Qualitative Data Analysis From Topics to Themes

Ron ChenailNova Southeastern UniversityPenn State Hershey, College of MedicineMay 8, 2013

¨ What is qualitative data analysis?¨ How do I conduct qualitative data analysis andpresent results?

¨Generic Design is whenresearchers use genericor “eclectic” qualitativemethods (e.g., opencoding, axial coding,constant comparison) toproduce conceptualcategories and themes.¨Designer Design iswhen researchers followdesigner or “namebrand” methodologiesto produce the mosttransformed results andfindings (e.g., thickdescriptions, groundedtheories, essences oflived experience).

¨¨¨¨¨¨¨Qualitative DataAnalysisQualitative ContentAnalysisThematic AnalysisQualitative InterviewingEthnographicPhenomenologicalCase Study¨¨¨¨¨EthnographyGrounded TheoryPhenomenologyNarrative AnalysisDiscourse Analysis

¨ Words¨ Feelings¨ Actions¨ Rituals¨ Experiences¨ Perspectives¨ Impressions¨ Events¨ Artifacts¨ Symbols

¨ Interviews and Transcripts¨ Observations and Fieldnotes¨ Documents¨ Pictures and Images¨ Audio and Visual Recordings

¨Qualitative data analysis is the rigorousprocess of selecting qualitatively distinct data,articulating the qualitative meaning ascribed tothose units, and commenting on the qualitativesimilarities and differences noted between andamong these distinct units of data.

¨The goal of qualitative data analysis can be todescribe, explain, and/or interpretingqualitative patterns in terms of words,numbers, matrices, pictures, sounds, or otherforms of representation.

¨ Human thoughts are combinational (simpleparts combine)¨ Human thoughts are recursive (parts can beembedded within parts)¨ Human languages are combinational¨ Human thoughts about human languages arerecursive¨ Qualitative Data Analysis is a recursivecombining of simple steps

¨ Metaphoric: Saying some thing about something¨ Coding: Drawing a distinction and naming it¨ Analysis: Making sense of how thesedistinctions are qualitatively different andqualitatively similar

Relationship 1, 2, 3, & 4Relationship 1, 2, & 3Idea4Idea3Relationship 1 & 2Idea 2 Idea 1

Composite 1, 2, 3, & 4Composite 1, 2, & 3Case4Case3Composite 1 & 2Case 2 Case 1

¨ Discovery-oriented (i.e., typically moreinductive than deductive analysis)¨ Iterative (i.e., the process can be changed basedupon the outcomes)¨ Circular (e.g., constant comparison leading theresearcher from the analysis to the field andback again)

¨ Transformation from data, to information, toknowledge, to wisdom¨ Transformation from separation to connection(e.g., coding to create discrete nodes and thenconnecting the nodes to create categories)

Closest

¨ Presenting data as if they were the findings¨ Reproducing interview data, case histories, orstories they had collected in a reduced formwith minimal or no interpretation of those data¨ Containing no analysis and no interpretation(Sandelowski & Barroso, 2003, pp. 909-910)

¨ Emphasizing nominal or categorical data, orlists and inventories of topics covered byresearch participants in interviews and focusgroups¨ Emphasizing inventories, frequencies, andpercentages of participants stating a topic, orenumerations of the topics themselves.¨ Investigators often introducing topics in theirinterview questions and/or derived from amanifest content analysis.(Sandelowski & Barroso, 2003, pp. 910-912)

¨ Conveying an underlying or more latentpattern or repetition discerned in the data¨ Containing the lowest level of abstraction¨ Offering more penetrating or nuanceddescriptions of experience, using either in vivoor everyday language, or themes or conceptsfrom existing empirical or theoretical literatureto label and/or organize their data(Sandelowski & Barroso, 2003, pp. 912-913)

¨ Presenting one or more concepts or themeseither developed in situ from the data orimported from existing theories or literatureoutside the study¨ Moving towards interpretively integratingportions of data¨ Extending the theoretical or other intellectualtradition from which they were imported and/or illuminating an experience(Sandelowski & Barroso, 2003, pp. 913-914)

¨ Transforming data to produce groundedtheories, ethnographies, or otherwise fullyintegrated explanations of some phenomenon,event, or case¨ Clarifying or elucidating conceptual orthematic linkages that re-present the targetphenomenon in a new way¨ Attending to relevant variations in both sampleand data (Sandelowski & Barroso, 2003, p. 914)

¨¨¨¨Target: The portion of the text that has caught yourinterestTag: The name you give to the target that reflects themeaning you ascribe to itTale: The story of the tag’s meaning in relationship tothe target in which you make the evidential qualities ofthe target in support of your tag overt.Thesis: The proposition that connects the patternsreflected across the tales of the tags’ relationships totheir targets (Chenail, 2010)

¨ Generate and Analyze the Data¨ Bricoleur¨ Preparation¨ Stamina

¨ Systematic yet creative¨ Focused yet reflective¨ Testimonial yet evidential¨ Subjective yet objective

¨ Codes, Categories, and Themes¨ Analysis and Memos¨ Quality Control

¨ Articulating the relationship between meaningand data (Codes), Codes (Categories) andCategories (Themes)¨ Internal Integrity (i.e., Is there a high degree ofhomogeneity across the individual codes oracross the coded units within the categories?)¨ External Integrity (i.e., Is there a high degree ofheterogeneity or differentiation between thearray of homogeneous codes, categories, orthemes?) (Chenail, 2008)

¨ Exhaustive System of Codes, Categories, andThemes (i.e., No significant meaningful featureof the phenomenon under study falls outside ofthe array)¨ Dual Planes of Focus: Horizontality (i.e.,category-to-category relationships) andVerticality (i.e., category-to-phenomenarelationships) (Chenail, 2008)

Clip OneTranscription• Code One• Code Two• Code Three• Code FourClip TwoTranscription• Code One• Code Two• Code Three• Code Four

Clip OneCategoryACategoryBCode OneCode TwoCodeThreeCodeFour

Clip OneClip TwoClipThreeCategory ACategory ACategory ACategory BCategory BCategory BCategory CCategory CCategory C

CategoryAClip OneCategoryBClip OneCategoryCClip OneClip TwoClip TwoClip TwoClip ThreeClip ThreeClip Three

Theme OneCategory A(CodedClips One,Two, &Three)Category B(CodedClips One,Two, &Three)Category C(CodedClips One,Two, &Three)

¨ Making Sense of the Codes, Categories, andThemes in Context (i.e., Research Question andLiterature)¨ Shifting from Information to Knowledge¨ Asking yourself the question, “What have Ilearned?”

¨ Memos are notes in which we make themeaning of our codes, categories, themes, andanalysis transparent¨ Memos are the start of the analysis and theresults¨ Memos create our audit trails

Qualitative Data AnalysisThe ThisThe ThatCode, Category, Concept,TheoryUnit of Analysis

Qualitative Data AnalysisThe ThisDescriptive CodeThe ThatUnit of AnalysisClinical Problem: The therapistnoted the client reported cominginto therapy because she sharedhaving anxiety as her clinicalproblem, that is, a reason forwhich someone seekspsychotherapy.Client came in with anxiety…

Qualitative Data AnalysisThe ThisIn Vivo CodeThe ThatUnit of AnalysisAnxiety: The therapist notedthe client reported coming intotherapy with anxiety.Client came in with anxiety…

¨ Instrumentation¨ Piloting¨ Transparency¨ Audit Trails¨ Constant Comparison¨ Member Checking¨ Peer or Expert Debriefing¨ Multiple Coders¨ Exemplars

¨ Data as Star¨ Juxtaposition¨ Exemplars for every assertion, pronouncement,or declaration (Chenail, 1995)

¨¨¨¨¨Compose your analytical memos so they couldbe cut and pasted into the body of your reportName and explain/describe the qualitativedifference in full sentencesCopy and paste the data excerpt into the memoExplain or describe how the data excerptexemplifies and evidences the meaning of yournoted qualitative difference (e.g., code, category,theme)Bottom Line: Your analysis is your report!

¨ Hedging¨ Avoid Errors of Deficiency and Exuberance byStaying Close to the Data¨ Develop a Rhythm(Chenail, 1995)

¨ Restricting as in setting or qualifyingconditions relative to the results being reportedto avoid extremes or limit meaning.¨ Fits with constructionist, constructivist,phenomenological, naturalistic, exploratory,descriptive, interpretive, post-modern, andcritical research approaches

¨¨¨¨Propositions to express results for consideration andnot “results as the facts” by using “such as,” “may,” or“perhaps”Possibilities to express a likelihood or certainty ofresults and not “results as the truth” by using “seem,”“appear,” “suggest,” “speculate,” or “imply”Approximations to express quantity, frequency, degree,and time of results by using “generally,”“approximately,” “most,” “some,” “many,” “few,” or“frequently”Conditions to express co-dependent or local situationsor circumstances relative to results by reporting “who,”“what,” “where,” “when,” “why,” and “how”

Although the findings of this study arecompromised by the small sample size and itsvariability, they suggest in a very tentative waythat there may be subtle differences in the waychoice is perceived, "exercised and burdened" [50](p. 237) in the aftermath of the detection of fetalimpairment. (Sandelowski & Jones, 1996, p. 359)

Although a much larger sample is required toverify that these are distinguishable patterns ofemplotments of choice, one very tentativeconclusion that may be drawn from this finding isthat some external circumstances may permit agreater variety of narrative constructions thanothers. (Sandelowski & Jones, 1996, p. 359)

Although the findings from this study are verytentative and must be verified, they do raiseinteresting questions for future research.(Sandelowski & Jones, 1996, p. 359)Sandelowski, M., & Jones, L. C. (1996). ‘Healing fictions': Stories of choosing inthe aftermath of the detection of fetal anomalies. Social Science & Medicine, 42(3),353-361.

¨¨Natural - The data are presented in a shape thatresembles the phenomenon being studied. Forinstance, if the data are excerpts from a therapysession, present them in a sequential order or inan order that re-presents the flow of the sessionitself.Most Simple to Most Complex - For sake ofunderstanding, start the presentation of datawith the simplest example you have found. Asthe complexity of each example or exemplarpresented increases, the reader will have abetter chance of following the presentation.Erving Goffman's work is a good example ofthis style.

¨ First Discovered/Constructed to LastDiscovered / Constructed - The data arepresented in a chronicle-like fashion,showing the course of the researcher'spersonal journey in the study. This style isreminiscent of an archeological style ofpresentation: What was the first "relic"excavated, then the second and so forth.

¨ Quantitative-Informed - In this scheme dataare presented according to strategiescommonly found in quantitative orstatistical studies. Data are arranged alonglines of central tendencies and ranges,clusters, and frequencies.

¨Theory-Guided - Data arrangement is governedby the researcher's theory or theories regardingthe phenomenon being re-presented in thestudy. For instance, a Marxist-informedresearcher might present data from a doctorpatientinterview in terms of talk which showswho controls the means for producinginformation in the interaction, talk whichillustrates who is being marginalized, and soforth. In clinical qualitative research, thisapproach is quite prevalent as cliniciansorganize the data in terms of theirunderstandings of how doctor-patient, ornurse-patient, and therapist-client interact.

¨ Narrative Logic - Data are arranged with aneye for storytelling. Researchers plot out thedata in a fashion which allows them totransition from one exemplar to another justas narrators arrange details in order to bestrelate the particulars of the story.¨ Most Important to Least Important or FromMajor to Minor - Like the journalistic styleof the inverted pyramid, the most important"findings" are presented first and the minor"discoveries" come last.

¨¨Dramatic Presentation - This one is theopposite of the inverted pyramid style. With thedramatic arrangement scheme, researchersorder their data presentation so as to save thesurprises and unforeseen discoveries for last.No Particular Order Order - As it sounds, dataare arranged with no particular, consciouspattern in mind, or the researcher fails toexplain how or why the data are displayed theway they are.Adapted from Chenail, R. J. (1995). Presenting qualitative data. The Qualitative Report, 2(3).Retrieved from http://www.nova.edu/ssss/QR/QR2‐3/presenting.html

¨¨¨¨¨¨List Section HeadingPresent the Distinction or FindingIntroduce the First Data Exemplar of this DistinctionDisplay the First Data Exemplar of this DistinctionComment on the First Data Exemplar as EvidenceMake Transition to Second Data Exemplar and Repeatthe Pattern Until the Closing of this Section

¨ Theme One¡ Introduce first category (i.e., Category A)¡ Introduce first exemplar (e.g., Category A’s CodeOne from Clip One)¡ Explain how the exemplar evidences a quality of firstcategory¡ Explain relationship between first category’squalities and the overall theme’s qualities¡ Introduce additional exemplars for first category¡ Explain how each evidences a different quality ofCategory A¡ Transition to second category

¨Theme One¡ Introduce second category (i.e., Category B)¡ Introduce first exemplar (e.g., Category B’s Code Onefrom Clip One)¡ Explain how the exemplar evidences a quality of secondcategory¡ Explain relationship between second category’s qualitiesand the overall theme’s qualities¡ Introduce additional exemplars for second category¡ Explain how each evidences a different quality ofCategory B¡ Transition to additional categories, conclude Theme One,and transition to Theme Two

¨ Theme Two¡ Repeat

¨ Maintain Coherence and Build Confidence¨ Stay Focused on the Research Question¨ The Role and Place of the Literature

¨ Chenail, R. J. (1995). Presenting qualitativedata. The Qualitative Report, 2(3). Retrievedfromhttp://www.nova.edu/ssss/QR/QR2‐3/presenting.html¨ Chenail, R. J. (2008). Categorization. In L. M.Given (Ed.), The SAGE encyclopedia of qualitativeresearch methods (Vol. 1, pp. 72-73). ThousandOaks, CA: Sage.

¨ Chenail, R. J. (2010). How to read and review abook like a qualitative researcher. TheQualitative Report, 15(6), 1635-1642. Retrievedfrom http://www.nova.edu/ssss/QR/QR15-6/readreview.pdf¨ Sandelowski, M., & Barroso, J. (2003).Classifying the findings in qualitative studies.Qualitative Health Research, 13(7), 905-923.

¨¨Chenail, R. J. (2012). Conducting qualitative dataanalysis: Reading line-by-line, but analyzing bymeaningful qualitative units. The Qualitative Report,17(1), 266-269. Retrieved fromhttp://www.nova.edu/ssss/QR/QR17-1/chenailline.pdfChenail, R. J. (2012). Conducting qualitative dataanalysis: Qualitative data analysis as a metaphoricprocess. The Qualitative Report, 17(1), 248-254.Retrieved from http://www.nova.edu/ssss/QR/QR17-1/chenail-metaphor.pdf

¨ Chenail, R. J. (2012). Conducting qualitativedata analysis: Managing dynamic tensionswithin, part one. The Qualitative Report, 17(T&LArt. 4), 1-6. Retrieved fromhttp://www.nova.edu/ssss/QR/QR17/chenail-tensions.pdf

¨ Online QDA:http://onlineqda.hud.ac.uk/index.php¨ The Qualitative Report:http://www.nova.edu/ssss/QR/¨ Graduate Certificate in Qualitative Researchhttp://www.nova.edu/ssss/QR/¨ Email me: ron@nova.edu

Patricia L. Munhall&Ronald J. ChenailJones and Bartlett2008

Ronald J. Chenail, PhD, is Professor of FamilyTherapy in theSchool of Humanities and Social Sciences at NovaSoutheastern University in Fort Lauderdale,Florida. From 2005 to 2011, he was the Editor-in-Chief of the Journal of Marital and Family Therapy,the flagship research journal of the AmericanAssociation for Marriage and Family Therapy.Since 1990, he is also the founding Editor-in-Chiefof The Qualitative Report, the world's first online,open access, transdisciplinary journal dedicated toqualitative research. He may be contacted atron@nova.edu