| , , ,

Judging Eyewitness Evidence


Vol. 104 No. 1 (2020) | A Clearer View | Download PDF Version of Article

Eyewitness evidence, in which a witness visually identifies the culprit, is a staple of criminal investigations. But its fallibility is notorious. As the National Academy of Sciences explained in an important 2014 report summarizing the scientific research on human visual memory, “it is well known that eyewitnesses make mistakes and that their memories can be affected by various factors including the very law enforcement procedures designed to test their memories.”1 Researchers have shown how eyewitness misidentification results in conviction of the innocent — and revealed the role that poorly designed and suggestive police procedures can play.2 Here, I examine the role that poorly designed court procedures can play.

Judges face the challenge of ensuring that lay jurors are not misled by eyewitness testimony at trial. A confident eyewitness is extremely powerful on the witness stand, and, in general, people place more weight on confident witnesses. A courtroom identification, of course, is not actually a reliable test of the witness’s memory of the crime itself: It is not difficult to see where the defendant had been sitting during the entire trial at counsel’s table, and the context predetermines that the witness will identify the defendant. And yet, despite this rather obvious set up, it is a powerful moment when a witness confronts the accused and states, with confidence, that the defendant is the one who did it. In contrast, the confidence of a witness at the time of an initial police lineup can provide valuable information about reliability. It is at that time that the witness actually does the cognitive work of making the identification. The witness’s memory may be relatively recent. And if police have not used improper procedures, the witness may be relatively free from suggestion. Yet, jurors may not focus on that lineup information, since the courtroom identification is more salient and dramatic.3 Overly relying on courtroom confidence as a proxy for a person’s accuracy can have serious consequences in the criminal justice system. One study examined defendants who spent an average of 15 years in prison only to have their convictions later overturned by DNA evidence. The eyewitnesses in those wrongful conviction cases, almost without exception, testified with complete certainty at the trials at which innocent people were convicted. However, at the time of initial police lineups, most displayed a lack of confidence in their identifications.4 In response, some courts have limited the use of courtroom identifications, at least in some circumstances, while others have attempted to use jury instructions or expert testimony to better explain the strengths and weaknesses of eyewitness evidence. Although several courts have drafted new and research-informed jury instructions, their efforts have not been shown to be effective. New research suggests more effective approaches for such instructions. But even with the best instructions, there remains great potential for confusion when explaining concepts relating to eyewitness memory, particularly when that explanation also requires addressing the issue of confidence. Ultimately, it would be preferable to simply prohibit the use of courtroom identifications — and limit the witness to describing what she observed at the scene and at the lineup.

On Paper: U.S. Supreme Court Rulings

Related content:

Eyewitness errors are not a new problem. As the Supreme Court has put it: “The vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification.”5 The law, in turn, has begun to respond to that evidence. The Supreme Court’s due process test for evaluating eyewitness evidence was set out in 1977 in Manson v. Brathwaite. That test asks, first, whether police used suggestive identification procedures, and second, whether any such suggestiveness can be excused based on a set of “reliability” factors.6 The “reliability” factors ask that the judge examine: (1) the eyewitness’s opportunity to view the defendant at the time of the crime; (2) the eyewitness’s degree of attention; (3) the accuracy of the description that the eyewitness gave of the criminal; (4) the eyewitness’s level of certainty at the time of the identification procedure; and (5) the length of time that had elapsed between the crime and the identification procedure.7 The Court more recently has held that when unreliability in eyewitness identifications is not due to intentional police action, it is not regulated under the Due Process Clause at all.8

The Supreme Court has also extended a right to counsel for certain procedures and commented broadly on concerns with certain types of identification procedures. In the 1967 case of United States v. Wade, the Supreme Court held that, once indicted, a person has a right under the Sixth Amendment to have a lawyer present at a lineup.9 However, that right does not extend to photo array procedures, which are far more commonly used today than live or in-person lineups.10 Regarding show-up procedures — the inherently suggestive process of presenting a single suspect to a witness — the U.S. Supreme Court noted in Stovall v. Denno that “[t] he practice of showing suspects singly to persons for the purpose of identification, and not as part of a lineup, has been widely condemned.”11 In that same case, the Supreme Court rejected any per se rule against the use of show-ups.12 The U.S. Supreme Court in Perry v. New Hampshire also held that viewings of a single suspect are not regulated by the Due Process Clause when officers do not intend to conduct an eyewitness identification procedure.13

In the Lab: Scientific Research

Supreme Court rulings carry great import, of course. But as the National Academy of Sciences Report has put it, “the best guidance for legal regulation of eyewitness identification evidence comes not from constitutional rulings, but from the careful use and understanding of scientific evidence to guide fact-finders and decision-makers.”14 At the time Manson v. Brathwaite was decided in 1977, little was known about what precisely had an impact upon the reliability of eyewitness identifications. Since then, a large body of research has called into question the validity of many of that case’s so-called “reliability” factors.15 Scientific evidence concerning human perception, vision, and memory provides a framework that can and should inform the collection and the use — both pretrial and at trial — of eyewitness evidence. In scientific terms, the law should take account of both estimator variables and system variables.16 Both types of variables can affect the memory of an eyewitness. Estimator variables are factors relating to the conditions of the crime-scene viewing, such as the lighting, the eyewitness’s eyesight and familiarity with the perpetrator, or race of the eyewitness and culprit. Estimator variables cannot be controlled by law enforcement. In contrast, system variables are factors associated with the procedures officers use to obtain identifications by an eyewitness. Studies have shown that certain such procedures can affect or even alter the memory of the eyewitness. Officers can affect this unintentionally. For example, just by asking an eyewitness to participate in an identification procedure, officers create an expectation that a suspect will be present (making it very important for officers to follow the standard practice of cautioning witnesses that a suspect may or may not be present). Moreover, an eyewitness naturally will be looking to the officer for guidance, reinforcement, and feedback. That is why the use of blinding, or ensuring that the officer administering the lineup does not know which is the suspect, is “central to the scientific method,” as the National Academy of Sciences Report emphasizes, because “it minimizes the risk that experimenters might inadvertently bias the outcome of their research, finding only what they expected to find.”17 Here I address the issue of the confidence of the identifying witness, which can be affected by both estimator and system variables. Eyewitness confidence becomes relevant both at the moment of identification and again at trial, when the witness is asked to make an in-court identification.

At the Lineup: Initial Confidence

It is crucial to document the confidence of an eyewitness at the time of an initial lineup procedure: It gives a clearer view of accuracy, and it provides a key data point in the case that such confidence becomes inflated over time. At trial, a confident eyewitness can be extremely powerful to jurors. However, that confidence may not correspond to reliability; the eyewitness may not in fact have been (as) sure when he identified the suspect at the time of the earlier eyewitness identification procedure. “At trial, an eyewitness’ artificially inflated confidence in an identification’s accuracy complicates the jury’s task of assessing witness credibility and reliability.”18 Although eyewitness memory and confidence are both malleable, they do not improve over time. Absent documentation of the confidence of an eyewitness, there may be no record that the confidence of an eyewitness has been enhanced over time. For that reason, leading scientific groups, including the National Academy of Sciences Committee, strongly recommend that the confidence of an eyewitness be carefully documented, in a manner that precludes influence by the officer conducting the procedure. Although scientists might prefer that confidence be recorded using a numerical scale, few agencies have followed such an approach, due to a concern that quantitative scores might be misunderstood in the courtroom. Instead, the approach has been to record confidence by asking an eyewitness to express it in his or her own words. The National Academy of Sciences recommends video-recording eyewitness identification procedures.19

On the Stand: In-Court Confidence

The Massachusetts Supreme Judicial Council and the Connecticut Supreme Court have ruled that no in-court identification is permitted if the out-of-court identification was suppressed as unduly suggestive.20 Massachusetts has also determined that in-court identifications can only be used when an identification is made for the very first time in court.21 That court explained, “[w]here, as here, a prosecutor asks a witness at trial whether he or she can identify the perpetrator of the crime in the courtroom, and the defendant is sitting at counsel’s table, the in-court identification is comparable in its suggestiveness to a showup identification.”22 Other courts have adopted a burden-shifting approach towards in-court identifications.23 These findings all stem from the acknowledgment that in-court identifications hold little probative value, since they ask the witness simply to point out someone already in the room and leave the witness no real alternatives to naming the defendant. Not only is the in-court identification not a test of the eyewitness’s memory, but in court, the eyewitness will predictably have inflated confidence in the identification. The entire exercise, however theatrical, is highly prejudicial and rather meaningless. In contrast, any prior identification at a properly con-ducted police lineup may provide probative information to the jury. I have argued that, more broadly, courts should not permit in-court identifications, as they are not actually a test of the eyewitness’s memory of the crime itself, and the eyewitness’s confidence will predictably be greater by the time of the in-court procedure.24

In the Jury Room: The Effect of Jury Instructions

If courts continue to permit in-court identifications, judges might try to blunt their potentially deleterious effect with cautionary jury instructions. But, as many now know, jurors face great difficulty following instructions delivered at the end of a trial regarding evidence presented earlier.25 Studies also find low levels of comprehension of jury instructions that (quite commonly) use technical or legalistic language.26 Many state and federal courts use standard instructions on eyewitness evidence pursuant to the D.C. Circuit’s ruling in United States v. Telfaire.27 But studies have found that these Telfaire instructions do not greatly affect how laypeople ultimately weigh eyewitness evidence.28

In recent years, several states have endorsed more detailed jury instructions on eyewitness evidence. In State v. Henderson, the New Jersey Supreme Court endorsed detailed jury instructions on the issue.29 The Massachusetts Supreme Judicial Council has recommended more concise jury instructions on eyewitness identification evidence, and other states have made more specific changes to their rules.30 Following the New Jersey Supreme Court’s ruling in Henderson, five studies responded by examining the effect of these far more lengthy and detailed instructions. In general, these studies show that the New Jersey instructions induced a generalized skepticism of both reliable and less reliable eyewitness evidence, but the instructions did not improve the accuracy of jurors’ decisions.31

In a study I worked on with colleagues, we similarly found that hearing the Henderson instructions after viewing testimony regarding eyewitness evidence had no measurable effect on verdict decisions. Nor did expert testimony. Instead, the courtroom confidence of an eyewitness had the greatest impact on laypeople. The mock juror’s perception of the eyewitness’s courtroom confidence explained the vast majority of the differences between jurors in their decisions whether to vote guilty or not-guilty in our mock trial scenario. Other important factors regarding the eyewitness testimony, such as the type of crime, the race of the witness, and the race of the defendant, in contrast, had no measurable impact on jurors. What the mock jurors reacted to most was the in-court confidence of the eyewitness.32

A Closer Look: Reason-Based Jury Instructions Affect Decision-Making

The most direct solution to this confidence problem is to bar courtroom identification altogether. Eyewitnesses can testify regarding what they observed and how they identified a person at a police lineup procedure, without making a separate courtroom identification. We tested that scenario on mock jurors and found that it reduced guilty-vote rates precisely where it should — where a witness was less confident at the time of the initial identification than she would have been at trial. However, for those courts determined to keep in-court eyewitness identifications, it is worth working to find the jury instructions that most effectively address the misleading effects of in-court expressions of confidence. In a new set of studies, new instructions were drafted based on a paradigm designed to provide research-based reasons why jurors should approach evidentiary questions differently. For both studies described below, we used a survey company to recruit 1,614 jury-eligible adults, census-representative by gender, race, age, and geographic region, who were paid to participate.33 What we found supports the view that telling jurors why is a useful tool.

Study One: Comparative Skepticism of Jurors

In the first study, all participants were shown a mock trial video of an eyewitness who describes complete confidence in the courtroom that the defendant was the assailant (she says: “I am one hundred percent sure”), but recalls her earlier uncertainty at the time of the police lineup (she says: “I was not sure at that time.”). Following that video, participants in the control condition received no judicial instructions. In the six study conditions, participants were shown a video of a judge providing one of six versions of instructions. One set of participants received an excerpt from the instructions in use in Massachusetts. They state that an eyewitness’s confidence “standing alone” is something that “may not be” reliable. The instructions do not provide reasons why that might be the case. A second set of participants viewed a video of a judge who provided a very brief didactic instruction asking participants to discount in-court confidence. The remaining four jury instructions were designed using a paradigm different from existing models. These instructions were designed to briefly explain why less weight should be placed on the testimony of an eyewitness who is not confident at the time of the initial lineup identification. Before the study participants viewed the video of the eyewitness, they heard judicial instructions that included explanations for why they should focus on the earlier confidence of an eyewitness: “Research shows that the memory of an eyewitness does not improve over time. The most reliable information about a witness’ level of confidence comes from the police lineup conducted before trial.” Only after stating these reasons did the judge provide a directive that they “should” discount the level of confidence that a witness expresses in the courtroom. Conviction rates decreased most, relative to the control condition, when participants were provided with some form of a reason to ignore the eyewitness’s in-court testimony. In other words, it appears that a strong way to ensure that laypeople discount evidence is to provide, via jury instructions, a reason for them to do so. Our participants were half as likely to convict the defendant when they received judicial instructions with a reason, as compared to the control condition. By contrast, we observed no meaningful change in conviction rates in the two scenarios where participants were not provided a reason: the Massachusetts instructions, and the didactic instructions telling jurors simply to disregard the eyewitness’s confidence.

Study Two: Comparative Sensitivity of Jurors

In a second study, we examined comparative sensitivity of jurors: whether different groups of mock jurors could give more weight to a more reliable eyewitness who was confident at both the lineup and a trial versus giving less weight to an eyewitness who was confident only at trial. We drafted somewhat revised reason-based jury instructions, with two key changes from Study One. First, all instructions, at the advice of the lawyers and judges we spoke to while drafting them, made clear throughout that jurors retain discretion to weigh evidence as they see fit. The instructions begin by saying: “It is up to each of you to decide how much weight you give to any witness’s testimony, as with any evidence admitted during this trial.” Second, because we had observed in pilots that jurors seemed confused by the fact that both the judge and the eyewitness were telling them about two identification procedures, the instructions laid out the context in a clear, step by step, fashion. As the judge explained:

I note that the eyewitness described two identifications. First, the eyewitness made an identification at the police station in a lineup. Second, the eyewitness made an identification here in court.

Importantly, we observed sensitivity effects.34 That is, these reason-based instructions helped jurors to distinguish between the more and less reliable eyewitnesses. However, only the more forceful instructions with reasons — telling jurors that they “should” as opposed to “may” discount courtroom confidence — had a significant effect on jurors’ sensitivity as to verdicts. These findings are consistent with the literature on judicial instructions, which indicate that technical or legal explanations provided without context may be ineffective or even counterproductive. The studies indicate the importance of telling jurors why they should discount information. Perhaps experts or lawyers, in addition to jury instructions, can similarly convey that information, and perhaps this reason-based approach can be applied more broadly. Future research will tell.


When cases involving eyewitnesses go to a criminal trial, jurors place great weight on eyewitness confidence, including in-court confidence, which does not necessarily correspond to accuracy. Some courts have begun to address this problem directly, by ruling that, at least under some circumstances, an eyewitness cannot make an in-court identification whatsoever. However, if judges continue to permit in-court identifications, the existing jury instructions have not been effective at remedying the prejudice of such in-court identifications. Two new studies suggest a different paradigm — in which jurors are told explicitly why to discount in-court identifications — may be more effective. Such an approach might also be used to counter common misconceptions about other types of evidence presented to juries in criminal cases. Far better, however, would be to eliminate the source of prejudice by barring in-court identifications altogether.

3. John T. Wixted & Gary Wells, The Relationship Between Eyewitness Confidence and Identification Accuracy: A New Synthesis, 18 PSYCHOL. SCI. IN THE PUB. INT. 10–65 (2017).
4. Id.
5. United States v. Wade, 388 U.S. 218, 228 (1967).
6. Manson v. Brathwaite, 432 U.S. 98, 114 (1977).
7. Id.
8. Perry v. New Hampshire, 565 U.S. 228, 245 (2012).
9. Wade, 388 U.S. at 235–36.
10. United States v. Ash, 413 U.S. 300, 321 (1973); Gary L. Wells & Deah S. Quinlivan, Suggestive Eyewitness Identification Procedures and the Supreme Court’s Reliability Test in Light of Eyewitness Science: 30 Years Later, 33 LAW & HUMAN. BEHAV. 1, 16 (2009) (a “large percentage of jurisdictions in the U.S. use only photographs and never use live lineups”).
11. Stovall v. Denno, 388 U.S. 293, 302 (1967).
12. Id.
13. Perry, 565 U.S. at 248.
14. NATIONAL RESEARCH COUNCIL, IDENTIFYING THE CULPRIT, supra note 1, at 5; see also PRESIDENT’S TASK FORCE ON 21ST CENTURY POLICING, FINAL REPORT OF THE PRESIDENT’S TASK FORCE ON 21ST CENTURY POLICING 23 (2015) (2.4 Recommendation encouraging adoption of identification procedures “that implement scientifically supported practices that eliminate or minimize presenter bias or influence”).
15. See, e.g., Gary L. Wells & Deah S. Quinlivan, Suggestive Eyewitness Identification Procedures and the Supreme Court’s Reliability Test in Light of Eyewitness Science: 30 Years Later, 33 LAW & HUMAN. BEHAV. 1, 16 (2009); Suzannah B. Gam-bell, The Need to Revisit the Neil v. Biggers Factors: Suppressing Unreliable Eyewitness Identifications, 6 WYO. L. REV. 189, 192 (2006).
16 Gary L. Wells, Applied Eyewitness-Testimony Research: System Variables and Estimator Variables, 36 J. PERSONALITY & SOC. PSYCHOL. 1546–57 (1978) (coining terms estimator and system variables).
18 Perry, 565 U.S. at 252 (Sotomayor, J., dissenting).
19 See NATIONAL RESEARCH COUNCIL, IDENTIFYING THE CUL-PRIT, supra note 1, at 74. There are few practical obstacles to doing so in the case of photo array procedures. Field identifications, such as show-ups, also can be recorded; body-cameras can make field recordings feasible.
20 Commonwealth v. Johnson, 45 N.E.3d 83 (Mass.2016); State v. Dickson, 141 A.3d 810 (Conn. 2016). For the argument that courts should not use “in-dependent source rules” to permit an in-court identification following suggestive out of court identifications, see Brandon L. Garrett, Eyewitnesses and Exclusion, 65 VAND. L. REV. 451 (2012).
21 Commonwealth v. Crayton, 21 N.E.3d 157 (Mass. 2014).
22 Id. at 166; see also United States v. Archibald, 734 F.2d 938, 941 (2d Cir. 1984) (“Any witness, especially one who has watched trials on television, can determine which of the individuals in the courtroom is the defendant . . . .”).
23 See State v. Hickman, 330 P.3d 551, 568 (Or. 2015) (“Courts considering the admissibility of first-time in-court identifications generally have placed the burden of seeking a prophylactic remedy on the defendant” (citing United States v. Brown, 699 F.2d 585, 594 (2d Cir. 1983), and United States v. Domina, 784 F.2d 1361, 1369 (9th Cir. 1986).
24 See generally Garrett, supra note 20.
25 Nancy K. Steblay, Harmon M. Hosch, Scott E. Culhane & Adam McWethy, The Impact on Juror Verdicts of Judicial Instruction to Disregard Inadmissible Evidence: A Meta-Analysis, 30 LAW & HUM. BEHAV. 469, 469 (2006).
26 Laurence J. Severance & Elizabeth F. Loftus, Improving the Ability of Jurors to Comprehend and Apply Criminal Jury Instructions, 17 LAW & SOC’Y REV. 153, 153–97 (1982).
27 United States v. Telfaire, 469 F.2d 552 (D.C. Cir. 1972).
28 Brian L. Cutler, Steven Penrod, & Hedy R. Dexter, Juror Sensitivity to Eyewitness Identification Evidence, 14 LAW & HUM. BEHAV. 185, 185–91 (1990).
29 State v. Henderson, 27 A.3d 872, 925–26 (N.J. 2011).
30 Commonwealth v. Gomes, 22 N.E.3d 897 (Mass. 2015); see also Young v. State, 374 P.3d 395 (Alas-ka 2016).
31 See Marlee K. Dillon, Angela M. Jones, Amanda N. Bergold, Cora Y.T. Hui & Stephen Penrod, Henderson Instructions: Do They Enhance Evidence Evaluation?, 17 J. FORENSIC PSYCHOL. RES. & PRAC. 1 (2017); Angela M. Jones, Amanda N. Bergold, Marlee K. Dillon & Steven D. Penrod, Comparing the Effectiveness of Henderson Instructions and Expert Testimony: Which Safeguard Improves Jurors’ Evaluations of Eyewitness Evidence?, 13 J. EXPERIMENTAL CRIMINOLOGY 29 (2017); Brandon L. Garrett, Alice Liu, Karen Kafadar, Joanne Yaffe & Chad S. Dodson, Factoring the Role of Eyewitness Evidence in the Courtroom (under revision, 2019); C.E. Laub, C.D. Kimbrough & Brian H. Bornstein, Mock Juror Perceptions of Eyewitnesses versus Earwitnesses: Do Safeguards Help?, 34 AM. J. FORENSIC PSYCHOL. 33 (2016); A.P. Papailiou, David V. Yokum, & Chris T. Robertson, The Novel New Jersey Eyewitness Instruction Induces Skepticism but Not Sensitivity, 10 PLoS ONE (2015), available at http://dx.doi.org/10.1371/journal. pone.0142695.
32 Garrett, et al., supra note 31.
33 All study materials and videos, as well as underlying data, are available on an Open Science Foundation website, https://osf.io/gep9v/.
34 The complete study results are being prepared for publication. The results discussed here are a general summary.