank," SociologicalForum, 8 (1993), 47-71.Kim, H. J., and P. S. Bearman, "The structure and dynamics of movement participation," AmericanSociological Review, 62 (1997), 70-93.Kirkman, B. L., and D. L. Shapiro, "The impact of cultural values on job satisfaction and organizationalcommitment in self-managing work teams: The mediating role of employee resistance," Academyof Management Journal, 44 (2001), 557-569.216

Klandermans, B., "The formation and mobilization of consensus," International Social MovementResearch, 1 (1988), 173-196.Klein, K. J., H. Tosi, and A. A. Cannella, "Multilevel theory building: Benefits, barriers, and newdevelopments," Academy of Management Review, 24 (1999), 243-248.Le Bon, G., The Crowd: A Study of the Popular Mind, London: Transaction Publishers, (1895/1995).Marwell, G., and P. Oliver, The Critical Mass in Collective Action: A Micro-Social Theory, New York:Cambridge University Press, (1993).Marx, G. T., and J. L. Wood, "Strands of Theory and Research in Collective Behavior," Annual Review ofSociology, 1 (1975), 363-428.Maslow, A., Motivation and personality, New York, NY: Harper, (1954).McAdam, D., "The Decline of the Civil Rights Movement," In Social Movements of the Sixties andSeventies, J. Freeman, ed., New York, NY: Longman (1982).—, "Recruitment to High-Risk Activism - the Case of Freedom Summer," American Journal ofSociology, 92 (1986), 64-90.McCrae, R. R., and P. T. Costa, "Updating Norman Adequate Taxonomy - Intelligence and PersonalityDimensions in Natural-Language and in Questionnaires," Journal of Personality and SocialPsychology, 49 (1985), 710-721.—, "Validation of the 5-Factor Model of Personality across Instruments and Observers," Journal ofPersonality and Social Psychology, 52 (1987), 81-90.—, "More Reasons to Adopt the 5-Factor Model," American Psychologist, 44 (1989), 451-452.McCrae, R. R., and O. P. John, "An Introduction to the 5-Factor Model and Its Applications," Journal ofPersonality, 60 (1992), 175-215.Monson, T. C., J. W. Hesley, and L. Chernick, "Specifying When Personality-Traits Can and CannotPredict Behavior - an Alternative to Abandoning the Attempt to Predict Single-Act Criteria,"Journal of Personality and Social Psychology, 43 (1982), 385-399.Mowday, R. T., and R. I. Sutton, "Organizational behavior: Linking individuals and groups toorganizational contexts," Annual Review of Psychology, 44 (1993), 195-229.Nagel, J., "Politics and the organization of collective action: The case of Nigeria, 1960-1975," PoliticalBehavior, 3 (1981), 87-116.Norman, W. T., "Toward an adequate taxonomy of personality attributes: Replicated factor structure inpeer nomination personality ratings," Journal of Abnormal and Social Psychology, 66 (1963),574-583.Oberschall, A., Social Conflict and Social Movements, Englewood Cliffs, N.J.: Prentice-Hall, (1973).Oliver, P., G. Marwell, and R. Teixeira, "A Theory of the Critical Mass. I. Interdependence, GroupHeterogeneity, and the Production of Collective Action," American Journal of Sociology, 91(1985), 522-556.Oliver, P. E., "Formal Models of Collective Action," Annual Review of Sociology, 19 (1993), 271-300.Oliver, P. E., and G. Marwell, "The Paradox of Group-Size in Collective Action - a Theory of the CriticalMass," American Sociological Review, 53 (1988), 1-8.Olson, M., The Logic of Collective Action, Cambridge: Harvard University Press, (1965).Park, R., The Crowd and the Public and Other Essays, Chicago, IL: University of Chicago Press, (1972).Pfeffer, J., "Management as symbolic action: The creation and maintenance of organizational paradigms,"In Research in organizational behavior,13, B. M. Staw and L. L. Cummings, eds., Oxford, UK:JAI Press (1981).Phillips, A. S., and A. G. Bedeian, "Leader-Follower Exchange Quality - the Role of Personal andInterpersonal Attributes," Academy of Management Journal, 37 (1994), 990-1001.Pillai, R., and J. R. Meindl, "Context and charisma: A "meso" level examination of the relationship oforganic structure, collectivism, and crisis to charismatic leadership," Journal of Management, 24(1998), 643-671.Quarantelli, E. L., and R. R. Dynes, "Organization as victim in mass civil disturbances," Issues inCriminology, 5 (1970), 181-193.217

Raja, U., G. Johns, and F. Ntalianis, "The impact of personality on psychological contracts," Academy ofManagement Journal, 47 (2004), 350-367.Rasmussen, E., Games and Information: An Introduction to Game Theory, Oxford: Basil Blackwell,(1989).Schein, E. H., "Culture: The missing concept in organization studies," Administrative Science Quarterly,41 (1996), 229.Schelling, T. C., Micromotives and macrobehavior, New York, NY: Norton, (1978).Seidler, J., K. Meyer, and L. Macgillivray, "Collecting Data on Crowds and Rallies - New Method ofStationary Sampling," Social Forces, 55 (1976), 507-519.Sherer, P. D., and K. Lee, "Institutional change in large law firms: A Resource dependency andinstitutional perspective," Academy of Management Journal, 45 (2002), 102-119.Smelser, N., The Sociology of Economic Life, Englewood Cliffs, N.J.: Prentice-Hall, (1963).Snow, D. A., E. B. Rochford, S. K. Worden, and R. D. Benford, "Frame Alignment Processes,Micromobilization and Movement Participation," American Sociological Review, 51 (1986), 464-481.Staw, B. M., "Dressing up like an organization: When psychological theories can explain organizationalaction," Journal of Management, 17 (1991), 805-819.Stewart, G. L., "Reward structure as a moderator of the relationship between extraversion and salesperformance," Journal of Applied Psychology, 81 (1996), 619-627.Tilly, C., From Mobilization to Revolution, Reading, MA: Addison-Wesley, (1978).Turner, R. H., and L. M. Killian, Collective behavior, Englewood Cliffs, NJ: Prentice-Hall, (1972).—, Collective behavior, Englewood Cliffs, NJ: Prentice-Hall, (1987).Tushman, M. L., and E. Romanelli, "Uncertainty, Social Location and Influence in Decision-Making - aSociometric Analysis," Management Science, 29 (1983), 12-23.Valenzuela, A., J. Srivastava, and S. Lee, "The role of cultural orientation in bargaining under incompleteinformation: Differences in causal attributions," Organizational Behavior and Human DecisionProcesses, 96 (2005), 72-88.Weller, J., and E. L. Quarantelli, "Neglected characteristics of collective behavior," American Journal ofSociology, 79 (1973), 665-685.Zimbardo, P., "The Human Choice: Individuation, Reason, and Order vs. Deindividuation, Impulse andChaos," In Nebraska Symposium on Motivation, W. J. Arnold and D. Levine, eds., Lincoln, NB:University of Nebraska Press (1969).218

Table 2Statistiques descriptives, corrélation et fidélité des échelles (internal reliabilities)Variables M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 151, Age 50,41 5,80 - ,18 ,03 ,06 ,07 ,15 -,05 -,06 -,10 -,07 -,10 ,02 ,01 ,01 ,112, Sexe - ,02 ,05 -,05 ,05 ,00 ,00 -,12 -,05 ,00 ,03 ,09 ,04 ,103, Éducation - ,01 ,00 -,04 -,05 ,07 -,04 ,01 -,02 -,03 ,00 -,01 -,034, Niveau - ,08 ,05 -,06 -,01 -,02 -,02 -,02 ,01 ,07 ,02 ,065, Habiletés 4,13 0,69 (,80) ,38 -,26 -,39 ,09 -,25 -,24 ,28 ,22 ,37 ,306, Autonomie 3,45 0,65 (,86) -,36 -,36 -,17 -,30 -,25 ,25 ,27 ,37 ,297, Conflit rôles 4,06 1,31 (,73) ,28 ,32 ,30 ,27 -,24 -,30 -,41 -,208, Ambigüité rôle 2,80 1,29 (,84) ,05 ,26 ,23 -,28 -,26 -,43 -,319, Charge de travail 4,24 0,62 (,81) ,19 ,11 -,02 -,08 -,12 -,0410, Conflit inter-groupe 2,50 0,77 (,83) ,49 -,20 -,23 -,31 -,3011, Conflit intra-groupe 2,09 0,84 (,79) -,20 -,18 -,29 -,2112, Justice interperson. 4,21 0,94 (,94) ,51 ,60 ,2813, Justice procédurale 2,55 0,94 (,86) ,49 ,2714, Soutien superviseur 3,65 0,75 (,90) ,3615, Engagement 4,83 1,36 (,84)Note : n = 1664. Les corrélation sont présentés au dessus de la diagonale. Les corrélations de .08 et au dessus sont significative à p ≤ 0,001. La fidélitéinterne des échelles (internal reliabilities) est présentée entre parenthèse le long de la diagonale.219

ASAC 2006Banff, AlbertaKevin TasaDeGroote School of BusinessMcMaster UniversityGreg SearsPublic Service Commission of CanadaAaron SchatDeGroote School of BusinessMcMaster UniversityDISPOSITIONAL AND CONTEXTUAL DETERMINANTS OF TEAMWORK BEHAVIOR INSELF-MANAGING TEAMS: A LONGITUDINAL CROSS-LEVEL ANALYSISWe developed and tested a longitudinal multilevel model in which individual-level andteam-level factors predicted individual teamwork behaviors. Four individual levelpredictors: conscientiousness, extraversion, core self-evaluation, and task-relevantknowledge, were associated with peer-rated teamwork behaviors. Furthermore, two teamlevelvariables: collective efficacy and relationship conflict, crossed levels and predictedteamwork behaviors. Aggregate levels of team behaviors, in turn, predicted multipleteam-level outcomes.Over the last decade, many organizations have begun instituting team-based work methods.Nearly half (48%) of all organizations in the United States report using some type of team structure intheir work activities (Devine, Clayone, Philips, Dunford & Melner, 1999). With the increasing usage andvisibility of teams in organizations, management researchers have sought not only to isolate theconditions in which team structures contribute to organizational performance, but also the centralcharacteristics that drive team performance (see Ilgen, Hollenbeck, Johnson, & Jundt, 2005). With respectto this latter theme, several studies have emerged reporting that team member personality traits aresignificant predictors of team performance. Barrick, Stewart, Neubert, and Mount (1998), for example,found that teams with higher aggregate levels of various Big Five Personality constructs (e.g.,extraversion, conscientiousness, agreeableness, emotional stability) receive higher supervisory ratings ofteam effectiveness. While this research has been helpful in signaling that team member personality traitscontribute to effective team performance, these studies have not clearly articulated how personality traitsimpact an individual’s direct contributions to the team. Likewise, although recent research has exploredthe relationship between personality traits and factors such as team roles (Stewart, Fulmer, & Barrick,2005) and organizational citizenship behavior (e.g., Borman, Penner, Allen, & Motowidlo, 2001), itremains unclear how specific personality traits relate to the display of individual teamwork behavior. Inaddition to the limited research in this area, very little work has investigated how contextual or team-levelfactors may impact the display of individual teamwork behavior. A better understanding of howindividual difference variables, including personality traits, and team-level factors influence teamworkbehavior will provide practical guidance regarding the design and implementation of team selection anddevelopment systems. This research will also furnish a more robust theoretical foundation for futureresearch exploring the determinants of team-level performance.Teamwork behaviors refer to individual activities that contribute to effective team processes.Consistent with Tasa, Taggar, and Seijts (2006), we expect that there are multilevel antecedentspredicting the extent to which a person engages in teamwork behaviors. Specifically, at the individuallevel, we posit that a person’s teamwork behavior will be, in large part, determined by their personalityand level of task-relevant knowledge. At the team level, we expect that collective efficacy and220

elationship conflict will cross levels and also determine a person’s teamwork behavior. In addition toaddressing these fundamental relationships, we in turn assess how teamwork behaviors, when aggregatedto the team level, influence subsequent collective efficacy beliefs and diverse measures of teameffectiveness. In response to Klein and Kozlowski’s (2000) call for researchers to adopt a multi-levelperspective in team research, the model we propose is cross-level in nature. Our model specifies crossleveleffects (Rousseau, 1985) whereby team-level constructs influence individual teamwork behavior,which in turn, influences team-level perceptions and performance (Lindsley, Brass, & Thomas, 1995).Theoretical Background and HypothesesTeamwork BehaviorWhat are the individual behaviors that contribute to team performance? After an extensive reviewof the teams literature, Stevens and Campion (1994) developed a typology of generic or transferableknowledge, skills, and abilities (KSAs) necessary for teamwork that focused specifically on the individualteam member. They identified two broad categories of teamwork KSAs: interpersonal KSAs (i.e., conflictresolution, collaborative problem-solving, and communication) and self-management KSAs (e.g., goalsettingand performance management, and planning and task co-ordination). While possessing theknowledge, skills and abilities to perform in a team may be important, it is the translation of the KSAsinto actual behaviors that ultimately contributes to team performance. Thus, in this research we rely onStevens and Campion’s typology as a guiding framework for assessing the visible teamwork behaviorsthat individual team members can engage in to contribute to team functioning. Since behaviors such asproblem-solving and performance management are fairly generic, or transportable (Cannon-Bowers,Tannenbaum, Salas, & Volpe, 1995), they should contribute to team performance across a range of teamtasks.Individual-level Antecedents of Teamwork BehaviorAt l`east four individual-specific factors should be related to the extent to which a person engagesin teamwork behavior. These factors are the personality constructs of extraversion, conscientiousness andcore self-evaluation, as well as an individual’s level of task-relevant knowledge.Extraversion. The “Big Five” personality construct of Extraversion has demonstrated consistentlinks to individual job performance for jobs requiring substantial social interaction (e.g., Barrick &Mount, 1991; Mount, Barrick, & Stewart, 1998). Individuals who are extraverted tend to be friendly,sociable, talkative, energetic, active, and optimistic (Costa & McCrae, 1992; Goldberg, 1990, 1992).Based on this pattern of traits, individuals who are extraverted may possess many of the interpersonalKSAs (e.g., communication, collaborative problem-solving, conflict resolution) identified by Stevens andCampion (1994). Indeed, preliminary research in team settings has shown that extraverts make a strongcontribution to the socioemotional component of team functioning (e.g., Barry & Stewart, 1997;Littlepage, Schmidt, Whisler, & Frost, 1995). Extraverts may also, however, make strong taskcontributions to the group. Barry and Stewart (1997), for instance, reported that extraverts (relative tointroverts) tend to possess higher group performance expectations and are more inclined to expressconcerns relating to product quality to the group.Conscientiousness. Conscientiousness is the Big Five personality construct that has emerged asthe strongest predictor of individual performance across occupations (Barrick & Mount, 1991; Barrick,Mount, & Judge, 2001). Individuals projecting higher levels of conscientiousness tend to be: “responsible,careful, persevering, orderly, cautious, planful, hardworking, and achievement-oriented” (Mount &221

Barrick, 1995, p. 164). Given that individuals high in conscientiousness exhibit higher levels of goaldirectedbehavior and work motivation (Barrick, Mount, & Strauss, 1993; Hough & Ones, 2002; Judge &Ilies, 2002), are more willing to take on extra work (Borman et al., 2001; Podsakoff, MacKenzie, Paine,& Bacharach, 2000) and tend to function effectively in unstructured roles (Barrick & Mount, 1993;Stewart, Casso, & Cardy, 1996), individuals higher in conscientiousness may be expected to demonstratea stronger task focus and self-management capability in the team. Indeed, Neuman and Wright (1999)reported that highly conscientious individuals tend to display higher levels of planning and taskcoordination in team settings. Likewise, Taggar, Hackett, and Saha (1999) found that conscientiousindividuals may be more inclined to take on a leadership role in the group.Core self-evaluation. Core self-evaluation refers to the fundamental, subconsciousconclusions individuals reach about themselves, other people, and the world (Judge, Locke, Durham, &Kluger, 1998). Four broad traits combine to form the core self-evaluation construct: self-esteem,generalized self-efficacy, emotional stability, and locus of control (e.g., Judge & Bono, 2001; Judge, Erez,Bono, & Thoresen, 2003). As reflected in these traits, the core self-evaluation construct, likeconscientiousness, contains strong work motivation and ability components (Judge, Erez, & Bono, 1998).While prior research suggests that facets of core self-evaluation, such as self-efficacy, may predictindividual teamwork behavior (Tasa, Taggar, & Seijts, 2006), we are unaware of research that hasinvestigated the relationship between the broader core-self evaluation construct and individual teamwork.Given the motivational and ability components underlying core self-evaluation, coupled with the strongaffective component also suffusing the construct (self-esteem and emotional stability), individuals higherin core self-evaluation may be expected to demonstrate more effective task-oriented and socioemotionalteamwork behavior.Task-relevant knowledge. Another factor that is likely to influence whether an individualdisplays teamwork behavior is his or her level of task-relevant knowledge. Marks, Mathieu, and Zaccaro(2001) argued that member expertise is utilized through group processes to determine emergent statessuch as collective efficacy. Similarly, research has shown that groups have an ability to recognize thetask-relevant expertise residing in their group (Henry, 1995), and that groups with a high level of taskrelevantknowledge outperform those with low levels of task-relevant knowledge (LePine, Hollenbeck,Ilgen, & Hedlund, 1997). Although it is far from clear how expertise is recognized and linked to teamperformance, it is likely that team members who know more about the team’s task and goal requirementswill contribute more to discussions about the team’s task and goals. Therefore, we posit that anindividual's task-relevant knowledge is related to individual teamwork behavior.The discussion to this point leads to the following hypothesis:Hypothesis 1: An individual’s teamwork behavior is predicted by his or her level of (a)extraversion, (b) conscientiousness, (c) core self-evaluation, and (d) task-relevant knowledge.Contextual Phenomena Relating to Teamwork BehaviorMowday and Sutton (1993: 198) characterized context as, “stimuli and phenomena that surroundand thus exist in the environment external to the individual, most often at a different level of analysis.” Anexamination of the contextual factors that influence teamwork behavior should explore the externalelements most likely to encourage, or discourage, the display of such behaviors. In this study we focus ontwo contextual factors that characterize the team as a whole, initial collective efficacy and the level ofrelationship conflict in the team.222

Initial collective efficacy. Bandura defined collective efficacy as “a group’s belief in theirconjoint capabilities to organize and execute the courses of action required to produce given levels ofattainments” (1997, p. 477). Collective efficacy is a prospective, rather than retrospective, judgment ofgroup capabilities. Since collective efficacy is a comprehensive judgment that is derived from knowledgeabout a team’s capabilities, resources, and constraints, it may exert a top-down influence on individualteamwork behaviors. More specifically, the extent to which individuals believe their team has thecapability to produce results or attain goals may affect the degree to which they themselves will bemotivated to demonstrate the behaviors that are required for effective team performance. Lindsley et al.(1995) suggested that such a cross-level effect is possible because efficacious teams create a context inwhich constructive individual behaviors are expected. Thus, we suggest that it is consistent with socialcognitive theory (Bandura, 1997) to expect that perceptions of the resources possessed by the teamshould be related to a member’s motivation to engage in teamwork behavior.Relationship conflict. Relationship conflict refers to disagreements related to interpersonalissues, political norms and values, and personal taste (Jehn, 1995, 1997). Such conflict has long beenknown to have the potential to harm group process and performance (i.e., De Dreu & Weingart, 2003).However, empirical findings with respect to the effect of relationship conflict on team performance havebeen mixed, with some reporting a significant negative relationship (i.e., Jehn & Mannix, 2001), andothers reporting no relationship (i.e., Pelled, Eisenhardt, & Xin, 1999). One source of confusion in thisbody of literature is the lack of research examining how relationship conflict at the team level influencesthe teamwork behavior of individual team members. Since relationship conflict has the potential todistract team members from the team’s task, and has also been shown to increase withdrawal oravoidance strategies (De Dreu & Weingart, 2001), we expect that people working in teams characterizedby high relationship conflict will be less likely to engage in teamwork behavior. The discussion to thispoint leads to the following hypothesis:Hypothesis 2: An individual’s teamwork behavior is predicted by the level of (a) initial collectiveefficacy, and (b) relationship conflict in the team.Team-level Teamwork Behavior, Collective Efficacy and Team PerformanceTo this point, we have focused on predictors of individual teamwork behavior. We also focus inthis research on the manner in which these behaviors lead to team performance outcomes. As we willexplain, individual teamwork behaviors are unlikely to have a direct impact on team performance.Instead, they form the basis for group judgments about overall team capability, which in turn have animpact on team performance. Thus, we expect that aggregated teamwork behavior is related to collectiveefficacy, which is related to team performance.Marks et al. (2001) suggested that emergent states such as collective efficacy are products of teamexperiences, both at the individual level and at the group level, and become new inputs to subsequentprocesses and outcomes. The reciprocal interplay among team interactions and emergent states at thecollective level is likely to start with the observable behaviors of individual team members. Seeing one’steammates perform behaviors that are generally recognized as helpful with respect to team performanceshould instill a sense of confidence about the team’s capability. Accordingly, we suggest that theemergence of collective efficacy is partly shaped by the aggregate level of individual teamwork behaviorsin the team.223

According to social cognitive theory, efficacy is a primary driver of the extent to whichindividuals or teams are likely to put forth the effort required to perform successfully (Bandura, 1986). Ameta-analysis by Gully, Incalcaterra, Joshi and Beaubein (2002) showed that the relationship betweencollective efficacy and team performance was positive and significant (ρ = .41). In contrast, few studieshave examined the relationship between collective efficacy and other dimensions of team effectiveness,such as the extent to which team members would like to continue to work together (Hackman, 1990). Weexpect that teams with a shared belief that they can perform at a high level will not only do so, they willalso develop a positive sense of mutual attraction. Therefore, we hypothesized that the level of collectiveefficacy, assessed after the team-members had been working together for several weeks, predicts twoteam outcomes – objective team performance and team viability. Thus, our final two hypotheses are:Hypothesis 3: Team-level teamwork behavior is related to collective efficacy.Hypothesis 4: Collective efficacy is related to (a) team performance, and (b) team viability.MethodsParticipantsParticipants were 471 business students enrolled in eight sections of an upper-level HumanResource Management (HRM) course. All eight sections were taught over a two year period by the sameinstructor, who was blind to the study’s hypotheses. Course content was identical across all eight sections.As part of the course requirements, participants were assigned to teams (N = 124) of three or fourindividuals. To maximize the breadth of within-team expertise and reduce the potential for self-selectionbias, teams were created by randomly drawing names from major-specific lists (e.g., finance, marketing,and engineering) in sequence. In total, 77% of the participants were enrolled in the business program, and23% were in a combined engineering and management program. The average age of the participants was21.4 years (SD = 1.0), and 57% of the participants were male.Although participants had to work on the simulation as part of a course requirement, participationin the study was voluntary. Participants received a 2% bonus mark in exchange for their self-reporteddata. The response rates to the questionnaires we distributed at weeks 2 and 7 were 88% and 90%respectively. Ten teams were dropped from the sample because two or fewer members responded to thesurveys. There were 92 four person groups and 22 three person groups in the final sample.Simulation and ProceduresA primary objective of the study was to assess the emergence of teamwork behaviors over time.We therefore studied the teams as they worked on a longitudinal simulation over an 8 week period. Thesimulation, called the Human Resource Management Simulation, was developed by Smith and Golden(2001), and began following the third week of the course. The pedagogical purpose of the simulation is tocreate a realistic context in which students must translate course content, covering such issues as selectionand reward systems, into team solutions and decisions to improve the performance of the HRMdepartment. All teams were told by the course instructor that their performance on the simulation wouldlargely be determined by the extent to which they were able to discern linkages among the various HRMtopics covered in the course. The course content that was most relevant to the simulation was presentedeither prior to group formation (e.g., Strategic HRM, Selection, & HRM Planning) or during the first fewweeks of the simulation (e.g., Compensation, Training, & Performance Appraisal). Furthermore,participants were told that the algorithms which determine performance on the simulation are carefullymodeled from current theories of human resources and management. A 25 page Player’s Manual224

containing information on the nature of the company, industry, and simulation procedures reinforcedthese instructions.Teams assumed the role of the human resource management department in a manufacturingorganization and were required to make weekly operational decisions on issues such as wage increases ordecreases, hiring, fringe benefits, performance appraisal, and training. Teams were required to submittheir decisions on a standardized form and these decisions were then entered into a computer program bythe course instructor. Each weekly set of decisions corresponded to one-quarter of a simulated fiscal year.Teams were given an equal yearly budget and thus had to make their decisions under constraints of finiteresources. Each team performed the simulation for 6 simulated quarters (or 1.5 fiscal years) over the 8week period. No simulation decisions were made during weeks 4 due to midterm exams and week 6 dueto other classroom activities.Performance feedback was returned within three days so that each team had ample time toprepare for the next decision period. Feedback was primarily quantitative, including such indicators asemployee morale, product quality, productivity, employee turnover, budget surpluses or deficits, and thenumber of grievances filed in the simulation. For each indicator, the teams were provided with theindustry average. More detail on how team performance was measured and presented to the participants isdescribed below in the Measures section.Each team’s performance on the HRM simulation accounted for 10% of their final course grade.The grade was shared equally by each team member. Individual team members had an incentive toactively participate because teams had an opportunity to complete, at the end of the simulation, ananonymous peer evaluation justifying why some members should not share the grade equally. Only a fewteams chose to do so. One-third of the class time in each week (50 minutes) was given to participants towork on the simulation. During a session at the end of the course in which we described the nature of thestudy, participants suggested, by an informal show of hands, that they spent between 30 and 90 minutesper week working on the simulation outside of class time.The simulation is appropriate for this study for several reasons. First, consistentwith Wood’s (1986) criteria of task complexity, the simulation is highly complex in terms of component,dynamic, and coordinative complexity. For example, team performance is not increased solely througheffort and persistence. For teams to do well they had to repeatedly monitor the relationship betweenchanges to a multitude of simulation-specific indicators and subsequent feedback from the simulationwith respect to performance. Second, since groups usually outperform individuals on heuristic problemsolvingtasks (Laughlin, Bonner & Miner, 2002; Kerr & Tindale, 2004), the longitudinal nature of thesimulation provided greater opportunity for team members to question assumptions, reframe problems,think about ideas and problems using novel approaches, stimulate avenues of thought that other groupmembers may not otherwise pursue, and successively improve ideas (Sosik, Kahai, & Avolio, 1998).Finally, the simulation enacts realistic and frequently occurring scenarios, in which groups must makeiterative decisions in the context of finite resources. The use of realistic tasks or simulations increases theextent to which results are generalizable to settings beyond the classroom or laboratory (Locke, 1986).MeasuresPersonality traits. Extraversion and Conscientiousness were assessed with 7 items each fromGoldberg’s (1999) International Personality Item Pool (IPIP) “Big Five” scale. A sample item forextraversion includes: “Don’t talk a lot” and a sample item for conscientiousness includes “Continue untileverything is perfect”. Using a 7-point Likert scale (from “strongly inaccurate” to “strongly accurate”),respondents indicated the extent to which each item accurately described themselves. Coefficient alphaswere .86 for extraversion and .76 for conscientiousness.225

Core self-evaluation was measured with the 12-item Core Self-Evaluation Scale (Judge, Erez,Bono & Thoresen, 2003). Scale items include: “When I try, I generally succeed” and “Overall, I amsatisfied with myself”. Participants were asked to indicate the extent to which they agreed with each itemin accordance with a 7-point Likert scale (from “strongly disagree” to “strongly agree”). Coefficient alphawas .84.Task-relevant knowledge. Since the course content was relevant to the performance of thesimulation, knowledge of the course content was used to represent task-relevant knowledge. As part ofthe course requirement, participants completed a multiple-choice exam covering topics in HRM at week 4of the simulation (week 7 of the course). A measure of task-relevant knowledge was created using thescore, out of 100, on the exam.Teamwork behaviors. Measures of teamwork behavior were taken near the end of the simulation(week 7 of the simulation) because by this time the participants had attained experience with the task andhad ample time to observe their fellow team members. Peer appraisals were used for assessing behaviors.Each team member assessed his or her peers on twelve items that were behavioral descriptors drawn fromthe Stevens and Campion (1994) framework. Sample items include: “take steps to ensure everyoneparticipate in group discussion,” and “remind other team members of the team’s goal.” Participants wereasked to rate the frequency with which each of their fellow team members engaged in the 12 behaviorsduring the simulation. Scale scores could range from 1 (“almost never”) to 7 (“almost always”). Thisprocedure is consistent with Neuman and Wright (1999) and Taggar and Seijts (2003). Item-scores wereaveraged to derive a composite behavioral measure for each individual. The coefficient alpha for thebehavioral measure was .93.To assess the level of inter-rater agreement on the behavioral ratings, we calculated the r wgcoefficient (James, Demaree, & Wolf, 1984) for each participant on each of the 12 behavioral items. Theaverage coefficient across all ratings and participants was .83 with a range of .36. There is debate aboutan appropriate cutoff point for r wg coefficients. In the present study over 95% of all r wg coefficients wereabove the commonly used .70 threshold (see Castro, 2002). We examined the cases in which the r wgscore was below .70 and found that low scores, usually just below.70, were not specific to a few teams orindividuals. Thus we did not omit individuals or teams from the sample. In conclusion, team membersappeared to agree on the degree to which other team members displayed teamwork behaviors.Collective efficacy. Collective efficacy was assessed twice, at weeks 2 and 7 of thesimulation. We measured collective efficacy at week 2, rather than at the time of group formation, toallow individuals to become familiar with their team members, their team roles, and the nature of thesimulation. The collective efficacy scale included 6 items drawn from Lester, Meglino and Korsgaard(2002). Scale scores could range from 1 (“to no extent”) to 7 (“to a great extent”). A sample item is, “towhat extent does your team believe it can be very productive?” The coefficient alpha for the collectiveefficacy measure was .88 at week 2 and .91 at week 7.Relationship conflict. Relationship was measured at week 2 using the common 4-item scale fromJehn (1995). A sample item is: “how much are personality clashes evident in your team?” Scale scorescould range from 1 (“none or never”) to 7 (“always”). The coefficient alpha for the scale was .90.Team outcomes: Performance and viability. Teams were instructed at the outset by the courseinstructor that their performance on the simulation would be determined by combining their ranking on 10indicators. The 10 indicators were: Absenteeism, Turnover Percentage, Accident Rate, Morale,Productivity, Quality Index, Unit Cost, Percentage of Female Employees, Percentage of Minority226

Employees, and Grievances. Using the quantitative scores provided by the simulation, each team’sperformance was ranked on each of the 10 indicators and then summed to provide a single item rankingscore. The 10 indicators were, on average, highly correlated (r = .56). This method drew participant’sattention to the performance of the overall HR department, rather than performance on only one or twoindicators. The measure of team performance was presented to all the teams publicly during class timeusing an overhead projector at weeks 2 and 8. Because performance levels relative to similar others arehelpful in forming accurate efficacy perceptions via social comparison processes (see Wood & Bandura,1989), we chose to report the ranking method of team performance in this study. We also note that thestudy’s conclusions do not change when we use the actual scores reported by the simulation rather thanthe rank measure of performance. We refer to the week 2 ranking as Initial Performance and the week 8ranking as Team Performance.Team viability was measured with items drawn from Hackman (1990). A sample item is: “I wishI had more time for socializing with other team members.” Scale scores could range from 1 (“stronglydisagree”) to 7 (“strongly agree”). The coefficient alpha of the scale was .82.Data Aggregation IssuesFour variables in this study, team-level teamwork behavior, collective efficacy, relationshipconflict, and team viability, were measured at the individual level and subsequently aggregated torepresent a group-level construct. To justify our procedure for treating variables measured at theindividual level as having group-level properties, we relied on Chan’s (1998) typology of compositionmodels. According to Chan (1998), a direct consensus model is a composition model in which individualratings of a higher level phenomenon are used to represent the higher level construct. To be theoreticallyvalid, direct consensus models must contain a suitable degree of within-group agreement (Chan, 1998;Gibson, Randel & Earley, 2000). As suggested by Bliese (2000), we computed intra-class correlations(ICC) to determine the reliability of the team-level teamwork behavior measure. For teamwork behavior,the ICC(1) coefficient, which represents the degree of variability in responses at the individual level thatis attributed to team membership, was .33. The ICC(2) coefficient, which represents the reliability of theteam-level means, was .71. The ICC(1) coefficients for the collective efficacy measures were .39 at week2 and .48 at week 7, and the ICC(2) coefficients were .72 and .78, respectively. For relationship conflict,the coefficients were .29 (ICC(1)) and .68 (ICC(2)). Finally, for team viability, the results were .34 and.75 respectively. In total, these results show that all 4 measures displayed an appropriate degree of withingroupagreement relative to between-group variance, and thus support aggregation of individual-leveldata to the team level (Klein & Kozlowski, 2000).ResultsTable 1 presents descriptive statistics and bivariate correlations among studyvariables at the individual and team levels. Each of the reported correlations is in the expected direction.Because the dependent variable in hypotheses 1 and 2 was at the individual level,and the independent variables were both individual and team level, we analyzed the data usinghierarchical linear modeling (HLM 5) (Bryk & Raudenbush, 1992). As suggested by Hofmann, Griffin,and Gavin (2000), we first tested a null model that had no predictors at either level 1 (the individual level)or level 2 (the team level). This null model is analogous to an analysis of variance in that it partitionsvariance in teamwork behavior into within- (Φ 2 ) and between-team (τ) components. The estimate ofbetween-team variance is the intra-class correlation coefficient and is computed as the proportion ofbetween-teams variance over the total variance (Φ 2 + τ). A test of the significance level of the level 2residual variance of the intercept (τ 00 = .28, p < .01) was significant. Also, the ICC(1) was .26, indicating227

that 26% of the variance in individual teamwork behavior resided between teams, and 74% of thevariance resided within teams. The significant between-team variance provided justification for testingthe full model.Hypotheses 1a to 1d predict that individual teamwork behavior is related toindividual-level variables and hypotheses 2a and 2b predict that teamwork behavior is related to twoteam-level variables. We estimated an HLM model in which the individual-level antecedents were thelevel 1 predictors and then regressed the intercept coefficient obtained from level 1 on the level 2variables, collective efficacy and relationship conflict. As Table 2 shows, extraversion (β = .16, p < .01),conscientiousness (β = .14, p < .05), core self-evaluation (β = .12, p < .05), and task-relevant knowledge(β = .02, p < .01) demonstrated significant relationships with individual teamwork behavior. Hence,hypotheses 1a to 1d were supported. Table 2 also shows that collective efficacy (γ = .45, p < .01) andrelationship conflict (γ = -.18, p < .01) predicted individual teamwork behavior, supporting hypotheses 2aand 2b.We also calculated the total variance explained by the full model (R 2 total) usingthe following formula (Bryk & Raudenbush, 1992): R 2 within-team × (1 – ICC(1)) + R 2 between-team × ICC(1).The four individual-level antecedents explained 22% of the within-team variance and collective efficacyand relationship conflict explained 55% of the between-team variance in teamwork behavior. Therefore,using the R 2 total formula, the full model explained 31% of the total variance in teamwork behavior.Hypothesis 3 states that aggregated teamwork behavior is significantly related tocollective efficacy and hypothesis 4 states that collective efficacy is positively related to teamperformance and team viability. As shown in Table 2, the correlation between team-level teamworkbehavior and collective efficacy measured at week 7 was significant (r = .51, p < .01). Table 2 also showsthat the correlations were significant between collective efficacy and team performance (r = .40, p < .01)and collective efficacy and team viability (r = .27, p < .01). Together, these results support the hypotheseslinking teamwork behavior to team outcomes.DiscussionA growing body of research suggests that the composition of individualdifferences in teams affects both team processes and team outcomes. However, few studies haveexamined the contextual and individual-level antecedents of teamwork behavior. Since teamworkbehavior encompasses individual actions that contribute to team functioning, the question as to howteamwork behavior develops over time has both theoretical and practical implications. Our study extendspast research by relating multilevel antecedents to teamwork behavior and the subsequent development ofefficacy beliefs and team performance.Theoretical ImplicationsThe first theoretical implication of the present study deals with the determinants of teamworkbehavior at the individual level. Core self-evaluation has only recently been identified and its relationshipto individual teamwork behavior, collective efficacy, and team performance has yet to be examined.Extraversion and conscientiousness have received extensive attention in the team literature; however, theprocess by which they influence collective efficacy and team performance has not been fully examined.Our results shed light on this process, by demonstrating that these personality characteristics, togetherwith task-relevant knowledge, predict individual-level teamwork behavior, which in turn, crosses levels topredict collective efficacy and team outcomes. Accordingly, our results suggest that awareness of teammembers’ knowledge and personality attributes is critical to understanding the evolution of teamperceptions and outcomes.228

A second contribution of this study is the finding that collective efficacy and relationship conflictappear to contextualize individual teamwork behavior. In this study, initial collective efficacy andrelationship conflict were significantly associated with the frequency with which individual teammembers engaged in such activities as goal monitoring and developing action plans. Therefore, it appearsthat individuals are more likely to be motivated to engage in teamwork behaviors when they findthemselves in efficacious groups and in groups free of relationship conflict.Finally, our results support Marks et al. (2001), who claimed that emergent states such ascollective efficacy are not processes in and of themselves because they do not describe how groupmembers interact while contributing to goal accomplishment. Rather, teamwork behaviors serve as inputsinto the collective development of perceptions of group capability, as illustrated by the positivecorrelation we observed between peer-rated individual-level teamwork behaviors and collective efficacy.Limitations and Future ResearchWe acknowledge several limitations. First, a common criticism of studies using student samplesis that the results may fail to generalize to business settings. Teams in business settings are likely to facedifferent, and perhaps more, constraints than the student teams in our study (e.g., budgets, time, policiesand turnover). It is possible that these constraints have an impact on individual teamwork behavior.However, we also note that these constraints will be included in collective efficacy perceptions, andthereby captured in our overall model. Nevertheless, the generalizability of our results must be assumedto be limited until empirical research explores other samples and other variables that may relate toteamwork behavior.Second, our results support the assertion that collective efficacy and relationship conflict play acontextual role in shaping individual teamwork behavior. It remains to be seen whether higher levelorganizational variables, such as organizational culture and perceived managerial support relate to theemergence of teamwork behavior. A recent study by Chen and Bliese (2002) found that leadershipclimate exerted a top-down influence on collective efficacy. Additional research examining top-downprocesses at a higher level than the team would not only enhance the generalizability of the currentliterature, but respond to Johns’s (in press) call to increase the level of contextual richness inorganizational research.Third, status differences among study participants were neither assigned nor assessed. Theparticipants in the study were instructed to play the role of a team member, but were not given specificroles. Nevertheless, our findings have important practical implications because in many situations teamsare given the responsibility to manage how their tasks are to be completed. Teams such as qualityimprovement or strategic decision making teams can frequently choose how they go about organizingthemselves to accomplish team goals. As the emerging literature on transactive memory indicates (i.e.,Austin, 2003), groups have an ability to organize themselves to take advantage of expertise. Futurestudies could explore how specific roles (e.g., monitor, encourager and devil’s advocate), whetherassigned or negotiated, affect the development of collective efficacy within the team and the subsequentperformance of the team. We also suggest that future research should explore whether teamworkbehaviors relate to collective efficacy in teams that have less opportunities for interaction (e.g., virtualteams).229

Practical ImplicationsOur finding that teamwork behavior was significantly related to collective efficacy speaks to thepractical importance of teamwork behavior and interventions that encourage such behavior. Mostcomplex political and organizational decisions are being made in groups because they frequentlyoutperform individuals in complex problem-solving tasks (Henry, 1995; Laughlin et al., 2002; Kerr &Tindale, 2004). Our results suggest that to capitalize on the potential advantages offered by groups,organizations would be wise to actively develop good teamwork behavior rather than focus solely ongroup outcome issues (improved efficiency or reducing error rates). We note that there are multiple waysto encourage the display of these behaviors. Training programs that teach appropriate teamwork behaviorare one way to increase teamwork skills. For example, Chen, Donahue, and Klimoski (2004) found that auniversity course geared towards teaching individual teamwork KSAs had a significant impact onindividual learning of the KSAs.In summary, our findings are consistent with the theory of collective structure, which states that “it is notthe collective construct, per se, that determines the behavior of individuals - rather, it is the individuals (orcollective) who determine the collective construct, and, through their actions, influence the behavior ofothers in the collective” (Morgeson & Hofmann, 1999: 251). Our results also suggest that teamworkbehavior at the individual level is likely to have a long-lasting effect on how well teams function andperform over time. Thus, understanding sources of teamwork behavior, and how these behaviors relate toteam functioning, is critically important for predicting the long-term prospects of the success of teams.Second, collective efficacy will change as a result of both performance feedback and the behavior of teammembers. Even a highly efficacious team can experience declining collective efficacy if its members donot continue to actively participate. Poor teamwork behavior is likely to result in a pattern of decreasingefficacy and performance over time, in what Lindsley et al. (1995) termed efficacy-performance spirals.On the other hand, such spirals can be positive and characterized by increasing levels of collectiveefficacy and performance.230

ReferencesAustin, J. R. (2003). Transactive memory in organizational groups: The effects of content, consensus,specialization and accuracy on group performance. Journal of Applied Psychology, 88, 866-878.Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. EnglewoodCliffs, N.J.: Prentice-Hall.Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W.H. Freeman.Barrick, M.R. & Mount, M.K. (1991). The big five personality dimensions and job performance: A metaanalysis.Personnel Psychology, 44, 1-26.Barrick, M.R. & Mount, M.K. (1993). Autonomy as a moderator of the relationships between the big fivepersonality dimensions and job performance. Journal of Applied Psychology, 78, 111-118.Barrick, M.R., Mount, M.K., & Judge, T.A. (2001). Personality and performance at the beginning of thenew millennium: What do we know and where do we go next? International Journal of Selectionand Assessment, 9, 9-29.Barrick, M.R., Mount, M.K., & Strauss, J.P. (1993). Conscientiousness and performance of salesrepresentatives: Testing of mediating effects of goal setting. Journal of Applied Psychology, 78,111-118Barrick, M.R., Stewart, G.L., Neubert, M.J., & Mount, M.K. (1998). Relating member ability andpersonality to work-team processes and team effectiveness. Journal of Applied Psychology, 83,377-391.Barry, B. & Stewart, G.L. (1997). Composition, process, and performance in self-managed groups: therole of personality. Journal of Applied Psychology, 82, 62-78.Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for dataaggregation and analysis. In K. J. Klein & S. W. Kozlowski (Eds.). Multilevel theory, research,and methods in organizations : Foundations, extensions, and new directions: 467-511. SanFrancisco: Jossey-Bass.Borman, W.C., Penner, L.A., Allen, T.D., & Motowidlo, S.J. (2001). Personality predictors of citizenshipperformance. International Journal of Selection and Assessment, 9, 52-69.Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysismethods. Newbury Park, CA: Sage.231

Cannon-Bowers, J. A., Tannenbaum, S. I., Salas, E., & Volpe, C. E. (1995). Defining competencies andestablishing team training requirements. In R. A. Guzzo & E. Salas (Eds.), Team Effectivenessand Decision Making in Organizations. San Francisco: Jossey-Bass.Castro, S. L. (2002). Data analytic methods for the analysis of multilevel questions: A comparison ofintraclass correlation coefficients, r wg(j) , hierarchical linear modeling, within- and betweenanalysis,and random group resampling. Leadership Quarterly, 13, 69-93.Chan, D. (1998). Functional relations among constructs in the same content domain at different levels ofanalysis: A typology of composition models. Journal of Applied Psychology, 83, 234-246.Chen, G. L., & Bliese, P. D. (2002). The role of different levels of leadership in predicting self- andcollective efficacy: Evidence for discontinuity. Journal of Applied Psychology, 87, 549-556.Chen, G. L., Donahue, L. M., & Klimoski, R. J. (2004). Training undergraduates to work inorganizational teams. Academy of Management Learning and Education, 3, 27-40.Costa, P.T. Jr. & McCrae, R.R. (1992). NEO-PI-R Professional Manual. Odessa, FL: PsychologicalAssessment Resources.De Dreu, K. W., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and teammember satisfaction: A meta-analysis. Journal of Applied Psychology, 88: 741-749.Devine, D.J., Clayone, L.D., Philips. J.L., Dunford, B.B., & Melner, S.B. (1999). Teams inorganizations: Prevalence, characteristics, and effectiveness. Small Group Research, 30, 678-711.Gibson, C. B., Randel, A. E., & Earley, P. C. (2000). Understanding group efficacy: An empirical test ofmultiple assessment methods. Group & Organization Management, 25, 67-97.Goldberg, L.R. (1990). An alternative “description of personality”: The big-five factor structure. Journalof Personality and Social Psychology, 59, 1216-1229.Goldberg, L.R. (1992). The development of markers for the big-five factor structure. PsychologicalAssessment, 4, 26-42.Goldberg, L.R. (1999). International Personality Item Pool. Scale available at: http://ipip.ori.org/ipip/.Gully, S. M., Incalcaterra, K. A., Joshi, A., & Beaubien, J. M. (2002). A meta-analysis of team-efficacy,potency, and performance: Interdependence and level of analysis as moderators of observedrelationships. Journal of Applied Psychology, 87, 819-832.Hackman, J. R. (Ed.). (1990). Groups that work (and those that don't): Creating conditions for effectiveteamwork. San Francisco: Jossey-Bass.Henry, R. A. (1995). Improving group judgment accuracy: Information sharing and determining the bestmember. Organizational Behavior and Human Decision Processes, 62, 190-197.Hofmann, D. A., Griffin, M. A., & Gavin, M. B. (2000). The application of hierarchical linear modelingto organizational research. In K. J. Klein & S. W. Kozlowski (Eds.). Multilevel theory, research,and methods in organizations : Foundations, extensions, and new directions: 467-511. SanFrancisco: Jossey-Bass.Hough, L.M. & Ones, D.S. (2002). The structure, measurement, validity, and use of personality variablesin industrial, work, and organizational psychology. In N. Anderson, S. Ones, H.K. Sinangil, & C.Viswesvaran (Eds.), International handbook of work and organizational psychology (Vol 1, pp.233-277). London, UK: Sage.Ilgen, D. R., Hollenbeck, J. R., Johnson, M., & Jundt, D. (2005). Teams in organizations: From inputprocess-outputmodels to IMOI models. Annual Review of Psychology, 56, 517-543.James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater reliability with andwithout response bias. Journal of Applied Psychology, 69, 85-98.Jehn, K. (1995). A multimethod examination of the benefits and detriments of intragroup conflict.Administrative Science Quarterly, 40: 256-282.Jehn, K. (1997). A qualitative analysis of conflict types and dimensions in organizational groups.Administrative Science Quarterly, 42: 530-557.Jehn, K. A., & Mannix, E. A. (2001). The dynamic nature of conflict: A longitudinal study of intragroupconflict and group performance. Academy of Management Journal, 44: 238-251.232

Johns, G. (in press). The essential impact of context on organizational behavior. Academy of ManagementReview.Judge, T.A. & Bono, J.E. (2001). Relationship of core self-evaluations traits – Self-esteem, generalizedself-efficacy, locus of control, and emotional stability – with job satisfaction and jobperformance: A meta-analysis. Journal of Applied Psychology, 86, 80-92.Judge, T.A., Erez, A., & Bono, J.E. (1998). The power of being positive: The relation between positiveself-concept and job performance. Human Performance, 11, 167-187.Judge, T.A. & Ilies, R. (2002). Relationship of personality to performance motivation: A meta-analyticreview. Journal of Applied Psychology, 87, 797-807.Judge, T.A., Locke, E.A., Durham, C.C., & Kluger, A.N. (1998). Dispositional effects on job and lifesatisfaction: The role of core evaluations. Journal of Applied Psychology, 83, 17-34.Judge, T.A., Erez, A., Bono, J.E., & Thoresen, C.J. (2003). The core self-evaluations scale: Developmentof a measure. Personnel Psychology, 56, 303-331.Kerr, N. L. & Tindale, R. S. (2004). Group performance and decision making. Annual Review ofPsychology, 55, 623-655.Klein, K. J., & Kozlowski, S. W. J. (2000). Multilevel theory, research, and methods in organizations :foundations, extensions, and new directions. San Francisco: Jossey-Bass.Laughlin, P. R., Bonner, B. L., & Miner, A. G. (2002). Groups perform better than the best individuals onLetters-to-Numbers problems. Organizational Behavior and Human Decision Processes, 88, 605-620.LePine J. A., Hollenbeck, J. R., Ilgen, D. R., & Hedlund, J. (1997). Effects of individual differences onthe performance of hierarchical decision-making teams: Much more than g. Journal of AppliedPsychology, 82, 803-811.Lester, S. W., Meglino, B. M., & Korsgaard, M. A. (2002). The antecedents and consequences of grouppotency: A longitudinal investigation of newly formed work groups. Academy of ManagementJournal, 45, 352-368.Littlepage, G.E., Schmidt, G.W., Whisler, E.W. & Frost, A.G. (1995). An input-process-output analysisof influence and performance in problem-solving groups. Journal of Personality and SocialPsychology, 69, 877-889.Lindsley, D. H., Brass, D. J., & Thomas, J. B. (1995). Efficacy - performance spirals: A multilevelperspective. Academy of Management Review, 20, 645-678.Locke, E. A. (1986). Generalizing from laboratory to field settings. Lexington, MA: Lexington Books.Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy ofteam processes. Academy of Management Review, 26, 356-376.Morgeson, F. P., & Hofmann, D. A. (1999). The structure and function of collective constructs:Implications for multilevel research and theory development. Academy of Management Review,24, 249-250.233

Mount, M.K. & Barrick, M.R. (1995). The big five personality dimensions: Implications for research andpractice in human resources management. In K.M. Rowland and G. Ferris (Eds.), Research inPersonnel and Human Resources Management (Vol. 13, pp. 153-200). Greenwich, CT: JAI Press.Mount, M.K., Barrick, M.R., & Stewart, G.L. (1998). Five-factor model of personality and performancein jobs involving interpersonal interactions. Human Performance, 11, 145-165.Mowday, R. T., & Sutton, R. I. (1993). Organizational behavior: Linking individuals and groups toorganizational contexts. Annual Review of Psychology, 44, 195-229.Neuman, G. A., & Wright, J. (1999). Team effectiveness: Beyond skills and cognitive ability. Journal ofApplied Psychology, 84, 376-389.Pelled, L., Eisenhardt, K., & Xin, K. (1999). Exploring the black box: An analysis of work groupdiversity, conflict, and performance. Administrative Science Quarterly, 44: 1-28.Podsakoff, P.M., MacKenzie, S.B., Paine, J.B., & Bacharach, D.G. (2000). Organizational citizenshipbehaviors: A critical review of the theoretical and empirical literature and suggestions for futureresearch. Journal of Management, 26(3), 513-563.Rousseau, D. M. (1985). Issues of level in organizational research: Multilevel and cross-levelperspectives. In L. L. Cummings & B. M. Staw (Eds.), Research in Organizational Behavior(Vol. 7, pp. 1-37). Greenwich CT: JAI Press.Smith, J. R., & Golden, P. A. (2001). Human Resource Management Simulation: Player's Manual.Upper-Saddle River, NJ: Prentice Hall.Sosik, J. J., Kahai, S. S., & Avolio, B. J. (1998). Transformational leadership and dimensions ofcreativity: Motivating idea generation in computer-mediated groups. Creativity Research Journal,11, 111-121.Stevens, M.J., & Campion, M.A. (1994). The knowledge, skill, and ability requirements for teamwork:Implications for human resource management. Journal of Management, 20, 503-530.Stewart, G.L., Carson, K.P., & Cardy, R.L. (1996). The joint effects of conscientiousness and selfleadershiptraining on employee self-directed behavior in a service setting. Personnel Psychology,49, 143-164.Stewart, G.L., Fulmer, I.S., & Barrick, M.R. (2005). An exploration of member roles as a multilevellinking mechanism for individual traits and team outcomes. Personnel Psychology, 58, 343-365.Taggar S., Hackett, R., & Saha, S. (1999). Leadership emergence in autonomous work teams:Antecedents and outcomes. Personnel Psychology, 52, 899-926.Taggar, S., & Seijts, G. H. (2003). Leader and staff role-efficacy as antecedents of collective efficacy andteam performance. Human Performance, 16, 131-156.Tasa, K., Taggar, S., & Seijts, G. H. (2006). The development of collective efficacy in teams: A multileveland longitudinal perspective. Unpublished Manuscript.Wood, R.E. 1986. Task complexity: Definition of the construct. Organizational Behavior and HumanDecision Processes, 37, 60-82.Wood, R.E., & Bandura, A. (1989). Impact of conceptions of ability on self-regulatory mechanismsandcomplex decision making. Journal of Personality and Social Psychology, 56, 407-415.234

Table 1Descriptive Statistics, Correlations, and Reliability CoefficientsVariables Mean SD 1 2 3 4 5 6Individual-level measures1. Extraversion 4.65 1.06 (.86) a2. Conscientiousness 4.89 .87 .03 (.76)3. Core self-evaluation 4.91 .85 .30** .21** (.84)4. Task-relevant knowledge 72.55 8.37 .02 .09 .22** --5. Teamwork behavior 4.90 .92 .17** .14** .21** .25** (.93)Team-level measures1. Average teamwork behavior 4.89 .55 --2. Relationship conflict 2.37 1.01 -.32** (.90)3. Collective efficacy (week 2) 5.75 .47 .47** -.16 --4. Collective efficacy (week 7) 5.79 .65 .52** -.14 .64** --5. Team performance 6.22 3.47 .02 -.08 .12 .40** --6. Team viability 4.07 .99 .46** -.47** .26** .27** .21* (.82)a For individual-level measures, n = 434; for team-level measures, n = 114. Numbers in parentheses arecoefficient alphas.* p < .05,** p < .01.Table 2HLM Results for Individual Teamwork Behavior aVariable Coefficient S.E. tIntercept 4.91** .04 106.73Extroversion (H1a) .16** .05 3.55Conscientiousness (H1b) .14* .06 2.49Core self-evaluation (H1c) .12* .05 2.29Task-relevant knowledge (H1d) .02** .01 3.89Collective efficacy - week 2 (H2a) .46** .11 4.26Relationship conflict (H2b) -.18** .04 -4.37aEntries are estimations of the fixed effects with robust standard errors.* p < .05** p < .01235

ASAC 2006Banff, AlbertaSamantha D. MontesP. Gregory IrvingSchool of Business and EconomicsWilfrid Laurier UniversityTHE MEDIATING ROLE OF TRUST IN PSYCHOLOGICAL CONTRACTBREACH EFFECTS: CLARIFYING THE RELATIONSPast research demonstrating that trust mediates the relations between psychologicalcontract breach and employee reactions might be criticized because breach was measuredusing difference scores and direct measures. Research in related areas suggests thatfindings based on such measures may be misleading. To demonstrate that trust is animportant underlying mechanism of breach effects, we employed an alternativemethodology that avoids the problems associated with the use of difference scores. Ourresults suggest that trust mediates the effects of breach, but only for certain types ofinducements.The psychological contract has been acclaimed an important topic in the study of the employmentrelationship. Psychological contracts refer to the perceived mutual obligations between employees andtheir employers (Rousseau, 1989). To date, a great deal of research on psychological contracts hasfocused on the attitudinal and behavioural consequences of psychological contract breach. Formally,breach is defined as a perceived discrepancy between what one’s organization promised to provide andwhat one actually received (Robinson & Rousseau, 1994). Numerous empirical investigations havedemonstrated that breach is associated with negative outcomes, both for individual employees (e.g., jobsatisfaction, Robinson & Rousseau, 1994; feelings of violation, Robinson, 1996) and for the organization(e.g., intentions to remain with the organization, Robinson, 1996; Turnley & Feldman, 2000). However,little empirical attention has been directed at the underlying mechanisms of these effects.Organizational trust has been touted as fundamental to understanding the negative consequencesof psychological contract breach (e.g., Buch & Aldridge, 1991; Rousseau, 1989). Further, there is someempirical evidence that trust serves an important mediating role in the relations between psychologicalcontract breach and employee attitudes and behavioural intentions (Robinson, 1996; Robinson &Morrison, 1995). However, recent criticisms directed at the measurement strategies (e.g., use of directmeasures and difference scores) typically employed in the study of psychological contract breach suggestthat our current understanding may be limited. In the present research, we employed an alternativemeasurement strategy to test the mediating role of trust in the relations among promised and deliveredinducements and several commonly studied outcome variables.Trust and Psychological Contract BreachAs defined in the psychological contract literature, trust refers to “one’s expectations,assumptions, or beliefs about the likelihood that another’s future actions will be beneficial, favourable, orat least not detrimental to one’s interests” (p.576; Robinson, 1996). Further, trust is thought to entail somewillingness to be vulnerable to the actions of another party (Mayer, Davis, & Schoorman, 1995). Theseaspects make trust especially relevant to psychological contracts because these agreements containimplicit promises exchanged between an employee and his or her employer. These sorts of implied236

agreements (e.g., organizational support in exchange for hard work and loyalty) lack the legal safeguardsassociated with explicit written contracts. That is, an employee’s confidence in the organization to fulfillthe promises set out in the psychological contract is not supported by the threat of legal action if thecontract is broken. Instead, the employee must trust that the organization will live up to its word.Indeed, researchers examining the role of psychological contracts in employee-employerrelationships have adopted trust as an important underlying mechanism that binds the two parties to theperceived agreement (Robinson, 1996; Rousseau, 1989, 1995; Rousseau & McLean Parks, 1993).Because trust is thought to be an integral aspect of the psychological contract, several authors havecontended that breach may have its negative effects on employee attitudes and behaviours through aviolation of trust (e.g., Buch & Aldridge, 1991; Rousseau, 1989). Robinson (1995) suggested thatemployees whose psychological contracts have been breached may feel that they cannot trust theirorganization to fulfill the remaining obligations of the psychological contract and, furthermore, that theorganization cares little about its employees. Robinson (1996) further suggested that, in such situations,this compromising of trust will lead to a decrease in employees’ motivation to contribute to theorganization, and an increase in behaviours intended to distance themselves from the organization (e.g.,intentions to quit). Indeed, those studying the psychological contract and trust have demonstrated thatbreach is negatively related to trust (e.g., Robinson & Rousseau, 1994) and that trust is positively relatedto important organizational outcomes like organizational citizenship behaviours (e.g., Rousseau &Tijoriwala, 1998). However, little published research has examined whether trust is, in fact, an underlyingmechanism of psychological contract breach effects. A notable exception to this is work done byRobinson (Robinson, 1996; Robinson & Morrison, 1995).In a comprehensive examination of the role of trust, Robinson (1996) tracked a group ofnewly hired managers over the course of two and a half years. At Time 1, respondentsindicated the extent of obligations that their employer had made at the time of hire. AtTime 2, participants were asked to indicate the degree to which their employer hadfulfilled the previously made obligations. Time 1 obligations and Time 2 fulfillment wereused to create a difference score measure of breach. Finally, at Time 3, participants’trust, performance, civic virtue, intentions to remain, and turnover were measured.Robinson’s results suggested that trust served to mediate the effects of breach on theseemployee behaviours.Robinson and Morrison (1995) reported analyses examining the mediating role of trust inthe relationship between breach and civic virtue. However, unlike the Robinson (1996)paper, in this case breach was assessed only with direct measures and separate analyseswere conducted for relational and transactional types of contracts. These analysesindicated that only relational contract breaches were negatively related to civic virtueand that trust only served to mediate the relationship between relational breaches andcivic virtue (i.e., trust did not mediate the effects of transactional contract breaches).Potential Shortcomings of Existing ResearchSeveral factors may limit the generalizability of Robinson’s (1996) Robinson andMorrison’s (1995) findings. First, because trust and the outcome variables weremeasured at the same point in time, these relations may have been inflated due tocommon method variance. Second, Robinson (1996) collapsed across two potentiallydistinct types of inducements (i.e., relational and transactional) in her measure of237

psychological contract breach. In doing so, an implicit assumption was made that trustmediates breach effects for both, relational and transactional, types of contracts.Relational contracts are described as relationship-oriented, long-lasting, and requiring agreat deal of emotional investment. In contrast, transactional contracts are short-term,monetizable exchanges between employee and employer and are not said to require anyemotional investment. Relational contracts typically include such obligations as jobsecurity and development opportunities whereas transactional contracts tend to includeobligations such as promotions and high pay. Because relational and transactional typesof contracts might be related to outcome variables in unique ways, it has been advisedthat, when studying psychological contracts, researchers avoid collapsing across thesepotentially distinct constructs (e.g., Coyle-Shapiro & Kessler, 2000). Indeed, whenRobinson and Morrison (1995) examined the separate effects of relational andtransactional contract breaches, they found evidence that trust only mediated the effect ofrelational breaches on civic virtue behaviour.Although not discussed in the Robinson & Morrison (1995) paper, there is reason tobelieve that trust may not serve to mediate the effects of transactional contract breaches.Relational and transactional contracts have been likened to Blau’s (1964) social andeconomic exchanges, respectively (Robinson, Kraatz, & Rousseau, 1994). Blau suggestedthat only social exchanges rely on trust. In a social exchange, each party must trust thatdelivered inducements will be reciprocated by the other party. In contrast, economicexchanges do not require that parties trust one another. Instead, economic exchangeshave legal safeguards that ensure obligations will be discharged. Thus, trust may bemore relevant to explaining the effects of relational contract breaches as compared totransactional contract breaches.These two issues raise some concern about interpretation of existing findings (i.e., Robinson,1996; Robinson & Morrison, 1995), however, a third, and potentially more serious problem, exists. Pastresearch relied on the use of difference scores (Robinson, 1996) and direct measures (Robinson &Morrison, 1995) to quantify breach. Use of these measures to assess discrepancies has receivedwidespread criticism in various literatures and has been shown to lead to problematic interpretation ofresults (e.g., Cronbach & Furby, 1970; Edwards, 1991, 1994, 2001; Edwards & Cooper, 1990; Irving &Meyer, 1994, 1995; Johns, 1981).A key issue in such criticism is that difference scores and direct measures confound the effects ofthe two separate component measures (Edwards, 2001). For example, in the study of psychologicalcontract breach, findings based on these measures of breach are interpreted in terms of a discrepancybetween promised and delivered inducements. This precludes any theorizing about the separate effects ofthese components themselves; results may be due to one or the other, or both components. Thus, thesemeasures serve to potentially mask important differences in the relations between each of the componentsand the outcome variable.Edwards (2001) also argued that the use of difference scores and direct measures of breach fail totest constraints that exist when assessing a discrepancy hypothesis. That breach is defined as adiscrepancy between two components, implies and algebraic difference score model. Edwards used thefollowing series of regression equations to illustrate the constraints imposed by this model.238

Equation 1 expresses the use of an algebraic difference score (between components X and Y) as apredictor of an outcome (Z):Expanding Equation 1 results in the following:Z = b 0 + b 1 (X - Y) + e (1)Z = b 0 + b 1 X - b 1 Y + e (2)As is apparent in its expanded form (i.e., Equation 2), two untested restrictions arise, namely, that thecoefficients for the two components are equal in magnitude and opposite in sign. Edwards cautioned thatsuch assumptions should not be made, but rather, treated as hypotheses to be tested. These hypotheses arenot testable when using direct measures or difference scores. Implicit assumptions made when using thesemeasures to assess discrepancies may lead to faulty interpretation of the relations among the variables. Incontrast to the constrained model, as depicted in Equation 3, if the effects of the two components on theoutcome are separately examined, no such assumptions are made.Z = b 0 + b 1 X + b 2 Y + e (3)If the component measures are entered as separate predictors in the analysis (i.e., Equation 3), theassumptions imposed by difference scores and direct measures can be tested. Consistent with thedefinition of psychological contract breach, evidence supporting a discrepancy hypothesis (e.g., apsychological contract breach effect) would be found if the regression coefficients for the components areapproximately equal in magnitude and opposite in sign (Edwards, 1991; Irving & Meyer, 1994).Present ResearchThe aforementioned limitations of research that has examined the mediating role of trust suggestthat the literature lacks strong empirical support for one of its fundamental contentions. Namely, thatnegative employee reactions following psychological contract breach are due to damaged trust. Becauseexisting tests of this hypothesis used difference scores and direct measures to assess breach, we sought totest the mediation hypothesis using an alternative strategy that avoids the problems associated with theuse of these measurement strategies. Specifically, we disentangled the effects of promised and deliveredinducements and examined the joint mediating role of trust. According to work by Edwards (e.g.,Edwards, 2001) we examined the pattern of relations for evidence that trust mediates the effects of breachon outcome variables. Following from Edwards’ logic, if trust serves to mediate the effects of breach onoutcome variables, we would expect that the indirect effects of promised and delivered inducements (viatrust) on outcomes would be approximately equal in magnitude and opposite in sign. Based on existingpsychological contract breach literature and findings, we hypothesized the following:Hypothesis: The indirect effects of promised and delivered inducements on the outcome variables will beapproximately equal in magnitude and opposite in sign.We also expected that support for this hypothesis would be found only for relational types ofcontracts. That is, based on the notion that transactional contracts are more economic in nature, we did notexpect trust to mediate the relations between transactional contract breach and the outcome variables.We tested the mediation hypothesis within the context of a four-month co-operative educationwork term. Using a 3-wave longitudinal design, we examined the mediating role of trust in the relationsamong promised and delivered inducements and several commonly studied outcome variables, namely,satisfaction, feelings of violation and employment intentions. In contrast to past work, employment239

intentions in this study did not refer to whether employees intended to remain in the employ of theorganization. Given the limited duration of the co-op employment relationship, we operationalizedemployment intentions as employees’ self-reported intentions to return to work for the employer in thefuture. For many reasons, students entering a co-op work term may intend to work for their co-op workterm employer again if given the opportunity. Over the course of the work term, however, students’experiences (e.g., psychological contract breach) might alter their intentions to work for that particularorganization again. Therefore, in the present study, rather than assess effects on overall employmentintentions, we assessed the effects of the predictors and mediator on change in employment intentions bycontrolling for initial employment intentions in our prediction of Time 3 employment intentions. In ouranalyses, we also controlled for participants’ initial levels of trust in the organization. According toexisting literature, negative employee attitudes and behaviours following psychological contact breachoccur purportedly because employees’ trust has been damaged. By controlling for individuals’ initial trustin our predictions of later trust, we were able to better capture and test this notion.Participants and ProcedureMethodStudents (51% female) entering a four-month Co-operative Education work term were recruitedvia email to participate in an anonymous web-based longitudinal study of their work term experiences.Respondents completed a web-based survey at three points in time: one month prior to the start of thework term (but after they had conversed with the employer), three months into the work term, and onemonth following the completion of the work term. Participant responses to each survey were matched byunique identification codes.The first survey included measures of promised inducements, initial trust in the organization, andinitial employment intentions, as well as measures of biographical and program-related information (e.g.,number of organizations applied to). A total of 522 individuals (51% women) completed the Time 1measures. The majority (90%) of respondents were between the ages of 19 and 22, had a self-reportedGPA between 75 and 89 (72%). Most secured jobs with their first or second choice organization (76%)and reported never having worked for that employer before (76%). Respondents were enrolled in a widevariety of degree programs, had varying degrees of co-op work term experience, and were employedwithin a wide variety of organizations.Three months into their work term, respondents were contacted by email and directed to thesecond survey which contained measures of delivered inducements and organizational trust. Of theindividuals who completed the first survey, 383 completed the second resulting in a response rate of 73%.Approximately one month later, after the work term had ended, respondents were sent a final emailmessage with a link to the last survey. This survey contained measures of feelings of violation,satisfaction, and employment intentions. Of those individuals who completed the second survey, 342completed the third yielding a Time 3 response rate of 89%. The overall response rate from Time 1 toTime 3 was 65%. No significant differences in demographic or experience variables were found amongthose who completed all three surveys and those who did not. However, all analyses were conducted onthe responses submitted by respondents who completed all three surveys.MeasuresPromised and delivered inducements. At Time 1, respondents were asked to rate the extent to whichtheir co-op work term employer promised to provide them with each of eight items on a 5-point scaleranging from 1 (Minimally or not at all) to 5 (To a very large extent).240

Originally, these and other similar items were drawn from Co-operative Education Programliterature that was created for the purpose of marketing a Co-op Program for university students.Interviews with Co-op Program administrators confirmed that participating organizations had agreed toprovide Co-op students with these inducements. For the most part, these items are quite similar to thoseused in general working populations (e.g., help in developing externally marketable skills) however, theywere somewhat tailored to ensure appropriateness for a co-op work term employee sample (e.g.,compensation that is consistent with other co-op placements in similar industries/geographical locations).A pilot test of these items was conducted at a neighbouring university. The results of a factoranalysis indicated that the items loaded onto three distinct factors: support, skill development, andcompensation. Somewhat consistent with traditional views, the psychological contracts of respondents inthe pilot study consisted of both relational and transactional elements (Coyle-Shapiro & Kessler, 2000;McLean Parks, Kidder, & Gallagher, 1998; Rousseau, 1990). Compensation items are considered moretransactional in nature, whereas, support and skill development items might be considered more relationalin nature (e.g., Rousseau, 1990). That support and skill development items loaded onto two distinctfactors is in line with previous work on psychological contract breach. Research has demonstrated thatitems related to training opportunities (e.g., skill development) do not consistently load onto relationaland transactional factors, and in some cases, load on to a distinct factor (Arnold, 1996; Coyle-Shapiro &Kessler, 2000).To minimize the length of the survey in the present study, from the pilot study we chose the threehighest loading items for support and skill development and the two existing compensation items. In totalthen, eight items were used to measure promised inducements in the study proper (see Table 2). At Time2, commensurate measures of delivered inducements were used to assess the extent to which employeesreceived each of these same items.Organizational trust. For purposes of comparison, the same scale that was used by Robinson (i.e.,Robinson, 1996; Robinson & Morrison, 1995) was used to measure trust in the present study at Time 1and Time 2. 1 To clearly specify the target of trust, “work term employer” was substituted for “employer”in each of the questions. All responses were made on a 5-point scale that ranged from 1 (Stronglydisagree) to 5 (Strongly agree).Employment intentions. At Time 1 and Time 3, respondents were asked to indicate the probability thatthey would accept future work term and permanent employment opportunities with their employingorganizations if given the chance. These two items were combined to form a measure of employmentintentions. Responses were made on an 11-point scale defined by probabilities (0% - Definitely wouldnot, 100% - Definitely would).Satisfaction. Four global items adapted from previous work on psychological contracts (e.g., Robinson &Rousseau, 1994) were used at Time 3. These items were combined into a global measure of satisfaction.Responses were made using a 5-point scale ranging from 1 (Strongly disagree) to 5 (Strongly agree).Feelings of violation. At Time 3 respondents completed a 4-item scale designed to measure feelings ofviolation (Robinson & Morrison, 2000). Responses were made on a 5-point scale ranging from 1(Strongly disagree) to 5 (Strongly agree).1To avoid confounding the trust measure with the related concept of organizational justice, one item that clearlymeasured perceived fairness was removed.241

Overview of AnalysesBecause we hypothesized that trust would mediate the joint effects of promised and deliveredinducements on outcomes, structural equation modelling (SEM) using AMOS Graphics (Arbuckle, 2003),in combination with the Joint Significance Test, was used to test the mediation hypothesis. In contrast tothe commonly used method of testing mediation effects (e.g., regression analyses combined with a SobelTest), the method used in the present research enables one to assess the significance of a mediator on thecombined effects of two predictor variables. In total, nine path models representing all combinations ofbreach type (i.e., support, skill development, and compensation) and outcome variables (i.e., satisfaction,violation, change in employment intentions) were estimated. As an example of the models tested, themodel used to assess the mediating role of trust in the support breach-satisfaction relationship is presentedin Figure 1. This model contends that promised and delivered support both directly and indirectly (viatrust) affect satisfaction. Given their potentially significant relationship with the outcome variables, gradepoint average (GPA) and how the organization ranked in terms of employees’ preferences (i.e.,preference) were used as controls in each analysis.Figure 1Example Path Model Used to Test Mediation HypothesisT1 Truste1e2PromisedSupportDeliveredSupportaaT2 TrustbSatisfactionGPAPreferenceConventional methods of testing indirect effects (e.g., Sobel and Goodman tests) have beencriticized on the basis of high type II error rates (MacKinnon, Lockwood, Hoffman, West, & Sheets,2002) and alternative methods (e.g., MacKinnon’s Z’ Test) have been criticized for having high type Ierror rates (MacKinnon et al., 2002). In a comparison of 14 methods for assessing the significance ofindirect effects (including the two aforementioned methods), MacKinnon et al. (2002) empiricallydemonstrated that the joint significance test is the best test of mediation (i.e., the method that producedthe best balance of type I and type II errors and the highest power). This test involves examining thesignificance of the paths from the predictor to the mediator (path a) and from the mediator to the outcome242

(path b) while controlling for the direct effects (see Figure 1). If both of these paths that comprise theindirect effect are found to be significant, evidence for mediation is said to exist.ResultsTable 1 presents the means, standard deviations, reliability coefficients, and correlations for allstudy variables.243

Table 1Means, Standard Deviations, Correlations, and Reliabilities a, bVariable 1 2 3 4 5 6 7 8 9 10 11 12 13 141. GPA2. Preference .013. Trust (T1) .17 -.16 (.79)4. Emp. Intentions (T1) .15 -.37 .24 (.78)5. Support Promised .10 -.14 .33 .26 (.76)6. Skill Dev. Promised .06 -.13 .02 .30 .43 (.75)7. Comp. Promised .17 -.07 .20 .22 .31 .17 (.81)8. Support Delivered .13 -.02 .19 .13 .33 .22 .15 (.76)9. Skill Dev. Delivered .17 -.08 .09 .25 .24 .44 .11 .59 (.74)10. Comp. Delivered .20 -.13 .21 .25 .16 .01 .59 .22 .23 (.84)11. Trust (T2) .15 -.07 .36 .03 .07 .02 .03 .49 .32 .12 (.82)12. Satisfaction .20 -.21 .23 .22 .10 .13 .02 .41 .47 .25 .42 (.92)13.Feelings of Violation -.17 .10 -.26 -.11 -.03 -.05 .04 -.38 -.32 -.14 -.55 -.65 (.90)14. Emp. Intentions (T3) .17 -.24 .16 .56 .13 .22 .11 .25 .45 .28 .27 .62 -.40 (.80)Means 4.87 1.61 4.22 7.12 3.58 3.27 3.54 3.15 3.11 3.42 4.24 3.88 1.49 6.83Standard Deviations 1.73 0.98 0.58 2.54 0.78 0.87 1.05 0.86 0.89 1.02 0.68 1.00 0.85 3.09a n = 293; alpha reliabilities are in parentheses along the diagonal.b For correlations greater than or equal to .10, p < .10; greater than or equal to .12, p < .05; greater than or equal to .15, p < .01244

Factor AnalysisConfirmatory factor analysis could not be used to assess the factor structure of the inducementitems because the model was unidentified. Instead, we conducted a principal components analysis withdirect oblimin rotation on the promised inducement items to determine whether the pattern of factorloadings was consistent with that found in the pilot sample. Consistent with the pattern found in the pilottest, three factors accounting for 72% of the variance had eigenvalues greater than one and examination ofthe scree plot supported this interpretation. As seen in Table 2, each item loaded highly on only onefactor. The items for each factor were combined to form the promised support, promised skilldevelopment, and promised compensation scales. Commensurate scales were then created for thedelivered inducements items.Table 2Factor Loadings for Promised InducementsFactorSkill Dev. Comp. SupportItemsAmple constructive feedback on performance .11 -.02 -.78Sufficient training and supervision -.15 -.01 -.86Support in attaining the highest possible levels of performance .17 .08 -.73Job assignments that enhance external marketability .88 -.04 .08Help to develop externally marketable skills .81 -.01 -.14An introduction to valuable employment contacts .73 .06 -.01Compensation that is consistent with other co-op placementsin similar industries/geographical locations-.03 .92 -.01Sufficient salary to help cover educational costs .02 .92 .01Eigenvalues 3.21 1.50 1.08Percent of Variance 40.13 18.79 13.56Note. Skill Dev. = Skill Development, Comp. = CompensationMediation TestsFit indices for models predicting satisfaction indicated that each model (support χ 2 = 4.45, df = 3,p = ns, CFI = .996, RMSEA = .038; skill development χ 2 = 4.16, df = 3, p = ns, CFI = .997, RMSEA =.034; compensation χ 2 = 5.32, df = 3, p = ns, CFI = .994, RMSEA = .048) fit the data very well. Likewise,fit was quite good for models predicting feelings of violation (support χ 2 = 5.29, df = 3, p = ns, CFI =.994, RMSEA = .047; skill development χ 2 = 3.96, df = 3, p = ns, CFI = .997, RMSEA = .031;compensation χ 2 = 5.77, df = 3, p = ns, CFI = .993, RMSEA = .052) and change in employment intentions(support χ 2 = 10.18, df = 4, p = .03, CFI = .987, RMSEA = .067; skill development χ 2 = 13.98, df = 4, p =.01, CFI = .979, RMSEA = .086; compensation χ 2 = 13.34, df = 4, p = .01, CFI = .981, RMSEA = .083).A decomposition of the effects in each of these models is presented in Table 3. As seen in that table,delivered inducements (support, skill development, and compensation) were significant predictors of eachoutcome variable. However, compensation was the only type of promise that seemed to predict the threeoutcome variables. Inconsistent with the conceptualization of breach as a discrepancy between promisedand delivered inducements, promised support and promised skill development did not significantlypredict satisfaction, feelings of violation or change in employment intentions. These outcome variables236

appear to be influenced more so by the inducements delivered than by a discrepancy between promisedand delivered inducements (i.e., psychological contract breach).Table 3Effects DecompositionPredictorsSupport Skill Development CompensationPromised Delivered Promised Delivered Promised DeliveredSatisfactionDirect Effect -.03 .29*** -.06 .42*** -.14* .25***Indirect Effect -.05** .12*** -.03* .09*** -.03 .05*Total Effect -.08 .41*** -.09† .51*** -.17* .30***ViolationDirect Effect .08† -.16** .06 -.19*** .15** -.15**Indirect Effect .09*** -.23*** .05* -.16*** .04 -.07*Total Effect .17* -.39*** .11* -.35*** .19* -.22***EmploymentIntentionsDirect Effect -.08 .14** -.06 .30*** -.10† .15**Indirect Effect -.03* .08*** -.02† .05** -.02 .03†Total Effect -.11† .22*** -.08 .35*** -.12 .18***Note. † p < .10, * p < .05, ** p < .01, *** p < .001. Effects are standardized path coefficients.As a first step, we used the Joint Significance test (MacKinnon et al., 2002) to test whether trustmediated the separate effects of promised and delivered inducements on outcome variables. Recall that,according to this test, evidence of mediation is said to exist if, while controlling for the direct effects, bothpaths contributing to the indirect effect (i.e., predictor to mediator and mediator to outcome) aresignificant. Table 4 summarizes the coefficients associated with the indirect effect paths in all analyses.As seen in this table, the mediator-outcome path (b) was significant in all cases. However, the predictormediatorpath (a) was not significant in the case of promised compensation for any of the outcomes.These findings suggest that trust mediated the separate effects of promised and delivered support and skilldevelopment on all three outcome variables. However, trust only mediated the effects of deliveredcompensation on these outcomes (i.e., trust did not mediate the effect of promised compensation on anyoutcome variable).237

Table 4Results of Joint Significance TestsPromisedInducements toTrust (T2)Path aDeliveredInducements toTrust (T2)Indirect Effect PathsPath bTrust (T2) toOutcome VariableInducement-OutcomeRelationshipSupport-Satisfaction -.19*** .48*** .25***Feelings of Violation -.19*** .48*** -.47***∆ Employment Intentions -.19*** .48*** .17***Skill Development-Satisfaction -.11* .33*** .26***Feelings of Violation -.11* .33*** -.48***∆ Employment Intentions -.11* .33*** .15***Compensation-Satisfaction -.07 .14* .37***Feelings of Violation -.07 .14* -.53***∆ Employment Intentions -.07 .14* .22***Note. Values are standardized path coefficients. * p < .05, *** p < .001Because we were interested in determining whether trust mediates the effects of breach and notsimply one or the other components that comprise breach, the indirect effects of promised and deliveredinducements were examined simultaneously. Following from Edwards’ (2001) rationale, if significantindirect effects for both of these components are approximately equal in magnitude and opposite in sign,this would support the hypothesis that the effects of breach on outcomes are mediated through trust. Asseen in Table 3, the significant indirect effects for promised and delivered support and skill developmenton all three outcome variables were opposite in sign, and, in most cases, were approximately equal inmagnitude. For feelings of violation, the indirect effects of delivered inducements appeared to besomewhat larger than the indirect effects of promised inducements. 2 This suggests that deliveredinducements may play a somewhat stronger role in these relations. Consistent with past work (e.g.,Robinson, 1996; Robinson & Morrison, 1995), these findings suggest that trust mediates the effects ofpsychological contract breach of support and skill development promises on satisfaction, feelings ofviolation, and change in employment intentions. However, trust did not appear to mediate the effects ofcompensation breach on any of the outcome variables. Rather, in this case, trust only functioned tomediate the effects of delivered compensation on satisfaction, feelings of violation, and change inemployment intentions. Together, these findings provide mixed support for past research examining themediating role of trust in psychological contract breach effects.2These mediation analyses do not provide the data necessary to determine if these paths are statisticallyequivalent.238

DiscussionAccording to existing theorizing regarding the psychological contract (e.g., Rousseau, 1995) andto the extant research (e.g., Robinson, 1996), an employee’s trust in the organization will be damaged ifthat employee perceives a discrepancy between promised and delivered inducements. That is, if theorganization fails to live up to the promises it made to the employee, the employee may no longer trustthe organization and may, because of that, develop negative feelings and attitudes toward the organizationand withdraw their contributions to the organization. This notion has received support from studies thatexamined the mediating role of trust (e.g., Robinson, 1996). However, the findings of those studies mightbe questioned given the methodology used (e.g., trust and outcome variables measured at the same pointin time, confounded effects of promised and delivered inducements on outcome variables).In the present study, we assessed trust and the outcome variables at separate points in time,separately studied three distinct types of contracts, and examined the separate and joint effects ofpromised and delivered inducements on trust and the outcome variables. Our results provided mixedsupport for the findings of past research (e.g., Robinson, 1996). Consistent with previous research, wefound evidence that trust mediates the relations between both support and skill development contractbreaches and satisfaction, feelings of violation, and change in employment intentions. In the case ofcompensation inducements, although the outcome variables were significantly influenced by a (direct)psychological contract breach of compensation promises, this effect did not appear to be mediated bytrust. We found that trust only mediated the effects of delivered compensation on outcome variables (i.e.,the effects of promised compensation were not mediated by trust).That trust did not mediate the joint effects of promised and delivered inducements in the case ofcompensation is not that surprising given existing theorizing about transactional contracts. As discussedpreviously, transactional contracts have been likened to economic exchange. Because safeguards are inplace to guarantee that these types of obligations are discharged, employees need not trust that thesepromises will be fulfilled. This would explain why trust did not mediate the effects of promisedcompensation on outcomes. Trust was only damaged (and only led to subsequent negative reactions) bythe amount of compensation delivered. This finding may not have surfaced in Robinson’s work (i.e.,Robinson, 1996; Robinson & Morrison, 1995) because of the manner in which she measuredpsychological contract breach.Potential LimitationsAlthough the present research aimed to improve upon previous research, it is not without itslimitations. The generalizability of these findings might be limited by several factors. First, the sampleused in this study was unique—co-operative education students employed on a four-month work term.Use of a student sample might be criticized based on students’ lack of experience in organizationalsettings. However, research (e.g., Barling, Kelloway, & Bremermann, 1991) suggests that most studentshave employment experience and that even those who might have limited experience tend to have adeveloped knowledge and attitudinal base similar to experienced workers. Although it is unknownwhether students in this study had extensive employment experience, 70% of the respondents indicatedthat they had completed one or more prior co-op work terms and, therefore, it is likely that theseindividuals responded in ways similar to what would be expected from more experienced employees.A related criticism concerns the employment status of this sample. Unlike much of the researchon psychological contract breach, participants in this study were employed for a limited duration. Moreconservative generalizations, therefore, might be made to contingent workers—individuals employed on alimited-term, contract basis. Like contingent workers, respondents in the present study were employed fora pre-determined amount of time. Although the content of the psychological contracts of contingent239

workers might differ from that of permanent full-time workers (e.g., promises regarding long-term jobsecurity), research suggests that contingent workers do perceive psychological contracts with theiremployers and that, at times, employers breach these contracts (McLean Parks et al., 1998). Thus,although employees in the present research were not full-time workers, their attitudes and behaviouralintentions as related to breach would not be expected to dramatically differ.Use of a contingent employee sample might, in fact, be seen as an additional strength of thepresent research. One might suggest that using a sample in which respondents are employed for a predeterminedamount of time has an advantage over studies that are conducted with samples of permanentemployees. In the former case, it is clear whether organizations have fulfilled their obligations. Whendealing with contingent workers, organizations have a limited time within which to fulfill the promisesthat they make to employees. When studying breach among permanent workers, promise fulfillmentbecomes less clear. Some employees may not have been employed long enough to have had theirpromises fulfilled. Thus, in the present study, it would have been fairly clear to employees whether or nottheir employers had lived up to their promises.ConclusionDespite the limitations of this research, our findings suggest caution is warranted in interpretingthe results of research that collapses across distinct contract forms and uses difference scores and directmeasures to assess psychological contract breach. By disentangling the effects of promised and deliveredinducements and by separating the unique types of psychological contracts, we found that interesting andimportant differences in the relations among promised and delivered inducements, organizational trust,and employee attitudes and behavioural intentions exist. Although our findings regarding compensationappear inconsistent with past empirical work (e.g., Robinson, 1996), they are consistent with currenttheorizing that relational and transactional contracts are distinct forms and these forms map well ontosocial and economic exchanges, respectively. Our results suggest that the widely accepted contention thatbreach has its negative effects on employee attitudes and behaviours through damaged trust in theorganization appears to be more relevant to relational, as opposed to transactional, types of contracts.240

ReferencesAjzen, I., & Fishbein, M. (in press). The influence of attitudes on behaviors. In D. Albarracin, B. T.Johnson, & M. P. Zanna (Eds.), Handbook of Attitudes. Mahwah, NJ: Lawrence ErlbaumAssociates.Arnold, J. (1996). The psychological contract: A concept in need of closer scrutiny? European Journal ofWork and Organizational Psychology, 5, 511-520.Barling, J., Kelloway, E. K., & Bremermann, E. H. (1991). Preemployment predictors of union attitudes:The role of family socialization and work beliefs. Journal of Applied Psychology, 76, 725-731.Blau, P. M. (1964). Exchange and power in social life. New York, NY: Wiley.Buch, K., & Aldridge, J. (1991). O.D. under conditions of organization decline. OrganizationDevelopment Journal, 9, 1-5.Conway, N., & Briner, R. B. (2003). Full-time versus part-time employees: Understanding the linkbetween work status, the psychological contract, and attitudes. Journal of Vocational Behavior,61, 279-301.Coyle-Shapiro, J. A-M., & Kessler, I. (2000). Consequences of the psychological contract for theemployment relationship: A large scale survey. Journal of Management Studies, 37, 903-930.Cronbach, L. J., & Furby, L. (1970). How should we measure “change” - or should we? PsychologicalBulletin, 74, 68-80.Edwards, J. R. (1991). Person-job fit: A conceptual integration, literature review, and methodologicalcritique. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial andorganizational psychology (Vol. 6, pp. 285-357). New York: Wiley.Edwards, J. R. (1994). The study of congruence in organizational behavior research: Critique and aproposed alternative. Organizational Behavior and Human Decision Processes, 58, 51-101.Edwards, J. R. (1996). An examination of competing versions of the person-environment fit approach tostress. Academy of Management Journal, 3, 292-339.Edwards, J. R. (2001). Alternatives to difference scores: Polynomial regression analysis and responsesurface methodology. In F. Drasgow & N. W. Schmitt (Eds.), Advances in measurement and dataanalysis (pp. 350-400). San Francisco: Jossey-Bass.Edwards, J. R., & Cooper, C. L. (1990). The person-environment fit approach to stress: Recurringproblems and some suggested solutions. Journal of Organizational Behavior, 11, 293-307.Irving , P. G., & Meyer, J. P. (1994). Reexamination of the met-expectations hypothesis: A longitudinalanalysis. Journal of Applied Psychology, 79, 937-949.Irving , P. G., & Meyer, J. P. (1995). On using direct measures of met expectations: A methodologicalnote. Journal of Management, 21, 1159-1175.Johns, G. (1981). Difference score measures of organizational variables: A critique. OrganizationalBehavior and Human Performance, 27, 443-463.Lambert, L. S., Edwards, J. R., & Cable, D. M. (2003). Breach and fulfilment of the psychologicalcontract: A comparison of traditional and expanded views. Personnel Psychology, 56, 895-934.MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison ofmethods to test mediation and other intervening variable effects. Psychological Methods, 7, 83-104.McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detecting interactions and moderatoreffects. Psychological Bulletin, 114, 376-390.McLean Parks, J., Kidder, D. L., & Gallagher, D. G. (1998). Fitting square pegs into round holes:Mapping the domain of contingent work arrangements onto the psychological contract. Journal ofOrganizational Behavior, 19, 697-730.Morrison, E. W., & Robinson, S. L. (1997). When employees feel betrayed: A model of howpsychological contract violation develops. Academy of Management Review, 22, 226-256.Olson, J. M., Roese, N. J., & Zanna, M. P. (1996). Expectancies. In E. T. Higgins & A. W. Kruglanski(Eds.), Social psychology: Handbook of basic principles (pp. 211-238). New York: Guilford.241

Pedhazur, E. J. (1997). Multiple regression in behavioral research: Explanation and prediction (3 rd ed.).Fort Worth, TX: Harcourt Brace.Raja, U., Johns, G. & Ntalianis, F. (2004). The impact of personality on psychological contracts.Academy of Management Journal, 47, 350-367.Rice, R. W., Phillips, S. M., & McFarlin, D. B. (1990). Multiple discrepancies and pay satisfaction.Journal of Applied Psychology, 75, 386-393.Robinson, S. L. (1996). Trust and breach of the psychological contract. Administrative Science Quarterly,41, 574-599.Robinson, S. L., Kraatz, M. S., & Rousseau, D. M. (1994). Changing obligations and the psychologicalcontract: A longitudinal study. Academy of Management Journal, 37, 137-152.Robinson, S. L., & Morrison, E. W. (1995). Psychological contracts and OCB: The effect of unfulfilledobligations on civic virtue behavior. Journal of Organizational Behavior, 16, 289-298.Robinson, S. L., & Morrison, E. W. (2000). The development of psychological contract breach andviolation: A longitudinal study. Journal of Organizational Behavior, 21, 525-546.Robinson, S. L., & Rousseau, D. M. (1994). Violating the psychological contract: Not the exception butthe norm. Journal of Organizational Behavior, 15, 245-259.Rousseau, D. M. (1989). Psychological and implied contracts in organizations. Employee Responsibilitiesand Rights Journal, 2, 121-139.Rousseau, D. M. (1990). New hire perceptions of their own and their employer’s obligations: A study ofpsychological contracts. Journal of Organizational Behavior, 11, 389-400.Rousseau, D. M. (1995). Psychological contracts in organizations: Understanding written and unwrittenagreements. Thousand Oaks, CA: Sage.Rousseau, D. M., & McLean-Parks, J. (1993). The contracts of individuals and organizations. Research inOrganizational Behavior, 15, 1-43.Rousseau, D. M., & Tijoriwala, S. A. (1998). Assessing psychological contracts: Issues, alternatives andmeasures. Journal of Organizational Behavior, 19, 679-695.Sheeran, P. (2002). Intention-behavior relations: A conceptual and empirical review. In W. Stroebe & M.Hewstone (Eds.), European review of social psychology, Volume 12: 1-36. Chichester, England:Wiley.Turnley, W. H., Bolino, M. C., Lester, S. W., & Bloodgood, J. M. (2003). The impact of psychologicalcontract fulfillment on the performance of in-role and organizational citizenship behaviors.Journal of Management, 29, 187-206.Turnley, W. H., & Feldman, D. C. (1998). Psychological contract violations during corporaterestructuring. Human Resource Management, 37, 71-83.Turnley, W. H., & Feldman, D. C. (1999). The impact of psychological contract violations on exit, voice,loyalty, and neglect. Human Relations, 52, 895-922.Turnley, W. H., & Feldman, D. C. (2000). Re-examining the effects of psychological contract violation:Unmet expectations and job dissatisfaction as mediators. Journal of Organizational Behavior, 21,25-42.242

ASAC 2006Banff, AlbertaPaul VarellaUniversity of CalgaryBRIDGING AND BONDING SOCIAL CAPITAL IN ORGANIZATIONAL GROUPS:THE DEVELOPMENT OF ASSESSMENT TOOLSThis paper examines the phenomena of social capital in formal, functional groups insideorganizations. It offers a consolidation of the different views of social capital that pertainsto that specific level of analysis. Building on such an understanding I propose approachesto measure the group bridging and bonding social capital. In view of the social networkperspective of the phenomena, I suggest a list of social positions for the positiongeneratormethodology and a set of investigation approaches for name-generator lists,which could be used for assessing the structural dimensions of group social capital.Finally, I developed an instrument aimed at assessing the cognitive dimensions of groupsocial capital. The process includes item pool generation, pre-testing with 181 subjects,exploratory factor analysis and instrument refining. Following, I implemented a datacollection with 347 members of organizational teams, refined the instrument viaexploratory and confirmatory factor analysis, and computed the reliability measures ofthe sub-constructs that comprise the cognitive dimensions of group social capital.Bridging and Bonding Social Capital in Organizational Groups:The Development of Assessment ToolsScholars have examined the positive impact of social capital in many societal activities, such asthe development of human capital (Coleman, 1988), the economic growth of countries (Fukuyama, 1995;Haan & Siermann, 1996), and the development of democracies and good governance (Evans, 1996;Newton, 1997; Putnam, 2000). At the organizational level, the literature suggests that social capitalinfluences the development of organizational intellectual capital (Nahapiet & Ghoshal, 1998), enhancesthe chances of firm survival (Pennings, Lee, & Witteloosttuijn, 1998), fosters organizational commitment,(Watson & Papamarcos, 2002), benefits individual careers (Burt, 1992), facilitates better access tofinancial capital (Florin, Lubatkin, & Schulze, 2003), and represents a source of competitive advantage(Kostova & Roth, 2003).Many scholars still regard social capital elusive and difficult to measure (Pennings & Lee, 1999;Portes, 1998; Woolcock, 1998). The goal of this paper is to propose a measurement approach forassessing the social capital of formal, organizational groups. We introduce a summary of the extentliterature and progress into some suggestions of assessments tools, as we develop an instrument formeasuring the cognitive social capital of these groups.The Theory of Social CapitalThe Development of Social Capital within Organizational GroupsSocial capital is a sociological concept defined in different and sometimes contradictory terms.Despite divergent viewpoints, there seems to be a general consensus for the basic notion of social capitalas the “ability of actors to secure benefits by virtue of membership in social networks or other socialstructures” (Portes, 1998, p.6). The general notion of capital, when applied in the social context, invokesthe idea that social units possess supportive goods, assets, or processes that facilitate the productive243

function. The idea is that there are specific social exchanges amongst social actors that promote collectiveeconomic performance. As Portes (1998, p. 7) describes it: “Whereas economic capital is in people’s bankaccounts and human capital is inside their heads, social capital inheres in the structure of the relationships.To possess social capital, a person must be related to others, and it is those others, not himself, who arethe actual source of his or her advantage”.It is generally accepted that social capital may be represented by two dimensions of social units:(1) cognitive and (2) structural (Adler & Kwon, 2002; Koka & Prescott, 2002; Nahapiet & Ghoshal,1998). The former is related to the social psychological aspects of the relationships among groupmembers, as illustrated by group norms, values, identification, and trust (Portes, 1998). The latter isassociated with the structure of the social networks, how social ties are distributed in space and time, thenature of these connections, and the assets that such connections are able to link (Burt, 1992; Lin, Cook,& Burt, 2001).Social capital results from social exchanges that evolve through two different conditions, namelystrong social ties and weak social ties (Fukuyama, 1995, 2000). The strong-tie milieu evolves in closerelationships, where connections are internal to social groups, and interaction is dense and constant(Coleman, 1988). Group members develop redundant and closed networks (e.g., high density of socialconnection), given the fact that members of a strong-tie community have to invest much time to developstrong, close social connections. Consequently, such social groups offer a wider array of opportunities forall members to know and support one other (Coleman, 1988; Granovetter, 1973). It is through strongsocial ties that communities develop strong social norms, pressures for conformity, and most importantly,trust and solidarity along with reciprocal identification and support of the membership (Portes &Sensenbrenner, 1993).The second mode, through weak ties, evolves within social contexts where connections are weakand the interaction is distant and sporadic. Reduced investment of time for its development allows peopleto develop a larger number of weak ties (Granovetter, 1973). Connections are not redundant and offer anextended reach to information. The restriction, however, is that such conduits are not able to transfercomplex knowledge (Hansen, 1999; Kostova & Roth, 2003). In such conditions, people profit from weakties as they are conducive to diversity and creativity, and they do not offer strong pressures for conformityor sanctions from other members of the social structure (Granovetter, 1973).The exploration of these two modes of social exchange, strong and weak ties, has led someauthors to adopt the notion of bonding social capital to describe the type that derives from strong ties, andbridging social capital as the alternative form, deriving from weak ties (Putnam, 2000). Bonding socialcapital enhances intragroup collaboration, behavioural control, trust, and high quality of relationshipsamong members of a close-knit community. On the other hand, bridging social capital enables groupmembers to span connections to other individuals, groups, and entities. Putnam (2000, p. 23) illustratesthe model above by stating: “Bonding social capital constitutes a kind of sociological superglue, whereasbridging social capital provides sociological WD-40. Bonding social capital, by creating strong in-grouployalty, may also create strong out-group antagonism.”Consequences of Social Capital on Organizational GroupsThe understanding of the positive influences of social capital in the organizational context can beenhanced if we use it as a metaphor for group advantage (Burt, 2001b). Initially, there are advantages thatresult from the relationships, synergies, and collaboration among members of a group, as they makeavailable to their peers the resources that they control (Bourdieu, 1986). Subsequently, social capitalreduces transactional costs because it helps to control behaviours that could be harmful to the group(Portes & Sensenbrenner, 1993; Putnam, 2000). In other words, social capital reduces transactional costs244

due to enforceable trust within the membership (Coleman, 1988; Portes, 1998). Finally, social capitalfacilitates the development of intellectual capital inside groups or organizations (Nahapiet & Ghoshal,1998), as it reduces the barriers of information flow (Hansen, 1999; Szulanski, 1996). However, there arealso risks associated with the development of social capital.There is much debate about the negative consequences that social capital can cause in groups,since it is possible that the processes that generate social capital may also promote outcomes that hinderoptimal socio-economic processes (Adler & Kwon, 2002; Portes, 1998; Portes & Landolt, 1996).Negative consequences come from the fact that high social capital groups have objectives that relateprimarily to their membership (Coleman, 1988). Even though individuals engage in behaviours thatgenerate high social capital and, consequently, facilitate collective action, there might be group-specificbehaviours that are accepted among members but cause negative externalities to the broader organization(Gabbay & Leenders, 1999; Portes, 2000). That is, the group could profit, but the overall organizationmight suffer.The negative influence of social capital development in groups can be reinforced by thestrengthening of bounded solidarity and the consequent constraining of the membership’s diversity ofthought. Questioning the group norms and values can be unpleasant and unacceptable. The discomfort invoicing differing opinions happens because members find protection in their group, and they may facesanctions or censure if they decide to behave differently from the group expectations (Pennings & Lee,1999; Portes, 1998). Groupthink, enforced compliance, lack of creativity, and inability to access neededexternal resources are the resulting unintended, but serious, consequences which can impact the survivalof an organization. Adler and Kwon (2002) describe how those consequences generate organizations thatresemble a collection of compartmentalized, inward groups.Consequently, the relationship between bonding social capital and group performance may takethe shape of an inverted-U curve (Frank & Yasumoto, 1998; Fukuyama, 2000; Gabbay & Leenders, 1999;Leenders & Gabbay, 1999; Pennings et al., 1998; Uzzi & Gillespie, 1999). In other words, groups mayhave not enough or too much bonding social capital, and both conditions are limiting. The formergenerates disconnected groups, with limited synergies and loss of opportunities for effective collaborativeaction (Adler & Kwon, 2002). The latter causes too much in-group pressure, conformity and isolation,thus resisting change and diversity. Another conclusion is that in addition to the development of bondingsocial capital, groups may need to develop bridging social capital to successfully develop connections toother groups within the organization (Burt, 2001a). This is a condition that can be more difficult to attainif the bonding social capital is too strong (Portes, 1998; Portes & Sensenbrenner, 1993).The conclusion is that social capital is an umbrella concept (Adler & Kwon, 2002). It has beendefined and examined at many different levels of analysis, e.g. nations, communities, organizations,groups, and individuals. It also includes a myriad of dimensions and specifications. The operationalizationof such a high-level, latent construct demands that empirical research makes use of precise definitions ofthe sub-constructs, and the utilization of clearly crafted instruments (Bagozzi, 1984).Social network methodology has offered solid and extensive examinations into the social capitalof organizations (Brass, Galaskiewicz, Greve, & Tsai, 2004). Such a methodology, however, does notexplore to the full extent of the cognitive aspects of the phenomenon. I consider now in more detail theissues of construct definition and assessment of the social capital of groups specifically in formal,functional organizational groups.245

Defining the Specific Constructs and Devising Measuring Approaches of Group Social CapitalI define social capital at the group level as the collection of assets, values, behaviours andprocesses, which members of an organizational group jointly exchange as they foster collaboration andsupport among themselves towards the achievement of their collective socio-economic goals. Socialcapital is an intangible attribute emerging from the relationships among members, which helps them topursue collective goals. Group members can access such capital because of their membership, but they donot own it individually.As discussed before, the development of group social capital will develop through the dualitiesof, strong and weak ties, and bonding and bridging social capital. Its assessment is possible throughcognitive and structural dimensions.Bridging Social Capital of GroupsGroups need to develop external reach to ensure the diversity of ideas and availability ofresources required for their long-term survival (Adler & Kwon, 2002). Burt (1992) has convincinglydemonstrated that external social connection is a major source of added value for organizational groups.As he puts it, group cohesion might be necessary for realization of value, but external connections arenecessary for adding value to such units. Consequently, despite the enhanced internal dynamics thatgroups enjoy from the cognitive dimensions of social capital, they cannot isolate themselves from theirbroader environment. The network positions that offer valuable, complementary resources and diversityare positions with a high degree of centrality and betweenness (Burt, 1992; Freeman, 1979). Betweennessis a form of network centrality, representing the socio-metric concept of being at the centre of the socialconnections within otherwise disconnected clusters in the network (Scott, 2000); it is the degree to whicha node connects other clusters in the organization.Specifically for bridging social capital (Lin, Fu, & Hsung, 2001) offers an extremely validapproach for assessing the social capital in the social structure of relationships, which seems to be quiteappropriate for assessing the social capital of organizational groups. He proposes the method of “positiongenerator” as an alternative to the usual “name generator” lists. He proposes that subjects would inform ifthey have strong or weak ties to a list of functions across the network, which could be potential sources ofresources. Accordingly, he recommends questioning informants if they have acquaintances (weak ties) orfriends, (strong ties) with a list of general members of the society; for instance, lawyers, politicians,trades, and etc.Different from name generator lists, subjects only inform the nature of the connections to otherfunctions in the society and not necessarily the people’s name in such positions. The assumption is thatthe exact identification of subjects is secondary. The idea is that specific members of a social unit occupypositions with social equivalence and are similar sources of resources, analogous to other members incorresponding positions across the network. He proposes that such a methodology has also someadvantages over name generator lists because subjects would be helped in remembering weak ties. Theposition generator would have the advantage to refresh informants’ memory about their weak ties in thesocial network.Measuring the Social Capital of Groups: Structural DimensionsThe work of Lin and colleagues (2001) does not suggest a valid list of social network positionsthat could be used when considering formal organizational groups. I implemented a simple qualitativeassessment of which social position could be sources of resources for organizational members. A Sampleof 45 working Commerce students were invited to offer a list of possible sources of resources for246

members of a group. Ten students volunteered to participate in the exercise. The compilation of the tenlists is consolidated below, based on the following guideline.When considering people outside the group of immediate co-workers, there are some peoplewhom group members can count on, if they need some help with their work. Below there is a list ofprofessionals who could be helpful for the group members’ work activities. Informants should state foreach professional category if the person they know is a friend, somebody that is close to you and who youknow really well (strong ties); or if she or he is only an acquaintance, somebody that they know onlysuperficially and that is not too close to them (weak ties); or if they know nobody in that category.1. Chief Executive Officer (CEO) or topexecutives of the organization.2. Owners of the company or members of theboard of directors.3. Personnel on the same organizational levelas you are, but in other areas/departments.4. Personnel, on the same level as your boss is,but not actually her/himself.5. Critical government officials (regulators,inspectors, agents, etc).6. Key people working for suppliers or serviceproviders for your department.7. Key people working for customers of yourdepartment.8. Supervisors and other front line employees,in your organization, outside your own areaof responsibility.9. Members of professional associations.10. Union representatives.11. Non Governmental Organizations (NGOs)from stakeholders with interests in youractivities.12. Members of the judiciary institutions (judge,lawyer, prosecutor, law enforcement, etc), incommunities related to your work activities.13. Professor or scientist at a university.14. Consultants.15. Politicians at the municipal level.16. Politicians at the state/province level.17. Politicians at the federal level.18. Media person (journalist, reporter, publicrelations, etc).19. Auditors.247

The list is a simple compilation for potential social positions within organizations that could beused in the position generator technique suggested by Lin et all (2001). The aggregate of each member’slink to those positions in the network would represent a valid metrics for the group’s bridging socialcapital. Not only bridging social capital benefits from the social network assessments. Social structuresthat are internal to formal, functional groups can also make use of existing social network methodologies,as follows.Bonding Social Capital of GroupsNature of internal social connections. Members of groups with high levels of social capital willenjoy higher physical and psychological proximity and will display a higher density of internal socialconnections. Physical proximity is one of the key characteristics of the social connections inside thegroups because it is largely through social interaction that social capital develops (Burt, 1992, 1997;Coleman, 1988; Lin, 2001). Membership interaction offers the opportunity for members to shareperspectives, exchange information, and collaborate; thus, developing the social ties that foster socialcapital (Putnam, 2000). In the organizational context, functional groups with higher levels of socialcapital will have more frequent personal interactions. Members will enjoy teamwork and portray a higherdensity of exchanges with colleagues. Meetings will be held more regularly, and joint actions are morefrequent.Second, there will be a higher frequency of social and informal encounters outside the workenvironment. Closed networks will enhance the generation of social capital because more frequentencounters make social capital more encompassing for a group (Coleman, 1988). In this way, membershave reinforcing connections (i.e., they have a chance to interact with group members in more than onesetting or fashion, which also makes the connections stronger). Increased social intercourse is particularlyimportant for social capital in organizational groups because in our broader society the scope and depth ofsocial contact has been reduced by the day-by-day pressures of modern life (Paul, Costley, Howell, &Dorfman, 2002; Putnam, 1995, 2000).Psychological proximity represents social connections of a stronger form, where not only face-tofaceinteractions exist, but they also reflect a richer and more supportive relationship. While physicalproximity and contact is a necessary condition for the development of social capital (Lin, 2001), it is notsufficient. Physical proximity in the organizational environment is less voluntaristic than it is in thebroader society because people may share the work environment without ever developing any personalattachment to their coworkers. Thus, members of a group with high levels of social capital will displaybehaviours that portray a shared appreciation of psychological proximity. They show concern for, andlook after, each other. Portes (1998) defined this phenomenon as bounded solidarity, describing howmembers of a group with high social capital are willing to help each other because of a process of sharedidentification and common objectives (van Knippenberg & Hogg, 2003).We can expect two sets of behaviour in such groups. First, as noted above, members in a highsocial capital group are more inclined to offer help to other group members. They offer material andpsychological support to help achieve collective objectives. Second, members of the group appreciate andtake advantage of the fact that they can count on others. Hence, high levels of social capital are associatedwith the propensity to reach out for help and support. In sum, besides the frequency of contact, high socialcapital groups enjoy deep and high quality interpersonal relationships.Physical proximity and psychological proximity could underlie the questions for assessing thestructure of a group’s internal social connections, via ego-centric name generator lists. I propose thatinstruments could consider the following questions.243

For physical proximity: enumerate members of your group of co-workers with whom you shareface-to-face interactions during your work activities. The face-to-face interaction means that a specific coworkerand you tend to be present at same meetings, to work side by side on some projects, to visit eachother’s work area, and may use some of your personal time to socialize together.For psychological proximity: enumerate members of your group of co-workers with whom youoffer support to each other in your work responsibilities. In other words, the specific co-worker and youclosely support each other to accomplish your goals in the organization. Both of you seek each other’sadvice and enjoy working collaboratively on projects, or helping each other with your individualresponsibilities.Building on these two assessments it is possible to make measurements of network density andcentrality measurements, among others. Based on that, there are opportunities of exploring furtherinsights about the social capital of organizational groups.Bonding Social Capital of Groups: Cognitive DimensionsBonding social capital evolves through the cognitive aspects of the group. There are specificattributes of groups that are associated with high levels of social capital as such groups exhibit strong andspecific norms and values. They achieve what Portes (1998) called enforceable trust, and what Coleman(1988) referred to as obligations and sanctions within an appropriable organization. These norms createan ambience of trustworthiness and empowerment, where members not only trust each other but also feelempowered to address non-conforming behaviours. The following is an explanation of the four supportivenorms of social capital.Reciprocity and willingness to delay repayments. Members are willing to engage and maintaintheir reciprocity exchanges through an informal and unwritten chit system whereby favours, information,or any other form of assistance is backed by a norm of reciprocity (Portes, 1998). Additionally, membershave little concern about contributing without immediate repayment because they believe other memberswill offer reciprocal help when they, in turn, need support from the group (Portes & Sensenbrenner,1993). There is an expectation that the aggregate value of the contributions of all members will be largerthan the sum of the individual inputs, and this is largely based on the shared trust in such groups(Granovetter, 1985).Remedies for non-conforming members. Free riding is a potential problem for groups with highlevels of reciprocity due to the unspecified time for repayment of the collaborative actions from members,and the sense of protection that social capital offers to them. Hence, those who benefit from currentofferings of the group could refuse to contribute back when called upon for their support. To prevent thisoutcome, along with expectations of reciprocity, groups develop norms to enforce their rules and topenalize free riders (Coleman, 1988; Portes, 1998; Putnam, 2000; Zhou & Bankston, 1996). Trust is thecommon underlying factor in group dynamics and is sustained through the existence of mechanisms topunish dysfunctional behaviour. Ostracism, alienation, and loss of reputation are possible examples ofsuch mechanisms. Individual members are motivated to abide by the group's rules due to their desire tosustain the benefits of good standing within the group, and their desire to avoid being ostracized andblacklisted.Conformity to group pressures. Norms not only work as controls for deviant behaviours but alsolimit individual expression and freedom (Portes & Sensenbrenner, 1993; Zhou & Bankston, 1996). Thereis an expectation that members willingly conform to the group expectations and suppress actions thatchallenge or contradict group norms and values. Such restrictions may even limit the abilities of members244

to build strong ties with those outside the group boundaries because such ties may be in conflict with thegroup’s values, leading to destructive in- and out-group relationships (Putnam, 2000).Meticulous processes for admission and socialization of new members. Embracing newindividuals to a group represents a major challenge to the norms that are agreed upon by all existingmembers. New individuals and ideas may challenge the status quo inside the group (Portes, 1998).Therefore, groups with high social capital tend to devise membership selection and socializationprocesses that ensure acculturation of the new members to group norms. They take hiring decisionsseriously and spend much time, both individually and as a group, to ensure that the new hires arecompatible with the group.An extensive literature search did not offered valid measurements for these specific norms ofsocial capital development inside organizations. Next, I introduce an instrument that constitutes acontribution towards such a goal.Measuring the Social Capital of Groups: Cognitive DimensionsInstrument Design. I implemented an instrument design in order to develop questionnaire itemsfor a Likert scale survey. I followed these steps in developing the instrument. (DeVellis, 1991).(1) Item generation based on the definition of the four constructs – reciprocity and willingnessto delay repayments, remedies for non-conforming members, conformity to grouppressures, and meticulous processes for admission and socialization of new members – andcue-sorting of the pool of generated items.(2) Pre-testing of the items and exploratory factor analysis of the results.(3) Review, editing and arrangement of the items.(4) Data collection with the desired sample(5) Statistical analysis using exploratory factor analysis, confirmatory factor analysis andreliability measurements of the final constructs.Step one, item generation, was implemented with 10 working Commerce students. They offered alarge number of possible items that were consolidated in a list of 41 items, associated with the four subconstructsdefining bonding social capital. A panel of three PhD students in Business Administrationworked independently sorting the 41 items back to the original constructs. According to the agreementbetween the three members, 28 items were included in the instrument, which were subsequently pretested.Table 1 includes the instrument for pre-testing.I applied the instrument to 34 groups in organizations. Thirty four working Commerce studentscollected the data with their co-workers, resulting in a sample of 182 subjects. Subsequently, I conductedan exploratory factor analysis that indicated the following. The optimal solution included only threefactors, which aggregately accounted for 33.96% of the variance in the model. The model with Varimaxrotation showed that the items originally under “meticulous processes for admission and socialization ofnew members” loaded with “conformity to group pressures”, and did not seem to indicate an additionalfactor, as suggested by the literature review. Organizational process, probably, interfere strongly with theprocess of selection, admission, and training/socialization of group members, making such dimension lessimportant for groups in the organizational context, than it would be in societies at large. Table 1 includesthe factor loadings of the three-factor solution.Supported by the exploratory factor analysis and the content of the items in each one of them, Idefined the following factors: Factor 1: Conformity, Factor 2: Reciprocity, and Factor 3: Enforcement.Based on the insights gained with the pre-testing, I produced the final instrument, which included the best245

items from the pre-testing and some added items, aligned with the underlying final three constructs. Table2 includes the final instrument.The questionnaire was applied via an Internet-delivered survey that was sent to all members of71 different teams across 7 organizations. In total 470 organizational group members received an e-mail,inviting their participation in the study. From that, 347 responses were received (73.8% response rate). Iconducted an initial exploratory factor analysis, which results are on Table 2. The analysis indicated thatthe best model includes only two factors, explaining 33.8% of the variance on all items; the three factormodel only enhanced the explained variation to 36.7%. For consistency and parsimonious reasons the twofactor solution seems to be the best approach.Factor one aggregates conformity and reciprocity, it is a factor that evokes the notion ofAllegiance. The data suggests a shared process of assistance and monitoring of the members needs andactions. I believe that it is more than reciprocity and denotes a process of exchanging favours and definingobligations among the membership. The results seem to agree with the suggestion in the literature thatsocial capital engenders collective action among members of a group in a way that the collective overallgroup objectives are shared and monitored among its members.Consistent with the pre-testing, the Enforcement Factor showed once again solid factor loadingsand consistency. This item indicates that group members could engage in punitive actions against nonconformingmembers, as suggested by the literature.Working only with the most consistent questionnaire items, see tables 2 and 3, I conducted aConfirmatory Factor Analysis using AMOS with the best items and the two factors as latent variables:Allegiance and Enforcement. The final items loaded nicely to two factors and the model fit wasreasonable. Although the model did not refute the null hypothesis (Chi-square = 617.963, d.f. = 251,P

offered an instrument for assessing the bonding social capital of organizational groups. The proposedinstrument suggests a model of two latent sub-constructs, Allegiance and Enforcement. While the formerdenotes the way people work together in groups with high social capital, the latter explains how membersdeal with members that do not follow group norms of social behaviour.Although the functional group seems to be an important building block of organizations, thisarticle leaves room for further analysis of informal organizational groups. It is my expectation that theseassessments tools could help future investigations connecting group social capital to other organizationalphenomena. For instance, the connections to group social capital and its internal cohesion, groupperformance, inter-group collaboration and creativity. It would also be interesting to examine howorganizational culture and team leaders influence the development of the group social capital.I believe to have shed some light on the elusive subject of how to measure social capital inorganizations. By integrating different methodology and employing the sociological constructs pertainingto social capital, I hope to offer a fresh new look into this fields of research.247

ReferencesAdler, P., & Kwon, S. (2002). Social capital: Prospectus for a new concept. Academy of ManagementReview, 27, 17-40.Bagozzi, R. (1984). A prospectus for theory construction in marketing. Journal of Marketing, 48, 11-29.Bourdieu, P. (1986). The forms of capital. In J. G. Richardson (Ed.), Handbook of Theory and Researchfor the Sociology of Education (pp. 241-258). New York: Greenwood Press.Brass, D. J., Galaskiewicz, J., Greve, H. R., & Tsai, W. (2004). Taking stock of networks andorganizations: A multilevel perspective. Academy of Management Journal, 47 Issue 6, 795, 723.Burt, R. S. (1992). Structural Holes: The Social Structure of Competition. Cambridge, MA: HarvardUniversity Press.Burt, R. S. (1997). The contingent value of social capital. Administrative Science Quarterly, 42, 339-365.Burt, R. S. (2001a). The social capital of structural holes. In M. F. Guillén, R. Collins, P. England & M.Meyer (Eds.), The New Economic Sociology: Developments in an Emerging Field. New York:Russell Sage Foundation.Burt, R. S. (2001b). Structural holes versus network closure as social capital. In N. Lin, K. Cook & R.Burt (Eds.), Social Capital Theory and Research (pp. 31-56). New York: Aldine de Gruyter Inc.Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94(Summer), 95-120.DeVellis, R. F. (1991). Scale development: Theory and applications (Vol. 26). Newbury Park, CA: SagePublications Inc.Evans, P. (1996). Government action, social capital and development: Reviewing the evidence onsynergy. World Development, 24, 1119-1132.Florin, J., Lubatkin, M., & Schulze, W. (2003). A social capital model of high-growth ventures. Academyof Management Journal, 46, 374-384.Frank, K. A., & Yasumoto, J. Y. (1998). Linking action to social structure: Social capital within andbetween subgroups. American Journal of Sociology, 104, 642-686.Freeman, L. C. (1979). Centrality in social networks: I. conceptual clarification. Social Networks, 1.Fukuyama, F. (1995). Trust: the Social Values and Creation of Prosperity. New York: Free Press.Fukuyama, F. (2000). Social capital and civil society. IMF Working Paper, WP/00/74, 1-18.Gabbay, S. M., & Leenders, R. T. (1999). Corporate social capital: The structure of advantage anddisadvantage. In R. T. Leenders & S. M. Gabbay (Eds.), Corporate Social Capital and Liability(pp. 1-16). Norwell, MA: Kluwer Academic Publishers.Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360-1380.Granovetter, M. S. (1985). Social structures and economic action: The problem of embeddedness.American Journal of Sociology, 91, 481-510.Haan, J., & Siermann, C. (1996). Political instability, freedom, and economic growth: Some furtherevidence. Economic Development and Cultural Change, 44, 339-350.Hansen, M. T. (1999). The search-transfer problem: The role of weak ties in sharing knowledge acrossorganizations sub-units. Administrative Science Quarterly, 44, 82-85.JAI Press.Kline, R. B. (1998). Principles and practice of structural equation modeling. New York: Guilford.Koka, B. R., & Prescott, J. E. (2002). Strategic alliances as social capital: A multidimensional view.Strategic Management Journal, 23, 795-816.Kostova, T., & Roth, K. (2003). Social capital in multinational corporations and a micro-macro model ofits formation. Academy of Management Review, 28, 297-317.Leenders, R. T. A. J., & Gabbay, S. M. (1999). Corporate Social Capital and Liability. Boston: KluwerAcademic.Lin, N. (2001). Building a network theory of social capital. In N. Lin, K. Cook & R. S. Burt (Eds.), SocialCapital: Theory and research (pp. 3-30). Hawthorne, NY: Aldine de Gruyter.248

Lin, N., Cook, K., & Burt, R. S. (2001). Social Capital: Theory and research. Hawthorne, NY: Aldine deGruyter.Lin, N., Fu, Y., & Hsung, H. (2001). The position generator: Measurement techniques for investigationsof social capital. In N. Lin, K. Cook & R. S. Burt (Eds.), Social capital: Theory and research (pp.57-84). Hawthorne, NY: Aldine de Gruyter.Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage.Academy of Management Review, 23, 242-266.Newton, K. (1997). Social capital and democracy. American Behavioral Scientist, 40, 575-586.Pennings, J. M., & Lee, K. (1999). Social capital of organization: Conceptualization, level of analysis,and performance implications. In R. T. A. J. Leenders & S. M. Gabbay (Eds.), Corporate SocialCapital and Liability (pp. 43-67). Boston: Kluwer Academic.Pennings, J. M., Lee, K., & Witteloosttuijn, A. (1998). Human capital, social capital, and firm dissolution.Academy of Management Journal, 41, 425-441.Pennings, J. M., & Lee, K. (1999). Social capital of organization: Conceptualization, level of analysis,and performance implications. In R. T. A. J. Leenders & S. M. Gabbay (Eds.), Corporate SocialCapital and Liability (pp. 43-67). Boston: Kluwer Academic.Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review ofSociology, 24, 1-24.Portes, A. (2000). The two meanings of social capital. Sociological Forum, 15(1), 1-12.Portes, A., & Landolt, P. (1996). The downside of social capital. The American Prospect, 26 (May -June), 18-22.Portes, A., & Sensenbrenner, J. (1993). Embeddedness and Immigration: notes on the social determinantsof economic action. American Journal of Sociology, 98, 1320-1350.Putnam, R. D. (1995). Tuning in, tuning out: The strange disappearance of social capital in America.Political Science and Politics, 28, 664-683.Putnam, R. D. (2000). Bowling Alone: The collapse and revival of American community. New York:Simon & Schuster.Scott, J. (2000). Social network analysis : a handbook (2nd ed.). London ; Thousands Oaks, Calif.: SAGEPublications.Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within thefirm. Strategic Management Journal, 17(Special Issue - Knowledge and the Firm), 27-43.Uzzi, B., & Gillespie, J. J. (1999). Corporate social capital and the cost of financial capital: Anembeddedness approach. In R. T. Leenders & S. Gabbay (Eds.), Corporate Social Capital andLiability (pp. 446-459). Boston: Kluwer Academic.Watson, G. W., & Papamarcos, S. D. (2002). Social capital and organizational commitment. Journal ofBusiness Psychology, 16, 537-552.Woolcock, M. (1998). Social capital and economic development: Towards a theoretical synthesis andpolicy framework. Theory and Society, 27, 151-208.Zhou, M., & Bankston, C. L. (1996). Social capital and the adaptation of the second generation: The caseof Vietnamese youth in New Orleans. In A. Portes (Ed.), The New Second Generation (pp. 197-220). New York: Russell Sage Foundation.249

Members of my Group:Table 1: Initial Instrument (For Pre-Testing) & Exploratory Factor AnalysisOriginal Instrument Factor 1 Factor 2 Factor 3Reciprocity and delayed repayments1. feel comfortable asking for support from one another. 0.456 0.3082. cover for each other in times of need. 0.275 0.327 0.1063. do not feel obligated to provide tit-for-tat, every time we help one another. 0.114 0.3184. exchange favours among ourselves. 0.164 0.435 0.2915. volunteer help to each other even when not asked. 0.429 0.5826. refrain from offering advice to each other. (reversed) 0.102 0.644 0.1427. avoid staying overtime to help each other out. (reversed) 0.283 0.295 -0.278Remedies for non-conforming members1. ostracize non-conforming members of the group. -0.121 -0.139 0.6242. openly challenge members who do not follow the group’s norms. 0.283 0.5993. know that we may suffer consequences if we do not perform according to thegroup’s expectations. 0.474 0.1144. do not invite to social gatherings members who are disloyal to the group. -0.164 0.5705. provide feedback to members who fall short of the group’s expectations. 0.625 0.178 0.2296. refuse to help non-conforming members of the group. -0.422 0.2647. inform each other when a member breaches the group’s norms. 0.224 0.4688. avoid collaborating with members who go against the group’s expectations. -0.267 0.404Conformity to group pressures1. believe that the attainment of the group goals are more important than theattainment of our individual objectives. 0.702 -0.1182. agree collectively upon guidelines and expectations of our behaviours. 0.611 0.295 -0.1003. seek the group’s approval for our individual actions. 0.506 0.1914. work towards the group’s goals, even if it might negatively affect our ownpersonal goals. 0.607 -0.125 -0.1425. fell obligated to conform to group pressures. 0.288 -0.5686. know what the group expects from each one of us. 0.590 0.1597. feel that personal beliefs are more important than the group’s norms(reversed) 0.224 0.154 -0.234Meticulous processes for admission and socialization1. value the ones among ourselves who have been in the group for longer time. 0.503 0.2192. spend considerable amount of time screening new recruits to our group. 0.397 -0.102 0.1593. prefer to hire people with whom we are acquainted, instead of admittingunknown people to the group. 0.252 -0.301 0.2294. celebrate career milestones of one another. 0.593 0.3055. make sure that we closely monitor the progress of new members of thegroup. 0.531 0.226 0.2806. do not offer informal mentoring opportunities to new hires. (reversed) 0.505 -0.131Extraction Method: Principal Component Analysis. Rotation Method: Varimax250

Members of my Group:Table 2: Final Instrument & Exploratory Factor AnalysisFinal Instrument Factor 1 Factor 2Conformity1. spend considerable amount of time screening new recruits to our group. 0.3642. make sure that we closely monitor the progress of new members of the group. 0.4923. celebrate career milestones of one another. 0.561 -0.1614. believe that the attainment of group goals are more important than the attainment ofour individual objectives. 0.562 -0.1775. agree collectively upon guidelines and expectations of our behaviors. 0.610 -0.1866. seek the group's approval for our individual actions. 0.424 0.1557. work towards the group's goals, even if it might negatively affect our own personalgoals. 0.603 -0.2618. know what the group expects from each one of us. 0.635 -0.2589. believe that we may suffer consequences if we do not perform according to thegroup's expectations. 0.289 0.14810. provide feedback to members who fall short of the group's expectations. 0.57811. value independent individual work rather than collective efforts. (reversed) 0.301 -0.179Reciprocity1. feel comfortable asking for support from one another. 0.539 -0.4062. cover for each other in times of need. 0.568 -0.3393. will do favours among ourselves. 0.1574. always find support from the group 0.646 -0.4195. volunteer help to each other, even when not asked. 0.560 -0.3846. willingly stay late to help each other out. 0.605 -0.1927. usually are not willing to help one another. (reversed) 0.466 -0.5458. people always come with excuses not to help each other. (reversed) 0.328 -0.53Enforcement1. ostracize non-conforming members of the group. -0.181 0.7382. openly criticize members who do not follow the group's norms. 0.6083. do not invite to social gatherings members who are disloyal to the group. -0.102 0.5134. make others in the group aware when a member breaches the group's expectations. 0.464 0.2065. avoid dealing with members who go against the group's expectations. -0.278 0.5386. turn a cold shoulder to members that don't do their part to help others. -0.22 0.6377. reprimand members that do not fit-in. 0.5918. avoid helping non conforming members of the group. -0.256 0.5639. members who ignores group norms loose the respect of the group. 0.42710. tolerate members that diverge from the accepted norms of the group. (reversed) 0.378Extraction Method: Principal Component Analysis. Rotation Method: Varimax251

Table 3: Confirmatory Factor AnalysisMembers of my Group:Latent Variable Items Factor LoadingsAllegiance1. spend considerable amount of time screening new recruits toour group. 0.298***2. make sure that we closely monitor the progress of new membersof the group. 0.469***3. celebrate career milestones of one another. 0.593***4. believe that the attainment of group goals are more importantthan the attainment of our individual objectives. 0.626***5. agree collectively upon guidelines and expectations of ourbehaviours. 0.652***6. seek the group's approval for our individual actions. 0.347***7. work towards the group's goals, even if it might negativelyaffect our own personal goals. 0.680***8. know what the group expects from each one of us. 0.705***9. provide feedback to members who fall short of the group'sexpectations. 0.539***10. value independent individual work rather than collective efforts.(reversed) 0.330***11. make others in the group aware when a member breaches thegroup's expectations. 0.361***12. cover for each other in times of need. 0.626***13. always find support from the group 0.774***14. volunteer help to each other, even when not asked. 0.663***15. willingly stay late to help each other out. 0.635***Enforcementostracize non-conforming members of the group. 0.777***openly criticize members who do not follow the group's norms. 0.519***do not invite to social gatherings members who are disloyal to thegroup. 0.542***avoid dealing with members who go against the group'sexpectations. 0.641***turn a cold shoulder to members that don't do their part to helpothers. 0.710***reprimand members that do not fit-in. 0.505***avoid helping non conforming members of the group. 0.629***members who ignores group norms loose the respect of the group. 0.399***tolerate members that diverge from the accepted norms of the group.(reversed) 0.313***Standardized Regression Weights*** p

ASAC 2006Banff, AlbertaMarjorie Armstrong-StassenSeung Hwan Lee (Student)University of WindsorTHE INFLUENCE OF RELATIONAL AGE ON OLDER EMPLOYEES’ PERCEIVEDTREATMENT 1Respondents with similar age (50+) supervisors perceived they were treated more fairlythan respondents with younger supervisors. Respondents in work groups with similar age(50+) members perceived they were treated with more respect than respondents in workgroups with members of varying ages. Respondents in organizations with a primarilyolder workforce perceived older employees in general were treated more fairly thanrespondents in organizations composed of a primarily younger or middle-aged workforce.Organizational and Relational DemographyStewman (1988, p. 198) noted that “Overall, the potential for organizational demography appearsgreat, especially for yielding new insights into organizational behavior.” Organizational demography isthe study of the composition of an organization in terms of its members’ attributes (Pfeffer, 1983;Lawrence, 1997). Pfeffer proposed that organizational demography is an important causal variable thataffects a number of intervening processes that, in turn, influence organizational outcomes such asturnover. According to Pfeffer (1985), an important feature of demographic analysis is the shift awayfrom examining individual demographic attributes to investigating compositional or relational elements ofthe demographic distributions within an organization. Relational demography refers to an individual’sdemographic characteristics relative to a referent other or group (Goldberg, 2005). Relational demographypredicts that individuals who are similar to referent others will experience more positive outcomes thanthose who are dissimilar to referent others. Unlike age which is an individual-level attribute, relative agerepresents an individual-within-the-group-level demographic unit because it focuses on attributes asaggregate-level rather than individual-level variables (Lawrence, 1997). Cleveland and Shore (1992)found that perceived relative age had the greatest impact for older employees. These researchers arguedthat perceived relative age measures need to be included in studies of older workers.This paper focuses on the relational age of the supervisor-subordinate dyad, work groupmembers, and the age composition of the organization’s workforce. The purpose of the study was toexamine the relationship between older employees’ perceived age similarity/dissimilarity to theirimmediate supervisor, the members of their work group, and their organization’s workforce and theirperceived treatment of older employees in general as well as how they themselves perceived they weretreated. We expected that the greater the perceived age dissimilarity between older employees and theirsupervisor, work group members and the organization’s workforce, the less likely they would perceiveolder employees as well as themselves to be treated fairly.Researchers have tended to focus on the age disparity between individuals and their immediatesupervisor. There has been less research conducted on the age disparity between individuals and themembers of their work group and between the individual and the organization’s workforce. Moreover,very few studies have examined the effects of perceived age similarity/dissimilarity for all three referentgroups—immediate supervisor, work group, and the organization—in the same study. Most studies have1This research project was supported by a grant from the Social Sciences and Humanities Research Council ofCanada.253

included only one of the referent groups. The present study also differs from previous studies because theparticipants are all aged 50 and older. With the aging of the workforce, a major challenge facingorganizations is the retention of their older employees. Therefore, the second purpose of the present studywas to determine the extent to which the relational age and perceived treatment variables influence theintention of older employees to remain in their organization.Perceived dissimilarity between a subordinate and his or her immediate supervisor has beenshown to negatively influence supervisors’ ratings of subordinates’ promotability (Barnes-Farrell, 1993;Shore, Cleveland & Goldberg, 2003), supervisors’ ratings of subordinates’ job performance (Shore et al.,2003; Tusui, Porter & Egan, 2002), the supervisor’s affect toward subordinates (Judge & Ferris, 1993;Tsui & O’Reilly, 1989), supervisors’ support and mentorship (Maurer, Weiss & Barbeite, 2003), trainingand development opportunities (Shore et al., 2003), job satisfaction and organizational commitment(Shore et al., 2003; Turban & Jones, 1988). Demographic dissimilarity among work group members hasbeen found to be associated with less work group cohesiveness (Riordan, 1997), less cooperative behavioramong group members (Chatman & Spataro, 2005), less communication between work group members(Zenger & Lawrence, 1989), and increased likelihood of leaving the work group and the organization(Jackson, Brett, Sessa, Cooper, Julin & Peyronnin, 1991; O’Reilly, Caldwell & Barnett, 1989; Tsui, Egan& O’Reilly, 1992; Wagner, Pfeffer & O’Reilly, 1984). Researchers have found that being older relativeto other organizational members is associated with greater turnover (McCain, O’Reilly & Pfeffer, 1983;Riordan, 1997), more absences (Riordan, 1997), less organizational commitment (Riordan, 1997), andgreater career plateauing and skills obsolescence (Rosen & Jerdee, 1990).Most of the research to date has examined the direct effects of demographic similarity/dissimilarity on outcome variables. There has been little consideration given to the mechanisms throughwhich demographic attributes predict outcomes (Lawrence, 1997). Lawrence proposed that demographicvariables act as organizational signals that acquire meaning. People use their organization’s agedistribution as a signal of how the organization values older employees. We suggest that the agesimilarity/dissimilarity of the supervisor-subordinate dyad, the work group, and the organization’sworkforce influences the perceived treatment of older employees in general as well as the perceivedtreatment of the individual. In other words, it is perceived treatment that mediates the relationshipbetween age dissimilarity and outcome variables such as turnover. Older employees who perceivethemselves to be older than their immediate supervisor, members of their work group, and organizationalmembers are more likely to perceive that not only are they themselves treated less fairly but also toperceive that the organization’s older employees as a group are treated less fairly than otherorganizational members. In a national survey of over 1200 adults in the U.S., Reynolds, Ridley and VanHorn (2005) reported that respondents believed that older workers in general are treated unfairly relativeto younger employees, and specifically when employers are making hiring, salary increase, promotion,and layoff decisions. Reynolds et al. concluded that “At a time when the number of workers 55 years andolder is growing rapidly, this survey finds widespread skepticism about the treatment of older employeesin the workplace.” (p. 27).The predictions between perceived age similarity/dissimilarity and perceived treatment are basedon the similarity-attraction theory, social identity theory, and the relative in-group worth hypothesis. Thesimilarity-attraction paradigm (Byrne, 1971) proposes that individuals who are similar are more attractedto each other and this greater attraction leads to a more favourable assessment of the referent other. Socialidentity theory (Tajfel & Turner, 1986; Turner, 1987) posits that individuals determine their socialidentity by categorizing themselves and others based on surface-level or visible demographiccharacteristics such as age, gender, and race. People categorize demographically dissimilar individuals asout-group members and demographically similar individuals as in-group members (Chatman & Spataro,2005). How individuals interact with each other, for example in terms of cooperation and communication,depends on these categorizations. Folger (1994) regarded justice as worth and suggested that both254

outcomes (distributive justice) and processes (procedural justice) convey messages about the extent towhich the employer, group members, and immediate supervisor regard the employee’s worth. Bruins,Platow and Ng (1995) proposed the relative in-group worth hypothesis which integrates social identitytheory with justice theory. According to Bruins et al., personal identity refers to one’s position withinone’s group in terms of how one is valued by one’s leader and by the group as a whole. These researcherssuggested that unfair treatment in the form of resource allocation reflects the perceived worth or value ofthe in-group relative to the out-group. We propose that when age dissimilarity exists between olderemployees and their immediate supervisor, within their work group, or within the organization, thefollowing will take place: (1) older employees will view themselves as different from those younger, i.e.,see themselves as members of the out-group and those younger as members of the in-group, (2) olderemployees will perceive that they are regarded as being of less worth or value to their supervisor, to theirworkgroup, and to the organization and therefore will perceive that they are treated less fairly, and (3)older employees will be less likely to remain with their organization.Hypothesis 1: Older employees with a younger supervisor will perceive significantly less fair treatmentthan older employees whose supervisor is the same age as they are.Hypothesis 2: Older employees with younger work group members will perceive significantly less fairtreatment than older employees in work groups with members the same age as they are.Hypothesis 3: Older employees in organizations with a primarily younger workforce will perceivesignificantly less fair treatment than older employees in organizations with a primarily olderworkforce.Hypothesis 4: Perceived fair treatment will be significantly positively related to older employees’intention to remain in the organization.Participants and ProcedureMethodQuestionnaire packets were mailed to 1500 people randomly selected from the membership list ofCARP, a not-for-profit national organization that represents the interests of people aged 50 and over.Each questionnaire packet contained two questionnaire booklets—one for people who were retired andone for people who remained in the workforce. The overall response rate was 45%. This paper focuses onthose respondents who had not yet retired.There were 291 participants who were actively engaged in the workforce. They ranged in agefrom 50 to 65 years with an average age of 58.19 years (SD = 3.59). They were employed in a wide rangeof organizations in both the private sector (manufacturing, high tech, finance, insurance, and retail) andthe public sector (public administration, health care, social services, and education). They had beenemployed in their current organization an average of 15.38 years (SD = 11.69) and in their present job anaverage of 9.96 years (SD = 8.99). Just under two-thirds (63%) were in management and professionalpositions, 21% were in support staff and administrative positions, 6.5% were in sales, 5.5% were inskilled trades, and 4% were in other types of job positions. There were 153 men and 137 women (1missing value) and the majority (78%) was married.Measures255

Perceived treatment of older employees in general. We assessed respondents’ perceptions ofhow well their organization was doing in retaining its older workers with a single item: “In general, howwould you rate how well your organization is doing in retaining employees 50 and over?” The responsecategories ranged from 1 (very poorly) to 5 (very well). We used a 3-item measure to assess the extent towhich respondents perceived their organization was providing older employees with training anddevelopment opportunities. Sample items are “Targeting 50+ employees for training to update current jobskills” and “Providing access to new technology that will assist 50+ employees in performing their job.”The reliability coefficient (Cronbach alpha) was .88. We used a 3-item measure to assess the extent towhich respondents perceived their organization was engaging in job design practices tailored to olderemployees. A sample item is “Providing challenging and meaningful tasks or assignments to 50+employees.” The coefficient alpha was .77. The 4-item measure of retirement practices includedencouraging later rather than early retirement and providing older employees with various options such asphased retirement, trial retirement, and retirement with callback. The coefficient alpha was .81. Theresponse categories for the training and development opportunities, job design practices, and retirementpractices ranged from 1 (not doing this at all) to 5 (highly engaged in doing this).Respondent’s perceived treatment. We assessed perceived lack of job challenge with the 6-itemJob Content Plateauing Scale developed by Milliman (1992) and adapted by Allen, Russell, Poteet andDobbins (1999). Sample items are “My job tasks and activities have become routine for me” and “I amchallenged by my job” (reverse scored). The response categories ranged from 1 (Strongly disagree) to 5(Strongly agree). A higher score reflects greater job plateauing. The coefficient alpha was .89. We usedthe 3-item Distributive Justice Scale from Kim, Price, Mueller and Watson (1996) to measurerespondents’ perceptions of how fairly they were rewarded. A sample item is “I am rewarded fairly inview of my experience.” The response categories ranged from 1 (strongly disagree) to 5 (strongly agree).The coefficient alpha was .96. We assessed respondents’ perceptions of how fair the procedures regardingdecisions about their job were with six items adapted from the Procedural Justice Scale developed byMoorman (1991) and Moorman, Blakely and Niehoff (1995). A sample item is “When decisions are madeabout my job, I feel that I am shown concern for my rights as an employee.” The response categoriesranged from 1 (strongly disagree) to 5 (strongly agree). The coefficient alpha was .94. Perceivedtreatment received from the members of one’s work group was assessed with seven items adapted fromthe Respect Scale developed by Tyler and Blader (2001). Sample items are “To what extent do the otherpeople that you work with value you as a member of your work group/unit” and “To what extent do theother people that you work with fail to appreciate your contributions to the job” (reverse scored). Theresponse categories ranged from 1 (almost never) to 5 (almost always). The coefficient alpha was .88.Intent to remain. The 3-item measure of intent to remain with the organization included oneitem from London and Howat (1978) “Barring unforeseen circumstances, I would remain in thisorganization indefinitely,” one item modified from the Intent to Remain scale developed by Lyons (1971)“If I were completely free to choose, I would prefer to continue working in this organization,” and oneitem developed for the study “I expect to continue working as long as possible in this organization.” Theresponse categories ranged from 1 (Strongly disagree) to 5 (Strongly agree). The coefficient alpha was.85.Relational age variables. To assess the relative age of their supervisor, respondents were asked“What is the age of your immediate superior in comparison to your age?” The responses were youngerthan you, about the same age as you, and older than you. To assess the relative age of their work group,respondents were asked “How would you describe the age composition of the other employees withwhom you work on a regular basis?” The response categories were primarily younger than you, about thesame age as you, primarily older than you, and composed of people of varying ages (older, younger, thesame age). To assess the relative age of their organization’s workforce, respondents were asked “Howwould you describe the age composition of your organization?” The responses were a primarily younger256

(18-34 years) workforce, a primarily middle-aged (35-49 years) workforce, a primarily older (50 andover) workforce, and a workforce that is evenly distributed across age groups.Demographic variables. The demographic variables included how long the respondents hadbeen a member of their organization (organizational tenure), how long the respondents had been amember of their workgroup (job tenure), job position, age, gender, and martial status.Data AnalysisWe conducted multivariate analyses of covariance (MANCOVA) with the relational demographicvariables as the independent variables, the perceived treatment variables as the dependent variables, andorganizational tenure, job tenure and age as the covariates. We used Tukey post hoc comparison tests toidentify significant differences among the relational age categories. To identify the significant predictorsof intent to remain with the organization, we used hierarchical regression with the relative age variables,which were effects coded for relative age of the workforce and work group, entered in step 1, perceivedtreatment of older employees in general entered in the second step, and perceived respondent treatmententered in the final step.ResultsOnly two respondents reported that their workgroup members were older than they were and just23 respondents indicated that their immediate supervisor was older than they were. Given these smallnumbers, these two categories were omitted from subsequent data analyses. This left three categories(younger, same age, varying ages) for the relative age of workgroup variable and two categories (younger,same age) for the relative age of immediate supervisor variable.There was a significant overall effect of relative age of supervisor for perceived treatment ofolder employees in general (F (4,233) = 3.86, p < .01) and for how respondents themselves perceived thatthey were treated (F (4,235) = 5.87, p < .001). The means, standard deviations and F-values are shown inTable 1. Respondents whose immediate supervisor was the same age as the respondent perceived thattheir organization was doing a significantly better job of retaining its older employees than respondentswhose immediate supervisor was younger than they were. Respondents whose immediate supervisor wasyounger than they were reported significantly more job plateauing, perceived they were treated less fairlywhen decisions were made about their job, and felt less respect from the members of their work groupthan respondents whose immediate supervisor was the same age as the respondent.There was no overall significant effect for the relative age of work group members for perceivedtreatment of older employees in general (F (8,506) = 1.93, p > .05) but there was a significant effect for howrespondents themselves perceived that they were treated (F (8,514) = 2.21, p < .05). The means, standarddeviations and F-values are shown in Table 2. The Tukey post hoc comparisons identified only onesignificant difference. Respondents whose work group members were the same age as the respondentperceived they were treated with significantly greater respect than respondents whose work groupmembers were of varying ages.257

Table 1Means, Standard Deviations and F-values for Relative Age of Immediate SupervisorYounger Same Age F-valuen = 195 n = 67Mean(SD)Mean(SD)Treatment of Older EmployeesRetention of older employees 3.21(1.14)Training opportunities for older employees 2.21(1.06)Job design practices for older employees 2.25(.96)Retirement options for older employees 1.37(.63)Treatment of IndividualJob Content Plateauing 2.50(.83)Perceived distributive justice 3.32(1.07)Perceived procedural justice 3.36(.89)Respect from work group members 3.76(.74)** p < .01 *** p < .0013.79(1.06)2.41(1.23)2.48(.98)1.49(.73)2.12(.75)3.31(1.06)3.83(.89)4.05(.78)15.30***1.603.192.3210.18** .05). The means, standard deviations, andANCOVA F-values for the relative age of the organization’s workforce are presented in Table 3.Respondents who indicated that their organization’s workforce was a primarily older workforce reportedthat their organization was doing a significantly better job of retaining its older employees thanrespondents in organizations with a primarily younger workforce. They also perceived their organizationto be providing significantly more training and development opportunities and retirement options forolder employees than respondents in organizations with a primarily younger or middle-aged workforceand to be significantly more likely to design jobs to accommodate the needs and preferences of olderemployees than respondents in organizations with a primarily middle-aged workforce. Respondents whoindicated that their organization’s workforce was evenly distributed across age groups reported that theirorganization was doing a significantly better job of retaining its older employees than organizations witha primarily younger workforce and to be significantly more likely to provide retirement options thanorganizations with a younger workforce. There was only one significant difference for the individual258

treatment variables. Respondents in organizations with a primarily middle-aged workforce weresignificantly more likely to report being job plateaued than respondents in organizations with a primarilyolder workforce or a workforce that was evenly distributed across age groups.Table 2Means, Standard Deviations and F-values for Relative Age of Work Group MembersYounger Same Age Varying Ages F-valuen = 189 n = 42 n = 57Mean(SD)Mean(SD)Mean(SD)Treatment of Older EmployeesRetention of older employees 3.28(1.12)3.50(1.26)3.65(1.08)2.93Training opportunities for olderemployees2.16(1.08)2.48(1.14)2.40(1.04)2.63Job design practices for olderemployees2.27(.98)2.41(.99)2.28(.94)1.11Retirement options for older employees 1.36(.64)1.44(.66)1.50(.75)2.20Treatment of IndividualJob Content Plateauing 2.48(.87)Perceived distributive justice 3.35(1.08)Perceived procedural justice 3.49(.91)Respect from work group members 3.83(.76)2.21(.68)3.37(1.14)3.40(.87)4.07 a(.68)Note. Means with the same superscripts are significantly different from each other.* p < .052.27(.79)3.21(1.06)3.46(.99)3.68 a(.81)2.17

their job, who felt that their organization was doing a good job of retaining its older employees, andwhose organization was encouraging later, rather than early, retirement and providing phased retirementoptions were significantly more likely to intend to remain with their organization than respondents whoperceived less procedural justice and a poor effort on the part of their organization to retain its olderemployees, and whose organization was not providing retirement options.Table 3Means, Standard Deviations and F-values for Relative Age of Organization’s WorkforceYoungerWorkforceMiddleagedWorkforceOlderWorkforceEvenlyDistributedWorkforcen = 45 n = 146 n = 44 n = 54Mean Mean Mean Mean(SD) (SD) (SD) (SD)F-valueTreatment of Older EmployeesRetention of older employees3.02 ab(1.14)3.26(1.10)3.71 a(1.19)3.79 b(1.07)4.20**Training opportunities for olderemployees2.13 c(1.09)2.13 d(1.06)2.73 cd(1.09)2.31(1.09)2.74*Job design practices for olderemployees2.30(.99)2.11 e(.93)2.64 e(1.03)2.48(.94)3.00*Retirement options for olderemployees1.25 fh(.58)1.28 g(.57)1.70 fg(.83)1.59 h(.71)6.53**Treatment of IndividualJob Content Plateauing 2.48(.79)Perceived distributive justice 3.23(1.15)Perceived procedural justice 3.47(.80)Respect from work groupmembers3.82(.78)2.52 ij(89)3.36(1.02)3.46(.95)3.76(.81)2.16 i(.78)3.08(1.26)3.44(.98)4.03(.71)Note. Means with the same superscripts are significantly different from each other.* p < .05 ** p < .012.16 j(.64)3.50(1.02)3.58(.87)3.93(.66)2.83*1.43

Table 4Hierarchical Regression Resultsβ β βYounger workforce -.19* -.08 -.08Middle-aged workforce -.27** -.16 -.15Mature workforce -.13 -.13 -.07Younger work group .03 .04 .02Middle-aged work group .01 .03 .04Supervisor age .03 -.05 -11R 2 .04F 1.65Retention of older employees .26*** .18*Training for employees .02 .02Job design ..06 -.04Retirement options .17* .15*R 2 .19∆ R 2 .15F 9.87***Job content plateauing -.04Perceived distributive justice .03Perceived procedural justice .25**Respect from work group .06R 2 .26∆ R 2 .07F 4.69**Note. The inclusion of organizational tenure, job tenure and age as control variables did not alter theseresults. For space reasons, the results for the control variables have been omitted. The correlation matrixis available from the first author.* p < .05 ** p < .01 *** p < .001261

DiscussionThis study addressed the research question: What is the effect of relational age on how olderemployees perceive they and others are treated by their immediate supervisor, the members of their workgroup, and their organization and what role do relational age and perceived treatment play in the retentionof older employees in the workplace? The findings indicate that age dissimilarity in the supervisorsubordinatedyad primarily influenced how fairly respondents perceived they themselves were treatedwhereas age dissimilarity with the organization’s workforce mainly had an effect on how fairlyrespondents perceived older employees in general were treated. When respondents had youngersupervisors, they reported their jobs to be significantly less challenging and meaningful (job contentplateauing), perceived they were treated significantly less fairly when decisions were made about their job(procedural justice), and felt that the members of their work group had significantly less respect for thejob they did and the contributions they made than respondents whose immediate supervisor was the sameage as the respondent. Compared with respondents in organizations with a primarily younger or middleagedworkforce, respondents employed in an organization with a primarily older workforce reported thattheir organization was doing a significantly better job of retaining its older employees, was significantlymore likely to be providing training and development opportunities tailored to older employees and to beproviding challenging and meaningful tasks as well as creating new roles for older employees such asmentoring and being put in charge of special projects, and was significantly more likely to be encouragingolder employees to delay their retirement and providing options such as phased retirement as a way ofretaining older employees.Hypothesis 2 received the least support of the four hypotheses. This hypothesis predicted thatrespondents in work groups with primarily younger members would perceive significantly less fairtreatment than respondents in work groups with members the same age as the respondent. There was onlyone significant difference and the difference was between respondents in work groups with members thesame age and respondents in work groups with members of varying ages. Respondents in work groupswith members of the same age as themselves reported that they received significantly greater respect fromtheir work group members than respondents in work groups whose members were of varying ages (older,younger, the same age). Group longevity may provide a possible explanation for the lack of support forhypothesis 2. The average length of time the respondents had been in their current work group was 10years indicating that many of the members of a work group had been together for some time. Harrison,Price and Bell (1998) and Pelled (1996) suggested that when group members have worked together for along time, age dissimilarity becomes less important than attitudinal dissimilarity. Over time, as workgroup members continue to interact with one another and to acquire more information about each other,they have fewer tendencies to categorize other members on the basis of highly visible demographicattributes and instead base their perceptions on the members’ attitudes, beliefs and values.Lawrence (1997) referred to “the black box of organizational demography” (p. 1). Lawrence washighly critical of the existing research on organizational demography because it ignored the interveningprocesses through which demographic predictors influence outcome variables. We proposed perceivedfair treatment as a possible mechanism through which relational age variables affect intention to remain inthe organization. In other words, relational age variables have an indirect effect on intention to remainthrough their relationship with perceived fair treatment. The results of the hierarchical regression provideat least some support for this. The effect of the age composition of the workforce is no longer significantwhen the perceived fair treatment variables are entered indicating that age composition of theorganization’s workforce influences intention to remain indirectly through perceived fair treatment.The one perceived fair treatment variable that showed no significant relationship to any of therelational age variables was perceived distributive justice. The items in the measure of perceiveddistributive justice referred to being rewarded fairly for the amount of effort put in, for the responsibilities262

the respondent had, and for the respondent’s experience. The findings indicate that relational age had animportant effect on the perceived fairness of processes (procedural justice) but not on the perceivedfairness of outcomes (distributive justice). In general, the more tenure employees have in an organization,the higher the rewards such as pay. This is one of the justifications used by organizations undergoingdownsizing to target older employees because having older employees leave is believed to result ingreater cost savings. The respondents in the present study had been employed with their current employeran average of 15 years suggesting that the respondents did perceive that they were paid more thanyounger employees. It is also possible that, given the secrecy that generally surrounds employeecompensation within an organization, respondents may not realize they are being paid less than otheremployees. Finally, it is possible that the reward system is less susceptible to bias and manipulation,especially in the case of unionized workers.The findings have several management implications. Employers need to determine the agecomposition of their workforce and to be aware of the issues of fair treatment that exist for their olderemployees when there is a primarily younger or middle-aged workforce. Younger managers need tounderstand that older employees may perceive that they are being treated unfairly by the manager. Toaddress this, younger managers should ensure that older subordinates’ jobs provide challenge and aremeaningful instead of boring and routine, that when decisions are made about an older subordinate’s jobthe reason for the decision is clearly communicated to the subordinate and the manager has treated allsubordinates in the same consistent manner. Younger managers also need to make sure that theythemselves as well as the members of the work group show respect for the contributions older employeesmake to the work group and to the job. Finally, older employees are more likely to stay with theirorganization if it is apparent that the organization is making a serious effort to retain its older employees,if the organization is encouraging older employees to delay their retirement, and if the organization isproviding older employees with desirable options such as phased retirement.The focus of our study was on older workers. Future research needs to be conducted to confirmthe generalizability of perceived fair treatment as an intervening mechanism in the relationship betweenrelational age and outcomes. It is possible that different types of intervening processes are more importantfor younger workers as well as other types of demographic attributes besides age. Some of the agecomposition groups were relatively small. The present study needs to be replicated with a much largersample of older workers. This study focused on intention to remain as the outcome variable of interestbecause this is an important variable in the challenge to retain older employees in the workplace.However, researchers need to examine the relationships of relational age and perceived fair treatmentwith other outcome variables such as job and career satisfaction.In summary, older employees with similar age supervisors perceived they were treated morefairly than did those with younger supervisors. Older employees in work groups with similar agemembers perceived they were treated with more respect than did those in work groups with members ofvarying ages. Older employees in organizations with a primarily older workforce perceived olderemployees in general were treated more fairly than did those in organizations composed of a primarilyyounger or middle-aged workforce. Future research is needed to replicate these findings with both olderworkers and workers of other ages and to identify other intervening processes.263

ReferencesAllen, Tammy D., Russell, Joyce E.A., Poteet, Mark L. & Dobbins, Gregory H., “Learning andDevelopment Factors Related to Perceptions of Job Content and Hierarchical Plateauing,”Journal of Organizational Behavior, 20, (1999), 1113-1137.Barnes-Farrell, Janet L., “Beyond Health and Wealth: Attitudinal and Other Influences on RetirementDecision-making,” in G.A. Adams & T.A. Beehr (eds.), Retirement: Reasons, Processes andResults, New York: Springer Publishing Company, 2003, pp. 159-187.Bruins, Jan, Platow, Michael J. & Ng, Sik Hung, “Distributive and Procedural Justice in Interpersonal andIntergroup Situations: Issues, Solutions, and Extensions, Social Justice Research¸ 8(1), (1995),103-121.Byrne, Donn, The Attraction Paradigm, New York: Academic Press, 1971.Chatman, Jennifer A. & Spataro, Sandra E, “Using Self-categorization Theory to Understand RelationalDemography-based Variations in People’s Responsiveness to Organizational Culture,” Academyof Management Journal, 48(2), (2005), 321-331.Cleveland, Jeanette N. & Shore, Lynn McFarlane, “Self- and Supervisory Perspectives on Age and WorkAttitudes and Performance,” Journal of Applied Psychology, 77, (1992), 469-484.Folger, Robert, “Workplace Justice and Employee Worth,” Social Justice Research, 7(3), (1994), 225-240.Goldberg, Caren B., “Relational Demography and Similarity-attraction in Interview Assessments andSubsequent Offer Decisions,” Group & Organization Management, 30(6), (2005), 597-624.Harrison, David A., Price, Kenneth H. & Bell, Myrtle P., “Beyond Relational Demography: Time and theEffects of Surface- and Deep-level Diversity on Work Group Cohesion,” Academy ofManagement Journal, 41(1), (1998), 96-107.Jackson, Susan E., Brett, Joan F., Sessa, Valerie I., Cooper, Dawn M., Julin, Johan A. & Peyronnin, Karl,“Some Differences Make a Difference: Individual Dissimilarity and Group Heterogeneity asCorrelates of Recruitment, Promotions, and Turnover,” Journal of Applied Psychology, 76,(1991), 675-689.Judge, Timothy A. & Ferris, Gerald R., “Social Context of Performance Evaluation Decisions,” Academyof Management Journal, 36, (1993), 80-105.Kim, Sang-Wook, Price, James L., Mueller, Charles W. & Watson, Thomas W., “The Determinants ofCareer Intent among Physicians at a US Air Force Hospital,” Human Relations, 49, (1996), 947-976.Lawrence, Barbara S., “The Black Box of Organizational Demography,” Organization Science, 8(1),(1997), 1-22.London, Manuel & Howat, Gary, “The Relationship between Employee Commitment and ConflictResolution Behavior,” Journal of Vocational Behavior, 13, (1978), 1-14.Lyons, Thomas F., “Role Clarity, Need for Clarity, Satisfaction, Tension, and Withdrawal,”Organizational Behavior and Human Performance, 6, (1971), 99-110.Maurer, Todd J., Weiss, Elizabeth M. & Barbeite, Francisco G., “A Model of Involvement in WorkrelatedLearning and Development Activity: The Effects of Individual, Situational, Motivational,and Age Variables,” Journal of Applied Psychology, 88, (2003), 707-724.McCain, Bruce E., O’Reilly, Charles A. & Pfeffer, Jeffrey, “The Effects of Departmental Demography onTurnover: The Case of a University,” Academy of Management Journal, 26, (1983), 626-641.Milliman, John F., Causes, Consequences and Moderating Factors of Career Plateauing, unpublisheddoctoral dissertation, University of Southern California, 1992.Moorman, Robert H., “Relationship between Organizational Justice and Organizational CitizenshipBehaviors: Do Fairness Perceptions Influence Employee Citizenship?” Journal of AppliedPsychology, 76, (1991), 845-855.264

Moorman, Robert H., Blakely, Gerald L. & Niehoff, Brian P., Does Perceived Organizational SupportMediate the Relationship between Procedural Justice and Organizational Citizenship Behavior?Paper presented at the annual meeting of the Academy of Management, 1995.O’Reilly, Charles A., Caldwell, David F., Barnett, William P., “Work Group Demography, SocialIntegration, and Turnover,” Administrative Science Quarterly, 34, (1989), 21-37.Pelled, Lisa Hope, “Demographic Diversity, Conflict and Work Group Outcomes: An Intervening ProcessTheory,” Organization Science, 7(6), (1996), 615-631.Pfeffer, Jeffrey, “Organizational Demography,” Research in Organizational Behavior, 5, (1983), 299-357.Pfeffer, Jeffrey, “Organizational Demography: Implications for Management,” California ManagementReview, XXVIII, (1985), 67-81.Reynolds, Scott, Ridley, Neil & Van Horn, Carl E., “A Work-filled Retirement: Workers’ ChangingViews on Employment and Leisure,” Work Trends Survey, 8(1), (2005), 1-38.Riordan, Christine M., “Advancing Relational Demography Theory: A Construct Validity Study of ThreeMeasures of Demographic Similarity,” paper presented at the annual meeting of the Academy ofManagement, Boston, (1997).Rosen, Benson & Jerdee, Thomas E., “Middle and Late Career Problems: Causes, Consequences, andResearch Needs,” Human Resource Planning, 13, (1990), 59-70.Shore, Lynn M., Cleveland, Jeanette N. & Goldberg, Caren B., “Work Attitudes and Decisions as aFunction of Manager Age and Employee Age,” Journal of Applied Psychology, 88, (2003), 529-537.Stewman, Shelby, “Organizational Demography,” Annual Review of Sociology, 14, (1988), 173-202.Tajfel, Henri & Turner, John, “The Social Identity Theory of Intergroup Behavior,” in S. Worchel & W.Austin (eds.), Psychology of Intergroup Relations, Chicago: Nelson-Hall, 1986, pp. 7-24.Tsui, Anne S., Egan, Terri D. & O’Reilly, Charles A., “Being Different: Relational Demography andOrganizational Attachment,” Administrative Science Quarterly, 37, (1992), 549-579.Tsui, Anne S., Porter, Lyman W., & Egan, Terri D., “When Both Similarities and Dissimilarities Matter:Extending the Concept of Relational Demography,” Human Relations, 55(8), (2002), 899-929.Tsui, Anne S. & O’Reilly, Charles A., “Beyond Simple Demographic Effects: The Importance ofRelational Demography in Superior-Subordinate Dyads,” Academy of Management Journal,32(2), (1989), 402-423,Turban, Daniel B. & Jones, A P., “Supervisory-Subordinate Similarity: Types, Effects, and Mechanisms,”Journal of Applied Psychology, 73, (1988), 228-234.Turner, John C., “A Self-categorization Theory,” in M. Hogg, P. Oakes, S. Reicher & M.S. Wetherell(eds.), Rediscovering the Social Group: A Self-categorization Theory, Oxford: Blackwell, 1987,pp. 42-67.Tyler, Tom R. & Blader, Steven L., “Identity and Cooperative Behavior in Groups,” Group Processes &Intergroup Relations, 4, (2001), 207-226.Wagner, W. Gary, Pfeffer, Jeffrey & O’Reilly, Charles A., “Organizational Demography and Turnover inTop-management,” Administrative Science Quarterly, 29, (1984), 74-92.Zenger, Todd R. & Lawrence, Barbara S., “Organizational Demography: The Differential Effects of Ageand Tenure Distributions on Technical Communication,” Academy of Management Journal,32(2), (1989), 353-376.265

ASAC 2006Banff, AlbertaAnthony E. CarrollJulian BarlingQueen’s School of BusinessDaniel G. GallagherJames Madison University, VirginiaPaul F. ClarkPennsylvania State UniversityClive J. FullagarKansas State UniversityLEADERSHIP DURING CRITICAL TIMES: TRANSFORMATIONAL LEADERSHIPINFLUENCES NEWCOMERS’ UNION ATTITUDES ELEVEN YEARS LATERThe influence of transformational leadership on union attitudes was investigatedlongitudinally in a sample of United States postal workers. One hundred and eighty ninemembers of the National Association of Letter Carriers (NALC) were studied over an 11year period. Multiple regression analyses showed that after controlling for demographicvariables, prior union attitudes and correlates of union attitudes, the union stewards’transformational leadership predicted members’ union attitudes eleven years later. Thesefindings highlight the enduring influence of transformational leadership when manifestedduring critical periods for employees and their organizations. Implications of the findingsfor theory and research on leadership and labor unions are discussed.266

ASAC 2006Banff, AlbertaLaurent M. LapierreUniversity of OttawaRick D. HackettMcMaster UniversityDISPOSITIONAL, RELATIONAL, AND ATTITUDINAL CORRELATES OFORGANIZATIONAL CITIZENSHIP BEHAVIOR: A TEST OF AN INTEGRATIVE MODELIn an effort to better understand why OCB and some of its strongest correlates areinterrelated with each other, we tested competing theoretical models that integrate OCB,trait conscientiousness, overall job satisfaction, and leader-member exchange (LMX)quality. Using meta-analytically based structural equation modeling, we found strongestsupport for a model representing the following theoretical process: Employees who aremore conscientious exhibit more frequent OCB; demonstration of this behavior enhancesthe quality of their relationship with their immediate supervisor, which affords themaccess to more satisfying job experiences; being more satisfied with their jobs, employeesthen tend to reciprocate the satisfying experiences provided by further demonstratingOCB. Our results suggest that the relationship between trait conscientiousness and overallsatisfaction may also be explained by variables other than OCB and LMX quality.Overall, our findings extend past research by showing support for a model positioningOCB as both cause and consequence of LMX quality and subsequent overall jobsatisfaction, and as a mediator explaining the relationship between trait conscientiousnessand LMX quality.267

ASAC 2006Banff, AlbertaNadine Quenneville (étudiante)Département de PsychologieUniversité du Québec à MontréalGilles SimardKathleen BenteinÉcole des Sciences de la GestionUniversité du Québec à MontréalPOURQUOI LES PRATIQUES DE GESTION DES RESSOURCES HUMAINES NE PEUVENTSE PASSER DU SUPPORT ET DE L’ENGAGEMENT DANS LA MOBILISATION DESEMPLOYES ?Cet article présente des résultats de recherche qui mettent en évidence l’effet de levier dusupport et de l’engagement dans la relation entre trois pratiques de gestion des ressourceshumaines et deux types de comportements de mobilisation. Étant donné l’avancement destravaux sur la question, nous avons mesuré deux cibles de support et d’engagement, soitle supérieur et l’organisation. De manière à éliminer les problèmes de variance commune,nous avons mesuré les comportements de mobilisation des employés (respect du contratde travail - RESCON et performance supérieure dans le poste - PERFSUP) à partir d’unquestionnaire destiné aux gestionnaires et les autres variables (pratiques RH, supportperçu de l’organisation et du supérieur, et engagement affectif envers l’organisation et lesupérieur) dans un questionnaire adressé aux employés. Les analyses ont été réalisées surun échantillon de 222 employés et 38 supérieurs de magasins de commerce au détailappartenant à la même chaîne canadienne. Les relations ont été établies avec LISREL 8.5en utilisant la méthode d’équations structurelles.Nos résultats montrent que le développement des compétences, la reconnaissance non-monétaireet le partage d’information sont positivement reliés à la perception de support organisationnel (PSO).Parmi ces trois pratiques, seul le partage d’information n’est pas associé à la PSO via la perception desupport du supérieur (PSS). Par contre, les relations entre la PSO, l’engagement affectif enversl’organisation (EA-ORG) et la PERFSUP sont significatives, au même titre que les relations entre la PSS,l’engagement affectif envers le supérieur (EA-SUP), et le RESCON et la PERFSUP. Finalement, le lienentre l’EA-SUP et les comportements de mobilisation de type respect du contrat de travail est plus fortque celui entre l’EA-ORG et ces mêmes comportements.268

ASAC 2006Banff, AlbertaNiti Singh (Doctoral Student)Narsee Monjee Institute of Management Studies,Mumbai, IndiaVenkat R. KrishnanXavier Labour Relations Institute, Jamshedpur, IndiaTRANSFORMATIONAL LEADERSHIP IN INDIA: DEVELOPING AND VALIDATINGA NEW SCALE USING GROUNDED THEORY APPROACHThis research developed and validated a measure of transformational leadership in Indiausing three studies. We followed grounded theory approach for generating the initialitem-pool (n=250). In the second study (n=379) a six-factor model emerged after anexploratory as well as confirmatory factor analysis was done. The factors show supportfor both universal and unique cultural dimensions. In the third study (n=202; 101 dyads)we found support for sound psychometric properties for the new scale, includingincremental validity.269

ASAC 2006Banff, AlbertaLaura ErskineMarshall School of BusinessUniversity of Southern CaliforniaHOW FAR DOES LEADERSHIP TRAVEL?DISTANCE AS A MULTIDIMENSIONAL CONCEPTCenturies after distance became a factor for individuals and organizations, the time hascome to develop a holistic understanding of the perceptions and organizational effectsthat could result. Two key factors are driving the need to more closely examine theconstruct of distance in organizations – the prominence of remote work and changes inthe structure and composition of organizations. I propose a reconceptualization ofdistance as a multidimensional perception held by individuals that may enhance ordiminish organizational outcomes such as individual performance, leadershipeffectiveness, and organizational identification. I suggest that distance is perceived in twointeractive dimensions. Structural distance comprises physical distance, channel ofcommunication, and frequency of communication while emotional distance includescomponents such as demographic distance, social distance, and psychological distance. Ianticipate that further use of this two-dimensional construct of distance will lead togreater recognition of its complexity and its effects on a wide variety of organizationalrelationships.270

ASAC 2006Banff, AlbertaRichard SoparnotProfesseur de Management Stratégique ESCEMLA COOPERATION INTERINDIVIDUELLE: UNE PERSPECTIVE PAR LA THEORIEDE L’APPRENTISSAGE, LE CAS DE LA GESTION DE PROJET DANS LE SECTEURAUTOMOBILELa coopération interindividuelle a été peu étudiée en sciences de gestion. Pourtant, ellepermet l’action unifiée des individus au service du développement de l’entreprise commeelle peut être assimilée à une ressource susceptible de fonder un avantage concurrentiel.Néanmoins, elle n’obéit pas à des lois universelles. Quels en sont les ressorts ? Commentl’instituer ou la maintenir ? Pour répondre à ces questions, nous avons observé desprocessus de gestion de projet au sein du département d’ingénierie des systèmes d’essaisdu groupe Renault. Cette étude met en évidence la double logique coopérative :calculatoire et idéologique. Elle souligne également son caractère dynamique en mettanten lumière les difficultés d’initialisation de la coopération interindividuelle.271

ASAC 2006Banff, AlbertaCéleste M. BrotheridgeESG-UQAMLE TRAVAIL ÉMOTIONNEL :ANALYSE CRITIQUE ET IMPLICATIONS PRATIQUES 1Cet article fait le bilan de la littérature au sujet du travail émotionnel pour examiner lesquestions suivantes : C’est quoi le travail émotionnel ? Quels sont les antécédents et lesconséquences du travail émotionnel ? Pour les employeurs ? Pour les employés ? Quellessont les implications plus larges du travail émotionnel ? Et, finalement, étant donné leprécédent, quelles pistes d’actions pourrons-nous suggérer pour les individus et lesorganisations ? Les réponses sont tirées de la littérature empirique et théorique depsychologie et sociologie.272

ASAC 2006Banff, AlbertaMichelle Harbour (étudiante)HEC MontréalDU COURAGE D’AGIR AU COURAGE D’ÊTREPROFILS DU COURAGE MANAGÉRIALUn modèle conceptuel du courage managérial est présenté. Issu de variables individuelleset situationnelles liées au jugement moral et à un contexte difficile ou risqué, il positionneles types de courage sur une matrice opposant l’intensité émotionnelle et le contrôle desémotions. Basé sur la structure cognitive de gestionnaires, ce modèle offre une définitionopératoire du courage managérial. Huit propositions de recherches sont suggérées dansune discussion sur les contributions théoriques, méthodologiques et pratiques.273

ASAC 2006Banff, AlbertaOlivier Doucet (étudiant)ESG-UQAMMichel TremblayHEC MontréalGilles SimardESG-UQAMLE RÔLE DU SUPÉRIEUR DANS LA RECONSTRUCTION DULIEN SOCIAL ENTRE L’EMPLOYÉ ET L’ORGANISATIONCette recherche évalue comment le support et la confiance permettent d’expliquer larelation entre les dimensions associées à la « full range leadership theory » etl’engagement affectif des employés. En plus de montrer que ces construits constituent desvariables intermédiaires importantes, nos résultats font ressortir que seule la dimensioncharismatique influence la confiance alors que plusieurs dimensions agissent sur lesupport.274

ASAC 2006Banff, AlbertaSylvie GuerreroEcole des Sciences de la GestionUniversité du Québec à MontréalOlivier HerrbachUniversité Montesquieu – Bordeaux IVTHE AFFECTIVE UNDERPINNINGS OF PSYCHOLOGICAL CONTRACTFULFILLEMENTThis paper is an empirical study of the link between the psychological contract andaffective states at work. We argue that perceived organizational prestige and contractviolation are the perceived organizational supportitive vs. negative affective mindset thatmay arise from the cognitive contract assessment, and test this assumption using amanager sample.275

ASAC 2006Banff, AlbertaGbemisola A. Abimbola (student)Zhehui Zhong (student)P. Robert DuimeringDepartment of Management SciencesUniversity of WaterlooEFFECTS OF PROBLEM STRUCTURE ON BEHAVIOUR ANDPERFORMANCE IN GROUP PROBLEM SOLVING: THEORY ANDEXPERIMENTThis paper presents a view of group problem solving based on Heider’s (1946; 1958)balance theory, which conceptualizes problem solving as a progression towardsincreasing structural balance. Balance theory provides a means of representing problemsin ways that capture both gestalt restructuring and incremental search properties ofproblem solving. Results of a group problem solving experiment are reported, whichinvestigates the effects of certain properties of problem structure on problem solvingbehaviour and performance.276

ASAC 2006Banff, AlbertaVincent RousseauÉcole de relations industriellesUniversité de MontréalCaroline AubéEstelle M. MorinHEC MontréalLE RÔLE DE L’AUTONOMIE AU TRAVAIL ET DU LIEU DE CONTRÔLE EN REGARD DELA SANTÉ PSYCHOLOGIQUE : EFFETS PRINCIPAUX ET D’INTERACTIONCette étude vise à vérifier (a) les effets principaux de l’autonomie au travail (autonomieprocédurale, autonomie temporelle et autonomie face aux critères) et du lieu de contrôleen regard de la santé psychologique (bien-être psychologique et détresse psychologique)et (b) leurs effets d’interaction. Les résultats obtenus à partir d’un échantillon composé de249 participants révèlent que les trois facettes de l’autonomie au travail sont reliéespositivement au bien-être psychologique et sont reliées négativement à la détressepsychologique. De plus, le lieu de contrôle agit à titre de modérateur notamment sur lesrelations impliquant l’autonomie face aux critères et les deux dimensions de la santépsychologique.277

ASAC 2006Banff, AlbertaTracy HechtConcordia UniversityBALANCING THE ROLES OF STUDENT, FAMILY MEMBER AND PAID EMPLOYEE:AN INVESTIGATION OF COPING, CONFLICT, FACILITATION, AND SATISFACTIONThe purpose of this study was to examine relations between coping, role conflict andfacilitation, and satisfaction. Participants (N = 284) were individuals balancing the rolesof employee, family member, and student. Results supported the hypothesis that conflictand facilitation have additive effects on satisfaction, and that conflict and facilitationmediate the relation between coping styles and satisfaction with life domains.278

ASAC 2006Banff, AlbertaMarylène GagnéJMSB Concordia UniversityWORK MOTIVATION AND COMMITMENT:MULTIDIMENSIONAL AND MULTIFOCI PERSPECTIVESToday’s most popular conceptualization of organizational commitment is the Three-Component Model of Organizational Commitment proposed by Allen and Meyer (1990),which includes affective, continuance, and normative commitment. Affectivecommitment represents an emotional attachment to, involvement in, and identificationwith the employer. Continuance commitment represents the perception that one has tostay with the employer because of the associated cost of leaving. Finally, normativecommitment is the perceived obligation to stay. There has been a multitude of studies thatsupports this theory, yet almost none has addressed the relation between organizationalcommitment and work motivation. This symposium will present the newest research onthe link between motivation and commitment, with different presentations focusing ondifferent and sometimes competing models of the relationship between motivation andcommitment. Moreover, recent research has started to examine the different foci(organizational and occupational) of commitment. Another goal of this symposium is topresent the latest research on the duality of commitment to different foci.279

ASAC 2006Banff, AlbertaMitchell J. NeubertBaylor UniversitySimon TaggarWilfrid Laurier UniversityPERCEPTIONS OF GROUP POTENCY AND PERSONALITY ON VOLUNTEER GROUPMEMBER SELLING BEHAVIORWe studied 388 volunteers loosely coupled in groups to sell memberships in a non-profitorganization. Consistent with economic models of altruism, we found individualperceptions of group potency to be negatively related to individual sales behavior (i.e.,making telephone calls and customer visits). Furthermore, individual member’sperception of group potency moderated the relationship between two personality traits(conscientiousness and extraversion) and individual sales behavior.280