M.B. Kovera, A.W. Gresham, E. Borgida, E. Gray & P.C.Does Expert Psyc - PDF document

M.B. Kovera, A.W. Gresham, E. Borgida, E. Gray & P.C.Does Expert Psychological Testimony Inform or InfluenceJuror Decision Making? A Social Cognitive Analysis.178 (1997).David F. Ross et al., Age Stereotypes, Communication Modality, and Mock JurorsÕ Perceptions of the Child WitnessERSPECTIVESON4.Michelle D. Leichtman & Stephen J. Ceci, The Effects ofStereotypes and Suggestions on PreschoolersÕ ReportsEVELOPMENTAL.568 (1995). Õm grateful to Dr. Martindale for introducing the reader to animportant and lively debate among practitioners and acade-mics over the relevance of recent research on childrenÕs sug-gestibility. In my article, I argued that therecent research on suggestibility was inspired by highly coer-cive interviewing techniques in widely publicized cases thatare not the norm in child sexual abuse investigations. Thesetechniques include telling children that they have been abused,telling children that a particular person is the abuser, and ask-ing children to imagine details regarding how abuse could havetaken place. Moreover, I argued that the research fails to mir-ror factors in real-world sexual-abuse cases that reduce thelikelihood that false allegations will occur. These factorsinclude the age of the child, childrenÕs reluctance to accuseloved ones of immoral acts, and childrenÕs embarrassmentregarding sexual topics. My goal was to alert judges, attorneys, and child-abuse pro-fessionals to the importance of carefully examining the meth-ods used by researchers before concluding that research appliesto a particular case. Certainly there are cases in which highlycoercive tactics have been used, and in which expert testimonyon the dangers of such tactics is justified. But it is just as cer-When expert testimony on the suggestibility of children isoffered, the court must consider whether the research theexpert will discuss sufficiently fits the facts of the case to behelpful to the jury.Jurors are likely to be influenced by experttestimony even if it fails to fit the facts of the case, in partbecause they are less adept at detecting a lack of fit, and in partbecause they naturally assume that an expert testifying for theMoreover, unlike casesinvolving adult eyewitnesses, most jurors come to court readyto doubt the reliability of a child witness.The need for screening potentially unhelpful and prejudi-cial expert testimony is accentuated when an expert knows lit-tle about the case in which he or she is testifying. As Dr.Martindale notes, experts routinely testify without familiariz-ing themselves with case-specific details, in order to retainobjectivity. Moreover, experts will almost inevitably make claims on the stand that they would be hesitant to make weretheir words subject to peer review and publication: as Dr.Martindale argues, Òan analysis of an expertÕs testimony pro-vides more information about the expertÕs performance underpressure than it does about the expertÕs findings.Ó The judgemust therefore scrutinize the applicability of the expertÕs find-ings before the expert is allowed to take the stand. If the expertis allowed to testify, the court should limit his or her discus-sion to research that applies to the case at bar. WHAT ARE THE RESEARCH FINDINGS?Dr. Martindale defends the applicability of the research toreal-world abuse cases and the relevance of expert testimonyon suggestibility without describing any of the research itself.In the hands of a less conscientious expert than Dr. Martindale,this argument can lead to mischief. Testifying experts willoften make blanket claims about the unreliability of children,or report the results of research without describing themethodology or the potential limits of the researchÕs applica-bility to the real world. LetÕs consider one of the most oft-cited studies demonstrat-ing the suggestibility of childrenÑthe Sam Stone study, which, a peer-reviewedThe study showed that a combination of suggestive interview-ing techniques led 72% of younger children to assert falselythat a stranger named Sam Stone had come to their preschooland committed various misdeeds. Children often embellishedtheir false stories with perceptual details and nonverbal ges-tures, making their reports highly credible. Experts often citethe study as evidence that children can be led to make false yetCloser examination reveals the lengths to which theresearchers worked to obtain false reports, and the extent towhich false allegations of sexual abuse are less likely in the realworld. One of the suggestive techniques the researchers usedwas Òstereotype induction,Ó which they analogized to negativeResearch assistants visited each child on four consecutiveweeks before Sam Stone came to the preschool, and providedthe child with details of 12 different misdeeds that the assis- Court Review - Fall 2001LetÕs Not Exaggerate the Thomas D. Lyon Amye R. Warren et al., ÒIt Sounds Good in Theory, But....ÓDo Investigative Interviewers Follow Guidelines Based on MemoryResearch?ALTREATMENT231 (1996). Michael Lamb,Kathleen Sternberg, and collegues at the National Institute ofies demonstrating how interviewers tend to ask closed-endedrather than open-ended questions. For one of the most recentKathleen J. Sternberg, Michael E. Lamb, Graham M. Davies, & Helen L. Westcott, Theory Versus ApplicationMaggie Bruck, Stephen J. Ceci, & Helene Hembrooke,Reliability and Credibility of Young ChildrenÕs Reports: FromResearch to Policy and Practice.136, 139 (notingincrease in suggestibility as one moves from yes/no questions toquestions that presuppose the truth of the suggested information,, ÒIs there a cabinet in the room?Ó to ÒIsnÕt there a cabinet inthe room?Ó to ÒIs the door open in the cabinet in the room?Ó). tants had purportedly witnessed Sam perform. Sam StoneÕsvidual children. Another suggestive technique the researchersused was suggestive questioning. For four weeks after SamÕsvisit, an interviewer questioned each child each week. In thefirst interview, the interviewer showed the child a soiled teddymight have done it. In the next three interviews the inter-viewer asked a series of highly suggestive questions. Thesequestions presupposed that Sam Stone had ripped the book orsoiled the teddy bear, did not give the child an opportunity todetails of the fictitious events. For example, ÒDid Sam StoneSam Stone got the bear dirty, did he do it by accident, or onpurpose?Ó These questions were asked regardless of whetherthe child affirmed or denied that Sam Stone had performed anymisdeeds. Ten weeks later, all children were interviewed in anonleading fashion. At that time, 72% of the three and fourThe high rates of false reports are impressive. But equallyimpressive are other details of the study that are often over-looked. First, the authors report Òdramatic developmentaltrendsÓ in childrenÕs susceptibility to the suggestive tech-The rate of false reports among the older preschool-ers, who were five to six years of age, was about half of that ofthe younger children. School-age children would be even lesslikely to succumb to the interviewerÕs pressures. One of themost consistent findings in the suggestibility literature is thatpreschool children are particularly vulnerable to suggestivequestioning, and preschool children predominate in recentresearch documenting the unreliability of childrenÕs testimony.Second, the study was unusual in that the final interviewcontained two questions mildly skeptical of the childrenÕsclaims. Asking children if they saw the events reduced thenumber of false reports by about half. Asking the children,ÒYou didnÕt really see him rip the book (or soil the bear), didyou?Ó cut the number by half again. Having been exposed tofour trials of stereotype induction and three trials of suggestivequestioning, 21% of the three and four year olds (and only 5%occurred.Dr. Martindale might respond that he is not reassured bythese numbers, and I would agree that any false allegation offor the child and his or her family. Yet these percentages exag- sexual abuse, compared toteddy bear. How close thenumbers get to zero is any-bodyÕs guess. important facts about sug-gestibility research. First,there are large age differ-ences in suggestibility. I am struck by how many experts appearto overlook the truism that just as preschoolers are much moresuggestible than school-age children, school-age children aremuch less suggestible than preschoolers. Second, children areboth suggestible counter-suggestible. Researchers often failto test the persistence of their suggestions; Sam Stone is anexception, and dramatically reduced the number of false reports.Third, much of the suggestibility research elicits false narrativesfrom young children by telling them that the events occurred (asopposed to merely asking them), and by providing them withdetails with which they can imagine the events. Indeed, theseare the techniques researchers have used to create substantialnumbers of false childhood memories in adults. But the issueis not whether children whether they arebeing led by current investigative methods. WHAT SORT OF INTERVIEWS ARE OCCURRING IN THELeichtman and Ceci asserted that the techniques in the SamStone study were based on Òreal-world forensic conditions.ÓIn my 1999 paper, I questioned the applicability of the SamStone study and other studies to the real world, pointing outthat the research on actual interviews had not documented thewidespread use of techniques such as stereotype induction.been documented was that interviewers were askingvery few open-ended questions, and relying heavily on closed-ended questions (yes-no questions and forced-choice ques-Closed-ended questions are often considered Òlead-ing,Ó and I believe they are being overused. In my presenta-tions to child interviewers, I emphasize the need for structuredinterview protocols and the potential benefits of rapport build-ing and greater use of open-ended questions as means ofincreasing information without reducing reliability. All thesame, closed-ended questions are far less leading than theThey are Fall 2001 - Court Review [T]he issue is notwhether childrencanbe led to makefalse allegations,but whether theyarebeing led bycurrent investigativemethods. Stephen J. Ceci and Richard D. Friedman, Children: Scientific Research and Legal Implications,.33, 46 (2000) (ÒThere is no serious doubt that directed ques-tioning will often be far more effective than requests for free recallin securing disclosure of abuse.Ó).Thomas D. Lyon, The New Wave in ChildrenÕs SuggestibilityResearch: A Critique.1004, 1046 (1999).11.Amye R. Warren, Sena Garven, Nancy E. Walker, & Cara E. Woodall, Setting the Record Straight: How Problematic Are ÔTypicalÕ(paper presented at the biennialmeeting of the American Psychology-Law Society, New Orleans,Louisiana, March 2000) (copy on file with author).12.For example, using the phrase Òdo you thinkÓ was classified as animagination or solve a mystery. less offensive than many of the techniques documented in thenotorious day-care cases like Kelly Michaels and the McMartincase. Closed-ended questions are also necessary in someAbused children are often quite reluctant to describeAfter my article appeared, Amye Warren and her colleaguesclusion? ÒWe believe that our results regarding the frequency ofproblematic techniques in ÔtypicalÕ child sexual abuse inter-views are encouraging. It appears that the assumptions madeby Lyon (1999) about the rarity of some of the most egregiousinterviewing practices (e.g. referring to what other people havesaid) in ÔtypicalÕ interviews may be well-founded.Óauthors issued several caveats: the interviews might not be rep-resentative of all interviews, and the researchers might haveMost importantly, theauthors stressed that the infrequent use of improper techniquesdoes not make an interview a good interview. Nevertheless, thestudy supported my basic assertion: studies like Sam Stoneexaggerate the suggestiveness of real-world interviews.In response to the relative infrequency of the highly sugges-tive techniques favored by suggestibility researchers, Dr.Martindale argues that the Òdeleterious effect of one strongsuggestion from an authoritative source is not likely to bediminished simply because it is followed by numerous non-suggestive questions.Ó I know of no research to support thisassertion, and Dr. Martindale does not offer any. The recentresearch on preschool childrenÕs suggestibility does not stopwith Òone strong suggestion.Ó It is the dogged persistence ofcoercive interviewers that reliably produces false narratives inyoung children. The cases that inspired the research involvedunrelenting suggestion over long periods of time by interview-ers utterly convinced that abuse had occurred. The mistakescommitted by investigative interviewers in less sensationalcases tend to be much more mundane. We clearly need to improve the quality of interviewing: weshould provide more training, more supervision, and moreresources. But the liberal receipt of expert testimony on theeffects of highly suggestive interviewing techniques onpreschool children is more likely to simply increase the num-ber of acquittals across the board than to improve interviewingI agree with Dr. Martindale that the recent research on chil-drenÕs suggestibility has done a lot of good. It has spawned several research programs aimed at improving the process bywhich children are interviewed about abuse. In appropriatecases, it can educate judges and jurors about the dangers ofhighly suggestive interviewing with young children. It haslargely silenced extremist claims that childrenÕs abuse allega-tions are never false or that children are no more suggestibleHowever, we must not forget that the extremist claims werefounded on overgenerous interpretations of research findingsurprisingly low rates of false reports among young children.Research highlighting high rates of error can easily lead to sim-ilarly unfounded claims about childrenÕs reliability. Theseclaims, in turn, can reinforce commonsense doubts about chil-drenÕs reliability and inherent reluctance to confront child sex-The solution is quite straightforward: judges must take careto assess the applicability of suggestibility research on a case-by-case basis. Experts seeking to testify must describe theresearch with sufficient specificity to allow the court to assesswhether the research fits the fact of the case. If it doesnÕt fit,you donÕt admit. Similarly, suggestibility research offered forother purposes (such as for assessing the reliability of hearsay)should be scrutinized with similar care. Judges should keep inmind the importance of the childÕs age, the suggestive influ-ences at issue, and the relationship of the child to the allegedoffender. Through judicious gatekeeping, extremist claimsabout suggestibility can be kept out of the courtroom.Thomas D. Lyon is a law professor at theLaw. He obtained his law degree from Harvard University School of Lawin 1987 and earned a Ph.D. in psychology fromStanford University in 1994. He has been on thefaculty at the USC law school since 1995, wherehe teaches evidence, psychology and law, andquantitative methods in law. Before joining USC, he served as aresearch associate at Harbor-UCLA Medical Center and was anattorney with the ChildrenÕs Services Division of the Los AngelesCounty CounselÕs office. Professor Lyon can be reached attlyon@law.usc.edu. Court Review - Fall 2001