Article Of Interest: Eyewitness Errors and Wrongful Convictions: Let’s Give Science a Chance, By David A. Sonenshein and Robin Nilson
Readers of this blog know of my interest in the inaccuracy of eyewitness identifications (see, e.g., here, here, here, here, and here) and efforts to increase their reliability (see, e.g. here). This interest came from the confluence of two events: First, I read about the results of The Illinois Pilot Program on Sequential Double-Blind Identification Procedures, which seemed to suggest that traditional lineup methods are uncannily accurate and that more modern safeguards actually lead to less reliable identifications. Thereafter, I saw a 60 Minutes piece about Jennifer Thompson’s (mis)identification of Ronald Cotton as her rapist and about the problems inherent in traditional lineup methods. This piece changed the way that I thought about eyewitness identifications, and I now play the piece in all of my Evidence and Criminal Procedure classes. The Cotton-Thompson case also provides the launching point for an excellent recent article about eyewitness (mis)identifications, Eyewitness Errors and Wrongful Convictions: Let’s Give Science a Chance, 89 Or. L. Rev. 263 (2010), by David A. Sonenshein and Robin Nilson, professors at the Temple University Beasley School of Law.
Supreme Court Case Law
After discussing the Cotton case, in Part I of the article, the authors lay out the Supreme Court’s case law on eyewitness identification. Specifically, they note that in its 1967 opinion in
Stovall v. Denno, the Court held that due process requires the suppression at trial of identifications that courts deem necessary but “unduly suggestive” under a “totality of the circumstances” test. An unnecessarily suggestive lineup or showup identification would be per se excluded.
But fast forward five years and substitute the Burger Court for the Warren Court, and you get Neil v. Biggers, in which
the Burger Court essentially overruled Stovall, holding that once a trial court found a lineup unnecessarily and impermissibly suggestive, due process required that a court apply a “totality of circumstances test” to demonstrate the reliability of the identification before admitting the identification at trial. The Court set out five factors that make up the totality of the circumstances: the witness’s opportunity to observe, the degree of attention paid by the witness, the accuracy of the witness’s initial description, the certainty of the witness’s lineup identification, and the length of time between the crime and the identification confrontation. The Court, however, deleted the term “unnecessarily suggestive” from the due process test, thereby placing its imprimatur on the admission of suggestive identifications even where the government chose, but was not forced by circumstance, to conduct a suggestive identification procedure.
Moreover, after another five years passed, the Court in Manson v. Brathwaite
reaffirmed that the Biggers five-prong test was required only when authorities had conducted an impermissibly suggestive lineup. The five prongs–henceforth known as the “Brathwaite factors”–are intended to ensure “reliability,” which, in the Brathwaite Court’s view, “is the linchpin in determining the admissibility of identification testimony.” Thus, after Biggers and Brathwaite, the Court would have the lower state and federal courts balance the degree of lineup suggestiveness against the five factors that the Court identified as those which could ensure reliability despite unnecessary and improper police suggestions.
The result: few identifications are deemed unnecessarily or impermissibly suggestive. But do the Brathwaite factors and traditional lineup methods still hold water on crash against the rocks of modern forensic science and psychology? The authors argue the latter.
Blinded Me With Science
In Part II, the authors begin by critiquing traditional lineup methods. These include:
(1) the fact that most jurisdictions use “simultaneous lineups” (where eyewitnesses view all suspects and fillers at once) rather than “sequential lineups” (where eyewitnesses view suspects and fillers seriatim, i.e., one at a time). The problem with the former technique is that it often leads to “relative judgment,” i.e., the eyewitness comparing everyone in the lineup and picking the person who most closely resembles the perpetrator, which leads to false positives;
(2) the fact that most jurisdictions use “non-double blind” lineups (where the administrator knows the suspect’s identity) rather than “double-blind” lineups (where the administrator does not know the suspect’s identity). The problem with the former is that implicit and/or explicit administrator bias can lead to false positive identifications; and
(3) the fact that most jurisdictions use non-blind administrators often leads to expressions of approval after “positive” identifications. The problem with this is that it leads to “confidence malleability,” wherein an eyewitness making an initially tentative identification becomes more confident in that identification after receiving (implicit and/or explicit) positive feedback from the administrator.
The authors then note that modern forensic science dispels four of the five Brathwaite factors:
(1) Opportunity to Observe: The Brathwaite analysis fails to account for (a) “Visual Hindsight Illusion,” pursuant to which eyewitnesses think that they recognize features of people that they think they know at distances (e.g., 150 feet) suggesting such recognition is impossible; and (b) the tendency of witnesses to overestimate how long they saw the perpetrator and underestimate the extent to which their view was obstructed;
(2) Attention: “Contrary to the simplistic Brathwaite analysis, modern forensic science has confirmed that the relationship between the degree of attention a witness pays and the reliability of his subsequent identification is subtle and often unpredictable. For instance, a witness who did not pay close attention to a perpetrator’s particular facial features, but instead simply looked at the perpetrator’s face, is more likely to make an accurate lineup identification. Although a witness who concentrated on the perpetrator’s facial features is more likely to provide police with an accurate description of the perpetrator, he is also more likely to make an inaccurate lineup identification;”
(3) Description Accuracy: “There is no significant correlation between the accuracy of a witness’s initial description of the perpetrator and the accuracy of the witness’s subsequent lineup identification. Scientists believe that the two types of memory involved–recognition and recall recognition–account for this lack of correlation. Recognition relates to the ability of a witness to describe someone whom he has seen before in a rapid and uninvolved process; recall recognition relates only to a witness’s ability to identify what he has seen after an intentional retrieval stage requiring some effort;”
(4) Certainty: “Historically, some experts believed that there is a significant (but not overwhelming) correlation between the certainty and accuracy of a witness’s identification. Unfortunately, as has been discussed, because the certainty of lineup identifications is so often the result of police encouragement, the validity of that correlation is questionable. In fact, the leading researchers have directly challenged the validity of eyewitness certainty as an indicator of accuracy.”
Thus, “[t]he ‘delay’ factor is the only Brathwaite factor that has earned the approval of the social science community as relevant to reliability.” to reliability.”
No Country for Old Laws
In part III of the article, the authors focus upon reform efforts at both the federal and state levels, including changes in New Jersey, Illinois, Wisconsin, Virginia, Maryland, and North Carolina, which “has adopted the most wide-ranging lineup reforms.” (which I detailed in this post). The authors then discuss Minnesota’s lineup reform pilot program and Illinois’ aforementioned counterpart, explaining why the latter was flawed (for more information, check out this link, which is included in the article).
In the Trenches
In part IV, the authors track the treatment of the the Brathwaite factors in state courts across the country. They initially note that “[i]n applying Brathwaite, the highest courts of some forty states have declined to broaden its protections, their authority to do so under their own interpretations of their own due process clauses notwithstanding.” On the other hand, some state supreme courts have found that the Brathwaite factors are inadequate. The Georgia and Connecticut Supreme Courts have decided to give strengthened cautionary instructions to jurors in connections with the introduction of eyewitness identification evidence, but the authors doubt their efficacy based upon a number of cognitive biases, such as reactance theory, belief perseverance, and ironic mental processes (e.g., if I tell you not to think of a white bear, what do you think about?). Moreover, some states have begun to allow defendants to present expert testimony on the inaccuracy of eyewitness identifications (with the Supreme Court of Utah recently reversing course and allowing such as testimony as evidence, as noted on this blog).
Conclusion
In the end, however, the authors recommend the more radical approach adopted by the Supreme Judicial Court of Massachusetts and the New York Court of Appeals. According to the authors,
We recommend that state courts follow the reasoning of the Supreme Judicial Court of Massachusetts and the New York Court of Appeals in returning to the Stovall test, which bars the admission of any identification that is unnecessarily suggestive (e.g., a showup in an emergency situation). In addition, this test allows the admission of other suggestive identifications only if the court is convinced by clear and convincing evidence that they and any resulting in-court identifications are reliable based on the scientific factors that provide real indicia of reliability. Rather than focus exclusively on the inadequate Brathwaite factors, courts should insist (except in extraordinary circumstances) on (1) double-blind lineup or photo array procedures, (2) the use of at least five fillers who resemble the suspect to a reasonable degree (e.g., height, weight, race or skin tone, and hair), (3) sequential lineups or photo arrays, (4) informing the witness that the suspect is not necessarily in any of the lineups, and (5) videotaping the lineup and the witness’s statements during the lineup procedure. In addition, state courts can and should require, under their state constitutions, the presence of defense counsel or some other person associated with the suspect at any lineup or photo array, irrespective of whether the identification procedure is conducted before or after the lodging of an indictment or other formal charge. Finally, in order to fully inform jurors of the counterintuitive information surrounding the identification process, and given its general acceptance in the field of psychology, all courts should admit properly qualified expert testimony on the manner in which the mind processes identification information.
From everything I have read about the inaccuracy of eyewitness identification and the high percentage of convictions reversed based upon mistaken identifications, I strongly agree with this approach and hope that more courts adopt this reasoning. Here’s hoping that those in power read this excellent article and begin enacting reforms to increase the validity of eyewitness identifications and decrease false positives. I asked the authors what led them to write the article, and Professor Sonenshein responded,
The genesis of the article is my following of the stories of innocent men convicted on wrongful identification, which along with erroneous confessions make up the bulk of wrongful convictions. As an Evidence teacher and scholar, I have been troubled by the resistance of many courts to admit expert testimony on the vagaries of eyewitness identification under either Rule 702 or Frye on the specious theory that jurors do not need assistance in assessing the probative value of eyewitness ids.
-CM