Become a Patron! 


Andrew E. Taslitz

Excerpted from: Andrew E. Taslitz, "Curing" Own Race Bias: What Cognitive Science and the Henderson Case Teach about Improving Jurors' Ability to Identify Race-tainted Eyewitness Error, 16 New York University Journal of Legislation and Public Policy 1049 - 1100 (2013) (319 Footnotes) (Full Article)


Eyewitness misidentifications are the most common contributor to wrongful convictions. The major types of identification procedures police use are live lineups (putting several persons in a row), photographic lineups (using a row of photographs), and live showups (asking a witness to identify a single person). Studies of Andrew E Taslitzactual criminal cases reveal that witnesses pick the suspect only forty-five percent of the time. Witnesses pick no one thirty-five percent of the time, and an innocent filler the remaining twenty percent. Most of this data is from the United Kingdom, which uses superior identification procedures to those in the United States. But archival studies in the United States reveal similar real-world numbers. If these numbers are correct, eyewitnesses who identify someone are choosing innocent persons as the alleged criminal culprit almost one-third of the time (20% out of 65% of identifications made).

Yet in real-world cases, we cannot know that the "suspect" is in fact guilty. Experimental studies reveal similarly troubling numbers *1052 albeit even when the actual perpetrator is present in the lineup. When the perpetrator is not present in the experimental lineup--akin to the real-world situation of the arrested suspect in fact being innocent-- eyewitnesses choose the incorrect person fifty percent of the time. Indeed, in the 250 cases of exoneration by DNA evidence in the United States, seventy-six percent involved eyewitness error. Although eyewitness testimony is not the only means by which to identify a criminal, eyewitnesses identify 77,000 criminal suspects in the United States every year. Of course, every time an innocent person is wrongly identified, the true perpetrator escapes justice.

The causes of eyewitness error are many and have been extensively studied. One important contributing factor to error is the "own race bias" (ORB) or "other race effect." This effect is straightforward: eyewitnesses of one race are more likely to misidentify innocent persons when those persons (and, of course, the perpetrator) are of another race. The ORB's existence has been repeatedly shown over decades via numerous study methodologies. The bias occurs across *1053 all races studied, though the bias might be somewhat worse when whites are identifying persons of other races. The effect does not mean that a cross-racial identification is necessarily wrong. But the risk of error is increased. If there are still other risks of error present, such as poor directions given to the witness, non-blind lineup methods (in which the witness is in the investigating officer's presence and can gauge his approval or disapproval of her tentative choices), officer feedback confirming a witness's choice, poor conditions for observing the perpetrator at the time of the crime, or a host of other factors, the ORB magnifies the existing chances of a mistake even further.

Police cannot control the respective races of the victim and suspect. There is thus, as of yet, no way to create an identification procedure that eliminates, reduces or adequately corrects for the ORB. Protecting against error stemming from that bias therefore requires improving the performance of the judges or jurors in determining whether the ORB has rendered the risk of eyewitness error such that the jurors should have a reasonable doubt about the suspect's guilt. In more common sense terms, can science devise procedures that make it more likely that jurors will acquit the factually innocent when the ORB is present?

Two major solutions have been suggested to improve jury performance: instructions cautioning jurors about the risks of mistake stemming from the ORB and expert testimony informing jurors about those risks in more vivid detail. The choice between these options, or the third one of using both remedies, was recently considered by the *1054 Supreme Court of New Jersey in State v. Henderson. The Henderson case quickly became a landmark in the development of the law on eyewitness identification because the Court, albeit by relying on the New Jersey Constitution, replaced the old Manson v. Braithwaite test articulated by the United States Supreme Court for suppressing unnecessarily suggestive eyewitness identifications as violative of due process. Manson held that such identifications were suppressible only when they created a very substantial likelihood of misidentification. The case identified several factors for courts to consider in determining the risk of unreliability of an identification.

But Manson has been widely criticized as incomplete and outdated in light of the scientific developments in the field of eyewitness research in the decades since the court decided the case. Henderson corrected many, indeed most, of Manson's errors; it articulated a wider array of scientifically-informed factors for courts to consider in determining the unreliability of an identification, which the case outlined with great specificity, partly in the apparent hope of encouraging police to update their procedures to avoid suppression. Moreover, the court in Henderson did so only after having appointed a special master who held hearings to collect the most thorough and complete judicial record yet on the factors affecting and correcting for eyewitness error.

The massive record in Henderson and the thoroughness of the court's analysis created hope that the United States Supreme Court would follow suit. The Supreme Court did not, however, and instead *1055 indicated that it may be reconsidering the wisdom of any due process restrictions on identification procedures. That high Court's reticence makes Henderson all that much more important as a potential model for reforming eyewitness procedures in the states.

Yet the Henderson Court's approach to the ORB ultimately proved enormously conservative in its results. The Court had recognized the ORB as a problem several years earlier and mandated giving a brief cautionary jury instruction on the point, but only in the subset of cross-racial identification cases where the identification was critical to the case and uncorroborated by other evidence. The Henderson Court concluded these "Cromedy instructions," named after the case in which they were mandated, or an improved variant of such, should be given in every cross-racial identification case. However, the *1056 Court instructed its Criminal Practice Committee and Committee on Model Criminal Charges to revise the general jury instructions on eyewitness identifications, and specifically to consider whether the Cromedy instructions should be revised in light of modern science.

Those committees ultimately concluded that no significant revision to the Cromedy instructions was necessary. A number of states have crafted similar cautionary instructions but the Cromedy instructions, although brief, are most likely the best, and certainly most well-known, variant.

Nevertheless, these instructions leave much to be desired. Expert testimony will likely rarely be used under the Henderson Court's approach to ameliorating the Cromedy instructions' weaknesses because the court declared, "with enhanced jury instructions, there will be less need for expert testimony." Moreover, the court praised instructions as the preferable remedy because they are authoritative (coming from the judge), spare jurors the task of choosing between battling experts, are cost-free, and "eliminate the risk of an expert invading the jury's role or opining on any eyewitness's credibility."

This brief article takes issue with both parts of the Henderson court's holding concerning the ORB. There is little, if any, research specifically on the value of the Cromedy instruction in helping the jury better assess eyewitness accuracy in light of the ORB. But there is analogous research on eyewitness jury instructions in other areas, *1057 and there is ample research concerning instructions to disregard evidence that suggests that the more-consistent use of the cautionary Cromedy instruction mandated by Henderson will not do an adequate job in compensating for the ORB. Part IA of this article explains the likely causes of the ORB and their relevance to crafting better jury instructions. Part IB discusses the analogous research just noted, drawing on its suggestion of criteria for drafting effective cautionary instructions--criteria that, this article argues, Cromedy fails. Those same criteria suggest ways to draft a superior instruction. Nevertheless, it is this article's position that even an improved instruction will often not be sufficient. Part II thus addresses why reliance on expert testimony should be expanded, while Part III summarizes this piece's main conclusions and suggests other specific reforms.

Professor of Law, Washington College of Law, American University; former prosecutor, Philadelphia, PA.; J.D., University of Pennsylvania 1981.