Gordon et al. (2001)
Gordon et al. (2001)
Gordon et al. (2001)
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Clinic<strong>al</strong> Child and Family Psychology Review, Vol. 4, No. 2, <strong>2001</strong>Children’s Testimony: A Review of Research on Memoryfor Past ExperiencesB<strong>et</strong>ty N. <strong>Gordon</strong>, 1,3 Lynne Baker-Ward, 2 and P<strong>et</strong>er A. Ornstein 1This review of children’s testimony focuses on research related to memory for past experiences.The aspects of the memory system that are involved in testimony are discussed and thedevelopment of autobiographic<strong>al</strong> memory is examined. Relevant research findings are summarizedin the context of an information-processing model of memory and the implicationsof this work for clinic<strong>al</strong> practice are outlined. We conclude that (1) under certain conditions,even very young children can remember and report past experiences with some accuracy oververy long periods of time; (2) substanti<strong>al</strong> and significant development<strong>al</strong> differences have beendemonstrated in children’s abilities to provide eyewitness testimony; (3) children can be influencedin a vari<strong>et</strong>y of ways to provide compl<strong>et</strong>e and elaborated reports of events that neveroccurred; and (4) even experts cannot <strong>al</strong>ways tell the difference b<strong>et</strong>ween true and f<strong>al</strong>se reports.KEY WORDS: children’s testimony; eyewitness memory; suggestibility; trauma.In recent years, children’s abilities to provide accurateaccounts of their experiences have been studiedextensively. This research has been fueled, in part,by concerns about the <strong>al</strong>legations of young children insever<strong>al</strong> high-profile cases of child sexu<strong>al</strong> abuse and theincreased frequency with which children are asked toprovide testimony in leg<strong>al</strong> s<strong>et</strong>tings. Although multipledimensions of individu<strong>al</strong> functioning, including manyaspects of socioemotion<strong>al</strong> and cognitive development(see Saywitz, in press), certainly affect the testimonyof young witnesses, memory is a basic prerequisite forchildren’s abilities to provide accounts of their experiences.Stated simply, children cannot report what theycannot remember (Ornstein, <strong>Gordon</strong>, & Baker-Ward,1992). Moreover, we argue that it is critic<strong>al</strong> for clinic<strong>al</strong>and leg<strong>al</strong> profession<strong>al</strong>s who work with child witnessesto understand the theor<strong>et</strong>ic<strong>al</strong> underpinningsof children’s memory/testimony performance as well1 Department of Psychology, University of North Carolina atChapel Hill, Chapel Hill, North Carolina.2 Department of Psychology, North Carolina State University,R<strong>al</strong>eigh, North Carolina.3 Address <strong>al</strong>l correspondence to B<strong>et</strong>ty N. <strong>Gordon</strong>, 433 NottinghamDrive, Chapel Hill, North Carolina 27514; e-mail: gordonjens@nc.rr.com.as the practic<strong>al</strong> implications of the research findings.Accordingly, our focus in this review of children’s testimonyis on their memory capabilities.In gener<strong>al</strong>, two broad approaches have beentaken in the study of children’s eyewitness memory.One approach is to examine the abilities ofchildren at different ages to accurately rememberand report their experiences (e.g., Baker-Ward,<strong>Gordon</strong>, Ornstein, Larus, & Clubb, 1993; Fivush, 1997;Ornstein, Baker-Ward, <strong>Gordon</strong>, & Merritt, 1997;Saywitz, Goodman, Nicholas, & Moan, 1991). This approachtypic<strong>al</strong>ly involves questioning the child at variousdelay interv<strong>al</strong>s about re<strong>al</strong>-life past experiences(e.g., visits to the doctor or the emergency room, tripsto a museum or amusement park) for which the d<strong>et</strong>ailsof the events in question can be specified. Theresults of this work have demonstrated clearly thatunder certain circumstances even very young childrenremember past events and can provide surprisinglyrich accounts of these experiences (Bahrick,Parker, Fivush, & Levitt, 1998). Moreover, thesememories often endure for very long periods of time(Howard, Osborne, & Baker-Ward, 1997; P<strong>et</strong>erson,1999; Schwarzmueller, Boyle, & Fivush, 1996). Significantdevelopment<strong>al</strong> differences <strong>al</strong>so have been1571096-4037/01/0600-0157$19.50/0 C○ <strong>2001</strong> Plenum Publishing Corporation
158 <strong>Gordon</strong>, Baker-Ward, and Ornsteindocumented, however, indicating that early elementaryschool children are more accomplished at thistask than are preschoolers. Older children, for example,provide more information in free rec<strong>al</strong>l, requirefewer specific prompts for compl<strong>et</strong>e and d<strong>et</strong>ailed reports,and forg<strong>et</strong> less information over time than doyounger children, especi<strong>al</strong>ly those below the age ofabout 4 years (Ornstein <strong>et</strong> <strong>al</strong>., 1997).A second research approach involves explorationof the many factors that influence—for b<strong>et</strong>teror worse—the accuracy of children’s eyewitnessmemory reports. Much of this work has involved exposingchildren to situations that provide an<strong>al</strong>ogs torobberies, abuse, and other crimes, and then manipulatingthe child’s experiences during the r<strong>et</strong>entioninterv<strong>al</strong> or varying the nature of the postevent interviewin ways that simulate aspects of forensic practice.Researchers have examined the effects of repeated interviewsor repeated questions within one interview—the impact of long delay interv<strong>al</strong>s b<strong>et</strong>ween the occurrenceof an event and a child’s subsequent testimony,the consequences of suggestive or misleading questions(or both), and stress or trauma (or both) as anevent is experienced, when a report is being made,or during both these situations (see Bruck, Ceci, &Hembrooke, 1998, for a review). The results of thisresearch have documented a number of factors thatcan greatly reduce the accuracy of children’s memoryreports. Moreover, some children can even beinduced to provide f<strong>al</strong>se information, claiming thatcertain events occurred when in fact they did not(see Ceci & Bruck, 1995). Indeed, the evidence suggeststhat young children, especi<strong>al</strong>ly preschoolers, aremore vulnerable to these types of suggestibility effectsthan are older children or adults. As a result ofthis work, specific recommendations for interviewingchildren involved with the leg<strong>al</strong> system have been proposed,with the go<strong>al</strong> of improving the accuracy of children’stestimony (e.g., American Academy of Childand Adolescent Psychiatry, 1997; American Profession<strong>al</strong>Soci<strong>et</strong>y on the Abuse of Children, 1990; Orbach<strong>et</strong> <strong>al</strong>., 2000; Poole & Lamb, 1998).Currently, research in the area of children’s testimonyis focused on gaining a b<strong>et</strong>ter understandingof the processes involved in remembering past eventsand the specific conditions under which some childrenmay be vulnerable to suggestion whereas othersare not. In this regard, this research is placedwithin the context of the rich theor<strong>et</strong>ic<strong>al</strong> frameworkand data provided by basic research in memory. Basicresearch indicates that remembering involves a seriesof information-processing steps, with each componentof the process affected by the nature of the tobe-rememberedmateri<strong>al</strong>, the conditions under whichremembering transpires, and the development<strong>al</strong> levelof the rememberer. Hence, to understand children’smemory and subsequent testimony, it is necessary toexamine both the processes through which informationis obtained, as well as the contents of memorystorage. Although a comprehensive review of memorydevelopment is well beyond the scope of this paper(see Schneider & Pressley, 1997, for an extendedtreatment), this review focuses on the theor<strong>et</strong>ic<strong>al</strong> andempiric<strong>al</strong> bases for understanding memory processesas they relate to children’s abilities to provide accurat<strong>et</strong>estimony. Thus, we begin with an overview ofthe aspects of the memory system that are involvedin children’s testimony. Next, we examine the emergenceof autobiographic<strong>al</strong> memory, a development<strong>al</strong>transition that defines the earliest age at which achild can be expected to provide leg<strong>al</strong> testimony. Themajor section of the paper provides a review of researchrelevant to understanding and facilitating children’stestimony, summarized within the context ofan information-processing framework. We concludewith a brief discussion of directions for future researchand the implications of this work for clinic<strong>al</strong> practice.THE TYPES OF MEMORY INVOLVED INCHILDREN’S TESTIMONYWhen children are involved as witnesses in leg<strong>al</strong>proceedings, they are asked to report events that transpiredmonths or even years previously. Hence, the extentand qu<strong>al</strong>ity of children’s testimony is d<strong>et</strong>erminedto a large extent by the r<strong>et</strong>rievability of informationin long-term memory. The long-term memory systemincludes two major representation<strong>al</strong> subsystems,declarative memory, defined as memory for facts andevents, and nondeclarative or procedur<strong>al</strong> memory,which includes stored representations of nonverb<strong>al</strong>actions or behavior<strong>al</strong> sequences (Bjorklund, 2000).In other words, declarative memory involves the r<strong>et</strong>entionof information, whereas procedur<strong>al</strong> memoryconcerns knowing how to accomplish specific tasks.Testimony typic<strong>al</strong>ly c<strong>al</strong>ls for the use of declarativememory, more specific<strong>al</strong>ly a type of declarative memorytermed episodic memory, or memory for informationthat can be linked to a particular occurrence.Because individu<strong>al</strong>s can be consciously aware of thecontents of episodic memory and can deliberately r<strong>et</strong>riev<strong>et</strong>he information, episodic memory is som<strong>et</strong>imesdescribed as explicit memory. In contrast, procedur<strong>al</strong>
Children’s Testimony 159memory is a type of implicit memory, it is automaticand must be assessed indirectly (Schacter, 1992). Althoughquestions som<strong>et</strong>imes arise about the role ofimplicit memory as the basis for evidence of childabuse (see Howe, 2000), testimony requires episodicmemory, which is expressed verb<strong>al</strong>ly.Episodic memory contains representations of ordinaryexperiences as well as unique events that becomepart of one’s life story. In gener<strong>al</strong>, the eventsreported by witnesses are not only referenced in timeand place, but are <strong>al</strong>so important occurrences for theindividu<strong>al</strong>. In this regard, testimony typic<strong>al</strong>ly c<strong>al</strong>ls fora type of episodic memory that is termed autobiographic<strong>al</strong>memory, defined by Nelson (1993, p. 61)as “specific, person<strong>al</strong>, long-lasting, and (usu<strong>al</strong>ly) ofsignificance to the self-system.” The development<strong>al</strong>emergence of autobiographic memory can reasonablybe considered to mark the earliest point at which achild can be expected to provide testimony, especi<strong>al</strong>lyin cases in which the child is the <strong>al</strong>leged victim.Another important characteristic of children’stestimony involves the nature of the events under investigation.Assuming that crimes were actu<strong>al</strong>ly committed,child witnesses are typic<strong>al</strong>ly victims of abuseor close observers of violent acts, often involving familymembers. They can be expected to have sufferedsome degree of trauma, and in many cases, may haveexperienced repeated abuse for some extended periodof time. As a consequence, in many instances theeffects of trauma on memory at both neurobiologic<strong>al</strong>and psychologic<strong>al</strong> levels are addition<strong>al</strong> influences onchildren’s testimony.A fin<strong>al</strong> consideration in examining children’s testimonyinvolves the flow of information through thememory system. Before witnesses can report events,they must first have encoded the information and establishedrepresentations in memory. Thus, changesin the memory representation that occur over timemust be understood in order to ev<strong>al</strong>uate children’scapacity to provide accurate testimony. Information<strong>al</strong>so must be r<strong>et</strong>rieved from long-term memory, andr<strong>et</strong>riev<strong>al</strong> is not <strong>al</strong>ways an automatic or perfect process.Hence, skills in monitoring what is in memory and inaccessing one’s own memory <strong>al</strong>so are important inunderstanding testimony.THE DEVELOPMENTAL EMERGENCE OFAUTOBIOGRAPHICAL MEMORYInvestigations conducted over the past 15 yearshave established that very young children havemuch b<strong>et</strong>ter memory capabilities than was previouslythought to be the case. Even infants and toddlers canencode, store, and r<strong>et</strong>rieve a great de<strong>al</strong> of informationabout the events they experience (Howe, 2000).Indeed, newborns can recognize voices and storiesto which they were exposed prenat<strong>al</strong>ly, as evidencedby modifications of their sucking patterns (DeCasper& Spence, 1986). Recent work with toddlers, usingelicited imitation tasks, in which action sequences thatproduce an event are demonstrated by an experimenterand subsequently reproduced by the children(see Bauer, 1995; Bauer, Hertsgaard, & Dow, 1994)has established that very young children form specificepisodic memories and r<strong>et</strong>ain them for long periodsof time. Such results, of course, strongly contradict the“tenacious and influenti<strong>al</strong> assumption” (Bauer, 1996,p. 39) that children cannot remember their own livesbefore the age of three or four (see <strong>al</strong>so Meltzoff, 1995;Rovee-Collier & Shyi, 1992).These findings regarding very early memory capabilitiescan be applied directly to children’s testimonyonly if two conditions are m<strong>et</strong>. First, earlymemories, which are demonstrated behavior<strong>al</strong>ly bypreverb<strong>al</strong> children, must subsequently be accessiblefor verb<strong>al</strong> reporting after language is established. Althoughit is clear that preverb<strong>al</strong> experiences may havelong-term effects on children’s behavior<strong>al</strong> responses,there is no way to directly link a particular behavior,even a manifestation of fear or anxi<strong>et</strong>y, with the d<strong>et</strong>ailsof a specific occurrence. Second, the accessible memoriesmust be autobiographic<strong>al</strong> and not just episodic innature. That is, they must represent the child’s memoryfor person<strong>al</strong> experiences that involve the individu<strong>al</strong>,rather than gener<strong>al</strong> rec<strong>al</strong>l of events without theincorporation of a person<strong>al</strong> context.The evidence regarding children’s abilities to verb<strong>al</strong>lyreport early experiences after they have acquiredlanguage is mixed. Bauer and her colleagues(Bauer, Kroupina, Schwade, Dropik, & Wewerka,1998), for example, found that 16- and 20-month-oldinfants demonstrated nonverb<strong>al</strong> evidence of memoryat a 6-month delayed assessment. Children who were20 months old at the time of exposure to the eventsmade utterances that provided verb<strong>al</strong> evidence of rec<strong>al</strong>l,whereas those in the younger group at the initi<strong>al</strong>sessions did not. It is interesting that productive vocabularyat the time of exposure to the event was notcorrelated with later verb<strong>al</strong> memory. The authors concluded,“Children who likely encoded events withoutthe benefit of language are capable of subsequent verb<strong>al</strong>mnemonic expression of them” (Bauer <strong>et</strong> <strong>al</strong>., 1998,p. 675). It appears, however, that there is a lower age
160 <strong>Gordon</strong>, Baker-Ward, and Ornsteinlimit for the maintenance of information in memoryafter long delays, at least when verb<strong>al</strong> access is required.Although Bauer <strong>et</strong> <strong>al</strong>. provide evidence for delayedrec<strong>al</strong>l of very early experiences among oldertoddlers, the extent to which their reports reflect autobiographic<strong>al</strong>memory is questionable. Howe (2000),in a discussion of this work, notes that evidence forautobiographic<strong>al</strong> memory would include reports ofaspects of the person<strong>al</strong> experience of the laboratoryvisit, not just the reconstruction of interactions withobjects. Bauer <strong>et</strong> <strong>al</strong>. (1998), <strong>al</strong>though pointing out thattheir research participants’ verb<strong>al</strong>izations includedspecific episodic information about the over<strong>al</strong>l event,acknowledge uncertainty as to the extent to which thereports represent autobiographic<strong>al</strong> rec<strong>al</strong>l.In contrast to the work of Bauer <strong>et</strong> <strong>al</strong>., P<strong>et</strong>ersonand Rideout (1998) clearly examined verb<strong>al</strong> reports ofan autobiographic<strong>al</strong> memory among very young children.The participants in this investigation were childrenb<strong>et</strong>ween 13 and 34 months of age who had had atraumatic injury requiring emergency medic<strong>al</strong> treatment.Verb<strong>al</strong> interviews were conducted with the childrenat 6-month interv<strong>al</strong>s until 2 years after their accidents.The presence of verb<strong>al</strong> narrative skills (definedas the ability to t<strong>al</strong>k about events that are removedin time and space from the immediate context) at th<strong>et</strong>ime of the event emerged as a d<strong>et</strong>erminant of subsequentverb<strong>al</strong> memory. Children who were at least 2years of age and who had narrative skills at the timeof the experience were able to report at least two centr<strong>al</strong>components of the event after a delay of 2 years.In contrast, <strong>al</strong>though slightly younger children whodid not demonstrate narrative abilities at the initi<strong>al</strong>assessment provided some verb<strong>al</strong> information aboutthe experience, the majority of these accounts werequite fragmentary and included substanti<strong>al</strong> levels ofinaccuracy. In addition, there was little evidence thatpreverb<strong>al</strong> memories could become verb<strong>al</strong>ly accessible.Only 2 of the 12 children in the youngest groupprovided verb<strong>al</strong> information about the experience atthe 12-month delayed interview, when they had acquirednarrative skills. Moreover, there were someconcerns that these children’s reports represented aspectsof family history conveyed by their parents afterthe event, rather than the r<strong>et</strong>ention of initi<strong>al</strong>ly encodedinformation.In summary, <strong>al</strong>though there is some emerging evidenc<strong>et</strong>hat young children can report aspects of experiencesthat are encoded before the ons<strong>et</strong> of productivelanguage, it seems unlikely that such memoriesare r<strong>et</strong>ained among children who are much youngerthan 2 years of age at the time of the experience, atleast over the very long delays that often characterizeleg<strong>al</strong> proceedings. Further, the information thatis provided appears to be fragmentary and accompaniedby inaccurate responses. For these reasons,it can be concluded that, at least in most instances,children cannot be expected to testify about eventsthat transpired before they were at least 2 years ofage.AN INFORMATION-PROCESSINGPERSPECTIVE ON CHILDREN’S TESTIMONYA simple conceptu<strong>al</strong> framework based on thestages of information-processing—that is, how informationis encoded, stored, and r<strong>et</strong>rieved—<strong>al</strong>lows on<strong>et</strong>o organize data relevant to the accuracy of children’stestimony. Because inaccuracies in rec<strong>al</strong>l can resultfrom disruptions at any of these stages, this inform<strong>al</strong>model can <strong>al</strong>so be used to understand the rangeof factors that can influence children’s remembering(<strong>Gordon</strong>, Schroeder, Ornstein, & Baker-Ward, 1995;Ornstein, Larus, & Clubb, 1991). This framework consistsof five broad themes: (1) not everything g<strong>et</strong>sinto memory, (2) what g<strong>et</strong>s into memory may varyin strength, (3) information in memory changes overtime, (4) r<strong>et</strong>riev<strong>al</strong> is not perfect, and (5) not everythingthat can be r<strong>et</strong>rieved is reported.Encoding: Not Everything G<strong>et</strong>s Into MemoryIn attempting to understand inaccuracies in children’smemory reports, it is important to keep in mindthat simple exposure to an event, even a s<strong>al</strong>ient person<strong>al</strong>experience, is not sufficient to insure compl<strong>et</strong>eencoding of the experience. In a study of memory fora routine pediatric examination (Baker-Ward <strong>et</strong> <strong>al</strong>.,1993), for example, children were interviewed immediatelyafter their checkups in order to obtain an estimateof the extent to which they encoded the experience.Even when strong r<strong>et</strong>riev<strong>al</strong> cues in the form ofvery specific questions were provided, rec<strong>al</strong>l was notperfect; the children reported 75%, 82%, and 92%, atages 3, 5, and 7, respectively, of the procedures thatcomprised the examination. Such encoding failuresmay arise from either selective attention in that individu<strong>al</strong>smay not notice some aspects of their experiencesas they transpire, or from the failure to transferinformation from short-term to long-term memory afterit enters the memory system.
Children’s Testimony 161Selective AttentionA number of factors have been shown to influenc<strong>et</strong>he likelihood that information will be attendedto and encoded. One important d<strong>et</strong>erminant of encodingis what one knows about the event before itoccurs. Knowledge affects how an individu<strong>al</strong> monitorsthe world, interpr<strong>et</strong>s events, and selectively attendsto certain types of stimuli while ignoring othertypes (Bjorklund, 1985; Chi & Ceci, 1987; Ornstein& Naus, 1985). A considerable body of evidence indicatesthat children’s understanding of the events towhich they are exposed will have a profound effecton what is encoded and stored in memory (Clubb& Ornstein, 1992; Nelson, 1986; Ornstein, Shapiro,Clubb, Follmer, & Baker-Ward, 1997; Ricci & Be<strong>al</strong>,1998). As an example, Goodman, Quas, Batterman-Faunce, Riddlesberger, and Kuhn (1997) found thatprior knowledge of a stressful medic<strong>al</strong> procedure predictedsubsequent memory performance of childrenaged 3 to 10 years, independently from the age of thechild.Given that knowledge in most, if not <strong>al</strong>l, domainsincreases with age, there should be comparabledevelopment<strong>al</strong> differences in the types of specificd<strong>et</strong>ails that are noticed and encoded. There is, however,little research that identifies specific age-relatedchanges in the content of children’s memories. Eisenand Goodman (1998) argue that what is memorable toany individu<strong>al</strong> child, regardless of age, and hence mostlikely to be encoded, is anything that is person<strong>al</strong>lysignificant to that child. Events or actions that affecta child’s sense of well-being, saf<strong>et</strong>y, or soci<strong>al</strong> acceptanceare considered to be person<strong>al</strong>ly significant andthus, more likely to be remembered (Goodman, Rudy,Bottoms, & Aman, 1990). Similarly, others (Bowers& Sivers, 1998; Howe, 2000) indicate that aspects ofan event are more likely to be encoded if they are“interesting” or “distinctive,” either because they areunexpected or emotion<strong>al</strong>ly arousing, to the child. Verytraumatic experiences, for example, may be rememberedvery clearly despite a lack of prior knowledgebecause of their distinctiveness. It has <strong>al</strong>so been noted,however, that high arous<strong>al</strong>, such as might occur duringa traumatic experience, results in a narrowing ofattention. Thus, many d<strong>et</strong>ails of such an experiencemay not be encoded because the child focuses on onlya few highly s<strong>al</strong>ient features (Bowers & Sivers, 1998).In this regard, it is not surprising that children remembercentr<strong>al</strong> features of even neutr<strong>al</strong> events b<strong>et</strong>ter thanmore peripher<strong>al</strong> features (Fivush, Gray, & Fromhoff,1987; Goodman, Hirschman, & Rudy, 1987). What iscentr<strong>al</strong> for a specific child depends on what is mostrelevant to that child, including the most threateningor most feared aspect of a traumatic experience.In situations in which children lack knowledgeabout an event, interactions with adults may compensatefor their development<strong>al</strong> limitation. Fivush(1998) argues that conversations about an event (asopposed to simply labeling things) b<strong>et</strong>ween childrenand adults that occur as the event unfolds are a critic<strong>al</strong>factor in d<strong>et</strong>ermining the features that childrenencode and remember. She suggests that these ongoingjointly constructed conversations provide childrenwith a b<strong>et</strong>ter understanding of their experiencesand help them to attend to important aspectsof events. Indeed, sever<strong>al</strong> researchers have demonstratedthis linkage b<strong>et</strong>ween adult–child conversationsand superior subsequent memory performance(Haden, Ornstein, Eckerman, & Didow, in press; Pipe,Dean, Canning, & Murachver, 1996; Tessler & Nelson,1994). Parents’ provision of information as eventsunfold may supplement children’s relatively limitedknowledge and hence play a centr<strong>al</strong> role in the children’scomprehension and subsequent memory fornovel experiences (Haden <strong>et</strong> <strong>al</strong>., in press).Failure of Encoded Information to ReachLong-Term StorageEven when individu<strong>al</strong>s attend to features oftheir experiences, this information may never enterlong-term storage. The information-processing modelposits a sequence of steps through which experienceis transferred to long-term memory. Perceptions of anevent are held very briefly in a nonlinguistic, sensorybasedstore, the sensory register. Unless sensory informationabout an event is immediately transferredto short-term memory, it will fade without having enteredconsciousness. Although there appear to be fewage related changes in the tot<strong>al</strong> capacity of the sensoryregister, the likelihood that information will be movedto the next level of information-processing increasessubstantively with age (Schneider & Bjorklund, 1998).This has been attributed to both changes in strategies,such as selective attention, and development<strong>al</strong>increases in processing speed.When perceptions enter consciousness, they areheld in short-term working memory (STWM). Thisstage of information-processing has been comparedwith RAM in a computer (Goldhaber, 2000). STWM,like RAM, can g<strong>et</strong> information from permanentmemory storage as well as extern<strong>al</strong> sources. Hence,
162 <strong>Gordon</strong>, Baker-Ward, and Ornsteinperceptions of new experiences can be transformedin short-term memory through integration with previouslystored materi<strong>al</strong>. An implication of the activeprocessing of information in STWM is that the storedrepresentation transferred to long-term memory isto some extent an interpr<strong>et</strong>ation of the experience,rather than a veridic<strong>al</strong> representation of the action asit transpired. That is, memory is not like a videotap<strong>et</strong>hat can be replayed at any time. Rather, in encodingan event, the individu<strong>al</strong> constructs a coherent story ofthe experience out of fragments of memories and perceptionsthat are combined or blended (Baker-Ward,Ornstein, & Principe, 1997; Bowers & Sivers, 1998).STWM increases substanti<strong>al</strong>ly with development.This increase, however, does not appear to beattributable to simple changes in the capacity of shorttermmemory storage (Schneider & Bjorklund, 1998).If this were the case, age-related improvements wouldbe observed in memory span regardless of the type ofinformation that is to be remembered. In contrast tosuch a domain-gener<strong>al</strong> improvement, memory span isdomain-specific; that is, it differs for an individu<strong>al</strong> onthe basis of his or her interest and knowledge in thecontent of the to-be-remembered information. Moreover,knowledge appears to affect memory span byincreasing speed of processing. Although neurologic<strong>al</strong>development (such as myelinization of areas ofthe cortex) is likely to have some effects on speedof processing, this important variable must be understoodin terms of experienti<strong>al</strong> as well as maturation<strong>al</strong>influences (Schneider & Bjorklund, 1998).It is clear from this brief overview that simpleexposure to an event, even a s<strong>al</strong>ient person<strong>al</strong> experience,is not sufficient to insure compl<strong>et</strong>e encodingof the experience. The likelihood that a d<strong>et</strong>ail of anevent will be encoded and subsequently become a partof permanent memory is influenced significantly by anindividu<strong>al</strong>’s prior knowledge and the nature of the experienceitself. Development<strong>al</strong> differences in memorycapacity must be interpr<strong>et</strong>ed within the context of achild’s knowledge in a particular domain and the extentto which the event in question is consistent withor distinct from that knowledge. Simply put, what we<strong>al</strong>ready know d<strong>et</strong>ermines, to a large extent, what wecan and cannot remember.Memory for Stressful and Traumatic ExperiencesGiven that child witnesses will undoubtedly beasked to remember experiences that are quite stressful,considerable research has been devoted to d<strong>et</strong>erminingthe effects of high levels of stress on theencoding and storage of information in memory (seeCicch<strong>et</strong>ti & Toth, 1998 for a review). Many of theseinvestigations have examined children’s memory formedic<strong>al</strong>ly indicated procedures that typic<strong>al</strong>ly invokesome level of distress among children, including emergencyroom treatment (P<strong>et</strong>erson & Bell, 1996) andurinary catherization-procedures (Goodman & Quas,1997; Merritt, Ornstein, & Spicker, 1994). Althoughthe relation b<strong>et</strong>ween stress and rec<strong>al</strong>l has been thesubject of considerable past debate, the accumulationof evidence now supports the conclusion that, when asignificant relation is reve<strong>al</strong>ed, higher levels of stressare predictive of lower levels of remembering (seeOrnstein, 1995). These findings are consistent withexpectations based on the Yerkes–Dodson law thatpostulates an inverted-U–shaped relation b<strong>et</strong>weenarous<strong>al</strong> and performance. In gener<strong>al</strong>, when stress ismoderate, rec<strong>al</strong>l may be enhanced; in contrast, whenstress is very high (or very low), memory performanceis debilitated (see Gold, 1987; Pezdek & Taylor, inpress).It should be noted that a considerable amountof complexity is masked by this gener<strong>al</strong> conclusionregarding the relation b<strong>et</strong>ween stress and rec<strong>al</strong>l.As noted earlier, indicators of stress and memoryperformance are not <strong>al</strong>ways associated, even whenpresumably stressful events are under investigation.Moreover, different indicators of stress are not <strong>al</strong>wayscorrelated with each other within the same investigationand the relation b<strong>et</strong>ween stress and memorymay differ across multiple assessments. Merritt <strong>et</strong> <strong>al</strong>.,for example, found that a fine-grained observation<strong>al</strong>measure of children’s distress during a painful invasivemedic<strong>al</strong> procedure was correlated with children’stot<strong>al</strong> rec<strong>al</strong>l for component features of the proceduresduring an interview conducted shortly after the experience.This measure, however, was not associatedwith memory performance as assessed 6 weeks afterthe procedure. Moreover, a physiologic<strong>al</strong> indicator ofstress obtained through s<strong>al</strong>ivary cortisol assays wasunrelated to the observation<strong>al</strong> data or to measures ofmemory.At this point in time, it appears clear that theeffects of stress on children’s memory must be interpr<strong>et</strong>edwithin the context of multiple influences onperformance. Individu<strong>al</strong> difference variables in areassuch as temperament (specific<strong>al</strong>ly reactivity to stress),psychopathology (e.g., depression or gener<strong>al</strong>ized anxi<strong>et</strong>y),coping style, and even parent<strong>al</strong> attachmentmay mediate the relation b<strong>et</strong>ween stress and memory(Goodman <strong>et</strong> <strong>al</strong>., 1997; Howe, 1998). In addition,
Children’s Testimony 163intervening experiences b<strong>et</strong>ween a stressful event anddelayed memory assessments, including relevant natur<strong>al</strong>lyoccurring or therapeutic adult–child conversations,may affect children’s understanding and consequently,the likelihood that specific d<strong>et</strong>ails of theexperience will be subsequently reported (Fivush,1998). Also, the constructive processes that characterizememory for everyday events apply to traumaticexperiences, as well. For example, in contrast to thewidespread belief that “flashbulb” memories for verys<strong>al</strong>ient events remain vivid and accurate after longdelays (see Brown & Kulik, 1977), prospective investigationsof children’s reports of the Ch<strong>al</strong>lengerexplosion reve<strong>al</strong>ed both forg<strong>et</strong>ting and reconstructiveerrors (Warren & Smartwood, 1992; <strong>al</strong>so seeMcCloskey, Wible, & Cohen, 1988, for similar findingswith adults).The person<strong>al</strong> significance of the emotionsaroused by an event is another important factor inunderstanding how well stressful experiences will beremembered. In a review of laboratory research examiningthe linkage b<strong>et</strong>ween emotion and memory,Bowers and Sivers (1998) conclude that there are twoconsistent findings. First, when the emotions that arearoused are relevant to or caused by an experience,memory for that information, particularly informationthat is perceived of as person<strong>al</strong>ly meaningful, isenhanced. Second, when the emotion aroused is irrelevantto the experience (such as in the case of testor performance anxi<strong>et</strong>y or chronic gener<strong>al</strong>ized anxi<strong>et</strong>y)memory is reduced or diminished. M<strong>al</strong>treatedchildren, for example, tend to focus their attention onaggressive stimuli and have difficulty screening outdistracting information, presumably because of theirongoing experience of abuse. This effect has obviousimplications for encoding and remembering aggressiveversus more neutr<strong>al</strong> information (Pollack,Cicch<strong>et</strong>ti, Klorman, & Brumaghim, 1997; Pollack,Cicch<strong>et</strong>ti, & Klorman, 1998).In addition, aspects of emotion<strong>al</strong> developmentaffect an individu<strong>al</strong>’s person<strong>al</strong> experience, such thatthe same event may induce very different degrees ofarous<strong>al</strong> for children with different histories. As anexample, children who demonstrate insecure parent–child attachment behavior have been shown to havehigher s<strong>al</strong>ivary cortisol levels (indicating greater distress)when faced with a stressful experience than dochildren who have a more secure attachment relationship(Hertsgaard, Gunnar, Erickson, & Nachmias,1995). Further, consistent with the gener<strong>al</strong> findingthat high levels of arous<strong>al</strong> negatively impact rec<strong>al</strong>l,children with insecure attachment designationsmake more errors when rec<strong>al</strong>ling a stressful medic<strong>al</strong>procedure than do more securely attached children(Goodman & Quas, 1997).Some questions remain about the extent to whichwork with an<strong>al</strong>og events such as medic<strong>al</strong> procedurescan be gener<strong>al</strong>ized to memory for the traumatic experiencesthat are the subject of leg<strong>al</strong> proceedings.Despite similarities in discomfort and bodily contact,there are <strong>al</strong>so important differences in parent<strong>al</strong>lysanctioned and medic<strong>al</strong>ly indicated treatments andthe violation and violence of abuse. Non<strong>et</strong>heless, recentwork on memory for trauma has investigatedwh<strong>et</strong>her or not the brain represents and stores traumaticexperiences differently than everyday autobiographic<strong>al</strong>memories (see Nadel & Jacobs, 1998). Thedata suggest that various aspects of an experience arerepresented in different parts of the brain. The amygd<strong>al</strong>ais particularly important in memory for emotion<strong>al</strong>lycharged events. In contrast, the hippocamp<strong>al</strong>formation functions to integrate event memories representeddiffusely in different brain areas. Stress hasa differenti<strong>al</strong> impact on these two separate areas andhence on <strong>al</strong>ternative aspects of explicit memory. Stressappears to enhance the function of the amygd<strong>al</strong>a, resultingin the strengthening of memories served by thisstructure. Alternatively, too little or too much corticosterone,the hormone produced by stress, appearsto disrupt the function of the hippocampus and hencereduces the likelihood that the d<strong>et</strong>ails of an experienceare integrated into a coherent memory. As aresult, memories for trauma may be represented asfragments, rather than as integrated event sequences(Nadel & Jacobs, 1998).Other work indicates that chronic stress may actu<strong>al</strong>lylead to changes in brain structure, specific<strong>al</strong>lyin areas associated with learning and memory andthat young children, because of their rapidly developingbrains, may be particularly vulnerable to thiseffect (Nelson & Carver, 1998). It is interesting thatwomen with histories of childhood physic<strong>al</strong> and sexu<strong>al</strong>abuse show elevated physiologic<strong>al</strong> responses tostress when compared with nonabused control participants(Heim <strong>et</strong> <strong>al</strong>., 2000). As research in the areaof brain function, trauma, and memory progresses,it may help to explain some of the unusu<strong>al</strong> symptomsof posttraumatic stress syndrome (i.e., intrusiverecollections of the trauma, nightmares, and flashbackmemories) and the often fragmentary and jumblednature of memory for highly traumatic experiences.Moreover, these findings are consistent withthe increased likelihood of generating f<strong>al</strong>se memoriesamong women with abuse histories (Bremner, Shobe,
164 <strong>Gordon</strong>, Baker-Ward, and Ornstein& Kihlstrom, 2000). Because constructive processesapply to memories for trauma as well as everyday experiences,events that are stored as fragments may beparticularly vulnerable to reconstructive error.Storage: What G<strong>et</strong>s Into Memory MayVary in StrengthGiven that d<strong>et</strong>ails of an event are encoded andstored, many factors can potenti<strong>al</strong>ly influence thestrength of the resulting trace in memory and consequentlythe ease with which information may ber<strong>et</strong>rieved at a later time. The extent to which informationis embedded in a coherent, well-organized knowledgestructure is one important d<strong>et</strong>erminant of thelikelihood that it can be subsequently r<strong>et</strong>rieved. Informationthat is less strongly elaborated within such asemantic n<strong>et</strong>work is <strong>al</strong>so more subject to suggestibility(Pezdek & Roe, 1995). Stronger representations maybe readily r<strong>et</strong>rieved, even in response to open-endedquestions, whereas weaker traces are more likely tobe forgotten or may require more specific questionsto be remembered. Sever<strong>al</strong> factors have been shownto influence the strength of the memory representationincluding wh<strong>et</strong>her one actively participates in anexperience or watches others, the age or development<strong>al</strong>status of the individu<strong>al</strong>, and the amount of exposur<strong>et</strong>o the events in question.Participant Versus ObserverConsiderable research has documented that childrenremember events in which they participate b<strong>et</strong>terthan those that they merely witness (Bauer <strong>et</strong> <strong>al</strong>.,1998). Active participation results in more d<strong>et</strong>ailedmemories (Baker-Ward, Hess, & Flanagan, 1990) aswell as increased resistance to suggestion or misleadinginformation (Rudy & Goodman, 1991), and thisis especi<strong>al</strong>ly true for preschool children. Bauer <strong>et</strong> <strong>al</strong>.(1998) argue that the superior memory that typic<strong>al</strong>lyresults from active participation may be a functionof the tendency to pay attention to and encode featuresof events that are most relevant to the self.This is consistent with the argument that person<strong>al</strong>lyexperienced events have greater trace strength becauseof the extent to which encoding benefits fromgreater knowledge. Thus, Bauer <strong>et</strong> <strong>al</strong>. hypothesiz<strong>et</strong>hat certain types of witnessed events, such as thoseinvolving violence or abuse of a significant other,might be remembered as well as other more benignparticipatory experiences. Supporting this possibility,Baker-Ward <strong>et</strong> <strong>al</strong>. (1990) found that the advantages ofparticipation over observation in children’s reportsof a laboratory play event were present when randomlyassigned classmates were observed, but werenot present when the observed individu<strong>al</strong>s were closefriends.AgeWith increasing age, there are correspondingchanges in a vari<strong>et</strong>y of cognitive functions that affectthe acquisition and storage of information in the memorysystem. Other influences being equ<strong>al</strong>, older childrenwill acquire more information from comparableexposure to an event and will maintain a strongermemory trace than will younger children. This effectcan be attributed to age-related changes in processingspeed as well as the availability of more efficientstrategies and an increased knowledge base (Ornstein<strong>et</strong> <strong>al</strong>., 1997). Moreover, it is likely that the age differencesin forg<strong>et</strong>ting that are commonly found in studiesof autobiographic<strong>al</strong> memory (e.g., Baker-Ward<strong>et</strong> <strong>al</strong>., 1993) reflect corresponding decreases in thestrength of the underlying memory representations.Thus, the strength of the memory representation diminishesless over time for older as compared withyounger children. As a result, stored representationsof experiences become more difficult to access overtime, especi<strong>al</strong>ly among younger individu<strong>al</strong>s (Howe &O’Sullivan, 1997).A recent theor<strong>et</strong>ic<strong>al</strong> perspective on children’smemory, Fuzzy-Trace Theory (e.g., Brainerd &Reyna, 1990; Reyna & Brainerd, 1995), postulates development<strong>al</strong>differences in the nature of stored eventrepresentations. Within the context of this framework,every experience is thought to result in the establishmentof multiple, independent memory traces.These representations can be ordered on a continuumranging from verbatim traces, which are fairlyexact representations of specific aspects of the event,to more imprecise “fuzzy traces,” which preserve onlythe gist of the experience. Although even young childrencan extract gist, they are biased toward encodingand r<strong>et</strong>rieving verbatim traces until the early elementaryschool years. In contrast, whereas olderchildren and adults <strong>al</strong>so store verbatim traces, theyare biased toward extracting gist. Because verbatimtraces decay more rapidly than gist does, preschoolersdemonstrate greater rates of forg<strong>et</strong>ting than do olderindividu<strong>al</strong>s.
Children’s Testimony 165Amount of ExposureVariations in the frequency and duration of exposur<strong>et</strong>o an event are associated with differences inthe strength of the resulting memory trace. Thus, inthe case of a single occurrence of an event, the longerthe exposure time to relevant features, the strongerwill be the resulting representation in memory. Traumaticevents to which one is exposed only brieflymay be exceptions to this rule because of their distinctiveness(Howe, 2000). Hence, aspects of unique,traumatic events may som<strong>et</strong>imes be well remembered(e.g., Bahrick <strong>et</strong> <strong>al</strong>., 1998), particularly when theseexperiences can be openly discussed (Fivush, 2000).Very young children may not have the requisite experiencefrom which to construct a coherent representationof this type of event, however, (Bauer <strong>et</strong> <strong>al</strong>., 1998)and as a result, their rec<strong>al</strong>l may be more fragmentedand less consistent than that of older children.Another factor associated with greater tracestrength is the rep<strong>et</strong>ition of an experience. All othercharacteristics being equ<strong>al</strong>, repeated exposures to astimulus will yield stronger representations and consequentlyb<strong>et</strong>ter rec<strong>al</strong>l (Marche, 1999; Pezdek & Roe,1995). Fivush and Hammond (1989), for example, exposed2-year-old children to a novel laboratory playevent, and provided h<strong>al</strong>f of the children with a reenactmentof the play scenario after a delay of twoweeks. Memory was assessed among <strong>al</strong>l the participants3 months after the initi<strong>al</strong> visit through a reenactmentof the previous playroom experience. Thechildren who received the addition<strong>al</strong> exposure to theevent, in comparison to those who did not, demonstratedmore rec<strong>al</strong>l at the fin<strong>al</strong> assessment. Similarly,Powell, Roberts, Ceci, and Hembrooke (1999) recentlyreported that children who experienced a repeatedevent were more accurate than those who wereexposed to one presentation when memory for thecomponents of the event that remained consistentacross rep<strong>et</strong>itions was examined.Fuzzy-Trace theory provides an explanation forthe benefici<strong>al</strong> effects of repeated exposure to eventson memory among young children. When an event isrepeated with some variation in d<strong>et</strong>ails, multiple verbatimtraces containing conflicting information areproduced. Because the verbatim traces are now inconsistent,young children may be encouraged to relymore on gist traces. This can result in less forg<strong>et</strong>tingbecause gist traces are maintained over longer periodsof time than are verbatim traces (Powell <strong>et</strong> <strong>al</strong>., 1999).It should be noted, however, that repeated exposuresto an event can have negative as well as positiveeffects on children’s memory performance. Childrenwho repeatedly experience an event form “scripts”(defined as generic representations of familiar events:Nelson, 1986) for the common features across theepisodes. Scripts enhance rec<strong>al</strong>l for the gener<strong>al</strong> structureof the experience at the expense of memoryfor particular episodes of the event (Hudson, 1990;Powell & Thomson, 1997). Thus, rec<strong>al</strong>l of repeated experiencesmay represent the child’s memory of what“usu<strong>al</strong>ly happens” rather than the d<strong>et</strong>ails of a specificepisode. Powell <strong>et</strong> <strong>al</strong>. (1999), for example, found lowerlevels of rec<strong>al</strong>l for features of an event that variedacross rep<strong>et</strong>itions than for features that were consistent.Based on an<strong>al</strong>ysis of the patterns of the children’sintrusion errors, they concluded that rep<strong>et</strong>ition mayincrease the likelihood of confusing what happenedwhen.Storage: The Status of Information in MemoryChanges Over TimeInformation that has been successfully encodedand stored in memory is not frozen. Rather, memoryis subject to a number of important influencesover time and the status of information in memorycan be <strong>al</strong>tered in the interv<strong>al</strong> b<strong>et</strong>ween the occurrenceof the event and the memory “test.” Stored informationcan be updated or modified, and the strength ofthe memory trace may increase or decrease. Thesechanges can occur through sever<strong>al</strong> processes with potenti<strong>al</strong>lydifferent consequences for the accuracy ofchildren’s subsequent reports. Both the passage oftime and prior knowledge exert a substanti<strong>al</strong> influenceon the underlying memory representation. Moreover,children may be exposed to a vari<strong>et</strong>y of experiences inthe time b<strong>et</strong>ween encoding and rec<strong>al</strong>l, some of whichact to strengthen memory, whereas others interferewith rec<strong>al</strong>l performance.The Length of the Delay Interv<strong>al</strong>As discussed in the preceding section, memorytraces can d<strong>et</strong>eriorate over time. Accordingly, themore closely the interview or testimony follows theevent, the greater the likelihood of obtaining accurateand compl<strong>et</strong>e accounts of the d<strong>et</strong>ails of children’sexperiences. Unfortunately, it is common for childwitnesses to provide testimony weeks, months, andeven years after the events in question. Hence, theeffects of the delay interv<strong>al</strong> on event memory is an
166 <strong>Gordon</strong>, Baker-Ward, and Ornsteinimportant consideration in ev<strong>al</strong>uating a child’s capacityto testify.Considerable research has shown that preschoolchildren’s rec<strong>al</strong>l of experienced events, both traumaticand nontraumatic, can be quite good even over relativelylong periods of time (Bauer <strong>et</strong> <strong>al</strong>., 1998; Fivush& Hammond, 1990; Fivush & Shukat, 1995; Howard<strong>et</strong> <strong>al</strong>., 1997). P<strong>et</strong>erson and Rideout (1998), for example,demonstrated that children who were at least26 months of age at the time of an accident<strong>al</strong> injuryand visit to the emergency room, accurately rec<strong>al</strong>ledthe d<strong>et</strong>ails of these experiences even after a 2-yeardelay. Similarly, Schwarzmueller <strong>et</strong> <strong>al</strong>. (1996) foundthat 8-year-old children could accurately rec<strong>al</strong>l eventsthat had occurred when they were as young as 3 1 /2years old. Despite this remarkable display of memory,there was evidence of forg<strong>et</strong>ting over these long delaysin each of these studies, consistent with forg<strong>et</strong>tingcurves that reflect children’s memory in gener<strong>al</strong> (Kail,1989; Schneider & Pressley, 1997). Moreover, othershave documented that the younger the child themore vulnerable he or she is to forg<strong>et</strong>ting over time(e.g., Baker-Ward <strong>et</strong> <strong>al</strong>., 1993; Brainerd, Kingman, &Howe, 1985; Goodman <strong>et</strong> <strong>al</strong>., 1990; Ornstein, <strong>Gordon</strong>,& Larus, 1992; Poole & White, 1993). The exception tothis appears to be long-term memory for particularlydistressful experiences. P<strong>et</strong>erson (1999), for example,found no age differences in forg<strong>et</strong>ting for an accident<strong>al</strong>injury among 2- to 13-year-old children.Prior KnowledgeIn gener<strong>al</strong>, as the interv<strong>al</strong> b<strong>et</strong>ween encodingand r<strong>et</strong>riev<strong>al</strong> increases, the memory trace becomesweaker, and it is increasingly likely that the informationin memory will be <strong>al</strong>tered as a result of priorknowledge. Research on scripts documents one wayin which memory can change as a result of priorknowledge. As memory for a particular episode fadesover time, children are likely to assume that whatusu<strong>al</strong>ly happens actu<strong>al</strong>ly occurred in this instance.Myles-Worsley, Cromer, and Dodd (1986), for example,demonstrated that over a 5-year period, children’smemories of events experienced in a preschool classincreasingly came to be reconstructions involving acombination of actu<strong>al</strong> remembered information andgener<strong>al</strong> knowledge about similar experiences. Thus,with the passage of time, the d<strong>et</strong>ails of a particularexperience may be forgotten and the information inmemory <strong>al</strong>tered to be more consistent with what achild knows usu<strong>al</strong>ly happens. To the extent that theparticular episode is consistent with the script, thereport may remain accurate <strong>al</strong>though d<strong>et</strong>ail may belost.Such reliance on event scripts, however, can leadto inaccurate reports when a specific experience isinconsistent with gener<strong>al</strong> expectations for an event.Ornstein and colleagues (Ornstein, Merritt <strong>et</strong> <strong>al</strong>.,1998) examined children’s memory for a speci<strong>al</strong>ly designedpediatric examination, in which some typic<strong>al</strong>procedures (e.g., checking the heart) were omittedand some novel components (e.g., measuring headcircumference) were included. Few intrusions wereobserved at the initi<strong>al</strong> interview; however, after a12-week delay, the children in some conditions spontaneouslyreported more than 20% of the expectedbut-omittedfeatures, while reporting essenti<strong>al</strong>ly noother type of f<strong>al</strong>se information. In this situation, thechange over time in the children’s accounts resultedin the inclusion of inaccurate information. Moreover,the f<strong>al</strong>sely reported actions, because they were reportedwithout prompting and included as much elaboratived<strong>et</strong>ail as the correctly reported components,m<strong>et</strong> criteria for credibility (see <strong>Gordon</strong> & Follmer,1994).Changes in Knowledge and BeliefsAn addition<strong>al</strong> type of change over time is observedwhen stored memories of past experiencesbecome more consistent with individu<strong>al</strong>s’ currentknowledge and beliefs. Knowledge gained at a laterdate (i.e., as a child develops and learns more abouthow the world operates) may influence rec<strong>al</strong>l long afterthe event in question has occurred (Ornstein <strong>et</strong> <strong>al</strong>.,1991). Similarly, individu<strong>al</strong>s whose attitudes chang<strong>et</strong>hrough either natur<strong>al</strong> or experimenter-provided experiencessubsequently report their initi<strong>al</strong> attitudesin a manner more consistent with their changedviews (see Ross, 1989). In these cases, however, itmust be recognized that what is remembered at <strong>al</strong>ater time represents a reinterpr<strong>et</strong>ation of the informationthat was origin<strong>al</strong>ly encoded into memory,and, as such, the d<strong>et</strong>ails rec<strong>al</strong>led may be substanti<strong>al</strong>ly<strong>al</strong>tered. Greenhoot (2000), for example, readkindergarten children stories that included ambiguousactions by a centr<strong>al</strong> character, and subsequentlyprovided information about the character’s typic<strong>al</strong>behavior. Experimenter-provided information abouta character’s typic<strong>al</strong> behavior affected kindergartenchildren’s subsequent memory of ambiguous actionsthat occurred within a previously presented story. On
Children’s Testimony 167the basis of the new information, the children <strong>al</strong>teredtheir previous reports of the story. Thus, it ispossible that experiences intervening b<strong>et</strong>ween youngwitnesses’ experiences and their subsequent leg<strong>al</strong> testimonycould provide an interpr<strong>et</strong>ative context thatmight similarly <strong>al</strong>ter recollections.If memories can be <strong>al</strong>tered through the provisionof addition<strong>al</strong> information, stereotyping can beexpected to produce even stronger effects. Leichtmanand Ceci (1995) found that children who often heard“Sam Stone” described as a clumsy person reportedthat he had broken the toys that were found to be damagedafter he made a brief and uneventful visit to theclassroom. Such reconstructive processes could obviouslyhave negative effects on the accuracy of children’stestimony. For example, young witnesses whoare told that their help is needed in keeping a badperson from hurting other children (as occurred inthe prosecution of Kelly Michaels; see Ceci & Bruck,1995), may selectively interpr<strong>et</strong> and report neutr<strong>al</strong> informationas consistent with the stereotype of the defendantas a “bad person.”It should be noted that the process of <strong>al</strong>terationto the stored representation could feasibly safeguardas well as threaten the accuracy of a subsequent report.For example, information encoded after an eventcould enhance a child’s understanding of the experienceby providing links b<strong>et</strong>ween component featuresor by adding elaborative d<strong>et</strong>ail to the representation.This modified event representation mightincrease trace strength and could facilitate r<strong>et</strong>riev<strong>al</strong>(see Baker-Ward <strong>et</strong> <strong>al</strong>., 1997).Exposure to Misleading InformationConsiderable research has demonstrated that individu<strong>al</strong>sexposed to information that is misleading orinconsistent with their experiences during the interv<strong>al</strong>b<strong>et</strong>ween encoding and r<strong>et</strong>riev<strong>al</strong> typic<strong>al</strong>ly performless well during memory interviews than do thosewho do not receive such information (e.g., Loftus,1979; Loftus & P<strong>al</strong>mer, 1974; Principe, Ornstein,Baker-Ward, & <strong>Gordon</strong>, 2000; Roberts <strong>et</strong> <strong>al</strong>., 1997).Exposure to misleading information can occur duringthe course of memory interviews (in the form ofsuggestive or very specific questions), before the interviewoccurs (e.g., conversations with parents or otherfamily members), or in-b<strong>et</strong>ween multiple interviews(e.g., some therapeutic procedures, television newsor newspaper reports, reading stories about similarevents).Despite the fact that <strong>al</strong>most everyone is to someextent vulnerable to this type of suggestion, preschoolchildren have been consistently found to be moreso than older children and adults (Ceci & Bruck,1995; Bruck <strong>et</strong> <strong>al</strong>., 1998; Bruck, Ceci, Francouer, &Barr, 1995; Loftus & Pickrell, 1995; Poole & White,1993). Indeed, Leichtman and Ceci (1995) demonstratedthat preschool children will provide elaborated<strong>et</strong>ails about things that did not happen whenthey are subjected to pre- and postevent suggestiveinformation. Moreover, some children will hold totheir misguided beliefs even in the face of attemptsby parents and experimenters to convince them thatthese things never happened (Ceci, Huffman, Smith,& Loftus, 1994; Levine, Stein, & Liwag, 1999). Bruck<strong>et</strong> <strong>al</strong>. (1995) extended this work by showing that misleadinginformation (i.e., the shot didn’t hurt; the childdidn’t cry much) provided at the time of a physic<strong>al</strong> examination,coupled with repeated interviews that containedsuggestions consistent with the misleading information,led to changes in 5-year-olds’ perceptionsof how much a previous innoculation had hurt, howmuch they had cried, and who had administered theshot.A vari<strong>et</strong>y of other factors have been shown toincrease the tendency to <strong>al</strong>ter children’s reports andpresumably their memory representations. The timingof exposure to misleading information is one example.Information that is inconsistent with the child’sexperience has been shown to be more d<strong>et</strong>riment<strong>al</strong>to memory accuracy, in the absence of repeated suggestiveinterviews, when it is provided just before thememory interview rather than earlier in the delay interv<strong>al</strong>(Marche, 1997; Warren & Lane, 1995). This reflectsthe fact that memories are more vulnerable todistortion when the memory trace has weakened withthe passage of time. That is, it is easier to recognize andreject information that is inconsistent with our experienceswhen that information is provided more closelyin time to the event in question. Indeed, the study byCeci, Huffman <strong>et</strong> <strong>al</strong>. (1994) indicates that suggestibilityeffects increase over the course of interviews thatare repeated over time.A second factor that has been shown to increasesuggestibility is the perceived authority or credibility(or both) of the person providing the misinformation.Children are more suggestible, for example, whenthe misleading information is presented by an adult,as opposed to another child (Ceci, Ross, & Toglia,1987), when the adult is perceived of as being morerather than less credible, knowledgeable, or authoritative(Simpson & Guttentag, 1996; Templ<strong>et</strong>on & Hunt,
168 <strong>Gordon</strong>, Baker-Ward, and Ornstein1997; Toglia & Ross, 1991), and when the misleadinginformation is provided by a familiar and trusted personversus a stranger (Jackson & Crockenberg, 1998).To what extent do inaccuracies in children’s reportsstem from changes in stored memory representationsfollowing exposure to inconsistent informationas opposed to compliance or other soci<strong>al</strong> demandcharacteristics? From the perspective of Fuzzy-TraceTheory, there is a cognitive basis for the persistence off<strong>al</strong>se memories (Brainerd, Reyna, & Brandse, 1995).As presented in this framework, information aboutevents is stored as both precise verbatim traces andas the gist of the experiences. Questions about eventsthat actu<strong>al</strong>ly occurred can cue the r<strong>et</strong>riev<strong>al</strong> of verbatimmemories, resulting in the definite recollection ofhaving experienced the action in question. In contrast,children who incorrectly accept misleading questions(i.e., produce f<strong>al</strong>se <strong>al</strong>arms) are likely to have r<strong>et</strong>rievedthe gist of an action that is similar to that referencedin the misleading probe. A vague feeling of familiaritywith the suggested action may result, because ofsome congruence b<strong>et</strong>ween the gist trace and the informationconveyed in the misleading question. Astime passes and verbatim traces become inaccessible,memory-based f<strong>al</strong>se <strong>al</strong>arms may be as stable as correctresponses. Hence, when misleading probes activaterelated gist traces, simply testing a child’s memory canresult in the creation of stable incorrect responses aswell as the maintenance of correct responses. The persistenceof incorrect responses in some circumstancesraises doubts about the v<strong>al</strong>idity of the use of consistencyacross interviews as a criterion for credibility(Brainerd & Mojardin, 1998).In a recent discussion of the underlying basesof suggestibility, Bruck and Ceci (1999) note thatsome cognitive distortion is likely, in light of development<strong>al</strong>changes in basic memory processes. Further,they note that in sever<strong>al</strong> investigations, childrenhave continued to provide f<strong>al</strong>se reports evenwhen they are asked to substantiate their claims orare given an addition<strong>al</strong> opportunity to respond correctly.They <strong>al</strong>so note, however, that previous f<strong>al</strong>sememories tend to fade over time when the suggestionshave ceased. Bruck and Ceci conclude by presentingthe hypothesis that “...a more d<strong>et</strong>ailed inspectionof children’s responses over time will reflecta more complex condition with a comingling ofsoci<strong>al</strong> (compliance) and cognitive (memory) factors...children may start out knowingly complying to suggestions,but with repeated suggestive interviews, theymay come to believe and incorporate the suggestionswith their memories” (p. 434). It is interesting to not<strong>et</strong>hat recent research indicates that these “f<strong>al</strong>se memories”may not persist over very long periods of time.Huffman, Crossman, and Ceci (1997) reported thatafter a 2-year delay, children “recanted” previouslyestablished f<strong>al</strong>se memories 77% of the time, whereasthey maintained accurate memories 78% of th<strong>et</strong>ime.R<strong>et</strong>riev<strong>al</strong>: R<strong>et</strong>riev<strong>al</strong> Is Not PerfectThe fin<strong>al</strong> phase of the memory process involvesr<strong>et</strong>riev<strong>al</strong> of the stored information. An origin<strong>al</strong> assumptionof the information-processing approachwas that, if information entered long-term memory,it remained there permanently and could be r<strong>et</strong>rievedat any time, assuming that an effective cuewas present (Shiffrin & Atkinson, 1969). In contrastto this view, it now appears that some memoriesare at least temporarily unrecoverable in theirorigin<strong>al</strong> form. Some explanations for r<strong>et</strong>riev<strong>al</strong> failuresemphasize information-processing d<strong>et</strong>erminantsof rec<strong>al</strong>l, including the organization of the event representationin memory and the absence of effectiver<strong>et</strong>riev<strong>al</strong> strategies. From this perspective, thechild’s role in a rec<strong>al</strong>l failure is passive; the lack ofr<strong>et</strong>riev<strong>al</strong> is attributed solely to the absence of theneeded information-processing components. Otherapproaches focus on psychologic<strong>al</strong> processes that activelyprohibit the r<strong>et</strong>riev<strong>al</strong> of experiences associatedwith psychologic<strong>al</strong> distress.Factors Associated With the Likelihood of R<strong>et</strong>riev<strong>al</strong>Assuming that a representation of an event remainsin memory, sever<strong>al</strong> major factors are importantin d<strong>et</strong>ermining wh<strong>et</strong>her the information will bereported during a memory interview. First, as notedearlier, the absence of an effective r<strong>et</strong>riev<strong>al</strong> cue isa widely accepted reason for r<strong>et</strong>riev<strong>al</strong> failure. Aninterviewer’s questions can be seen as representingone type of r<strong>et</strong>riev<strong>al</strong> cue. In order to be effective, acue must be part of the context that was present atthe time of encoding (Tulving & Thompson, 1973).The encoding context involves multiple dimensions,including information that was presented <strong>al</strong>ong withthe targ<strong>et</strong> materi<strong>al</strong>, prior memories and knowledg<strong>et</strong>hat were activated during the encoding of the targ<strong>et</strong>information, the physic<strong>al</strong> environment in which encodingoccurred, and the individu<strong>al</strong>’s intern<strong>al</strong> state atthe time of encoding. As an example, simply being
Children’s Testimony 169interviewed in the room in which an experience tookplace, in comparison to being interviewed in a neutr<strong>al</strong>location has been shown to increase 5- to 7-yearoldchildren’s rec<strong>al</strong>l (Priestley, Roberts, & Pipe, 1999,Experiment 1).Another factor that influences r<strong>et</strong>riev<strong>al</strong> is theextent to which the representations of aspects ofthe event are embedded in an extensive knowledgestructure (Bjorklund, 1987). Such a knowledge structureis an important d<strong>et</strong>erminant of trace strength.In addition, when information is incorporated ina rich knowledge base, a greater number of cuesare effective in activating the memory of the targ<strong>et</strong>materi<strong>al</strong>.An addition<strong>al</strong> influence on the likelihood of r<strong>et</strong>riev<strong>al</strong>is the distinctiveness of the event under consideration(see Howe, 2000). Distinctiveness is definedas the extent to which to-be-remembered informationstands out from a background context. Whenone item in a list of to-be-remembered words representsa different semantic category than the remainingitems, for example, the distinctive item is memorizedmore quickly and r<strong>et</strong>ained longer than theothers. This phenomenon, which has been studied extensivelyin adults, is described as the von RestorffEffect. Howe argues that distinctiveness is equ<strong>al</strong>lyimportant in natur<strong>al</strong> environments, <strong>al</strong>though this phenomenonis difficult to investigate in a controlled manner.In an extensive investigation of 2- to 13-year-oldchildren’s memory for an injury and resulting hospit<strong>al</strong>treatment, P<strong>et</strong>erson (1999) provides evidence for theimportance of distinctiveness in memory. Two yearsafter the event, the children rec<strong>al</strong>led more d<strong>et</strong>ails regardingthe injury than the treatment. P<strong>et</strong>erson notedthat the majority of the children had visited the emergencyroom on multiple occasions, and argues that theuniqueness of the injury contributed to the greater rec<strong>al</strong>lof this component of the event, as compared withthe emergency room procedures.Fin<strong>al</strong>ly, the extent to which children can use r<strong>et</strong>riev<strong>al</strong>strategies, defined as planful operations usedto access stored information (Schneider & Bjorklund,1998), affects the likelihood of rec<strong>al</strong>l. As reviewed bySchneider and Bjorklund, the ability to actively searchthe long-term memory store effectively develops relativelylate in childhood. Hence, preschoolers and evenyounger elementary school-aged children cannot beexpected to spontaneously use r<strong>et</strong>riev<strong>al</strong> strategies inrec<strong>al</strong>ling experiences, and their reports may consequentlyprovide a re<strong>al</strong> underestimation of what theyknow about an event. To some extent, an effectiveinterviewer may compensate in part for this development<strong>al</strong>limitation by providing young witnesses withspecific directions for searching their memories (seeFisher & Geiselman, 1992).Hidden Memories of Childhood AbuseAn emotion<strong>al</strong>ly charged controversy surroundsthe possibility that childhood sexu<strong>al</strong> abuse may notbe rec<strong>al</strong>led because of the operation of psychologic<strong>al</strong>defense mechanisms and that the memoriesmay be subsequently recovered during psychotherapyor in the context of other experiences (cf. Alpert,Brown, & Courtois, 1998; Ornstein, Ceci, & Loftus,1998). A full discussion of this complex issue is wellbeyond the scope of this paper (for reviews, seePope & Brown, 1996; Putnam, 1997; Roediger &Bergman, 1998). It should be noted, however, thatchildren’s responses to painful experiences have clearimplications for information-processing, regardless ofwh<strong>et</strong>her or not the concept of psychologic<strong>al</strong> repressionis accepted. Dissociation (i.e., isolating the selffrom a painful thought or experience) is very commonamong children, particularly preschoolers, andis thought to be an adaptive mechanism for copingwith stress. To the extent that this distancing curtailsthe encoding of an experience, memory failurescan result (Eisen & Goodman, 1998). In thisinstance, components of the event simply do notexist in memory. Dissociation could <strong>al</strong>so limit accessto representations that exist in memory andare hence potenti<strong>al</strong>ly r<strong>et</strong>rievable. This could ariseif memories for traumatic experiences exist as relativelyisolated representations (for example, becausememories for the traumatic experience arenot linked to other memories involving the critic<strong>al</strong>components of the event) and hence the number ofcues that can gain access to the representation is reduced.In contrast to dissociation, Eisen and Goodman(1998) suggest that it is <strong>al</strong>so possible that individu<strong>al</strong>smay banish memories of threatening experiencesfrom consciousness after they have been encoded.Because the memory was actu<strong>al</strong>ly encoded in theseinstances, it is possible that it may be subsequentlyr<strong>et</strong>rieved when the appropriate cues are presented.Hence, adults who have “repressed” painful memoriesmay be stimulated to rec<strong>al</strong>l the childhood traumawhen they encounter a related experience, perhapsone associated with rearing their own children. It isimportant to note, however, that adults can be inducedto construct pseudomemories of events that did not
170 <strong>Gordon</strong>, Baker-Ward, and Ornsteinoccur (Loftus, 1993), and the risk of this happeningwhen suggestive “memory recovery” techniques (e.g.,hypnosis, guided imagery, journ<strong>al</strong>ling) are used isvery high (see Lindsay & Read, 1994, for an an<strong>al</strong>ysisof the therapeutic practices that are most likely toresult in the generation of such pseudomemories).Despite many areas of disagreement about hiddenmemories of early trauma, the APA WorkingGroup on Investigation of Memories of ChildhoodAbuse (1998), a group of psychologists representingwidely disparate positions on the issue of hiddenmemories of early trauma, outlined the following areasof agreement: (1) Most people who were abusedas children remember <strong>al</strong>l or at least part of what happenedto them; (2) It is possible to remember instancesof abuse that have been forgotten for a long time; (3) It<strong>al</strong>so is possible to construct convincing pseudomemoriesof things that did not occur; and (4) There is muchy<strong>et</strong> to be learned about the process of rememberingchildhood trauma.R<strong>et</strong>riev<strong>al</strong>: Not Everything That Can BeR<strong>et</strong>rieved Is ReportedEven when information can be r<strong>et</strong>rieved frommemory, it may not <strong>al</strong>ways be reported. Children ofelementary school age may be hesitant to discloseactions they associate with embarrassment or shame(Saywitz <strong>et</strong> <strong>al</strong>., 1991). Preschoolers’ reports <strong>al</strong>so maybe constrained by the desire to avoid disclosing uncomfortableexperiences. In addition, development<strong>al</strong>limitations and individu<strong>al</strong> differences among childrenin such areas as language, temperament, emotion<strong>al</strong>status, and intelligence can influence the reporting ofinformation during an interview (Saywitz, in press),particularly among preschool children. Hence, in theabsence of optim<strong>al</strong> interviewing, younger children’sreports can underestimate the amount of informationthey may have in memory. Moreover, differences ininterviewer behavior and the context of the interviewitself <strong>al</strong>so affect the accuracy and compl<strong>et</strong>eness of children’smemory reports (Ceci & Bruck, 1995; Poole &Lamb, 1998).Language DevelopmentA child’s level of language comp<strong>et</strong>ency, at th<strong>et</strong>ime of both encoding and r<strong>et</strong>riev<strong>al</strong>, can have animportant impact on memory performance, particularlythat of the preschool child (Saywitz, in press;W<strong>al</strong>ker & Warren, 1995). With increases in age, thereare corresponding changes in the ability to use narrativestructure to report what can be remembered(Mandler, 1990). It is likely that some of the agerelatedimprovement in children’s rec<strong>al</strong>l (as discussedearlier) stems from these increases in narrative skill(P<strong>et</strong>erson & Rideout, 1998).Because preschool children’s limited verb<strong>al</strong> skillsare reflected in their difficulty in reporting <strong>al</strong>l of theinformation they may remember, efforts have beenmade to enhance their reports through the use of dollsor other props. The results of this research have beenconsistent in indicating that the use of these types ofcontextu<strong>al</strong> supports is not effective in improving theaccuracy of children’s memory reports (<strong>Gordon</strong> <strong>et</strong> <strong>al</strong>.,1993; Greenhoot, Ornstein, <strong>Gordon</strong>, & Baker-Ward,1999; Steward & Steward, 1996). Indeed, Bruck <strong>et</strong> <strong>al</strong>.(1995) found that using anatomic<strong>al</strong>ly d<strong>et</strong>ailed dollsto interview 3-year-olds about a physic<strong>al</strong> examinationresulted in greatly increased f<strong>al</strong>se reports of genit<strong>al</strong>touching and other inappropriate sexu<strong>al</strong> actions.Similarly, Goodman <strong>et</strong> <strong>al</strong>. (1997) reported that useof dolls and other props increased inaccuracy amongyounger children. They <strong>al</strong>so found, however, that useof dolls elicited more correct information than did freerec<strong>al</strong>l among older children (see <strong>al</strong>so <strong>Gordon</strong> <strong>et</strong> <strong>al</strong>.,1993).Children’s level of language comp<strong>et</strong>ency <strong>al</strong>so canaffect their ability to understand and respond to thesyntactic constructions, or the vocabulary (or both)used by interviewers (Carter, Bottoms, & Levine,1996; Saywitz, Nathanson, & Snyder, 1993). Childrenmay not be aware of when they do not understandcomplex questions and typic<strong>al</strong>ly do not report thesecomprehension failures (Carter <strong>et</strong> <strong>al</strong>., 1996). Y<strong>et</strong>, childrenusu<strong>al</strong>ly try to answer any question, even thos<strong>et</strong>hey do not understand. As Saywitz and colleagues(Saywitz & Snyder, 1993; Saywitz <strong>et</strong> <strong>al</strong>., 1993) pointout, children may think that they understand a termwhen, in fact, they do not (e.g., <strong>al</strong>legations = <strong>al</strong>ligators).Saywitz and Snyder (1993) demonstrated thatschool-age children can be trained to monitor theircomprehension of complex questions and indicatewhen they do not understand. This work and that ofothers (e.g., Gee, Gregory, & Pipe, 1999) indicates thatpreinterview training designed to discourage guessingand compliance can successfully enhance the memoryperformance of trained school-aged children as comparedwith untrained children. The extent to whichthis type of training is effective for preschool childrenis not y<strong>et</strong> known, however.
Children’s Testimony 171Person<strong>al</strong>ity/Temperament CharacteristicsIndividu<strong>al</strong> differences in temperament have beenshown to influence children’s rec<strong>al</strong>l performance.<strong>Gordon</strong> <strong>et</strong> <strong>al</strong>. (1993), for example, found that childrenwho were more at ease in new situations (adaptability)than were their peers provided more information,both verb<strong>al</strong>ly and nonverb<strong>al</strong>ly, about a physic<strong>al</strong> examination.Similar findings were reported by Merritt<strong>et</strong> <strong>al</strong>. (1994) in their study of young children’s rec<strong>al</strong>lof the d<strong>et</strong>ails of an invasive radiologic<strong>al</strong> procedure.In this study, both adaptability and the tendency toapproach others (as opposed to withdraw<strong>al</strong> or shyness)were found to correlate strongly with rec<strong>al</strong>l. Itis likely that these and other person<strong>al</strong>ity characteristicsinteract with the interview s<strong>et</strong>ting and process toinfluence rec<strong>al</strong>l performance. That is, children who aremore outgoing and adaptable adjust b<strong>et</strong>ter to beinginterviewed and, as a result, are able to r<strong>et</strong>rieve moreinformation from memory. Further, it seems reasonabl<strong>et</strong>o expect that the performance of younger childrenwould be more affected by temperament thanthat of school-aged children, who have had extensiveexperience with structured s<strong>et</strong>tings that require themto regulate their behavior.Another individu<strong>al</strong> difference variable that mayaffect memory performance is suggestibility. Thereis some evidence that suggestibility may be viewedas a person<strong>al</strong>ity “trait” rather than a solely development<strong>al</strong>factor (Clarke-Stewart, Thompson, & Lepore,1989). In <strong>al</strong>most <strong>al</strong>l studies of suggestibility, a substanti<strong>al</strong>proportion of even the youngest children do notreport f<strong>al</strong>se information. In the study by Leichtmanand Ceci (1995), for example, even under the mostsuggestive conditions one quarter of the 3- to 4-yearoldsand two-thirds of the 5- to 6-year-olds resistedsuggestion.Research conducted by Candel, Merckelbach,and Muris (2000) suggests that individu<strong>al</strong> differencesin suggestibility can be reliably measured in children.Using the Bonn Test of Statement Suggestibility,they found that younger children had higher suggestibilityscores than did older children. Moreover,within age groups, children who were judged by theirteachers as being more suggestible had higher scoresthan did those who were judged to be not particularlysuggestible. Similarly, using the GudjonssonSuggestibility Sc<strong>al</strong>e (Gudjonsson, 1989) to measureindividu<strong>al</strong> differences in interrogative suggestibility,Richardson and Kelly (1995) found that suggestibilitywas significantly negatively correlated both withintelligence and rec<strong>al</strong>l performance in 10- to 16-yearoldboys.The ch<strong>al</strong>lenge for researchers is to begin to identifythose characteristics that distinguish children whoare more suggestible from those who are not. Workin this area has just begun (e.g., Goodman & Quas,1997; Ornstein <strong>et</strong> <strong>al</strong>., 1997; Quas, Qin, Schaaf, &Goodman, 1997) with a focus on cognitive factors suchas event knowledge and source monitoring abilities,and soci<strong>al</strong>/emotion<strong>al</strong> variables including attachmentstyles, self-esteem, parenting styles, and the presenceor absence of psychopathology. Children with highself-esteem, for example, have been found to be moreresistant to misleading or suggestive questions whencompared with those with lower self-esteem (Howie& Dowd, 1996; Vrij & Bush, 2000). Other workindicates that self-esteem may interact with the age(or cognitive development<strong>al</strong> level or both) of the childin d<strong>et</strong>ermining individu<strong>al</strong> differences in suggestibility.Preschool children, who do not y<strong>et</strong> have a welldeveloped sense of self, may be more influenced bydeveloping cognitive abilities, such as the ability todistinguish the sources of their memories and to simultaneouslyconsider more than one aspect of a problem,than by self-esteem. In contrast, self-esteem mayplay a more important role in individu<strong>al</strong> differences insuggestibility among older children (Mazzoni, 1998;Muir-Broaddus, King, Downey, & P<strong>et</strong>ersen, 1998;Welch-Ross, Diecidue, & Miller, 1997). Indeed, it maybe that individu<strong>al</strong> and age-related differences in thesebasic cognitive abilities underlie the increased susceptibilityto suggestion that is consistently found amongpreschool children (Templ<strong>et</strong>on & Wilcox, 2000)Cognitive FactorsAs children develop, they learn strategies for organizingmateri<strong>al</strong> and r<strong>et</strong>rieving it from memory. Asa result, their narrative accounts of who, what, why,where, and when become more d<strong>et</strong>ailed, organized,and coherent. In addition, as children learn how narrativesare structured, they are b<strong>et</strong>ter able to r<strong>et</strong>rievespecific information about an event, such as its s<strong>et</strong>ting,participants, conversation, affective states, andconsequences (Mandler, 1990; Saywitz <strong>et</strong> <strong>al</strong>., 1993).On a very basic level, it is not surprising that intelligenceis associated with the ability to provide reliableand accurate testimony (Geddie, Fradin, & Beer,2000; Michel, <strong>Gordon</strong>, Ornstein, & Simpson, 2000;Richardson & Kelly, 1995).
172 <strong>Gordon</strong>, Baker-Ward, and OrnsteinA particularly interesting cognitive factor thatcontributes to individu<strong>al</strong> differences in suggestibilityis the ability of children to distinguish the sourcesof their memories, c<strong>al</strong>led source monitoring. As Ceciand Huffman (1997) point out, memories of our experiencespotenti<strong>al</strong>ly come from a vari<strong>et</strong>y of sources(e.g., thinking, t<strong>al</strong>king, hearing, and reading aboutthe event, as well as seeing it, participating in it ordoing both). If children are not able to differentiateamong these various sources of information, they willbe more susceptible to error and suggestions, possiblymisattributing information obtained from othersources to their actu<strong>al</strong> experience. Considerable researchhas examined this topic and some consistentfindings have emerged. Preschool children, for instance,have particular difficulty with most aspects ofthis cognitive task, whereas school-aged children areable to distinguish what they said from what someoneelse said (Foley, Johnson, & Raye, 1983), and whatthey did from what someone else did (Foley & Johnson,1985). However, these older children do not performas well as adults in distinguishing memories ofactivities they performed from those that were imagined(Foley & Ratner, 1998).The Memory InterviewJust as children must be able to understandthe questions asked, their memory performance <strong>al</strong>sorequires some comprehension of what is required inthe interview process itself (Poole & Lamb, 1998). Asan example, Steward and Steward (1996) found thatmany children in their study of memory for a medic<strong>al</strong>examination, did not understand what they weresupposed to t<strong>al</strong>k about until the researchers explicitlytold them. Children <strong>al</strong>so may switch the topic ofconversation in the middle of an interview, withoutwarning and unbeknownst to the interviewer (Poole& Lindsay, 1996). Furthermore, when interviews arerepeated, children may assume that the interviewer<strong>al</strong>ready knows the answers to the questions, and somay fail to provide sufficient information for the interviewerto make sense out of the response (Saywitz,1995). These characteristics of children’s “interviewee”skills can easily lead to misinterpr<strong>et</strong>ation of thechild’s statements, and may <strong>al</strong>so contribute to the lackof consistency that is often noted in the reports ofyoung children (Fivush & Hammond, 1990; <strong>Gordon</strong>& Follmer, 1994).Many aspects of the memory interview have beenshown to influence the accuracy and compl<strong>et</strong>eness ofchildren’s r<strong>et</strong>riev<strong>al</strong>. Research has examined ways toimprove testimony by enhancing children’s skills ineffectively r<strong>et</strong>rieving information from memory, andby changing interviewers’ behaviors to minimize biasand hence to reduce suggestibility. In an attempt tocompensate for children’s lack of understanding ofthe interview process, research has focused on incorporatingm<strong>et</strong>acognitive factors into training programsdesigned to maximize the rec<strong>al</strong>l of children (Fisher& McCauley, 1995; Gee <strong>et</strong> <strong>al</strong>., 1999; Saywitz, 1995).Preinterview training in providing the types of informationand the level of d<strong>et</strong>ail required for a forensicinterview, in resistance to misleading questions,and in comprehension monitoring has been shownto improve the performance of school-aged children(Saywitz & Snyder, 1993) but not that of preschoolchildren (Robinson & Briggs, 1997).Fisher and McCauley (1995) describe a procedurec<strong>al</strong>led the Cognitive Interview that was designedto increase the compl<strong>et</strong>eness and accuracy of eyewitnessreports (see Poole & Lamb, 1998 for a compl<strong>et</strong>edescription of this procedure). The CognitiveInterview has been demonstrated to substanti<strong>al</strong>ly improv<strong>et</strong>he eyewitness accounts of adults when comparedwith a standard police interview (e.g., Fisher,Geiselman, Raymond, Jurkevich, & Warhaftig, 1987).Modifications to the Cognitive Interview format havebeen made to accommodate the speci<strong>al</strong> needs of childrenand use of this modified version has been shownto improve memory reports of school-aged children(McCauley & Fisher, 1995) and children with ment<strong>al</strong>r<strong>et</strong>ardation (Milne & Bull, 1996). Despite the documentedeffectiveness of the Cognitive Interview overother m<strong>et</strong>hods, Poole and Lamb (1998) argue for cautionin its use in forensic interviews with childrenwhere it is not possible to know precisely what achild experienced. This caution stems, in part, fromthe component of the interview that asks the child tovisu<strong>al</strong>ize the context of his or her experience. Pooleand Lamb argue that the act of visu<strong>al</strong>ization may resultin f<strong>al</strong>se memories among children, a problem forwhich Ceci, Loftus, Leichtman, and Bruck (1994) providecompelling empiric<strong>al</strong> evidence. It remains to beseen wh<strong>et</strong>her the Cognitive Interview is effective withpreschool children.The most important d<strong>et</strong>erminant of the successof the interview is the behavior of the individu<strong>al</strong>conducting the interview. Children’s acquiescenc<strong>et</strong>o interviewers’ suggestions is a well-documentedphenomenon (see Ceci & Bruck, 1993, 1995, for reviews).Although it is clear that younger children areparticularly vulnerable to misleading or suggestive
Children’s Testimony 173questions, especi<strong>al</strong>ly under conditions in which thepower of the interviewer is asserted, the mechanismsunderlying suggestibility remain in question.One critic<strong>al</strong> way in which interviewers influencechildren’s reports is the extent to which they are biased.Interviewers who hold strong beliefs about whatthey think happened can influence children to providememory reports that are consistent with thesebeliefs, regardless of the fact that these beliefs maybe wrong (Bruck <strong>et</strong> <strong>al</strong>., 1998; Ceci & Huffman, 1997).According to Bruck <strong>et</strong> <strong>al</strong>. (1998), biased interviewerspotenti<strong>al</strong>ly influence children’s reports in sever<strong>al</strong>ways. These include (1) failure to pursue <strong>al</strong>ternativeexplanations for a child’s statements; (2) use of selectiveattention to and reinforcement of responses thatare consistent with their a priori beliefs; (3) failure toexplore children’s inconsistent or bizzare statements;(4) overuse of specific or leading questions aimed atconfirming the interviewer’s hypothesis; (5) introducinginformation about the suspected perp<strong>et</strong>rator thatportrays him or her in a negative way; (6) inappropriateuse of anatomic<strong>al</strong>ly d<strong>et</strong>ailed dolls or other props;and (7) inappropriate use of guided imagery or memorywork.In a study demonstrating the powerful impactof interview bias, Ceci and his colleagues (White,Leichtman, & Ceci, 1997) provided an experiencedinterviewer with information about an event experiencedby 3- to 4-year-old and 5- to 6-year-old children.Some of this information was accurate and somewas not. When the children were interviewed 1 monthlater, 34% of younger children and 18% of the olderchildren confirmed the f<strong>al</strong>se information. Two monthslater, the notes of the first interviewer were givento a second individu<strong>al</strong> and the children were reinterviewed.At this time, the children not only continuedto consent to f<strong>al</strong>se information but did sowith increased confidence and embellishment. Anotherexample of interviewer bias is provided by astudy reported by Roberts and Lamb (1999) in whichtranscripts of actu<strong>al</strong> investigative interviews with 3-to 14-year-old children who had made <strong>al</strong>legations ofsexu<strong>al</strong> or physic<strong>al</strong> abuse were an<strong>al</strong>yzed. Roberts andLamb found that interviewers often distorted the d<strong>et</strong>ailsprovided by the children. Moreover, the childrenrarely corrected these distortions and when they werenot corrected, the interviewers continued to misrepresentthe d<strong>et</strong>ails throughout the remainder of the interview.Other work suggests that soci<strong>al</strong> pressure (e.g.,asking children to endorse their peers’ disclosures,inviting speculation, repeating previously answeredquestions) and selective reinforcement may have agreater impact on children’s suggestibility than simplyasking misleading or suggestive questions (Garven,Wood, M<strong>al</strong>pass, & Shaw, 1998).Repeated InterviewsIn the absence of suggestive questions or misleadinginformation, multiple interviews can haveboth positive and negative effects on memory performance.On the positive side, repeated interviewscan som<strong>et</strong>imes enable the child to rec<strong>al</strong>l informationnot rec<strong>al</strong>led earlier (<strong>Gordon</strong> & Follmer, 1994) andunder some conditions may “innoculate” the childagainst forg<strong>et</strong>ting (for reviews, see Poole & White,1995; Warren & Lane, 1995). In some sense, then, repeatedinterviews may serve to maintain informationin memory, in a process that is an<strong>al</strong>ogous to the constructof “reinstatement” by Campbell and Jaynes’(1966). For Campbell and Jaynes, parti<strong>al</strong> rep<strong>et</strong>itions(e.g., repeated interviews) of an initi<strong>al</strong> event during adelay can serve to reinstate information that ordinarilywould be forgotten. P<strong>et</strong>erson’s (P<strong>et</strong>erson, 1999)study of children’s memory for accident<strong>al</strong> injuries indicatesthat the reinstatement effect may be particularlybenefici<strong>al</strong> when the information to be rememberedis less “memorable,” whereas there may be littleeffect for highly s<strong>al</strong>ient materi<strong>al</strong>.Other research suggests that more than just simplerep<strong>et</strong>ition is necessary to reinstate memory. Pooleand White (1995), for example, state that the timingand number of repeated interviews may be importantin d<strong>et</strong>ermining their effects. It is possible that an interviewcarried out immediately after the event in questionmay be more effective in maintaining memorythan one that takes place some time later, particularlyif the r<strong>et</strong>elling of the event involves the opportunityto consolidate one’s understanding of the event.Indeed, this may explain why P<strong>et</strong>erson (1999) foundthat children’s memory for a traumatic injury was sowell maintained over a 2-year period. In her study,parents reported many conversations about the injurywith grandparents, other relatives, and friends immediatelyafter the injury. P<strong>et</strong>erson argues that traumaticevents become part of the child’s and family’s “lifestory” and as such are typic<strong>al</strong>ly discussed at length immediatelyafter the event, essenti<strong>al</strong>ly providing a reinstatementof the features of the experience. In contrastto these positive effects, Warren and Lane (1995)argue that with multiple interviews over time, errorsmade at one time are likely to be repeated at a latertime, intrusion errors are more likely, and guesses or
174 <strong>Gordon</strong>, Baker-Ward, and Ornsteinspeculation that occur during an early interview tendto be incorporated into memory as fact in subsequentinterviews. Brainerd and Reyna (1996) found thateven a single interview consisting of specific yes/noquestions that occurred immediately after an eventresulted in f<strong>al</strong>se memory responses on delayed testsamong 5- and 8-year-olds. Further, Ceci <strong>et</strong> <strong>al</strong>. (1994)found that merely asking preschool children to thinkabout wh<strong>et</strong>her or not both true and f<strong>al</strong>se events hadever happened led to increased claims over the courseof 11 nonsuggestive interviews that the children hadexperienced events that had never happened. Moreover,over time the children provided increasingly d<strong>et</strong>ailedand coherent narratives describing these f<strong>al</strong>seevents. Ceci <strong>et</strong> <strong>al</strong>. suggest that these children actu<strong>al</strong>lycame to believe they had experienced these f<strong>al</strong>seevents, that is that the f<strong>al</strong>se events became part ofthe children’s memory system. Evidence for this assumptionwas provided by the experimenter’s inabilityto “debrief” the children about the fact that theevents never happened. It should be noted, however,that in a follow-up investigation 2 years later, the children“recanted” their earlier f<strong>al</strong>se reports more than77% of the time while maintaining high levels of assentsto questions about true experiences (Huffman<strong>et</strong> <strong>al</strong>., 1997). The authors interpr<strong>et</strong> the findings as suggestingthat f<strong>al</strong>se memories, which are not elaboratedthrough family discussions and other natur<strong>al</strong>ly occurringinteractions, may fade in the absence of repeatedquestioning.One of the purposes of conducting multiple interviewsin leg<strong>al</strong> cases involving children is to givechildren the opportunity to disclose information orexperiences that they might be reluctant to discuss.Indeed, considerable work has documented that childrenoften provide more d<strong>et</strong>ails or new informationwhen repeatedly interviewed. Unfortunately, itis <strong>al</strong>so true that children provide more f<strong>al</strong>se or inaccurateinformation at later interviews (Bruck, Ceci,& Hembrooke, 1997; S<strong>al</strong>mon & Pipe, 1997). Further,if misleading or suggestive information is introducedduring the course of any one of multiple interviews,this information may become an actu<strong>al</strong> part of the individu<strong>al</strong>’smemory for the event (Bruck <strong>et</strong> <strong>al</strong>., 1995).Bruck <strong>et</strong> <strong>al</strong>. (1997) provide compelling evidenc<strong>et</strong>hat repeated suggestive interviews lead childrento provide more compl<strong>et</strong>e, elaborated, and spontaneousdescriptions of events that never occurred andthat these descriptions can be indistinguishable fromthose regarding experiences that the children actu<strong>al</strong>lyhad.The Interview ContextAnother aspect of the memory interview thathas been shown to influence children’s reports is thes<strong>et</strong>ting or context of the interview. P<strong>et</strong>ers (1991),for example, suggests that a stressful, potenti<strong>al</strong>lyconfrontation<strong>al</strong> interview can result in impairedperformance. Saywitz and Nathanson (1993) demonstratedthat children’s perceived stress while beinginterviewed in a mock courtroom interfered withtheir rec<strong>al</strong>l. Thus, simply making the child feelcomfortable and at ease during the interview canfacilitate r<strong>et</strong>riev<strong>al</strong> by reducing the child’s level ofstress (Goodman, Bottoms, Schwartz-Kenney, &Rudy, 1991; Hovde & Strapp, 1999). Poole and Lamb(1998), however, point out that when working withyoung children, it is not <strong>al</strong>ways easy to b<strong>al</strong>ance thecomp<strong>et</strong>ing demands of the investigative interview.Reducing the child’s stress, for example, can involveprovision of age-appropriate toys or activities, whichmay interfere with the need to maintain the child’sfocus of attention on the task at hand and thegathering of objective, reliable information.SUMMARY AND IMPLICATIONS FORFUTURE RESEARCH, CLINICAL PRACTICE,AND PUBLIC POLICYThe preceding review of the memory researchleads to sever<strong>al</strong> gener<strong>al</strong> conclusions regarding children’sability to remember their experiences and provideaccurate and reliable eyewitness testimony:1. Even very young children remember past experiencesover very long periods of time, particularlywhen they have some prior knowledgeof the events in question.2. When they are interviewed in a nonsuggestive,nonleading manner and are not exposedto misinformation, children older than age3 can provide reasonably accurate and relativelycompl<strong>et</strong>e accounts of these experiences.3. There are robust age differences in children’sabilities to provide eyewitness testimonythat most likely reflect development<strong>al</strong>changes in cognitive and m<strong>et</strong>acognitivefunctioning. In gener<strong>al</strong>, school-aged childrenprovide more accurate information, are moreconsistent, and require fewer specific promptsin rec<strong>al</strong>l when compared with preschoolers.Moreover, older children are less vulnerable
Children’s Testimony 175to the effects of misinformation than are theiryounger peers.4. Children can be influenced in a vari<strong>et</strong>y of waysto provide compl<strong>et</strong>e and elaborated narrativesabout experiences that did not occur tosuch an extent that experts cannot tell thesef<strong>al</strong>se reports from those that are true. It is suspectedthat many children actu<strong>al</strong>ly come tobelieve that these f<strong>al</strong>se events re<strong>al</strong>ly did occur.Directions for Future ResearchDespite the large volume of research on children’stestimony that has appeared in the past 20years, important questions remain to be addressed,especi<strong>al</strong>ly in the broad area of individu<strong>al</strong> differences.Research is needed to identify how individu<strong>al</strong> differencesin specific cognitive and soci<strong>al</strong>/emotion<strong>al</strong> domainsinfluence children’s performance at <strong>al</strong>l phasesof the memory process. For example, differences inhow children cope (e.g., active vs. avoidant styles) withstressful experiences may play a role in the amountand type of information that is encoded and thus, whatis rec<strong>al</strong>led at a later time. Specific<strong>al</strong>ly, the use of dissociationas a coping strategy among children warrantsfurther investigation. Are some children more pron<strong>et</strong>o use this m<strong>et</strong>hod of coping than are others? Do dissociationand other styles of coping result in systematicmemory failures?Research <strong>al</strong>so is needed to delineate the specificcharacteristics that distinguish children who differin their vulnerability to suggestion, especi<strong>al</strong>ly thosewithin the same age groups. Although we know thatpreschoolers are more suggestible than are schoolagedchildren, for example, any individu<strong>al</strong> 4-year-oldmay be more or less vulnerable to incorporating inconsistentinformation into his or her report than areage peers. Further, it is particularly important to beginto understand the unique characteristics of childrenwho have been abused and the impact thesecharacteristics might have on the memory process.Abused children are more likely to be emotion<strong>al</strong>lydistressed, to have low self-esteem, to come from dysfunction<strong>al</strong>families, and to experience considerablymore chronic life stress than are nonabused children.Research could help to d<strong>et</strong>ermine what role, if any,these factors, as well as others, play in increasing achild’s vulnerability to memory distortion.Although the study of individu<strong>al</strong> differences isof interest in its own right, this research <strong>al</strong>so can informclinic<strong>al</strong> practice in important ways. For instance,research in this area might lead to the developmentof instruments designed to assess the relevant characteristicsand abilities of individu<strong>al</strong> children prior toconducting a forensic interview. The results of suchassessment could then aid in the interpr<strong>et</strong>ation of thechild’s report.Implications for Clinic<strong>al</strong> PracticeThe memory research reviewed raises sever<strong>al</strong> issuesregarding clinic<strong>al</strong> practice that merit some discussion.First, guidelines for interviewing young childrenthat take into account many of the preceding researchfindings have been proposed (American Academy ofChild and Adolescent Psychiatry, 1997; Lamb, 1994;Poole & Lamb, 1998). These guidelines spell out m<strong>et</strong>hodsfor conducting an appropriate forensic interviewthat are well accepted. For example, most peopleunderstand the importance of relying primarily onopen-ended rather than specific questions, conductingthe interview as quickly as possible, and minimizingthe possibility of repeated interviews. Whatis not clear, however, is to what extent individu<strong>al</strong>swho conduct these interviews are able to follow thepublished guidelines during the course of an actu<strong>al</strong>interview. Perhaps the next step in the process of developinga “technology of interviewing” is to producestructured interviews such as those used in the diagnosisof child psychopathology. Although structuredinterviews have many disadvantages, this type of interviewwould <strong>al</strong>low systematic incorporation of questionsdesigned to explore <strong>al</strong>ternative explanations forthe child’s <strong>al</strong>legations. Moreover, the accuracy of informationobtained through use of a structured interviewcould be tested against other interview m<strong>et</strong>hods(see for example Michel, <strong>2001</strong>).Second, it is not clear that the results of researchon children’s testimony are filtering down fromthe laboratory to clinic<strong>al</strong> and leg<strong>al</strong> profession<strong>al</strong>s onthe front lines of interviewing children. Researchersmust view dissemination of their findings as one extremelyimportant aspect of their work. Clinicianscan be taught to use the research findings to an<strong>al</strong>yzeand ev<strong>al</strong>uate the facts of individu<strong>al</strong> cases accordingto the specific aspects of memory research thatare most relevant. In one case, for example, issuesof source monitoring might be important, whereas inanother, negative stereotyping might have influencedthe child’s report. Ev<strong>al</strong>uation of a child’s <strong>al</strong>legations
176 <strong>Gordon</strong>, Baker-Ward, and Ornsteinshould include d<strong>et</strong>ailed examination of <strong>al</strong>l occasionsduring which the child was asked to discuss the eventsin question, including those in inform<strong>al</strong> s<strong>et</strong>tings (e.g.,with parents or other family members), particularlywhen these discussions preceded the forensic interview.Given that memory is a constructive process,the most reliable piece of “data” available to the clinicianis the child’s first report of his or her experiences.Hence, d<strong>et</strong>ailed information (e.g., what wassaid, to whom it was said, wh<strong>et</strong>her it was spontaneousor prompted, and the context of this conversation)should be gathered regarding the circumstances surroundingthis report. This type of an<strong>al</strong>ysis ensures thatthe clinician’s ultimate judgment as to if or by whom achild has been abused, or both, is based on systematicreview of <strong>al</strong>l the possibilities.Third, <strong>al</strong>though aspects of traumatic experiencesmay remain vivid in memory because of their distinctiveness,it is important for clinicians to understandthat these memories are subject to the same reconstructiveprocesses as are memories for everydayexperiences. Hence, the risk of creating f<strong>al</strong>se memoriesthrough suggestive questioning, biased interviewing,the use of certain therapeutic techniques such ashypnosis and repeated visu<strong>al</strong>ization, or a combinationof these, is great (Lindsay & Read, 1994). Becaus<strong>et</strong>here is no reliable m<strong>et</strong>hod for distinguishing“true” from “f<strong>al</strong>se” memories in the absence of extern<strong>al</strong>corroborative evidence, clinicians who workwith clients (both adults and children) whose symptomsor <strong>al</strong>legations suggest the possibility of childhoodtrauma are urged to exercise great cautionin pursing this possible interpr<strong>et</strong>ation. Explanationsfor the symptoms or <strong>al</strong>legations other than abuseshould <strong>al</strong>ways be examined. Moreover, cliniciansshould inform their clients of the possibility of f<strong>al</strong>sememories.Fourth, it is evident that interviewing childrenwho are <strong>al</strong>leged to have been abused is a very complexendeavor, requiring highly skilled profession<strong>al</strong>s. Theneed for training should be obvious. Unfortunately,there is little evidence that even the most intensiv<strong>et</strong>raining programs have much effect on the behavior ofinterviewers in the long term (Aldridge & Cameron,1999; Warren <strong>et</strong> <strong>al</strong>., 1999). Research is needed to d<strong>et</strong>ermin<strong>et</strong>he most effective m<strong>et</strong>hods for training andproviding ongoing supervision of potenti<strong>al</strong> interviewers.It is possible that training potenti<strong>al</strong> interviewersbefore they begin to practice and become “s<strong>et</strong> in theirways” would be more effective than r<strong>et</strong>raining experiencedinterviewers. Moreover, ways of maintainingtraining effects over the long term must be examined.These might include “booster sessions,” provision ofongoing supervision, or both.Fin<strong>al</strong>ly, given that it is so difficult to change thebehavior of experienced interviewers and in light ofthe possibility of the generation of f<strong>al</strong>se memories, itis critic<strong>al</strong> that <strong>al</strong>l interviews with potenti<strong>al</strong> child witnessesbe recorded. There is considerable controversyabout the issue of recording interviews, and many cliniciansare resistant to doing so for fear that their interviewingskills will be “dissected” in court. It is only byhaving a concr<strong>et</strong>e record of the interview, however,that the child’s responses can be ev<strong>al</strong>uated in lightof the prompts used by the interviewer to elicit them.Videotape recording is particularly important to ev<strong>al</strong>uatesubtle interviewer behaviors, such as differenti<strong>al</strong>reinforcement or systematic ignoring of informationreported, that have been shown to contribute to children’sf<strong>al</strong>se statements.Some Policy ImplicationsResearchers clearly have a role in influencinglegislation that d<strong>et</strong>ermines how child witnesses willbe handled in leg<strong>al</strong> proceedings. Although some importantsteps have been taken, much remains to bedone. For example, having a child appear in personto testify can be an advantage or a disadvantage, dependingon the circumstances of the case and wh<strong>et</strong>herone is the prosecutor or defense attorney. Consider,for example, the case of a child who was <strong>al</strong>leged tohave been abused at age 5, but does not appear incourt to testify until age 10. Because changes in thechild’s development<strong>al</strong>/cognitive status and the potenti<strong>al</strong>for events that influence the child’s memory duringthe 5-year interv<strong>al</strong> are enormous, the qu<strong>al</strong>ity andperceived credibility of this child’s testimony wouldclearly be different if she had testified at age 5. Theresearch reviewed above suggests that a videotapedinterview made when the child was 5 years old wouldprovide a b<strong>et</strong>ter, more accurate means of having thechild testify than having the 10-year-old appear inperson. Moreover, young children often have difficultyproviding adequate testimony because of thesurroundings in which they are asked to do so. Avideotape record made in a more comfortable, childfriendlys<strong>et</strong>ting would provide b<strong>et</strong>ter information inmany cases.A related concern is the use of expert witnessesin leg<strong>al</strong> procedings (see Ceci & Hembrooke, 1998, fora review). In the past, experts have become caught inthe adversari<strong>al</strong> process that often characterizes leg<strong>al</strong>
Children’s Testimony 177proceedings with the result that it is becoming increasinglydifficult to g<strong>et</strong> researchers to agree to performthis function. Ide<strong>al</strong>ly, experts should be able to functionas “friends of the court” rather than being forcedto <strong>al</strong>ign themselves with the prosecution or defense.The expert’s responsibility then becomes to inform oreducate the jury about the relevant aspects of memoryresearch. As an example, in the case describedearlier, the videotape recording of the 5-year-old’s interviewmight have been compl<strong>et</strong>ed sever<strong>al</strong> monthsafter the initi<strong>al</strong> <strong>al</strong>legations came to light. During thistime, numerous things with the potenti<strong>al</strong> to interferewith the accuracy of the child’s report could have occurred.The role of the expert would be to describefor the court the aspects of the memory process thatwould enable the jury to b<strong>et</strong>ter ev<strong>al</strong>uate the child’svideotaped testimony.Although continuing research is needed to builda knowledge base that can more fully inform practice,the existing literature represents significant advancesin the understanding of children’s testimony over thepast decade. At present, the empiric<strong>al</strong> work on children’sreports of person<strong>al</strong>ly experienced events succeedsnot only in describing age-related changes butin identifying some caus<strong>al</strong> explanations for development<strong>al</strong>and situation<strong>al</strong> differences in performance aswell. A major ch<strong>al</strong>lenge for psychologists, includingboth researchers and clinicians, is to utilize this knowledgemore fully. Collaboration within the professionis needed to apply and disseminate existing finding aswell as to design further investigations that can effectivelyinform practice. This work can help to optimizechildren’s participation in leg<strong>al</strong> proceedings.REFERENCESAldridge, J., & Cameron, S. (1999). Interviewing child witnesses:Questioning techniques and the role of training. Applied Development<strong>al</strong>Science, 3, 136–147.Alpert, J. L., Brown, L. S., & Courtois, C. A. (1998). Symptomaticclients and memories of childhood abuse: What the traumaand child sexu<strong>al</strong> abuse literature tells us. Psychology, PublicPolicy, and Law, 4, 941–995.American Academy of Child and Adolescent Psychiatry. (1997).Practice param<strong>et</strong>ers for the forensic ev<strong>al</strong>uation of children andadolescents who may have been physic<strong>al</strong>ly or sexu<strong>al</strong>ly abused.Journ<strong>al</strong> of the American Academy of Child and AdolescentPsychiatry, 36, 423–442.American Profession<strong>al</strong> Soci<strong>et</strong>y on the Abuse of Children. (1990).Guidelines for psychosexu<strong>al</strong> ev<strong>al</strong>uation of childhood sexu<strong>al</strong>abuse in young children. Chicago, IL: Author.APA Working Group on Investigations of Memories of ChildhoodAbuse. (1998). Fin<strong>al</strong> conclusions of the American Psychologic<strong>al</strong>Association Working Group on Investigation of Memoriesof Childhood Abuse. Psychology, Public Policy, and Law, 4,933–940.Bahrick, L., Parker, J. F., Fivush, R., & Levitt, M. (1998). The effectsof stress on young children’s memory for a natur<strong>al</strong> disaster.Journ<strong>al</strong> of Experiment<strong>al</strong> Psychology: Applied, 4, 308–331.Baker-Ward, L., <strong>Gordon</strong>, B. N., Ornstein, P. A., Larus, D. M.,& Clubb, P. A. (1993). Young children’ s long-term r<strong>et</strong>entionof a pediatric examination. Child Development, 64, 1519–1533.Baker-Ward, L., Hess, T. M., & Flanagan, D. A. (1990). The effectsof involvement on children’s memory for events. CognitiveDevelopment, 5, 55–70.Baker-Ward, L., Ornstein, P A., & Principe, G. F. (1997). Reve<strong>al</strong>ingthe representation: Evidence from children’s reports of events.In P. van den Broek, P. Bauer, & T. Borg (Eds.), Development<strong>al</strong>spans in event comprehension and representation: Bridgingfiction<strong>al</strong> and actu<strong>al</strong> events (pp. 79–107). Hillsd<strong>al</strong>e, NJ: Erlbaum.Bauer, P. J. (1995). Rec<strong>al</strong>ling past events from infancy to early childhood.Ann<strong>al</strong>s of Child Development, 11, 25–71.Bauer, P. J. (1996). What do infants rec<strong>al</strong>l of their lives? Memoryfor specific events by one- to two-year-olds. AmericanPsychologist, 51, 29–41.Bauer, P. J., Hertsgaard, L. A., & Dow, G. A. (1994). After 8 monthshave passed: Long-term rec<strong>al</strong>l of events by 1- to 2- year-oldchildren. Memory, 2, 353–382.Bauer, P. J., Kroupina, M. G., Schwade, J. A., Dropik, P. L.,& Wewerka, S. S. (1998). If memory serves, will language:Later verb<strong>al</strong> accessibility of early memories. Development andPsychopathology, 10, 655–680.Bjorklund, D. F. (1985). The role of conceptu<strong>al</strong> knowledge in thedevelopment of organization in children’s memory. In C. J.Brainerd & M. Pressley (Eds.), Basic processes in memory development(pp. 103–142). New York: Springer-Verlag.Bjorklund, D. F. (1987). How age changes in knowledge contribut<strong>et</strong>o the development of children’s memory: An interpr<strong>et</strong>ativereview. Development<strong>al</strong> Review, 7, 93–130.Bjorklund, D. F. (2000). Children’s thinking: Development<strong>al</strong> functionand individu<strong>al</strong> differences (3rd ed.). Belmont, CA:Wadsworth.Bowers, G. H., & Sivers, H. (1998). Cognitive impact of traumaticevents. Development and Psychopathology, 4, 625–654.Brainerd, C. J., Kingman, J., & Howe, M. L. (1985). On the developmentof forg<strong>et</strong>ting. Child Development, 56, 1103–1119.Brainerd, C. J., & Mojardin, A. H. (1998). Children’s and adults’spontaneous f<strong>al</strong>se memories: Long-term persistence and mer<strong>et</strong>estingeffects. Child Development, 69, 1361–1377.Brainerd, C. J., & Reyna, V. F. (1990). Gist is the grist: Fuzzy trac<strong>et</strong>heory and the new intuitionism. Development<strong>al</strong> Review, 10,3–47.Brainerd, C. J., & Reyna, V. F. (1996). Mere memory testing createsf<strong>al</strong>se memories in children. Development<strong>al</strong> Psychology,32, 467–478.Brainerd, C. J., Reyna, V. F., & Brandse, E. (1995). Are children’sf<strong>al</strong>se memories more persistent than their true memories? Psychologic<strong>al</strong>Science, 6, 359–364.Bremner, J. D., Shobe, K. K., & Kihlstrom, J. F. (2000). F<strong>al</strong>se memoriesin women with self-reported childhood sexu<strong>al</strong> abuse: Anempiric<strong>al</strong> study. Psychologic<strong>al</strong> Science, 11, 333–337.Brown, R., & Kulilk, J. (1977). Flashbulb memories. Cognition, 5,73–99.Bruck, M., & Ceci, S. J. (1999). The suggestibility of children’s memory.Annu<strong>al</strong> Review of Psychology, 50, 419–439.Bruck, M., Ceci, S. J., Francoeur, E., & Barr, R. J. (1995). “I hardlycried when I got my shot!”: Influencing children’s reportsof a pediatric examination involving genit<strong>al</strong> touching. ChildDevelopment, 66, 193–208.Bruck, M., Ceci, S. J., & Hembrooke, H. (1997). Children’s reportsof pleasant and unpleasant events. In D. Read and S. Lindsay(Eds.), Recollections of trauma: Scientific research and clinic<strong>al</strong>practice (pp. 199–219). New York: Plenum Press.
178 <strong>Gordon</strong>, Baker-Ward, and OrnsteinBruck, M., Ceci, S. J., & Hembrooke, H. (1998). Reliability andcredibility of young children’s reports: From research to policyto practice. American Psychologist, 53, 136–151.Campbell, B. A., & Jaynes, J. (1966). Reinstatement. Psychologic<strong>al</strong>Review, 73, 478–480.Candel, I., Merckelbach, H., & Muris, P. (2000). Measuring interrogativesuggestibility in children: Reliability and v<strong>al</strong>idity ofthe Bonn Test of Statement Suggestibility. Psychology, Crimeand Law, 6, 61–70.Carter, C. A., Bottoms, B. L., & Levine, M. (1996). Linguistic and socioemotion<strong>al</strong>influences on the accuracy of children’s reports.Law and Human Behavior, 20, 335–358.Ceci, S. J., & Bruck, M. (1993). Suggestiblity of the child witness:A historic<strong>al</strong> review and synthesis. Psychologic<strong>al</strong> Bull<strong>et</strong>in, 113,403–439.Ceci, S., & Bruck, M. (1995). Jeopardy in the courtroom: A scientifican<strong>al</strong>ysis of children’s testimony. Washington, DC: AmericanPsychologic<strong>al</strong> Association.Ceci, S. J., & Hembrooke, H. (1998). Expert witnesses in child abusecases: What can and should be said in court. Washington, DC:American Psychologic<strong>al</strong> Association.Ceci, S. J., & Huffman, M. L. C. (1997). How suggestibile arepreschool children? Cognitive and soci<strong>al</strong> factors. Journ<strong>al</strong> ofthe American Academy of Child and Adolescent Psychiatry,36, 948–956.Ceci, S. J., Huffman, M. L. C., Smith, E., & Loftus, E. W. (1994).Repeatedly thinking about non-events: Source misattributionsamong preschoolers. Conscious Cognition, 3, 388–407.Ceci, S. J., Loftus, E. F., Leichtman, M. D., & Bruck, M. (1994). Thepossible role of source misattributions in the creation of f<strong>al</strong>sebeliefs among preschoolers. Internation<strong>al</strong> Journ<strong>al</strong> of Clinic<strong>al</strong>Experiment<strong>al</strong> Hypnosis, 42, 304–320.Ceci, S. J., Ross, D. F., & Toglia, M. P. (1987). Age differences insuggestibility: Narrowing the uncertainties. In S. J. Ceci, M.P. Toglia, & D. F. Ross (Eds.), Children’s eyewitness testimony(pp. 79–91). New York: Springer-Verlag.Chi, M. T. H., & Ceci, S. J. (1987). Content knowledge: Its role, representation,and restructuring in memory development. In H.W. Reese (Ed.), Advances in child development and behavior(Vol. 20, pp. 91–142). New York: Academic Press.Cicch<strong>et</strong>ti, D., & Toth, S. (Eds.). (1998). Risk, trauma, and memory.[Speci<strong>al</strong> issue]. Development and Psychopathology, 10(4).Clarke-Stewart, A., Thompson, W. C., & Lepore, S. (1989, April).Manipulating children’s interpr<strong>et</strong>ations through interrogation.In G. S. Goodman (Chair), Can children provide accurate eyewitnessreports? Symposium conducted at the bienni<strong>al</strong> me<strong>et</strong>ingof the Soci<strong>et</strong>y for Research in Child Development, KansasCity, MO.Clubb, P. A., & Ornstein, P. A. (1992, April). Visiting the doctor:Children’s differenti<strong>al</strong> r<strong>et</strong>ention of individu<strong>al</strong> components ofthe physic<strong>al</strong> examination. In D. Bjorklund & P. A. Ornstein(Chairs), Children’s memory for re<strong>al</strong>-world events: Implicationsfor testimony. A symposium presented at the Conferenceon Human Development, Atlanta, GA.DeCasper, A. J., & Spence, M. J. (1986). Prenat<strong>al</strong> matern<strong>al</strong> speechinfluences newborns’ perception of speech sounds. Infant Behaviorand Development, 9, 133–150.Eisen, M. L., & Goodman, G. S. (1998). Trauma, memory, and suggestibilityin children. Development and Psychopathology, 10,717–738.Fisher, R. P. & Geiselman, R. E. (1992). Memory-enhancing techniquesfor investigative interviewing: The Cognitive Interview.Springfield, IL: Thomas.Fisher, R. P., Geiselman, R. E., Raymond, D. S., Jurkevich, L. M.,& Warhaftig, M. L. (1987). Enhancing enhanced eyewitnessmemory: Refining the cognitive interview. Journ<strong>al</strong> of PoliceScience and Administration, 15, 291–297.Fisher, R. P., & McCauley, M. R. (1995). Improving eyewitnesstestimony with the Cognitive Interview. In M. S. Zaragoza,J. R. Graham, G. C. N. H<strong>al</strong>l, R. Hirschman, Y. S. Ben-Porath(Eds.), Memory and testimony in the child witness (pp. 141–159). Thousand Oaks, CA: Sage Publications.Fivush, R. (1997). Event memory in childhood. In N. Cowan (Ed.),The development of memory in childhood (pp. 139–162). Sussex:Psychology Press.Fivush, R. (1998). Children’s recollections of traumatic and nontraumaticevents. Development and Psychopathology, 10, 699–716.Fivush, R. (2000). Accuracy, authority, and voice. In P. H. Miller& E. K. Scholnick (Eds.), Toward a feminist development<strong>al</strong>psychology (pp. 85–105). New York: Routledge.Fivush, R., Gray, J., & Fromhoff, F. A. (1987). Two-year-olds t<strong>al</strong>kabout the past. Cognitive Development, 2, 393–409.Fivush, R., & Hamond, N. R. (1989). Time and again: Effects ofrep<strong>et</strong>ition and r<strong>et</strong>ention interv<strong>al</strong> on 2-year-olds’ event rec<strong>al</strong>l.Journ<strong>al</strong> of Experiment<strong>al</strong> Child Psychology, 47, 259–273.Fivush, R., & Hammond, N. R. (1990). Autobiographic<strong>al</strong> memoryacross the preschool years: Toward reconceptu<strong>al</strong>izing childhoodamnesia. In R. Fivush & J. A. Hudson (Eds.), Knowingand remembering in young children (pp. 223–248). New York:Cambridge University Press.Fivush, R., & Shukat, J. R. (1995). Content, consistency, and coherenceof early autobiographic<strong>al</strong> rec<strong>al</strong>l. In M. S. Zaragoza,J. R. Graham, G. C. N. H<strong>al</strong>l, R. Hirschman, Y. S. Ben-Porath(Eds.), Memory and testimony in the child witness (pp. 5–23).Thousand Oaks, CA: Sage Publications.Foley, M. A., & Johnson, M. K. (1985). Confusions b<strong>et</strong>ween memoriesfor performed and imagined actions. A development<strong>al</strong>comparison. Child Development, 56, 1145–1155.Foley, M. A., Johnson, M. K., & Raye, C. L. (1983). Age-relatedchanges in confusion b<strong>et</strong>ween memories for thoughts andmemories for speech. Child Development, 54, 51–60.Foley, M. A., & Ratner, H. H. (1998). Distinguishing b<strong>et</strong>ween memoriesfor thoughts and deeds: The role of prospective processingin children’s source monitoring. British Journ<strong>al</strong> of Development<strong>al</strong>Psychology, 16, 465–484.Garven, S., Wood, J. M., M<strong>al</strong>pass, R. S., & Shaw, J. S., III.(1998). More than suggestion: The effect of interviewing techniquesfrom the McMartin Preschool case. Journ<strong>al</strong> of AppliedPsychology, 83, 347–359.Geddie, L., Fradin, S., & Beer, J. (2000). Child characteristics whichimpact accuracy of rec<strong>al</strong>l and suggestibility in preschoolers: Isage the best predictor? Child Abuse and Neglect, 24, 223–235.Gee, S., Gregory, M., & Pipe, M. E. (1999). “What colour isyour p<strong>et</strong> dinosaur?” The impact of pre-interview training andquestion type of children’s answers. Leg<strong>al</strong> and Criminologic<strong>al</strong>Psychology, 4, 111–128.Gold, P. E. (1987). Swe<strong>et</strong> memories. American Psychologist, 75,151–155.Goldhaber, D. E. (2000). Theories of human development: Integrativeperspectives. Mountain View, CA: Mayfield.Goodman, G. S., Bottoms, B. L., Schwartz-Kenney, G., & Rudy,L. (1991). Children’s memory for a stressful event: Improvingchildren’s reports. Journ<strong>al</strong> of Narrative and Life History, 1,69–99.Goodman, G. S., Hirschman, J., & Rudy, L. (1987, April). Children’stestimony: Research and policy implications. In S. Ceci(Chair), Children as witnesses: Research and soci<strong>al</strong> policy implications.Symposium conducted at the me<strong>et</strong>ing of the Soci<strong>et</strong>yfor Research in Child Development, B<strong>al</strong>timore, MD.Goodman, G. S., & Quas, J. A. (1997). Trauma and memory: Individu<strong>al</strong>differences in children’s recounting of a stressful experience.In N. Stein, P. A. Ornstein, C. J. B. Tversky, & C. J.Brainerd (Eds.), Memory for everyday and emotion<strong>al</strong> events(pp. 267–294). Hillsd<strong>al</strong>e, NJ: Erlbaum.Goodman, G. S., Quas, J. A., Batterman-Faunce, J. M.,Riddlesberger, M. M., & Kuhn, J. (1997). Children’s reactionsto and memory for a stressful event: Influences of age,
Children’s Testimony 179anatomic<strong>al</strong> dolls, knowledge and parent<strong>al</strong> attachment. AppliedDevelopment<strong>al</strong> Science, 1, 54–75.Goodman, G. S., Rudy, L., Bottoms, B. L., & Aman, C. (1990). Children’sconcerns and memory: Issues of ecologic<strong>al</strong> v<strong>al</strong>idity inthe study of children’s eyewitness testimony. In R. Fivush & J.A. Hudson (Eds.), Knowing and remembering in young children(pp. 249–284). New York: Cambridge University Press.<strong>Gordon</strong>, B. N., & Follmer, A. (1994). Development<strong>al</strong> issues in judgingthe credibility of children’s testimony. Journ<strong>al</strong> of Clinic<strong>al</strong>Child Psychology, 23, 283–294.<strong>Gordon</strong>, B. N., Ornstein, P. A., Nida, R. E., Follmer, A., Crenshaw,M. C., & Albert, G. (1993). Does the use of dolls facilitatechildren’s memory of visits to the doctor? Applied CognitivePsychology, 7, 1–16.<strong>Gordon</strong>, B. N., Schroeder, C. S., Ornstein, P. A., & Baker-Ward,L. (1995). Clinic<strong>al</strong> implications of research on memory development.In T. Ney (Ed.), True and f<strong>al</strong>se <strong>al</strong>legations of childsexu<strong>al</strong> abuse: Assessment and case management (pp. 99–124).New York: Brunner/Mazel.Greenhoot, A. F. (2000). Remembering and understanding: Theeffects of changes in underlying knowledge on children’s recollections.Child Development, 71, 1309–1328.Greenhoot, A. F., Ornstein, P. A., <strong>Gordon</strong>, B. N., & Baker-Ward, L.(1999). Acting out the d<strong>et</strong>ails of a pediatric check-up: The impactof interview condition and behavior<strong>al</strong> style on children’smemory reports. Child Development, 70, 363–380.Gudjonsson, G. H. (1989). Compliance in an interrogative situation:A new sc<strong>al</strong>e. Person<strong>al</strong>ity and Individu<strong>al</strong> Differences, 10, 535–540.Haden, C. A., Ornstein, P. A., Eckerman, C. O., & Didow, S. M.(in press). Mother–child conversation<strong>al</strong> interactions as eventsunfold: Linkages to subsequent remembering. Child Development.Heim, C., Newport, J., Heit, S., Graham, Y. P., Wilcox, M., Bons<strong>al</strong>l,R., Miller, A. H., & Nemeroff, C. B. (2000). Pituitary-adren<strong>al</strong>and autonomic responses to stress in women after sexu<strong>al</strong> andphysic<strong>al</strong> abuse in childhood. Journ<strong>al</strong> of the American Medic<strong>al</strong>Association, 284, 592–597.Hertsgaard, L., Gunnar, M., Erickson, M. R., & Nachmias, M.(1995). Adrenocortic<strong>al</strong> responses to the strange situation in infantswith disorganized/disoriented attachment relationships.Child Development, 66, 110–1106.Hovde, A. M., & Strapp, C. M. (1999, April). The impact of interviewereye level on amount and accuracy of children’s rec<strong>al</strong>l.Poster presented at the bienni<strong>al</strong> me<strong>et</strong>ing of the Soci<strong>et</strong>y forResearch in Child Development, Albuquerque, NM.Howard, A. N., Osborne, H. L., & Baker-Ward, L. (1997, April).Childhood cancer survivors’ memory of their treatment afterlong delays. Poster presented at the bienni<strong>al</strong> me<strong>et</strong>ing of theSoci<strong>et</strong>y for Research in Child Development, Washington, DC.Howe, M. L. (1998). Individu<strong>al</strong> differences in factors that modulatestorage and r<strong>et</strong>riev<strong>al</strong> of traumatic memories. Development andPsychopathology, 10, 681–698.Howe, M. L. (2000). The fate of early memories: Development<strong>al</strong>science and the r<strong>et</strong>ention of childhood experiences. Washington,DC: American Psychologic<strong>al</strong> Association.Howe, M. L., & O’Sullivan, J. T. (1997). What children’s memoriestell us about rec<strong>al</strong>ling our childhoods: A review of storage andr<strong>et</strong>riev<strong>al</strong> processes in the development of long-term r<strong>et</strong>ention.Development<strong>al</strong> Review, 17, 148–204.Howie, P. M., & Dowd, H. J. (1996). Self-esteem and the perceivedobligation to respond: Effects on children’s testimony. Leg<strong>al</strong>and Crinimologic<strong>al</strong> Psychology, 1, 197–209.Hudson, J. A. (1990). Constructive processing in children’s eventmemory. Development<strong>al</strong> Psychology, 26, 180–187.Huffman, M. L., Crossman, A. M., & Ceci, S. J. (1997). “Are f<strong>al</strong>sememories permanent?” An investigation of the long-term effectsof source misattributions. Consciousness and Cognition:An Internation<strong>al</strong> Journ<strong>al</strong>, 6, 482–490.Jackson, S., & Crockenberg, S. (1998). A comparison of suggestiblityin 4-year-old girls in response to parent<strong>al</strong> or stranger misinformation.Journ<strong>al</strong> of Applied Development<strong>al</strong> Psychology,19, 527–542.Kail, R. V. (1989). The development of memory in children (2nded.). New York: Freeman.Lamb, M. E. (1994). The investigation of child sexu<strong>al</strong> abuse: An interdisciplinaryconsensus statement. Expert Evidence, 2, 151–156.Leichtman, M. D., & Ceci, S. J. (1995). The effects of stereotypesand suggestions on preschoolers’reports. Development<strong>al</strong>Psychology, 31, 568–578.Levine L. J., Stein N. L., & Liwag, M. D. (1999). Remembering children’semotions: Sources of concordant and discordant countsb<strong>et</strong>ween parents and children. Development<strong>al</strong> Psychology, 35,790–801.Lindsay, D. S., & Read, J. D. (1994). Psychotherapy and memoriesof childhood sexu<strong>al</strong> abuse: A cognitive perspective. AppliedCognitive Psychology, 8, 281–338.Loftus, E. F. (1979). Eyewitness testimony. Cambridge, MA: HarvardUniversity Press.Loftus, E. F. (1993). The re<strong>al</strong>ity of repressed memories. AmericanPsychologist, 48, 518–537.Loftus, E. F., & P<strong>al</strong>mer, J. C. (1974). Reconstruction of automobiledestruction: An example of the interaction b<strong>et</strong>ween languageand memory. Journ<strong>al</strong> of Verb<strong>al</strong> Learning and Verb<strong>al</strong> Behavior,13, 585–589.Loftus, E. F., & Pickrell, J. (1995). The formation of f<strong>al</strong>se memories.Psychiatric Ann<strong>al</strong>s, 25, 720–725.Mandler, J. M. (1990). Rec<strong>al</strong>l and its verb<strong>al</strong> expression. In R.Fivush & J. A. Hudson (Eds.), Knowing and remembering inyoung children (pp. 317–330). New York: Cambridge UniversityPress.Marche, T. A. (1997, April). Timing of misleading information influencespreschoolers’ susceptibility to misinformation effect.Poster presented at the Bienni<strong>al</strong> Me<strong>et</strong>ing of the Soci<strong>et</strong>y forResearch in Child Development, Washington, DC.Marche, T. A. (1999). Memory strength affects reporting of misinformation.Journ<strong>al</strong> of Experiment<strong>al</strong> Child Psychology, 73,45–71.Mazzoni, G. (1998). Memory suggestibility and m<strong>et</strong>acognition inchild eyewitness testimony: The roles of source monitoring andself-efficacy. European Journ<strong>al</strong> of Psychology of Education, 13,43–60.McCauley, M. R., & Fisher, R. P. (1995). Facilitating children’s eyewitnessrec<strong>al</strong>l with the revised Cognitive Interview. Journ<strong>al</strong> ofApplied Psychology, 80, 510–516.McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a speci<strong>al</strong>flashbulb-memory mechanism? Journ<strong>al</strong> of Experiment<strong>al</strong>Psychology: Gener<strong>al</strong>, 117, 171–181.Meltzoff, A. N. (1995). What infant memory tells us about infantileamnesia: Long-term rec<strong>al</strong>l and deferred imitation. Journ<strong>al</strong> ofExperiment<strong>al</strong> Child Psychology, 59, 497–515.Merritt, K. A., Ornstein, P. A., & Spicker, B. (1994). Children’smemory for a s<strong>al</strong>ient medic<strong>al</strong> procedure: Implications for testimony.Pediatrics, 94, 17–23.Michel, M. K. (<strong>2001</strong>). The effects of interviewing m<strong>et</strong>hod on children’srec<strong>al</strong>l. Unpublished doctor<strong>al</strong> dissertation, Departmentof Psychology, University of North Carolina at Chapel Hill.Michel, M. K., <strong>Gordon</strong>, B. N., Ornstein, P. A., & Simpson,M. A. (2000). The abilities of children with ment<strong>al</strong> r<strong>et</strong>ardationto remember person<strong>al</strong> experiences: Implications fortestimony. Journ<strong>al</strong> of Clinic<strong>al</strong> Child Psychology, 29, 453–463.Milne, R., & Bull, R. (1996). Interviewing children with mild learningdisability with the Cognitive Interview. Issues in Criminologic<strong>al</strong>and Leg<strong>al</strong> Psychology, 26, 44–51.Muir-Broaddus, J., King, T., Downey, D., & P<strong>et</strong>ersen, M. (1998).Conservation as a predictor of individu<strong>al</strong> differences in
180 <strong>Gordon</strong>, Baker-Ward, and Ornsteinchildren’s susceptibility to leading questions. PsychonomicBull<strong>et</strong>in and Review, 5, 454–458.Myles-Worsley, M., Cromer, C. C., & Dodd, D. H. (1986). Children’spreschool script reconstruction: Reliance on gener<strong>al</strong> knowledgeas memory fades. Development<strong>al</strong> Psychology, 22, 22–30.Nadel, L., & Jacobs, W. J. (1998). Traumatic memory is speci<strong>al</strong>.Current Directions in Psychologic<strong>al</strong> Science, 7, 154–157.Nelson, C. A., & Carver, L. J. (1998). The effects of stress andtrauma on brain and memory: A view from development<strong>al</strong>cognitive neuroscience. Development and Psychopathology,10, 793–809.Nelson, K. (1986). Event knowledge: Structure and function in development.Hillsd<strong>al</strong>e, NJ: Erlbaum.Nelson, K. (1993). The psychologic<strong>al</strong> and soci<strong>al</strong> origins of autobiographic<strong>al</strong>memory. Psychologic<strong>al</strong> Science, 4, 1–8.Orbach, Y., Hershkowitz, I., Lamb, M. E., Sternberg, K. J., Esplin,P. W., & Horowitz, D. (2000). Assessing the v<strong>al</strong>ue of structuredprotocols for forensic interviews of <strong>al</strong>leged abuse victims. ChildAbuse and Neglect, 24, 733–752.Ornstein, P. A. (1995). Children’s long-term r<strong>et</strong>ention of s<strong>al</strong>ientperson<strong>al</strong> experiences. Journ<strong>al</strong> of Traumatic Stress, 8, 581–605.Ornstein, P. A., Baker-Ward, L., <strong>Gordon</strong>, B. N., & Merritt, K. A.(1997). Children’s memory for medic<strong>al</strong> experiences. AppliedCognitive Psychology, 11, S87–S104.Ornstein, P. A., Ceci, S. J., & Loftus, E. F. (1998). Adult recollectionsof childhood abuse: Cognitive and development<strong>al</strong> perspectives.Psychology, Public Policy, and Law, 4, 1025–1051.Ornstein, P. A., <strong>Gordon</strong>, B. N., & Baker-Ward, L. (1992). Children’smemory for s<strong>al</strong>ient events: Implications for testimony. In M.L. Howe, C. J. Brainerd, & V. F. Reyna (Eds.), Development oflong-term r<strong>et</strong>ention (pp. 135–158). New York: Springer-Verlag.Ornstein, P. A., <strong>Gordon</strong>, B. N., & Larus, D. M. (1992). Children’smemory for a person<strong>al</strong>ly experienced event:Implications fortestimony. Applied Cognitive Psychology, 6, 49–60.Ornstein, P. A., Larus, D. M., & Clubb, P. A. (1991). Understandingchildren’s testimony: Implications of research on the developmentof memory. In R. Vasta (Ed.), Ann<strong>al</strong>s of childdevelopment (Vol. 8, pp. 145–176). London: Jessica KingsleyPublishers.Ornstein, P. A., Merritt, K. A., Baker-Ward, L., <strong>Gordon</strong>, B. N., Furtado,E., & Principe, G. (1998). Children’s knowledge, expectation,and long-term r<strong>et</strong>ention. Applied Cognitive Psychology,12, 387–405.Ornstein, P. A., & Naus, M. J. (1985). Effects of the knowledgebase on children’s memory strategies. In H. W. Reese (Ed.),Advances in child development and behavior (Vol. 19, pp. 113–148). Orlando, FL: Academic Press.Ornstein, P. A., Shapiro, L. R., Clubb, P. A., Follmer, A., & Baker-Ward, L. (1997). The influence of prior knowledge on children’smemory for s<strong>al</strong>ient medic<strong>al</strong> expeirences. In N. Stein,P. A. Ornstein, C. J. Brainerd, & B. Tversky (Eds.), Memoryfor everyday and emotion<strong>al</strong> events (pp. 83–111). Hillsd<strong>al</strong>e, NJ:Erlbaum.P<strong>et</strong>ers, D. P. (1991). The influence of stress and arous<strong>al</strong> on the childwitness. In J. Doris (Ed.), The suggestibility of children recollections(pp. 60–76). Washington, DC: American Psychologic<strong>al</strong>Association.P<strong>et</strong>erson, C. (1999). Children’s memory for medic<strong>al</strong> emergencies:2 years later. Development<strong>al</strong> Psychology, 35, 1493–1506.P<strong>et</strong>erson, C., & Bell, M. (1996). Children’s memory for traumaticinjury. Child Development, 67, 3045–3070.P<strong>et</strong>erson, C., & Rideout, R. (1998). Memory for medic<strong>al</strong> emergenciesexperienced by 1- and 2-year-olds. Development<strong>al</strong>Psychology, 34, 1059–1072.Pezdek, K., & Roe, C. (1995). The effect of memory trace strengthon suggestibility. Journ<strong>al</strong> of Experiment<strong>al</strong> Child Psychology,60, 116–128.Pezdek, K., & Taylor, J. (in press). Memory for traumatic eventsby children and adults. In M. L. Eisen, G. S. Goodman, &J. A. Quas (Eds.), Memory and suggestibility in the forensicinterview. Mahwah, NJ: Lawrence Erlbaum and Associates.Pipe, M. E., Dean, J., Canning, J., & Murachver, T. (1996, July). Narratingevents and telling stories. Paper presented at the secondInternation<strong>al</strong> Conference on Memory, Albano, It<strong>al</strong>y.Pollak, S. D., Cicch<strong>et</strong>ti, D., & Klorman, R. (1998). Stress, memory,and emotion: Development<strong>al</strong> considerations from the studyof child m<strong>al</strong>treatment. Development and Psychopathology, 10,811–828.Pollak, S. D., Cicch<strong>et</strong>ti, D., Klorman, R., & Brumaghim, J. T.(1997). Cognitive brain event-related potenti<strong>al</strong>s and emotionprocessingin m<strong>al</strong>treated children. Child Development, 68,773–787.Poole, D. A., & Lamb, M. E. (1998). Investigative interviews ofchildren: A guide for helping profession<strong>al</strong>s. Washington, DC:American Psychologic<strong>al</strong> Association.Poole, D. A., & Lindsay, D. S. (1996, June). Effects of parent<strong>al</strong> suggestions,interviewing techniques, and age on young children’sevent reports. Paper presented at the NATO Advanced StudyInstitute, Recollections of Trauma: Scientific research and clinic<strong>al</strong>practice, Port de Bourgenay, France.Poole, D. A., & White, L. T. (1993). Two years later: Effects ofquestion rep<strong>et</strong>ition and r<strong>et</strong>ention interv<strong>al</strong> on the eyewitnesstestimony of children and adults. Development<strong>al</strong> Psychology,29, 844–853.Poole, D. A., & White, L. T. (1995). Tell me again and again: Stabilityand change in the repeated testimonies of children and adults.In M. S. Zaragoza, J. R. Graham, G. C. N. H<strong>al</strong>l, R. Hirschman,Y. S. Ben-Porath (Eds.), Memory and testimony in the childwitness (pp. 24–43). Thousand Oaks, CA: Sage Publications.Pope, K. S., & Brown, L. S. (1996). Recovered memories of abuse:Assessment, therapy, forsenics. Washington, DC: AmericanPsychologic<strong>al</strong> Association.Powell, M. B., Roberts, K. P., Ceci, S. J., & Hembrooke, H. (1999).The effects of repeated experience on children’s suggestibility.Development<strong>al</strong> Psychology, 35, 1462–1477.Powell, M. B., & Thomson, D. M. (1996). Children’s rec<strong>al</strong>l of an occurrenceof a repeated event: Effects of age, r<strong>et</strong>ention interv<strong>al</strong>,and question type. Child Development, 67, 1988–2004.Priestley, G., Roberts, S., & Pipe, M. E. (1999). R<strong>et</strong>urning to thescene: Reminders and context reinstatement enhance children’srec<strong>al</strong>l. Development<strong>al</strong> Psychology, 35, 1006–1014.Principe, G. F., Ornstein, P. A., Baker-ward, L., & <strong>Gordon</strong>, B.N. (2000). The effects of intervening experiences on children’smemory for a physic<strong>al</strong> examination. Applied CognitivePsychology, 14, 59–80.Putnam, F. (1997). Dissociation in children and adolescents. A development<strong>al</strong>perspective. NY: Guilford.Quas, J. A., Qin, J., Schaaf, J., & Goodman, G. S. (1997). Individu<strong>al</strong>differences in children’s and adults’ suggestibility and f<strong>al</strong>seevent memory. Learning and Individu<strong>al</strong> Differences, 9, 359–390.Reyna, V. F., & Brainerd, C. J. (1995). Fuzzy-trace theory: An interimsynthesis. Learning and Individu<strong>al</strong> Differences, 7, 1–75.Ricci, C. M., & Be<strong>al</strong>, C. R. (1998). Child witnesses: Effect of eventknowledge on memory and suggestibility. Journ<strong>al</strong> of AppliedDevelopment<strong>al</strong> Psychology, 19, 305–317.Richardson, G., & Kelly, T. P. (1995). The relationship b<strong>et</strong>weenintelligence, memory and interrogative suggestibility in youngoffenders. Psychology, Crime and Law, 1, 283–290.Roberts, K. P., & Lamb, M. E. (1999). Children’s responses wheninterviewers distort d<strong>et</strong>ails during investigative interviews. Leg<strong>al</strong>and Criminologic<strong>al</strong> Psychology, 4, 23–31.Roberts, K. P., Lamb, M. E., Sternberg, K. J., Beresford, J.,Domenici-Lake, P. L., & Heiges, K. (1997, April). The effectof a delay on the incorporation of postevent information intochildren’s eyewitness memory. Paper presented at the bienni<strong>al</strong>me<strong>et</strong>ing of the Soci<strong>et</strong>y for Research in Child Development,Washington, DC.
Children’s Testimony 181Robinson, J., & Briggs, P. (1997). Age trends and eye-witness suggestibilityand compliance. Psychology, Crime and Law, 3,187–202.Roediger, H. L., III, & Bergman, E. T. (1998). The controversy overrecovered memories. Psychology, Public Policy, and Law, 4,1091–1109.Ross, M. (1989). The relation of implicit theories to the constructionof person<strong>al</strong> histories. Psychologic<strong>al</strong> Review, 96, 341–357.Rovee-Collier, C., & Shyi, G. (1992). A function<strong>al</strong> and cognitivean<strong>al</strong>ysis of infant long-term r<strong>et</strong>ention. In M. L. Howe, C. J.Brainerd, & V. F. Reyna (Eds.), Development of long-term r<strong>et</strong>ention(pp. 3–55). New York: Springer-Verlag.Rudy, L., & Goodman, G. S. (1991). Effects of participation onchildren’s reports: Implications for children’s testimony. Development<strong>al</strong>Psychology, 27, 527–538.S<strong>al</strong>mon, K., & Pipe, M. E. (1997). Providing props to facilitateyoung children’s event rec<strong>al</strong>l: The impact of a one year delay.Journ<strong>al</strong> of Experiment<strong>al</strong> Child Psychology, 65, 261–292.Saywitz, K. J. (1995). Improving children’s testimony: The question,the answer, and the environment. In M. S. Zaragoza, J. R.Graham, G. C. N. H<strong>al</strong>l, R. Hirschman, Y. S. Ben-Porath (Eds.),Memory and testimony in the child witness (pp.113–140). ThousandOaks, CA: Sage Publications.Saywitz, K. J. (in press). Development<strong>al</strong> underpinnings. In H.Westcott, G. Davies, & R. Bull (Eds.), Children’s testimonyin context. New York: Wiley.Saywitz, K. J., Goodman, G. S., Nicholas, E., & Moan, S. (1991).Children’s memories of physic<strong>al</strong> examinations involving genti<strong>al</strong>touch: Implications for reports of child sexu<strong>al</strong> abuse. Journ<strong>al</strong>of Consulting and Clinic<strong>al</strong> Psychology, 59, 682–691.Saywitz, K. J., & Nathanson, R. (1993). Children’s testimony andtheir perceprions of stress in and out of the courtroom. ChildAbuse and Neglect, 17, 613–622.Saywitz, K. J., Nathanson, R., & Snyder, L. (1993). Credibility ofchild witnesses: The role of communicative comp<strong>et</strong>ence. Topicsin Language Disorders, 13, 59–78.Saywitz, K. J., & Snyder, L. (1993). Improving children’s testimonywith preparation. In G. Goodman & B. Bottoms (Eds.),Child victims, child witnesses: Understanding and improvingtestimony (pp. 117–146). New York: Guilford Press.Schacter, D.L. (1992). Understanding implicit memory: A cognitiveneuroscience approach. American Psychologist, 47, 559–569.Schneider, W., & Bjorklund, D. F. (1998). Memory. In W. Damon(Series Ed.) & D. Kuhn & R. S. Siegler (Vol. Eds.), Handbookof child psychology: Vol. 2. Cognition, perception, and language(pp. 467–521). New York: Wiley.Schneider, W., & Pressley, M. (1997). Memory development b<strong>et</strong>ween2 and 20 (2nd ed.). New York: Springer-Verlag.Schwarzmueller, A., Boyle, P., & Fivush, R. (1996, March). Autobiographic<strong>al</strong>memory from age 3 through 8. Poster presented atthe Conference on Human Development, Birmingham, AL.Shiffrin, R. M., & Atkinson, R. C. (1969). Storage and r<strong>et</strong>riev<strong>al</strong> processesin long-term memory. Psychologic<strong>al</strong> Review, 76, 179–193.Simpson, C., & Guttentag, R. (1996, March). Suggestibility ofchildren’s remembering: Effects of interviewer credibility.Poster presented at the Conference on Human Development,Birmingham, AL.Steward, M. S., & Steward, D. S. (with L. Farquhar, J. E. B. Myers, M.Reinhart, J. Welker, N. Joye, J. Driskill, & J. Morgan) (1996).Interviewing young children about body touch and handling.Monographs of the Soci<strong>et</strong>y for Research in Child Development,61(4/5, Seri<strong>al</strong> No. 248).Templ<strong>et</strong>on, L. M., & Hunt, V. H. (1997, April). The effects of misleadinginformation and level or authority of interviewer onchildren’s eyewitness memory. Poster presented at the bienni<strong>al</strong>me<strong>et</strong>ing of the Soci<strong>et</strong>y for Research in Child Development,Washington, DC.Templ<strong>et</strong>on, L. M., & Wilcox, S. A. (2000). A t<strong>al</strong>e of two representations:The misinformation effect and children’s developingtheory of mind. Child Development, 71, 402–416.Tessler, M., & Nelson, K. (1994). Making memories: The influenceof joint encoding on later rec<strong>al</strong>l by young children. Consciousnessand Cognition, 3, 307–326.Toglia, M. P., & Ross, D. F. (1991, April). Children’s suggestibility:The role of biaser credibility. Paper presented at the bienni<strong>al</strong>me<strong>et</strong>ing of the Soci<strong>et</strong>y for Research in Child Development,Seattle, WA.Tulving, E., & Thomson, D. M. (1973). Encoding specificity andr<strong>et</strong>riev<strong>al</strong> processes in episodic memory. Psychologic<strong>al</strong> Review,80, 352–373.Vrij, A., & Bush, N. (2000). Differences in suggestibility b<strong>et</strong>ween5–6 and 10–11 year olds: The relationship with self confidence.Psychology, Crime and Law, 6, 127–138.W<strong>al</strong>ker, A. G., & Warren, A. R. (1995). The language of the childabuse interview: Asking the questions, understanding the answers.In T. Ney (Ed.), True and f<strong>al</strong>se <strong>al</strong>legations of child sexu<strong>al</strong>abuse: Assessment and case management (pp. 153–162). NewYork: Brunner/Mazel.Warren, A. R., & Lane, P. (1995). Effects of timing and type ofquestioning on eyewitness accuracy and suggestibility. In M.S. Zaragoza, J. R. Graham, G. C. N. H<strong>al</strong>l, R. Hirschman, Y. S.Ben-Porath (Eds.), Memory and testimony in the child witness(pp. 44–60). Thousand Oaks, CA: Sage Publications.Warren, A. R., & Smartwood, J. (1992). Development<strong>al</strong> issues inflashbulb memory research: Children rec<strong>al</strong>l the Ch<strong>al</strong>lengerevent. In E. Winograd & U. Neisser (Eds.), Affect and accuracyin rec<strong>al</strong>l: The problem of flashbulb memories (pp. 95–12). NewYork: Cambridge University Press.Warren, A. R., Wood<strong>al</strong>l, C. E., Thomas, M., Nunno, M., Keeney,J. M., Larson, S. M., & Stadfeld, J. A. (1999). Assessingthe effectiveness of a training program for interviewingchild witnesses. Applied Development<strong>al</strong> Science, 3, 128–135.Welch-Ross, M. K., Diecidue, K., & Miller, S. A. (1997). Youngchildren’s understanding of conflicting ment<strong>al</strong> representationpredicts suggestibility. Development<strong>al</strong> Psychology, 33, 43–53.White, T. L., Leichtman, M. D., & Ceci, S. J. (1997). The good,the bad, and the ugly: Accuracy, inaccuracy, and elaborationin preschoolers’ reports about a past event. Applied CognitivePsychology, 11, S37–S54.