Two stories of eyewitness error


	Print or Email This Page November 2003, Page p24 Two stories of eyewitness error By *James M. Doyle* Then the prosecutor argues, “You can believe an eyewitness,” and the defense lawyer simply counters, “No, you can’t,” the prosecutor wins. Even worse, the research indicates that when the prosecutor’s eyewitness is wrong, the prosecutor wins anyway. Of the first 100 wrongful convictions proven by DNA technology, over 80 percent relied to an important extent on sincere, confident, mistaken eyewitnesses.1�Jurors (and judges) believe eyewitnesses, or at least they find it hard to admit that they don’t believe them. An enormous (and still growing) body of scientific psychological knowledge addresses this problem, but it hasn’t been easy to mobilize the research to show jurors that a particular eyewitness has made a mistake in a particular case. This follows from the nature of the psychological studies. Experiments can tell you that under certain conditions an eyewitness will be mistaken 80 percent of the time, but no experiment — or group of experiments — can tell you whether this�case is one of the 80 percent or one of the 20 percent. This fundamental statistical, probability-driven, nature of the psychological knowledge is no barrier to reforming investigative procedures and reducing the number of errors by preventing them before they happen. Working on a “public health,” prevention model, the National Institute of Justice has published its influential Eyewitness Evidence: A Guide for Law Enforcement. The New Jersey Attorney General’s office has instituted a state-wide program requiring research-developed, double-blind, sequential identification procedures. Other courts, even some prosecutors, have begun (at a glacial pace) to show interest in these new models.2 But trials look back in time: trials try to determine whether an eyewitness mistake has occurred in the past, and the statistical nature of psychological knowledge presents a very difficult challenge to that using the science in that effort. One reason that trial judges are hostile to expert testimony on eyewitness performance is that they recognize that the eyewitness experts are incapable of determining whether this eyewitness is wrong or right. Judges are afraid that the experts will only raise the general level of skepticism about all eyewitnesses — taking the correct eyewitnesses down with the mistaken ones. Nothing makes a trial judge more nervous than the vision of the rape or robbery victim being judged against a “Super-Reasonable-Doubt” standard because of the testimony of some professor of psychology. Even when expert psychological witnesses are�allowed to testify, it can’t be guaranteed that they will have the “silver bullet” impact lawyers seem to expect. Jurors can still feel that they are being subjected to a blizzard of data which tells them that eyewitnesses are sometimes unreliable, but leaves them with the questions “Unreliable compared to what?” and “Which ones?” or “When?” Fortunately, lurking in the ever-growing pile of data is a strategic approach to eyewitness cases. This approach can make it more difficult to exclude expert testimony and make expert testimony more effective when it is admitted. But most importantly, it can infiltrate some of the psychological principles into the legal decision-making process even when no expert testifies. Three stages, two stories Research indicates that an eyewitness identification has three stages: a witness must perceive the criminal, store the memory trace, and then retrieve it when making an identification. The research can point out relatively reliable conditions under which to perceive, store, and retrieve information. Research also reveals the opposites of those conditions, factual contexts which are likely to degrade perception, storage, and retrieval to dangerous degrees. The defense lawyer’s job is to identify the reliable conditions, contrast them with the conditions presented in the case on trial, and encourage the jurors to base a reasonable doubt on the differences. To get to “not guilty” you have to free the jurors from their fear of blaming the sincere and sympathetic eyewitness, and convince them to use their own critical judgment. A look at the recent research reveals two stories that will help trial lawyers organize their attacks on eyewitness reliability. They aren’t silver bullets, but then there are no silver bullets in eyewitness cases. These two stories can be particularly useful in shaping evidence for and making arguments to jurors—there’s another mountain of data which addresses the effectiveness of the story model3�of trials. But these stories can also be effective in designing coherent legal arguments that convince skeptical trial judges to utilize their discretion to permit the defense to integrate the findings of psychological science into the trial. How mistakes happen: ‘forced choice’ story Most jurors have had experiences which help them to see that an eyewitness might be unable to remember and recognize someone. That may not be enough to save an innocent defendant. The problem for the defense lawyer is that this eyewitness says that he does recognize someone, and not just any someone; the eyewitness says he recognizes your someone. Even though juror “commonsense” concedes that people can forget things, jurors generally will not know that witnesses can remember the wrong�thing. If you can’t explain that, you lose, and you wave goodbye to your client as the marshals’ bus pulls away from the curb. The psychological process of “relative judgment” — the “tendency to select the person who looks most like the offender relative to the other members of the lineup”4�— explains why an eyewitness might choose an innocent defendant. Relative judgments are not a problem when the real culprit is in the lineup; no one looks “more like” the culprit than the culprit does himself. But when the culprit is absent, and everyone in the lineup is innocent, eyewitnesses show a dangerous tendency to chose the “looks most like” innocent option. A well-known experiment illustrates this phenomenon. Iowa State Professor Gary Wells staged the same crime 200 times for 200 witnesses. The witnesses were then separated into two groups, asked to view a lineup, and carefully warned that the real culprit may or may not be present. Then, one half of the witnesses viewed a six-person lineup in which the actual perpetrator was present. Of these first 100 witnesses, 54 picked the right man; 21 made no choice; 13 picked the wrong man. What happened when the real culprit was removed? What should�have happened was that the 54 percent of the witnesses who chose the real culprit would simply migrate to the “no choice” column when the real culprit was removed, resulting in a “no choice” total of 75 percent. But that did not happen. Instead the “no choice” column, which should have grown by 54 (to 75 percent) rose only by 11, (to 32 percent.) Even with the real culprit absent 68 percent of the witnesses still chose someone, and 38 chose the same�someone: the jeopardy of the “looks most like” innocent person was almost three times higher when the actual culprit was removed than it had been when the real culprit was present. The significance of this—and many similar — findings for criminal practice cannot be overstated. It explains not only a mechanism for eyewitness mistake, but also why several eyewitnesses would make the same�mistake — why several other witnesses would “corroborate” a mistaken identification. This “Forced Choice” story provides a new theory of the case: now the defense can argue not “the eyewitness has made a freakish mistake,” or “eyewitnesses can’t be trusted,” but rather, “this witness naturally chose the man who looked most like the robber.” Can this story be told to judges and jurors without an expert witness? Telling the ‘forced choice’ story Some theories of the case fit in a category that jurors will accept immediately, others in a category that they will reject just as quickly. The “forced choice” story falls in a middle ground. It is not the first thing that the jurors would think of on their own, but it is not so counter-intuitive that the jurors will necessarily reject it out of hand. It is a recognizable story about a multiple choice test in which “none of the above” is not offered as an option. Still, the forced choice story has to be presented carefully: it is a mistake to try to jam it down the jury’s throat. Many lawyers would try just that: Q. You knew they called you down there to pick their suspect, didn’t you? A. Yes, No, Maybe. If the witness disagrees, the lawyer then tries to argue his or her way back to even. That seldom works, and even if the witness eventually agrees, the impact is muted. It is probably better to take some time to establish in front of and with the jury that as a matter of common sense the witness must have known that the police had their suspect in the photo array: Q. This call came from Det. Carroll? A. Yes. Q. You met him on the night of the shooting? A. Yes. Q. And talked to him? A. Yes. Q. At your home? A. Yes Q. He met your kids? A. Yes. Q. He knew you worked late? A. Yes. Q. He listened carefully to what you had to say? A. He seemed to. Q. And you told him how upset you were by what you’d seen? A. Yes. Q. And how upset you were every time you talked about it? A. Yes. Q. But you’d spoken to him several times before he called you in to look at these six pictures? A. Yes Q. And you knew he needed your help? A. Yes. Q. And you knew he’d been gathering information? A. Yes Q. But before that Monday night he never asked you to come to the station before? A. No. Q. So you knew something had changed? A. I guess. Q. Or he wouldn’t have called you in? A. Yes. Q. And as soon as you got there he showed you some pictures? A. Yes. Q. Not a whole book of pictures, just a group of six? A. Right. An advantage of this approach is that by now it may no longer be necessary to risk asking, “So you believed he had a suspect in his group of six?”; the argument is there to be made as a matter of inference already. Now, if it is necessary to ask the crucial question, asking it will be safer since the jury is likely to discount a “No” answer because it no longer seems to make sense. Not only that, getting a “No” answer has become less likely because while the common sense has been building for the jury, it has been building for the witness too. A cop or other professional witness who sees where you are going may try to head you off. A lay eyewitness who sees where you are going will be concerned mostly with not looking like a fool. Still, sometimes you have to drive the point home: Q. So you thought Det. Carroll didn’t call you down to waste your time? A. No. Q. And you didn’t think he showed you this array as a trick question? A. No. Q. You thought he’d developed a suspect? A. Yes. Q. And you thought his suspect was in this group of six? A. Yes. Q. No one told you otherwise? A. No. A similar, step-by-step approach might help to make the psychologists’ “forced choice” findings seem like common sense to the jurors even without a psychologist’s testimony. Q. You’re a real estate broker? A. Yes. Q. Went to high school? A. Yes. Q. Went to college? A. Yes. Q. Took a lot of tests? A. Yes. Q. Sometimes they had true/false questions? A. Yes. Q. One statement, and you said, true or false? A. Yes. Q. And some times they were multiple choice questions? A. Yes. Q. Now, you never saw Photo #3 by itself? A. No. Q. Or saw this Photo #6 of my client by itself? A. No. Q. Never said “true/false” to any single picture by itself? A. No. Q. Sometimes when you took multiple choice tests they gave you four choices? A. Yes Q. Sometimes five? A. Yes. Q. And this group of photos looked more like multiple choice? A. Yes. Q. And sometimes when you took multiple choice tests they told you that one of the choices was “None of the above?” A. Yes. Q. But you didn’t think Det. Carroll dragged you in to waste your time? A. No. Q. And he didn’t tell you that “none of the above was one of the choices?” A. No. Again, it may not be necessary to follow this line to its end in order to have an argument to make. By going one step at a time you have made your common sense the jurors’ common sense. But whether or not it seems important to ask a clinching “So you thought that the cops had a suspect in this array?” And by “looping” the term “compare” through a series of questions to the eyewitness, you may have another vehicle for driving the “relative judgment” point home: Q. He gave you a group of six pictures to compare? Q. When you compared�Number Two to Number Four, he seemed older? The key idea is that the photographs are being compared to each other, not to the eyewitness’s memory of the robber. The “forced choice” story can also provide a context which turns “prosecution facts” into defense facts. For example, it is generally easy to get a witness or a police officer to agree that the eyewitness, when presented with the photo array was scrupulously careful and took his or her time. But the longer the eyewitness takes, the better the argument that what we confront here is a comparison�— a multiple choice test, a process of elimination, not a recognition. The fact is, the research indicates that the longer the time-to-decision, the less likely that the identification will be accurate. Having the “forced choice” story in mind during investigation and pretrial discovery helps to fuel the cross-examination of the civilian eyewitness. It can also help in preparing to cross-examine the police. A few questions during investigative interviews or at a pretrial motion to suppress hearing, for example, might open up a safe line of trial cross-examination of the cop: Q. Detective, the witness seemed to take this seriously? A. Very. Q. No joking, no clowning? A. No. Q. You showed him the whole array? A. Yes. Q. And he seemed to study the whole array? A. Yes. Q. He was allowed to take his time? A. Yes, counsel. Q. You didn’t rush him? A. No. Q. And he didn’t rush? A. No. Q. He seemed to compare all the pictures? A. Yes. The “relative judgment” research has also exposed reliable identification practices which could have been followed in your case, but probably were not. At least since the Department of Justice published its Eyewitness Evidence: A Guide for Law Enforcement5 in 1999, it has been widely known that the police could mitigate the “forced choice” problem by explicitly instructing witnesses, as the Guide suggests, that the real perpetrator might or might not be in the array. The Guide also explains the “sequential” (one photograph or person at a time) method of showing an array or lineup, a method which researchers believe helps to eliminate the “relative judgment” tendency which is implicit in the traditional “simultaneous” (all suspects and fillers at once) method. Some judges might even allow you to question the police directly6 about their failure to follow these steps, using the Guide on analogy to a learned treatise. But whether the Guide itself is ever brandished or even mentioned, defense lawyers can use its ideas to show that there were a number of things that the police could easily have done, but did not do to avoid a “forced choice” situation. The “forced choice” story can also help to organize the legal arguments for the judge too. “Forced choice” can explain why a second eyewitness identification doesn’t necessarily corroborate the first in any meaningful way. That in turn can show why excluding an expert psychological witness will not necessarily be — or be seen by a reviewing court to be — harmless error. Where the trial judge has excluded expert psychological testimony the prejudice caused by that decision can be illuminated by requesting cautionary instructions on the evaluation of eyewitness testimony which emphasize the persistence of the “forced choice” potential. Courts which have been willing to instruct jurors, for example, that a lineup is more reliable than a one-on-one showup should be willing to instruct them that a “sequential” photo array is more reliable than a “simultaneous” one, especially when the defense has offered to prove just that through expert testimony, and the judge has excluded the evidence. An advantage of the “forced choice” story is that it reinforces a “no villains” defense to the eyewitness case. The witness did what any conscientious person would have done if the real robber wasn’t immediately recognizable: sorted through the “looks most like” process. The police didn’t try to frame or finger the defendant; they went by the book. It isn’t their fault if — as you and the jury have discovered together — their book is defective, or at least out of date. Contaminated trace evidence The “forced choice” story can explain why an honest and unbiased eyewitness might choose someone whom he or she had never seen before; the “trace evidence” story shows why that witness would persist in that choice and seem very confident in doing so. The two most problematic aspects of jury “common sense” about eyewitnesses are the jurors’ belief that eyewitnesses might forget but could never invent, and that a witness’ confidence is a strong indicator of a witness’s accuracy. In fact, the sensory traces left in memory are, just like blood, drug, or semen evidence, subject to contamination. The crucial differences in the memory trace evidence are that if the memory trace is contaminated, there is no uncontaminated sample of the trace left to test, and the contamination can never be proved or disproved by further testing. Contaminates in adulterated blood can be revealed by later lab work; no equivalent process is available when memory has been contaminated. Some of the most significant findings in recent psychological research can be found in a series of studies done by Gary Wells and Amy Bradfield in which simple feedback after an identification (“Good, you identified the suspect”) was enough to dramatically boost the confidence of eyewitnesses who made the wrong choice in a lineup.7 An insidious effect of this confidence-boosting feedback was that as confidence increased the substance of the witness’ account also improved, almost as if to explain the increased levels of confidence. After feedback, more witnesses remember a good view, remember paying more attention to the culprit while witnessing the event, remembering watching the culprit for a longer period of time, remember better perception of facial details, and remember their decision to chose the defendant from a lineup as an easier one. The memory traces had been radically contaminated by the feedback. Virtually all of the factors which the courts rely on to demonstrated eyewitness reliability,8�in other words, can be distorted by brief feedback to witnesses, even when the feedback comes days after the identification. Now and then you might run into a case where the police greet the identification of their suspect at a photo array with applause, or admit that they said “good, you got the suspect,” and that will make the trace evidence story easy to tell. You can point to the early general description or “looks like” identification from the night of the crime, point to the “never forget that face” trial testimony, and assign the cause of the change to the contaminating applause. But even when the feedback is not so explicit — in fact even when the feedback is nothing more than good manners — you can begin to show the jury why the witness’s confidence is a product of the contamination, not of the original memory. Q. You chose Number 7 as the robber? A. Yes. Q. And you said that Number 7 “looked like” the robber? A. Yes. Q. And here in court this morning, you chose my client, Mr. Ritchie as Number 7? A. That’s him, I’m certain. Q. When you took all those multiple choice tests in school did you do well? A. Usually. Q. You cared about your grades? A. Yes. Q. And your parents cared about your grades? A. Yes. Q. And you looked forward to getting your papers back? A. Yes. Q. And after you made this pick, you were curious about how you had done? A. Yes. Q. Did anyone ask you to go back and try again? A. No. Q. Did anyone tell you the process was over? A. No. Q. In fact, Det. Carroll said you’d be hearing from him? A. Yes. Q. And you did? A. Yes. Q. He took you to see the Assistant D.A.? A. Yes. Q. And they showed you the same picture, “Number 7,” in that meeting? A. Yes. Q. And you learned that you were going forward on that basis? A. Yes. Q. And you prepared to come to court and talk about that? A. Yes. Q. The police and prosecutors treated you politely? A. Yes. Q. They never challenged your choice of Number 7? A. No. Q. And you didn’t believe that Det. Carroll would put you up on the witness stand to embarrass yourself? A. No. Q. And by this morning you knew that he agreed with your choice? A. Yes. Q. Or you wouldn’t be here? A. Right. Q. But when you saw this picture for the first time, the best you could say was that it “looked like” the robber? A. That’s what I said. The “trace evidence” story should allow the defense to focus the jury’s scrutiny on that first “looks like” identification, because nothing happened between the “looks like” choice and the trial which could have improved the witness’s memory except contamination. There is no basis for anyone — even a juror relying on “common sense” without scientific input —- to assume that memory independently gets better�as time passes. The “trace evidence” story, even more than the “forced choice” story, can be helpful in explaining the dangers of eyewitness error to judges as well as to jurors. The psychological findings supporting the trace evidence story radically undermine the whole structure of “reliability” centered criteria which Manson v. Braithwaite and its progeny have used to rescue suggestive identifications.9 When the procedures are suggestive, and there may have been post-identification feedback, the psychological evidence plainly shows that it is a mistake to rely on, to take one Manson�criterion for example, the witness’s subjective certainty, for reassurance that the identification is reliable. The fact is, this line of argument may well enhance the opportunity for the admission of expert testimony in pretrial suppression litigation, where issues of juror competence and the “aura of expertise” surrounding psychological testimony can provide no excuse for excluding expert psychologists. Beyond that, the “trace evidence” metaphor can help to show judges why they should consider mobilizing any number of traditional judicial weapons against contaminated evidence. After all, it is always the burden of the proponent of evidence which is vulnerable to contamination to prove that the evidence has not been contaminated. The “trace evidence” story can help courts who are accustomed to focusing entirely on what happens before an eyewitness chooses a suspect to recognize that what happens after the choice can have a catastrophic impact on the integrity of the evidence. If the first description on the night of the crimes was limited to height and weight, should the elaborate trial description be excluded as contaminated? When feedback has boosted confidence, should a motion in limine excluding reference10�to confidence in questions or argument be considered? Should a cautionary instruction on confidence be given? Should the standard instruction’s references to confidence or certainty be redacted? Tools The “forced choice” and “trace evidence” stories are only two of many which emerge from the research on eyewitness performance. Telling those stories — even without an expert — is the defense lawyer’s role. The researchers have given us some tools, but we have to pick them up and use them. Notes 1. Jim Dwyer, Barry Scheck & Peter Neufeld, Actual Innocence�(2000). 2. David Feige describes a successful campaign to institute those error-prevention procedures through litigation. David Feige, ‘I’ll Never Forget That Face’: The Science And Law of the Double-Blind, Sequential Lineup, 26 Champion�28 (2002). 3. Pennington & Hastie, Cognitive Theory of Juror Decision Making: The Story Model, 13 Cardozo L. Rev. 519 (1991) 4. Wells, G.L., The Psychology Of LineupIdentifications, 14 J.Applied Social Psychol. 89 (1984). A basic research foundation for all of the arguments set out in this article can be supplied by pursuing the links posted on Professor Well’s Web site: http: //www.psychology.iastate.edu/faculty/gwells/homepage. 5. United States Department of Justice, National Institute of Justice, Eyewitness Evidence: A Guide for Law Enforcement (1999), http://www.ncjrs.org/pdfiles1/nij /178240.pdf. 6.Doyle, Larson & DiTraglia, The Eyes Have It—Or Do They? 16 Crim.Just.�12 (2001). 7. Gary Wells & Amy Bradfield, ‘Good, you identified the suspect’: Feedback to eyewitnesses distorts their reports of the witnessing experience, 83 J.Applied Psychol. 138 (1998); Gary Wells & Amy Bradfield, Distortions in eyewitness recollections: Can the postidentification-feedback effect be moderated? 10 Psychol.Sci. 138 (1999); Gary Wells, Elizabeth Olson & Steve Charman, Distorted Retrospective Eyewitness Reports as Functions of Feedback and Delay, 9 J. Experimental Psychol., Applied 42 (2003. 8. Manson v. Braithwaite, 432 U.S. 98 (1977). Michael Headley, Note: Long on Sustance: Short on Process: An Appeal For Process Long Overdue in Eyewitness Lineup Procedures, 53 Hastings L.J. 681 (2002). 9. Id. 10. Elizabeth Loftus & James Doyle, Eyewitness Testimony: Civil and Criminal, (LEXIS Lawbooks, 3d ed. 1997) 320-326, presents a fuller discussion of these suggestions.
	National Association of Criminal Defense Lawyers (NACDL) 1660 L St., NW, 12th Floor, Washington, DC 20036 (202) 872-8600 • Fax (202) 872-8690 • assist@nacdl.org