Cross-Racial Identifications: Solutions to the “They All Look Alike” Effect

Introduction

On a late summer evening in August of 1997, Nathan Brown was in his apartment rocking his young daughter to sleep when the police knocked on his door.¹ Louisiana Man Is Freed from Prison after DNA Results Prove His Innocence in Attempted Rape Conviction , Innocence Project Blog (Jun. 4, 2014, 2:35 PM), http://www.innocence project.org/news-events-exonerations/louisiana-man-is-freed-from-prison-after-dna-resultsprove-his-innocence-in-attempted-rape-conviction [hereinafter Louisiana Man ]. The police sought Brown, one of a few Black men in his apartment complex, after a young White woman said she had been assaulted by a shirtless Black man wearing black shorts with strong body odor walking through the complex’s courtyard.² Id. Minutes later the police took Brown outside and put him in the patrol car for a one-on-one “showup.”³ Id . A showup is the “presentation of a single suspect to a witness on the street shortly after the crime.” Keith A. Findley, Judicial Gatekeeping of Suspect Evidence: Due Process and Evidentiary Rules in the Age of Innocence , 47 Ga. L. Rev. 723, 768 (2013). They brought him out by himself to see the victim wearing black shorts without a shirt, and she quickly identified him as her attacker, even though he lacked a strong body odor.⁴ Louisiana Man , supra note 1. The victim explained later that she believed he had showered right after the attack, meaning he was her attacker.⁵ Id .

The victim again identified Brown as her attacker at trial.⁶ Id . Though Brown took the stand in his own defense and testified that he was home at the time of the attack caring for his “fussy infant daughter”—an alibi corroborated by four of his family members—he was convicted of attempted aggravated rape and sentenced to twenty-five years in prison on the basis of the victim’s identification alone.⁷ Id . In June 2014, Brown was exonerated of the crime when DNA evidence revealed that he could not have been the attacker.⁸ Id . The DNA evidence was an exact match to a seventeen-year-old Black male who had been living within blocks of the apartment complex where the victim had been attacked.⁹ Id. Nevertheless, Brown spent nearly seventeen years in prison for a crime that he did not commit.¹⁰ Id .

Brown’s story is one of many false convictions that have occurred because of mistaken eyewitness identifications. Innocence Project statistics show that mistaken identifications account for approximately seventy-two percent of the 321 wrongful convictions¹¹ An Examination of Why Innocent People Are Locked Up , Innocence Project , http:// www.innocenceproject.org/news-events-exonerations/an-examination-of-why-innocent-peo ple-are-locked-up (last visited Oct. 11, 2015). in the United States that have since been overturned by DNA evidence.¹² The Causes: Eyewitness Misidentification , Innocence Project , http://www.inno cenceproject.org/understand/Eyewitness-Misidentification.php (last visited Dec. 29, 2014). Of the 330 people exonerated through DNA evidence, forty percent are the result of mistaken crossracial identification,¹³ DNA Exonerations Nationwide , Innocence Project 1-2 (Sept. 3, 2015, 12:30 PM), www.innocenceproject.org/Content/DNA_Exonerations_Nationwide.php. which “occurs when an eyewitness is asked to identify a person of another race.”¹⁴State v. Cromedy, 727 A.2d 457, 461 (N.J. 1999), abrogated by State v. Henderson, 27 A.3d 872 (N.J. 2011). The police procedures used in investigating the attack and convicting Brown likely exacerbated the fallibility of the cross-racial identification.¹⁵ See Brian L. Cutler & Margaret Bull Kovera , Evaluating Eyewitness Identification 37 (2010); Brandon L. Garrett , Convicting the Innocent: Where Criminal Prosecutions Go Wrong 49-50, 52-53 (2011).

Part I of this Note will examine “own-race bias”¹⁶ See Cutler & Bull Kovera , supra note 15. and its effect on eyewitness identification. Part II will discuss Supreme Court precedent on eyewitness identification, which fails to adequately protect defendants who are on trial as a result of eyewitness misidentifications.¹⁷ See Manson v. Brathwaite, 432 U.S. 98, 112-13 (1977) (stating that it would be “Draconian” to reverse a case using a reliable eyewitness identification that was obtained using an “unnecessarily suggestive identification procedure”); Perry v. New Hampshire, 132 S. Ct. 716, 730 (2012) (holding “that the Due Process Clause does not require a preliminary judicial inquiry into the reliability of an eyewitness identification when the identification was not procured under unnecessarily suggestive circumstances arranged by law enforcement”); United States v. Ash, 413 U.S. 300, 321 (1973) (holding that the “Sixth Amendment does not grant the right to counsel at photographic displays conducted by the Government for the purpose of allowing a witness to attempt an identification of the offender”); Stovall v. Denno, 388 U.S. 293, 297 (1967) (refusing to retroactively apply two cases that would “deter law enforcement authorities from exhibiting an accused to witnesses before trial for identification purposes without notice to and in the absence of counsel”); Cutler & Bull Kovera , supra note 15; Garrett , supra note 15, at 53- 54. Part II will also survey how state courts are addressing the problems of cross-racial identification in the absence of Supreme Court guidance.¹⁸ See State v. Madison, 536 A.2d 254 (N.J. 1988); see also State v. Henderson, 27 A.3d 872, 917 (N.J. 2011). Part III will then examine proposed police practice reforms that can help to stop mistaken identifications as the result of the cross-race effect.¹⁹ See Eyewitness Identification Reform , Innocence Project , http://www.innocence project.org/understand/Eyewitness_Misidentification_Reform.php (last visited Dec. 29, 2014); Dep’t of Justice: Nat’l Inst. of Justice , Eyewitness Identification: A Guide for Law Enforcement 1 (1999); Dan Simon , In Doubt: The Psychology of the Criminal Jus – tice Process 82 (2012); Nyerere Davidson, Research in Brief: Eyewitness Identifications: A National Survey on Procedures , The Police Chief (2013), http://www.policechiefmagazine.org/ magazine/index.cfm?fuseaction=print_display&article_id=3060&issue_id=92013; Int’l Ass’n of Police Chiefs & Dep’t of Justice , Nat’l Summit on Wrongful Convictions : Building a Systemic Approach to Prevent Wrongful Convictions 18 (2013); Gar – rett , supra note 15, at 52. Finally, Part IV will argue that federal courts should have a per se rule granting an expert witness the ability to testify to the unreliable nature of cross-racial identifications in cases where the defendant is represented by a public defender or by private counsel under the Criminal Justice Act and where the cross racial identification is the primary evidence used against the defendant.

I. Own-Race Bias and Its Impact on Eyewitness Testimony

This section will examine own-race bias and its effect on eyewitness identification. A study conducted in 1988 asked whether people are more accurate in identifying members of their own racial or ethnic group.²⁰ Cutler & Bull Kovera , supra note 15, at 37-38. In the study, White, Black, and Hispanic customers were asked to visit a convenience store in El, Paso Texas, and interact with White, African-American, and Hispanic clerks.²¹ Id . The customers engaged with the clerks in odd ways, such as paying for their purchases entirely with pennies.²² Id . at 38. Two to three hours later, an investigator asked the clerks to identify the customers from a photo array.²³ Id. The White clerks identified fifty-three percent of the White customers correctly, forty percent of the African-American customers correctly, and thirty-four percent of the Hispanic customers correctly.²⁴ Id. African-American clerks identified sixty-four percent of the African-American customers correctly, fifty-five percent of the White customers correctly, and forty-five of the Hispanic customers correctly.²⁵ Id. Hispanic clerks identified fifty-four percent of the Hispanic customers correctly, thirty-six percent of the White customers correctly, and twenty-five percent of the African-American customers correctly.²⁶ Id.

This study is indicative of the phenomenon of own-race bias. Ownrace bias suggests that “witnesses are more accurate at recognizing samerace perpetrators than other-race perpetrators.”²⁷ Id. at 37. The mainstream scientific community has accepted own-race bias as fact.²⁸ See, e.g. , State v. Henderson, 208 N.J. 208, 267 (2011) (citing Christian A. Meissner & John C. Brigham, Thirty Years of Investigating the Own-Race Bias in Memory for Faces: A Meta- Analytic Review , 7 Psychol. Pub. Pol’y & Law 3, 21 (2001)). As the New Jersey Supreme Court noted in State v. Henderson, one “meta-analysis . . . involving thirty-nine studies and nearly 5,000 identifications confirmed” the existence of own-race bias.²⁹ Id. According to researchers June Chance and Alvin Goldstein of the University of Missouri, Columbia, “All reviewers, with one exception . . . [have] concluded that the other-race effect is easily replicated and substantially affects subject accuracy in recognizing faces.”³⁰June E. Chance & Alvin G. Goldstein, The Other Race Effect and Eyewitness Identification , in Psychol. Issues in Eyewitness Identification 153, 155 (Siegfried L. Sporer ed., 1996).

Own-race bias is particularly disconcerting in the context of eyewitness identifications, as the jurors who are meant to judge the credibility of these identifications may not understand the complex issues surrounding them.³¹Sheri Lynn Johnson, Cross-Racial Identification Errors in Criminal Cases , 69 Cornell L. Rev. 934, 946-49 (1984). Eyewitness accounts often persuade juries, as “jurors tend to believe eyewitness accounts even in extremely doubtful circumstances.”³² Id. at 946. Further, many potential jurors do not know about or believe in the existence of own-race bias.³³ Id. A study conducted at the University of Washington showed that only fifty-eight percent of college students believed that it is more difficult to identify someone of a different race than their own.³⁴ Id. at 947. Another study indicated that eighty-four percent of prosecutors believe that eyewitness identifications are “probably correct” ninety percent of the time.³⁵John P. Rutledge, They All Look Alike: The Inaccuracy of Cross-Racial Identifications , 28 Am. J. Crim. L. 207, 209 (2001). Given the fact that prosecutors—trained legal professionals—do not see the problems with eyewitness identifications, it is ambitious to believe that lay people serving on juries will think more critically about these identifications.

Chance and Goldstein, two prominent researchers in the field of eyewitness psychology, have noted that without a jury instruction or expert testimony about the problems surrounding cross-racial identification, most juries will not initiate a conversation about the credibility of a cross-racial identification when evaluating the evidence.³⁶Chance & Goldstein, supra note 30, at 172. They suggest that many judges are reluctant to allow a discussion in court about how frequently misidentifications are made because they believe that juries are already aware of own-race bias.³⁷ Id. Many judges think this because “‘they all look alike to me’ has a long history of being a one-line gag used by both Black and White stand-up comedians.”³⁸ Id. However, as Chance and Goldstein point out, many jurors may be uncomfortable starting a discussion about the issues surrounding cross-racial identifications without prompting from a judge or expert because they do not want to be seen as “harbor[ing] racist views.”³⁹ Id. Therefore, without expert testimony or jury instructions on the topic, most jurors will not initiate a discussion about likelihood of error in a cross-racial identification in their deliberations.⁴⁰ Id. at 173. Many judges are also reluctant to exclude identifications or to comment on their reliability because they feel that doing so would interfere with the traditional function of the jury.⁴¹ See, e.g. , Dana Walsh, Note, The Dangers of Eyewitness Identification: A Call for Greater State Involvement to Ensure Fundamental Fairness , 36 B.C. Int’l & Comp. L. Rev. 1415, 1440 (2013). This is a dangerous conundrum, for without comment or information about the problems surrounding cross-racial identifications, juries may be deciding cases without all of the facts they need to evaluate this type of eyewitness testimony.⁴² See Elizabeth F. Loftus & James M. Doyle , Eyewitness Testimony: Civil and Criminal §1-5, §1-6 (1997) (describing studies that support the notion that juries place extra emphasis on eyewitness testimony).

Many scholars believe that understanding the cause of own-race bias will help resolve cross-racial identification issues in criminal trials.⁴³ See Chance & Goldstein, supra note 30, at 156. While no scientific study has concluded the cause of own-race bias, there are three major hypotheses. The first is that some groups of faces may be inherently more difficult to identify and to remember.⁴⁴ Id. Most studies suggest this is unlikely to be the cause, however, because it would mean that some ethnic groups have more variability in their faces, which has been proven untrue.⁴⁵ Id. One study has even shown that African-Americans have more variability and distinct characteristics in their faces, which would mean that this racial group would be easier to identify.⁴⁶ Id. Given the pervasive issues of cross-racial identifications in the criminal justice system, the theory that African-Americans are easier to identify has proven to be untrue.⁴⁷ Id. ; see also Eyewitness Misidentification , supra note 12.

The second hypothesis is that attitudes and stereotypes about certain ethnicities impair or enhance an individual’s ability to identify people of certain racial backgrounds.⁴⁸ See Chance & Goldstein, supra note 30, at 156. Attitudes are also unlikely to explain the own-race bias because attitude encompasses too many other factors.⁴⁹ Id. at 157.

The third hypothesis is that familiarity with certain ethnic groups may affect one’s ability to identify them more easily.⁵⁰ Id. at 158. While many scholars believe that familiarity is the most likely explanation for own-race bias, this has not been consistently seen in field or laboratory studies.⁵¹ Id. at 159. Nevertheless, more recent studies have shown “positive associations between subject reports of contact and other-race face recognition,” meaning that people who spend more time with people of different races more easily identify people of other races.⁵² Id. at 160. This partially supports the third hypothesis: despite these studies, it is unclear what type of contact and how much of it is needed to improve other-race facial recognition.⁵³ Id.

Without more information about the cause or causes of own-race bias, courts and law enforcement are left to determine how to best ensure that it does not taint eyewitness identifications. This is concerning because courts do not have many laws outside of constitutional rights to guide them, as discussed in the following section, and because law enforcement have not been trained on the issues surrounding cross-racial identification, as discussed in Part III.

II. How Courts are Handling the Cross-Race Effect

Though own-race bias is well documented and accepted in the scientific community, courts have been slow to accept these findings and to incorporate them into criminal procedure.⁵⁴ See Stephen J. Saltzburg, Report to House of Delegates , 104 D American Bar Ass’n Criminal Justice Section , http://www.americanbar.org/content/dam/aba/publish ing/criminal_justice_section_newsletter/crimjust_policy_am08104d.authcheckdam.pdf (noting that only five states have added a jury instruction regarding cross-racial identifications as of 2008). Part II will discuss Supreme Court precedent on eyewitness identification, which fails to adequately protect defendants who are on trial as a result of eyewitness misidentifications.⁵⁵ See Manson v. Brathwaite, 432 U.S. 98, 113 (1977); Perry v. New Hampshire, 132 S. Ct. 716, 720 (2012); United States v. Ash, 413 U.S. 300, 321 (1973); Stovall v. Denno, 388 U.S. 293, 296 (1967); Cutler & Bull Kovera , supra note 15; Garrett , supra note 15. Part II will also survey how state courts are addressing the problems of cross-racial identification in the absence of Supreme Court guidance.

A. Supreme Court Precedent

As Justice William Brennan once wrote, “[T]here is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says ‘That’s the one!’”⁵⁶Watkins v. Sowders, 449 U.S. 341, 352 (1981) (Brennan, J., dissenting) (citing Eliza – beth F. Loftus , Eyewitness Testimony (1979)). Given the persuasive nature of identifications, one would think that the Supreme Court would have established a protective jurisprudence that regulates suggestive eyewitness identification procedures carefully, but that is not the case. The Court’s eyewitness identification jurisprudence focuses on suggestive police practices,⁵⁷ See Manson v. Brathwaite, 432 U.S. 98 (1977) (involving the suggestive use of a single-photograph identification procedure). but it has not addressed the scientific issues that go beyond the suggestiveness of the procedures, such as own-race bias.

In Manson v. Brathwaite, the Supreme Court held that the Due Process Clause of the Fourteenth Amendment includes the right to be free from suggestive eyewitness identification procedures.⁵⁸ Id. at 113 (“The standard, after all, is that of fairness as required by the Due Process Clause of the Fourteenth Amendment.”) (citing Manson v. Brathwaite , 432 U.S. 98 (1977)). Therefore, pre- and post-indictment lineups, showups, and photo arrays are analyzed under a due process inquiry.⁵⁹Perry v. New Hampshire, 132 S. Ct. 716, 720 (2012) (“This Court has recognized . . . a due process check on the admission of eyewitness identification, applicable when the police have arranged suggestive circumstances leading the witness to identify a particular person as the perpetrator of a crime.”); see also Ash , 413 U.S. at 321 (finding that a due process inquiry of the photographic display remains open on remand). The inquiry begins with an analysis of whether the identification was police-orchestrated⁶⁰ Perry , 132 S. Ct. at 720. and whether it was unnecessarily suggestive.⁶¹ Manson , 432 U.S. at 107. (“[T]he first inquiry was whether the police used an impermissibly suggestive procedure in obtaining the out-of-court identification.”). In Perry v. New Hampshire, for instance, the Court held that the victim’s identification of the defendant by spotting him through a window could not be excluded on due process grounds because the police had not set up the identification or asked the victim to identify the defendant.⁶² Perry , 132 S. Ct. at 720. In another opinion, the Court provided the following examples of situations that it would consider unnecessarily suggestive: when only one person is shown, when the suspect is handcuffed, when the suspect is the only person of his race, and when the police praise the witness’s choice.⁶³ See Stovall v. Denno, 388 U.S. 293, 295 (1967). While detailing the procedures that make an identification unnecessarily suggestive, the Court said that in emergencies, these procedures may be allowed based on the circumstances.⁶⁴ See id. at 302 (explaining that where the only witness is potentially going to die in surgery, police can bring the defendant to do a showup in her hospital room). For example, if the victim is injured and may die before identifying the suspect in a lineup, these procedures would not violate a defendant’s due process rights.⁶⁵ See id.

If the court determines that an out-of-court identification is policeorchestrated and thereby unnecessarily suggestive, the trial court can still allow the in-court identification. In Manson, the Court undercut its protection of defendants from false identifications by permitting “reliable” incourt identifications.⁶⁶Manson v. Brathwaite, 432 U.S. 98, 111 (1977). Yet, the Court has given very little guidance to trial courts about what constitutes a “reliable” identification.⁶⁷ See id . at 107. Consequently, trial judges, who are not well versed in the science of eyewitness identifications, allow into evidence many eyewitness identifications that are actually unreliable.⁶⁸After the initial suggestive procedure leads to a mistaken identification, many witnesses affirm the identification in their minds and become more certain of it. See Garrett , supra note 15, at 49. One study found that trial records of 125 cases of false convictions clearly showed police contamination of the eyewitness identifications. Brandon L. Garrett , Eyewitnesses and Exclusion , 65 Vand. L. Rev. 451, 475 (2012). Yet the identifications were allowed to proceed in court. See Manson v. Brathwaite, 432 U.S. 98, 111 (1977).

Even if a method of out-of-court identification is inherently suggestive, a trial judge can determine that the in-court identification is reliable and allow the witness to identify the defendant in court. A judge can only begin to examine whether there is a substantial likelihood of mistaken identification if that judge determines that an identification procedure was police-orchestrated and unnecessarily suggestive.⁶⁹Perry v. New Hampshire, 132 S. Ct. 716, 720 (2012); see also id. at 112-14. In Manson v. Braithwaite, the Court provided factors that lower courts should consider in determining whether there has been a substantial likelihood of misidentification: the eyewitness’s opportunity to view the suspect, the degree of attention paid by the eyewitness, the accuracy of the description as compared to what the suspect actually looks like, the level of certainty the witness has displayed, and the time between the crime and the identification.⁷⁰ Manson , 432 U.S. at 114.

There are two main issues with current Supreme Court case law regarding eyewitness identifications. The first is that a court will not even look at the reliability of the identification unless the procedure was policeorchestrated and unnecessarily suggestive.⁷¹ Perry , 132 S. Ct. at 720. As a result, the scientific findings about the cross-race effect will not even be considered unless the police used an unnecessarily suggestive technique.⁷² See id. Nevertheless, the crossrace effect will influence eyewitness testimony regardless of police procedure. The second issue builds upon the first—under current Supreme Court jurisprudence, trial judges can determine that an identification is reliable even if suggestive police practices were used, but these judges cannot determine that the identification is unreliable without suggestive police practices.⁷³ See Garrett , supra note 15, at 53.

The Supreme Court has not provided guidance for cross-racial identifications, and it has not required that trial judges consider these issues when determining if an identification is reliable.⁷⁴ See Manson v. Brathwaite, 432 U.S. 98, 114 (1977). Trial judges will only look into the reliability of an identification if the procedure was unnecessarily suggestive, meaning that the courts are solely concerned with initial police behavior.⁷⁵Perry v. New Hampshire, 132 S. Ct. 716, 720, 730 (2012). The factors concerning the reliability of eyewitness identifications will not be analyzed unless the police methods are of the nature the Supreme Court termed unnecessarily suggestive.⁷⁶ Id . at 730. The courts will not question the reliability of an eyewitness identification unless the police used an inappropriate method of identification.

B. State Court Approaches

While the Supreme Court has not addressed the issues surrounding cross-racial identification, the New Jersey Supreme Court has led the way in eyewitness identification reform. Before State v. Henderson, New Jersey used the due process standard that the Supreme Court set forth in Manson v. Braithwaite.⁷⁷State v. Madison, 536 A.2d 254 (N.J. 1988). Then in State v. Henderson, the New Jersey Supreme Court considered thirty years’ worth of scientific studies that had been conducted since Manson and determined that the current protections were inadequate.⁷⁸State v. Henderson, 27 A.3d 872, 917-18 (N.J. 2011).

Before ruling in Henderson, the New Jersey Supreme Court appointed a “Special Master” to sift through all the current evidence on eyewitness identifications and present a report on his/her findings.⁷⁹After initially granting certiori and hearing oral argument on the case, the New Jersey Supreme Court determined that it needed more information about the science of eyewitness identification. Id. at 877. It remanded the case to a Special Master that it appointed, a retired New Jersey Court of Appeals judge. Id. at 884. The Special Master was tasked with evaluating scientific and other evidence about eyewitness identification on remand and made a determination about whether the research was credible. Id. at 877. In that report, the Special Master determined that “[t]he science abundantly demonstrates the many vagaries of memory encoding, storage, and retrieval; the malleability of memory; the contaminating effects of extrinsic information; the influence of police interview techniques and identification procedures; and the many other factors that bear on reliability of eyewitness identifications.”⁸⁰ Id. at 916.

Based on these findings, the court determined that the “Manson/ Madison [test] does not adequately meet its stated goals: it does not provide a sufficient measure for reliability, it does not deter, and it overstates the jury’s innate ability to evaluate eyewitness testimony.”⁸¹ Id. at 918. Moreover, the New Jersey Supreme Court determined that there were five significant problems with the current approach: (1) defendants must show that police procedures were “impermissibly suggestive” before courts can consider reliability; (2) three of the five factors that are used to determine reliability depend on self-reporting by the eyewitnesses, which may not be reliable; (3) the test may unintentionally reward suggestive police practices because the more suggestive the practice, the more reliable a witness will seem; (4) suppression is the only remedy, and most judges are hesitant to give such a harsh punishment; and (5) the test asks that judges look at the “totality of the circumstances” when making a reliability determination, but most only use the five factors because there is no other guidance regarding what else should be considered when examining the totality of the circumstances.⁸² Id. at 918-20.

Given these issues, the court revised the five-factor reliability test to include an inquiry into all of the system and estimator variables⁸³ Id . at 921. A system variable is one that the legal system has control over, and an estimator variable is one that the legal system cannot control. Id. at 895. that help determine eyewitness reliability.⁸⁴ Id. at 918-20. The revised New Jersey test has four steps.⁸⁵ Id . at 920. First, the defendant must “show some evidence of suggestiveness that could lead to a mistaken identification.”⁸⁶ Id. To show suggestiveness, the defendant must point out issues with system variables, including blind administration, pre-identification instructions, lineup construction, the feedback provided to the witness, and recording confidence, multiple viewings, simultaneous versus sequential lineups, composites, or showups. Id. at 920-21. Once the defendant has met his initial burden, the State is required to show proof of reliability, using both system and estimator variables.⁸⁷ Id. at 920. System variables are within the state’s control. Id. at 904-07. Estimator variables are beyond the state’s control and relate to the “incident, the witness, or the perpetrator.” Id. They include stress, weapon focus, duration, distance and lighting, witness characteristics, characteristics of the perpetrator, memory decay, race-bias, private actors, and speed of identification. Id. The defendant is ultimately responsible for proving that there is a substantial likelihood of misidentification by showing evidence of system and estimator variables, and he can do so through cross-examination and the presentation of his own witnesses.⁸⁸ Id. at 920. Finally, after weighing all of the evidence, the trial judge can either decide to suppress the identification or provide “appropriate, tailored jury instructions” that inform jurors that they should give weight to the identification on the basis of its credibility.⁸⁹ Id.

Most importantly, the New Jersey Supreme Court said that race-bias, or a cross-racial identification, should be considered when a judge is weighing the credibility of an identification under its new test.⁹⁰ Id. at 921. The court allows defendants to present evidence about the flaws of cross-racial identifications through the use of expert testimony and in cross-examination and in closing arguments.⁹¹ Id. at 925. Furthermore, the court requires a jury instruction on cross-racial identification “whenever cross-racial identification is in issue at trial.”⁹² Id. at 926. This means that the jury will be educated about the issues surrounding cross-racial identifications whenever one is presented at trial. This instruction and the other reforms outlined above are immensely helpful in ensuring that juries are aware of the issues surrounding these types of identifications.

Not all state supreme courts have agreed that it is their role to reform eyewitness identifications. For example, the Washington Supreme Court held in State v. Allen that a defendant’s due process rights are not violated if the trial court refuses to instruct jurors on cross-racial eyewitness identifications.⁹³State v. Allen, 294 P.3d 679, 687 ( Wash. 2013) (en banc) (“We decline to adopt a general rule requiring the giving of a cross-racial instruction in cases where cross-racial identifi- The Supreme Court of Washington noted that a defendant has other options for exposing the issues surrounding cross-racial identifications, including the “right to effective assistance of counsel, who can expose the unreliability in eyewitness’ testimony during cross examination and focus the jury’s attention on the fallibility of eyewitness identification during opening and closing arguments.”⁹⁴ Id. at 685. Most jurisdictions follow this approach.⁹⁵ See Saltzburg, supra note 54, at 12.

As of 2008,⁹⁶These are the most recent comprehensive statistics available on states that require a jury instruction on cross-racial identifications. See Commonwealth v. Bastaldo , 32 N.E.3d 873 (2015). the only states that require or authorize a jury instruction on cross-racial identifications are California, Utah, Massachusetts, and New Jersey.⁹⁷Saltzburg, supra note 54, at 12-14. These states have modeled their jury instructions on the model instructions provided by Judge Bazelon in his concurring opinion in United States v. Telfaire:

In this case the identifying witness is of a different race than the defendant. In the experience of many it is more difficult to identify members of a different race than members of one’s own. If this is also your own experience, you may consider it in evaluating the witness’s testimony. You must also consider, of course, whether there are other factors present in this case which overcome any such difficulty of identification. For example, you may conclude that the witness has had sufficient contacts with members of the defendant’s race that he would not have greater difficulty in making a reliable identification.⁹⁸United States v. Telfaire, 469 F.2d 552, 561 (D.C. Cir. 1972) (Bazelon, J., concurring).

No federal circuits have endorsed Judge Bazelon’s instructions as of 2008.⁹⁹Saltzburg, supra note 54, at 11.

III. Police and Federal Agency Practices and Reforms

The Innocence Project, an organization that works to overturn wrongful convictions and one of the leading voices in the movement to reform both police and courtroom eyewitness identification procedures, notes on its website that “[d]espite solid and growing proof of the inaccuracy of traditional eyewitness ID procedures . . . traditional eyewitness identifications remain among the most commonly used and compelling evidence brought against criminal defendants.”¹⁰⁰ Eyewitness Identification Reform , Innocence Project (June 10, 2015, 4:57 PM), http://www.innocenceproject.org/free-innocent/improve-the-law/fact-sheets/eyewitness-iden tification-reform. Given the importance the criminal justice system places on eyewitness evidence, and the fact that many states have been hesitant to implement reforms, police agencies should undertake eyewitness identification reform. This Part will examine current police practices and evaluate proposed police practice reforms that can help to stop mistaken identifications as the result of the cross-race effect.

A. Current Police Practices

The Department of Justice took notice of the problems surrounding eyewitness identifications in 1999, and it commissioned a guide to eyewitness identification.¹⁰¹ Dep’t of Justice Eyewitness Identification Guide , supra note 19. The product of collaboration among thirty-four prosecutors, police officers, and experts in the field of eyewitness identification, the Department of Justice Guide was designed to aid “a jurisdiction shaping its own protocols” surrounding eyewitness identification.¹⁰² Id . at iii-iv. The Eyewitness Guide explains that it “is not intended to state legal criteria for the admissibility of evidence;”¹⁰³ Id . at 2. the project was simply about reforming police procedures without affecting the legal standard for the admissibility of eyewitness evidence.¹⁰⁴ Id.

The Eyewitness Guide provides detailed instructions for how law enforcement should handle eyewitness identifications, starting with the initial report of the crime and ending with the actual identification procedure.¹⁰⁵ Id. at 11. While the Eyewitness Guide did not recommend the use of a double-blind procedure, the Guide recommended this procedure in a 2013 joint report with the International Association of Police Chiefs.¹⁰⁶ Int’l Ass’n of Police Chiefs , supra note 19. Nevertheless, the Eyewitness Guide is especially impressive because it seeks to incorporate the “growing body of psychological knowledge regarding eyewitness evidence” into its procedures, something it notes most police agencies and courts have not yet done.¹⁰⁷ Dep’t of Justice Eyewitness Identification , supra note 19, at 1.

The Eyewitness Guide also provides detailed procedures for creating photo- and live-suspect lineups to help witnesses give more accurate identifications.¹⁰⁸ Id . These procedures include having a minimum of four fillers, or non-suspects, for live lineups and five fillers for photo lineups and ensuring that these fillers generally fit the description of the suspect or resemble the suspect.¹⁰⁹ Id. at 29-31. These guidelines are meant to ensure that the suspect does not “unduly stand out” of the lineup.¹¹⁰ Id. at 29. Once the lineup is created, the Eyewitness Guide details instructions on conducting the procedure, starting with how the investigator is to instruct the witness.¹¹¹ Id. The focus of the instruction should be that “the purpose of the identification procedure is to exculpate the innocent as well as to identify the actual perpetrator.”¹¹² Dep’t of Justice Eyewitness Identification , supra note 19, at 31. The Eyewitness Guide also explains that the witness should be instructed that the person who committed the crime may not be in the lineup and assured that the police will continue to investigate “regardless of whether an identification is made.”¹¹³ Id. at 32.

The Eyewitness Guide is most innovative in its recommendations for the actual identification procedure. It provides instructions for police officers to conduct either sequential or simultaneous lineups.¹¹⁴ Id. at 33-36. Perhaps most notably, the Eyewitness Guide suggests that the police record any identification results and witness’ statements of certainty after making identifications,¹¹⁵ Id. at 33. as studies have shown witnesses who pick a person quickly and indicate that they are highly confident in their choice are considerably more accurate.¹¹⁶ Simon , supra note 19. Therefore, getting a written record of the witness’s confidence at the time of the procedure would aid those assessing the reliability of the identification. Furthermore, the Eyewitness Guide is clear that the investigator should be careful not to say anything to the witness after the identification procedure that would influence the witness’s confidence in her decision.¹¹⁷ Dep’t of Justice Eyewitness Identification , supra note 19, at 34.

Despite the existence of the Eyewitness Guide and the importance of eyewitness identifications, most police agencies have no written procedures or trainings on conducting identifications.¹¹⁸ See Davidson, supra note 19. The National Institute of Justice conducted a national survey of police agencies regarding their eyewitness identification procedures in September 2013.¹¹⁹ Id. Photo lineups, where the suspect is displayed along with other fillers through photographs, were used by 94.1 percent of agencies, though 64.3 percent of those agencies did not have a written policy on how to conduct such lineups.¹²⁰ Id. Showups, where the suspect is shown to the witness in person without fillers, were used by 61.8 percent of agencies, and 76.9 percent of those agencies had no written policy about how to conduct them.¹²¹ Id. Composites, where police use a sketch artist or program to create a picture of the suspect based on a witness’s recollection, were used by 35.5 percent of those surveyed, 90.6 percent of which lacked a written policy on the matter.¹²² Id. Finally, live lineups were used by 21.4 percent of agencies and 84 percent had no written policy to inform their use.¹²³ Id. Furthermore, only 68 percent of agencies that use photo lineups and 44 percent of agencies that use live lineups provide training to their officers on how to conduct them, according to this survey.¹²⁴ Id. The lack of written procedures and training is disturbing, given the influence of eyewitness identifications on convictions.

While some may argue that police officers do not need to be trained in eyewitness identification procedures and the problems that surround them,¹²⁵ See, e.g. , Steven E. Clark, Eyewitness Identification Reform: Data, Theory, and Due Process , Perspectives on Psychol. Science 7, 279 (May 2012). one study showed why training is so desperately needed.¹²⁶ Simon , supra note 19, at 76. The study surveyed 500 police officers and found that only 30 percent were aware of the steep drop in the memory of eyewitnesses with the passage of time, with more than 50 percent of participants believing that it was possible to conduct a proper lineup after a witness had been exposed to a biased lineup procedure.¹²⁷ Id. These results suggest that police officers are largely unaware that poor procedures can contaminate eyewitness testimony. Even an officer acting in good faith may not realize the impact that a certain procedure may have on a witness. Therefore, training and written procedures are necessary to ensure that eyewitness identifications are conducted properly.

B. Proposed Innocence Project Reforms

While the Eyewitness Guide was an important step in eyewitness identification reform, it did not go far enough in its recommendations. The Innocence Project has compiled a list of suggested reforms that can help improve the accuracy of eyewitness identifications.¹²⁸ Eyewitness Identification Reform , supra note 19. The Innocence Project suggests some similar reforms to the Eyewitness Guide, such as instructing the witness that the suspect may not be present, compiling the lineup to have fillers that resemble the witness’s description, taking statements about the witness’s confidence immediately following the lineup procedure, and documenting the lineup.¹²⁹ Id. The Innocence Project also recommends two additional precautions: (1) double-blind procedures and (2) the sequential presentation of lineups.¹³⁰ Id.

Double-blind procedures mean that neither the administrator nor the witness knows who the suspect is during the lineup.¹³¹ Id. Administering double-blind lineups ensures that the administrator cannot provide unintentional suggestive cues to the witness while conducting the lineup.¹³² Id. This is perhaps the most important reform that police agencies can undertake in changing their eyewitness identification practices. Though there has been limited scientific research done regarding double-blind procedures, some studies suggest that when lineup administrators are aware of the identity of the suspect, witnesses are more likely to indicate that person.¹³³ Simon , supra note 19, at 74. Though the administrators and the witnesses are often unaware of the suggestiveness, the administrators may subconsciously influence the witnesses by encouraging them to take another look if they do not pick someone out initially or by removing photos of suspects more slowly when the witnesses do not pick them out during sequential photo lineups.¹³⁴ Id. Though administrators with knowledge of the suspects may use these approaches subconsciously instead of with bad faith, the problem of influencing the outcome still remains. This problem is removed when the administrator is unaware of the suspect.

The Innocence Project also recommends that police agencies use a sequential lineup instead of a simultaneous lineup,¹³⁵A simultaneous lineup is when the eyewitness views a lineup or photo array in which all of the individuals are shown at the same time, whereas a sequential lineup is when photographs or individuals are presented to the eyewitness one at a time. Eyewitness Identification Reform , supra note 19. as sequential lineups have been shown to improve accuracy because witnesses cannot make a relative, as opposed to an absolute, judgment.¹³⁶ Simon , supra note 19, at 71. When an eyewitness views a simultaneous lineup, she is choosing who in the lineup looks most like the person she saw as compared to all of the people in the lineup. Id. In a sequential lineup, she has to look at each person on his own and answer whether that is the person she saw before she can See the next suspect. Id. This eliminates the problem of comparing the people in the lineup and simply choosing who looks most similar to the suspect. Id. Eyewitnesses will just choose the best option when presented with multiple people at once. See id. With sequential lineups, where the witness is required to say yes or no before moving on to the next picture or person, witnesses are required to make absolute judgments.¹³⁷ Id. When witnesses are shown a simultaneous lineup, researchers believe that they choose the person that most resembles the perpetrator.¹³⁸ Id. As witnesses are already inclined to choose someone in a lineup, a simultaneous lineup enables them to choose the person who seems to most resemble the perpetrator.¹³⁹ See id. Since the problems of cross-racial identifications already influence witnesses’ decisions,¹⁴⁰ Id. at 63. they should not be swayed by relative judgments as well. One study has shown that the use of sequential lineups decreases false identifications without decreasing the rate of suspect identifications.¹⁴¹ Id.

C. Other Ways to Improve the Accuracy of Cross-Racial Identifications

While the above reforms to eyewitness identification procedures will certainly help with the fallibility of cross-racial identifications, and eyewitness identifications in general, there are three additional reforms that police agencies can implement to specifically help with the fallibility of crossracial identifications: (1) create photo and live lineups with proper fillers, (2) reform procedures around composite sketches, and (3) use eyewitness identification experts to train their officers about the problems of crossracial identifications.

Much of the literature surrounding cross-racial identification centers on the problems that witnesses have identifying suspects of another race.¹⁴² See, e.g. , Chance & Goldstein, supra note 30, at 154-55. Perhaps equally problematic, however, are the issues that police officers have constructing live and photo lineups involving individuals of another race.¹⁴³ See id. at 169. Own-race bias applies equally to the people who are charged with constructing these lineups.¹⁴⁴ See id. A study conducted by Brigham and Ready supports this premise.¹⁴⁵ Id. The study asked subjects to create two lineups for two different suspects, one African-American and one White.¹⁴⁶ Id. Participants were given eighty photos and were tasked with creating lineups so that the foils matched the general appearance of the suspect.¹⁴⁷ Id. The study found that the subjects chose fewer photos as resembling the suspect of their own race and more photos as resembling the suspect of another race, meaning that the fillers were not as accurate for the suspect of the opposite race.¹⁴⁸ Id. This exercise demonstrates how own-race bias can affect even the composition of lineups. When possible, police agencies should have an officer of the same race as the suspect create the lineups to ensure that the fillers are not suggestive.¹⁴⁹ Id. at 173. While this recommendation is not always feasible, especially for small precincts, it could mitigate the issues of cross-racial identifications.

Emerging technology could help to resolve own-race bias in police officers. Computer programs may soon be able to create lineups for officers from state or nationwide databases, and fillers could be pulled based on the description of the suspect.¹⁵⁰ Simon , supra note 19, at 86-89. Such programs could eliminate human error or bias.¹⁵¹ Id. Police agencies and legislatures should look into putting such programs into practice to eliminate the human bias inherent in the construction of lineups.

Another way that police agencies can help limit the fallibility of cross-racial identifications is to reform their procedures regarding composite sketches. While composites are a very useful police tool, many studies suggest that they are often incorrect and that they taint the witness’s memory.¹⁵² See State v. Lawson, 291 P.3d 673, 703 (Or. 2012). Furthermore, “the connection between face recognition memory and verbal processes has been shown to be weak or nonexistent.”¹⁵³Chance & Goldstein, supra note 30, at 164 (noting that good face describers are not more skilled at remembering faces than bad face describers). As a result of the difficulty of describing the features of a perpetrator aloud, composites are often flawed.¹⁵⁴ Garrett , supra note 15. The process of working with a sketch artist or computer program can also distort the witness’s memory of the perpetrator.¹⁵⁵ See id. (detailing how in one exoneree’s case a witness said that it was only “after the composite” that she began “looking and putting the pieces together myself of what I remembered of that person’s face,” suggesting that the composite distorted her memory). In the context of cross-racial identifications, at least one study suggests that some witnesses focus mainly on the race of the perpetrator and not on her features.¹⁵⁶ See Chance & Goldstein, supra note 30, at 164 (One study found that “[n]ine out of 10 Scottish subjects, in contrast to 3 out of 10 African subjects, weighted the racial identity dimension most heavily in their judgments.”). These witnesses make for less than ideal witnesses in the context of composite sketches.¹⁵⁷ Id. Therefore, police agencies should use composites sparingly in cross-racial identifications.

Finally, while expert testimony about cross-racial identification is left to the discretion of the trial courts,¹⁵⁸See infra Part IV. experts can inform police departments how to develop better eyewitness identification procedures and training sessions.¹⁵⁹Chance & Goldstein, supra note 30, at 173. They can point out the flaws in certain lineups and help departments draft guidelines. While it is not feasible for all police departments to have an eyewitness identification expert on their staffs, experts could be involved in national initiatives, like the Department of Justice Guide and the National Summit on Wrongful Convictions.¹⁶⁰ See Dep’t of Justice , supra note 19, at 34; Int’l Ass’n of Police Chiefs , supra note 19, at 29. Manuals informed by experts can assist departments of all sizes in creating procedures that will lead to more accurate identifications.

IV. Expert Testimonyfor Indigent Defendants

Circuit courts are increasingly allowing expert testimony on eyewitness testimony issues,¹⁶¹ See, e.g. , United States v. Langan, 263 F.3d 613, 621-22 (6th Cir. 2001). but there are still some that do not allow it.¹⁶² See, e.g. , United States v. Smith, 148 F. App’x 867, 872 (11th Cir. 2005). The Sixth Circuit discussed the growing acceptance of experts on this topic in United States v. Langan:

[Expert] testimony has been allowed in with increasing frequency where the circumstances include “cross-racial identification, identification after a long delay, identification after observation under stress, and [such] psychological phenomena as . . . unconscious transference.” Nonetheless, each court to examine this issue has held that the district court has broad discretion in, first, determining the reliability of the particular testimony, and, second, balancing its probative value against its prejudicial effect.¹⁶³United States v. Langan, 263 F.3d 613, 621-22 (6th Cir. 2001) (internal citations omitted). The Sixth Circuit referred to opinions of the Fourth, Seventh, and Eighth Circuits in this passage. See id.

The practice of allowing eyewitness identification experts is growing, but there are still many hurdles.¹⁶⁴ See, e.g. , United States v. Jones , 762 F. Supp. 2d 270, 275 ( D. Mass. 2010), aff’d, 689 F.3d 12 (1st Cir. 2012). Given this deference to the trial courts, many Federal District Courts still do not allow expert testimony regarding the cross-racial effect because it typically “consists of generalized notions regarding the flaws of eyewitness testimony.”¹⁶⁵ See id. Moreover, evidentiary rules require that the expert testimony pertain to the specific issues of the case.¹⁶⁶ See Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 589 (1993) (“[U]nder the Rules [of Evidence] the trial judge must ensure that any and all scientific testimony or evidence admitted is not only relevant, but reliable.”). Federal courts should establish a per se rule granting an expert witness on the unreliable nature of cross-racial identification in cases where the defendant is represented by a federal defender or indigent counsel and where the cross-racial identification is the primary piece of evidence against the defendant.

The issue of cross-racial identification as it relates to indigent defendants requires a per se exception, as juries should be informed of the issues of such identifications to afford these defendants due process. As previously indicated, eyewitness testimony and courtroom identifications are extremely influential.¹⁶⁷ See, e.g. , Watkins v. Sowders , 449 U.S. 341, 352 (1981) (Brennan, J., dissenting). Juries typically do not know about the issues surrounding such testimony.¹⁶⁸Chance & Goldstein, supra note 30, at 172. To ensure fair trials, courts must give juries some guidance about how much weight to give this testimony, especially when it is the primary piece of evidence connecting a defendant to a crime.

Even in courts that allow expert testimony, indigent defendants are less likely to receive its benefit because they cannot afford to hire experts.¹⁶⁹ See John M. West, Expert Services and the Indigent Criminal Defendant: The Constitutional Mandate of Ake v. Oklahoma, 84 Mich. L. Rev. 1326, 1326-28 (1986) (quoting United States v. Johnson , 238 F.2d 565, 572 (1956) (Frank, J., dissenting)) (“‘The best lawyer in the world cannot competently defend an accused person if the lawyer cannot obtain existing evidence crucial to the defense, e.g., if the defendant cannot pay the fee of an investigator to find a pivotal missing witness or a necessary document, or that of an expert accountant or mining engineer or chemist. . . . In such circumstances, if the government does not supply the funds, justice is denied the poor—and represents but an upper-bracket privilege.’”). The Criminal Justice Act, which provides counsel for indigent defendants, outlines an additional application that indigent defendants must submit to obtain an expert:

Counsel for a person who is financially unable to obtain investigative, expert, or other services necessary for adequate representation may request them in an ex parte application. Upon finding, after appropriate inquiry in an ex parte proceeding, that the services are necessary and that the person is financially unable to obtain them, the court . . . shall authorize counsel to obtain the services.¹⁷⁰18 U.S.C. § 3006A(e)(1) (2006).

In the context of eyewitness testimony, the Ninth Circuit has held that “[a]ny weaknesses in eyewitness identification testimony can ordinarily be revealed by counsel’s careful cross-examination of the eyewitnesses.”¹⁷¹United States v. Labansat, 94 F.3d 527, 530 (9th Cir. 1996). Therefore, to bring in expert testimony, “the defendant must establish why such cross-examination is inadequate and why an expert is required.”¹⁷²United States v. Pitts, 346 F. App’x 839, 841-42 (3d Cir. 2009). Courts that have addressed this issue often follow Pitts in holding that expert testimony cannot come in because the unreliability of cross-racial identification can be established through cross-examination.¹⁷³ See, e.g. , United States v. Rodriguez-Felix, 450 F.3d 1117, 1128-29 (10th Cir. 2006) (“affirm[ing] the district court’s evidentiary rulings denying expert psychological testimony on the general reliability of eyewitness identification”).

Cross-examination is insufficient to protect an indigent defendant when the identification is the primary piece of evidence against him. Given the weight that juries place on these identifications, they should be alerted to the issues surrounding them. When a court-certified expert is allowed to testify to these problems, juries will be more likely to bring them up during deliberations and to weigh identifications accordingly.¹⁷⁴ See Am. Bar Ass’n , Report to the House of Delegates , 2008 A.B.A. Sec. Crim. Just. 15 (citing People v. Beckford, 532 N.Y.S.2d 462, 465 ( S. Ct. Kings Cty. 1988)) (“Jurors are more apt to comfortably discuss racial differences without fear of discord in the jury room when they have received testimony from an expert considering the possible influence of racial differences as affecting the accuracy of the identification. Also, they argue that the possibility of error in crossracial identifications is not within the ordinary knowledge of many jurors.”).

Conclusion

Although cross-racial identifications have consistently proven to be problematic in the context of criminal prosecutions, the Supreme Court and most state courts have refused to place limits on the use of these types of identifications. While eyewitness identifications are an important tool for police and prosecutors to use to solve crimes, reforms are needed to ensure that their use is proper and promotes accurate identifications. State courts have started to implement some reforms, and police departments can supplement these measures with practices like double-blind procedures and the sequential presentation of lineups, which are designed to ensure that the correct suspect is apprehended. As Justice Harlan famously wrote in In re Winship, “[I]t is far worse to convict an innocent man than to let the guilty go free.”¹⁷⁵ In re Winship, 397 U.S. 358, 372 (1970) (Harlan, J., concurring). With Justice Harlan’s foundational principle in mind, Congress and the federal courts should re-examine their rules regarding expert testimony and cross-racial identifications and to provide expert testimony for those who cannot otherwise afford it. These measures would help prevent mistaken identifications and thereby prevent the wrongful convictions that inevitably follow.