AMERICAN BAR ASSOCIATION
CRIMINAL JUSTICE SECTION
REPORT TO THE HOUSE OF DELEGATES
RESOLVED, That the American Bar Association adopts the American Bar Association Statement of Best Practices for Promoting the Accuracy of Eyewitness Identification Procedures dated August 2004.
FURTHER RESOLVED, That the American Bar Association urges federal, state, local and territorial governments to reduce the risk of convicting the innocent, while increasing the likelihood of convicting the guilty
, by adopting the following principles:
1. Police and prosecutors craft detailed guidelines for conducting lineups and photospreads in a manner that maximizes their likely accuracy;
2. Police and prosecutors receive periodic training on how to implement the above-referenced guidelines,
3. Police and prosecutors receive periodic training on non-suggestive techniques for interviewing witnesses;
4. Internal mechanisms be created within police departments and prosecutors’ offices to periodically update such guidelines to incorporate advances in social scientific research and in the continuing lessons of practical experience; and
5. Every set of guidelines should address at least the subjects, and should incorporate at least the social scientific teachings and best practices, set forth in the American Bar Association Statement of Best Practices for Promoting the Accuracy of Eyewitness Identification Procedures dated August 2004.
FURTHER RESOLVED, That the American Bar Association, to improve the ability of juries and judges to make fully informed trial decisions concerning the accuracy of eyewitness identifications, urges federal, state, local and territorial governments to reduce the risk of convicting the innocent, while increasing the likelihood of convicting the guilty, by adopting the following principles:
1. Courts should have the discretion, where appropriate in an individual case, to allow a properly qualified expert to testify both pretrial and at trial on the factors affecting eyewitness accuracy; and
2. Whenever there has been an identification of the defendant prior to trial, and identity is a central issue in a case tried before a jury, courts should consider exercising their discretion to use a specific instruction, tailored to the needs of the individual case, explaining the factors to be considered in gauging the accuracy of the identification.
AMERICAN BAR ASSOCIATION STATEMENT OF BEST PRACTICES FOR PROMOTING THE ACCURACY OF EYEWITNESS IDENTIFICATION PROCEDURES DATED AUGUST, 2004
A. General Guidelines for Administering Lineups and Photospreads
1. Whenever practicable, the person who conducts a lineup or photospread and all others present (except for defense counsel, when his or her presence is constitutionally required) should be unaware of which of the participants is the suspect;
2. Eyewitnesses should be instructed that the perpetrator may or may not be in the lineup; that they should not assume that the person administering the lineup knows who is the suspect; and that they need not identify anyone, but, if they do so, they will be expected to state in their own words how certain they are of any identification they make or the lack thereof;
B. Foil Selection, Number, and Presentation Methods
1. Lineups and photospreads should use a sufficient number of foils to reasonably reduce the risk of an eyewitness selecting a suspect by guessing rather than by recognition;
2. Foils should be chosen for their similarity to the witness's description of the perpetrator, without the suspect's standing out in any way from the foils and without other factors drawing undue attention to the suspect;
3. The advisability of either a sequential lineup or photospread (showing one person or photo to a witness at a time, with the witness being asked to identify or not identify each person or photo immediately after it is presented) or a simultaneous lineup or photospread (showing a witness all lineup members or photographs at the same time) should be carefully considered;
4. Police departments and prosecutors should be urged to participate in properly-designed comparative field experiments in which one group of police districts in a city or county uses simultaneous lineup and photospread methods while another group of police districts uses sequential methods;
C. Recording Procedures
1. Whenever practicable, the police should videotape or digitally video record lineup procedures, including the witness’s confidence statements and any statements made to the witness by the police;
2. Absent videotaping or digital video recording, a photograph should be taken of each lineup and a detailed record made describing with specificity how the entire procedure (from start to finish) was administered, also noting the appearance of the foils and of the suspect and the identities of all persons present.
3. Regardless of the fashion in which a lineup is memorialized, and for all other identification procedures, including photospreads, the police shall, immediately after completing
the identification procedure and in a non-suggestive manner, request witnesses to indicate their level of confidence in any identification and ensure that the response is accurately documented.
D. Immediate Post-Lineup or Photospread Procedures
1. Police and prosecutors should avoid at any time giving the witness feedback on whether he or she selected the "right man" -- the person believed by law enforcement to be the culprit.
I. Introduction: Illustrating the Problem
On June 5, 1999, Calvin C. Johnson, Jr. was released from prison after having served more than 15 years of a life sentence for rape.1 Johnson was released because he had recently been exonerated by DNA evidence. Johnson’s conviction had been based largely on a flawed eyewitness identification.
The rape victim, Ms. Mitchell, had selected Johnson’s black-and-white photo from a photospread that included a number of full color pictures. But Ms. Mitchell selected someone other than Johnson during a live lineup. Johnson was clean-shaven in the photospread, but his work identification photos taken around the time of the rapes showed him sporting a very full, bushy beard. He still had the beard at the time of the lineup. The lineup was held about one week after the crime, far too soon after the rape for him to have had sufficient opportunity to grow a full beard in the interim. Yet Ms. Mitchell had told the police that her assailant was either clean-shaven or sported some “stubble.”
The rape took place mostly in darkness (there was some light from the nearby bathroom shining into the bedroom), with Ms. Mitchell passing in and out of consciousness. Ms. Mitchell was white, while her assailant was African-American, as was Johnson. The police reported finding a single African-American pubic hair on Ms. Mitchell’s body, a hair that police forensics examiners twice concluded could not have been Johnson’s.
Ms. Mitchell had, at the request of the police, attended a preliminary hearing on another rape charge against Johnson, watching as Johnson was there identified in open court as a rapist. The two rapes were so similar that the police believed that the same man had committed both crimes. Yet Johnson was later acquitted of the second rape, with that victim’s father actually congratulating Johnson because, after hearing the evidence, the father believed that Johnson was innocent of the crime.
When Ms. Mitchell identified Johnson at the trial that would eventually lead to his conviction, Ms. Mitchell claimed at one point that she was so upset at the lineup that she purposely identified the wrong man. She also changed her story, now saying at trial that her assailant “might have had a beard.” At another point, she said, “I just wanted to pick someone out [of the lineup] and get out of there.” Johnson offered alibi witnesses to further challenge the victim’s testimony. Nevertheless, the jury convicted the entirely innocent Johnson.
When this error was finally brought to light, the prosecutors’ office announced that too much time had passed to determine who the real rapist was. Forensics sciences professor Greg Hampikian later explained: “The DA has more pressing needs than to reinvestigate a sixteen-year-old case, especially without an available victim; meanwhile, someone has gotten away with rape.”2
Although there were numerous likely causes of Calvin Johnson’s wrongful conviction, flawed eyewitness identification was a chief contributor. Cross-racial identifications, like that made by Ms. Mitchell, the research shows, are less trustworthy than intra-racial ones; the victim had little opportunity to observe her assailant; she misidentified someone as her attacker at the lineup; and her testimony was tainted by her attendance at a hearing in another related case, all of which happened in the face of forensics evidence excluding Johnson as a suspect.3 Johnson’s conviction starkly illustrates how entirely innocent persons can be convicted when condemned by confident eyewitnesses in good faith fingering the wrong man.
The reliability of eyewitness identification is frequently questionable, as this Report will explain, even under circumstances in which the police do a much better job than they did with Calvin Johnson. Nor is Johnson’s case unusual. Numerous high-profile cases of exonerations where the innocent were convicted based substantially upon inaccurate eyewitness testimony have made their way into the media.4
The most notorious of the recent cases was that of Anthony Porter, who was once but a few days from execution and whose experience eventually led to a complete re-examination of the death row process in Illinois.5 Other notorious cases have been the subject of recent best-selling or well-received books.6 Perjured or compelled eyewitness testimony is part of the problem and is addressed in a related paper.7 The subject of this Report, however, is mistaken eyewitness testimony, and its status has been concisely summarized by award-winning journalist Stanley Cohen, who notes that many criminal cases commonly include the sorts of factors that wrongly took away Calvin Johnson’s freedom:
It is difficult to counter [a] mistaken identification offered in good faith by a witness who actually saw the accused. But even when the sole intent of the witness is to abet the judicial process, eyewitness accounts have been found to be generally unreliable. The original identification is often made under unfavorable conditions; the witness was likely to be a good distance away from the accused who was possibly shrouded in darkness; the glimpse of a suspect was likely a fleeting one, perhaps no more than a second or two; observations made in extreme circumstances, when adrenaline is running high, tend to be untrustworthy. When a defendant is convicted solely on the basis of such testimony, the possibility of error is exceptionally high.8
Cohen’s point is not to suggest that eyewitnesses are routinely wrong - - an extreme position that would flatly require exclusion of most such testimony from trial.9 Rather, Cohen apparently argues that the risk of error is so high that safeguards are needed to minimize that risk.10 The state of the research into the causes of, and cures for, eyewitness error is luckily sufficiently advanced that there is widespread agreement on some ways that we can do better now.11 In other areas, there is a dispute about whether the research has gone far enough to justify implementing certain new procedures without more data.12 This Report summarizes the state of, and lessons learned from, that research. The Report concludes that the research unequivocally supports: (1) using “double-blind” procedures in which no one involved in administering a lineup or photospread knows who is the suspect; (2) carefully instructing eyewitnesses not to assume that the right person is in the line or spread; (3) increasing the number of “foils” in the line and selecting them to match the particular eyewitness’s description of the perpetrator; (4) the witness’s reciting in her own words how confident she was in her selection; and, whenever practicable, (5) videotaping or digitally video recording a lineup. The Report further concludes that powerful research mandates wider use of special jury instructions and expert testimony on eyewitness identification problems to assist factfinders in fairly evaluating the evidence in appropriate cases. However, concerns about the maturity of the research and its dependence on simulations rather than fieldwork caution against a too-ready embrace of one new procedure, “sequential” lineups or photospreads, in which foils and the suspect are presented to the witness one-at-a-time instead of, as is currently done, in a single simultaneous presentation of all the participants.13 This Report does recommend, however, that the accuracy and practicability of the promising sequential techniques should be tested in comparative field studies in which some police districts use the new method while others do not, an approach similar to that recently implemented in Illinois by statute.14 Greater detail about these proposals is contained in the Resolution on Eyewitness Identification attached to this Report.15
Part II of this Report examines the causes of eyewitness error, while Part III summarizes the data relevant to our suggested improvements for conducting lineups and photospreads. Part IV explores the data on ways to enhance the jury’s ability better to gauge the quality of eyewitness testimony, with Part V summarizing other reform efforts and stating this Report’s conclusions.
II. The Causes of Eyewitness Error
A. Factors Affecting Identification Accuracy
The sorts of factors that can lead eyewitnesses into or out of sin are routinely grouped into five categories, specifically, those concerning witness characteristics, perpetrator characteristics, the nature of the event (the crime) itself, post event experiences, and witnessing or testifying factors:
1. Witness Characteristics: Neither the eyewitness’s sex, race, nor ethnicity, nor his intelligence (if within normal range), belief in having strong face-recognition skills, personality, or expectation of a future recall or recognition test have any influence on his ability accurately to identify the perpetrator. However, very young children do poorer than older ones or adults at recognizing strangers and are more susceptible to suggestion, while the elderly may have information – recall and face-recognition - - disadvantages. Witnesses intoxicated at either the time of the crime or during a later interview respectively have greater encoding and accurate recall problems.16
2. Perpetrator Characteristics: Perpetrators with distinctive appearances, such as unusual hairstyles, tattoos, or scars, are more easily recognized than are the less distinctive. Cross-racial identifications are generally inferior to within-race identifications.17
3. Event Factors: The longer the crime, the more time effectively to encode information, thus enhancing memory. Visible weapons (“weapons focus”), however, draw a witness’s attention to, for example, the gun or knife, thus reducing accuracy in describing people, things, or events. Moderate levels of stress-induced physiological arousal enhance memory performance but low or high arousal levels harm performance.18
4. Post-Event Factors: The greatest memory decline occurs shortly after the crime, but memory degradation continues as more time passes. If an eyewitness commits to an identification of a mug shot, the witness is likely to identify the same person at later photo arrays, lineups, or trials, whether or not the suspect is the perpetrator. “Unconscious transference” is also a significant problem “in which an eyewitness is familiar with the suspect from some event other than the crime (perhaps, for example, because both occasionally use the same subway station), does not recall why he knows the suspect, and therefore assumes that he knows the suspect because the suspect is the perpetrator.”19
5. Testimonial Factors: These factors concern the relationship between the quality of eyewitness testimony and the accuracy of identification. Counter-intuitively, a mismatch between an eyewitness’s description of a perpetrator and the appearance of the suspect is often not an appropriate reason to doubt the witness’s accuracy. This is so because of the difference between ”recall” - - retrieving information from memory - - and “recognition,” simply recognizing the right answer when someone else presents it to you. Research reveals that the quality of the recall process of describing the perpetrator is only weakly related to the accuracy of the recognition process of identifying a lineup or photospread suspect. Similarly, inconsistency among multiple perpetrator descriptions given by a single witness can be caused by variation in interview methods, interviewer expectations, or other factors, but is, in any event, not a good predictor of identification accuracy.
Also counter-intuitively, there is a weak association between the eyewitness’s confidence in the accuracy of an identification and its true accuracy. Confidence is also malleable and can be raised or lowered by post-crime events such as investigating officer feedback that the witness “picked the right man.”20
Countering these factors’ influence can be done at two separate stages: the input stage in which identification events are implemented by law enforcement and the processing stage in which the judicial system must evaluate the accuracy of identifications resulting during input.
III. Improving Inputs: Police Procedures and Human Memory
A. The Victim or Eyewitness’s Initial Report of a Crime and Pre-Identification Interviews
Suggestion can inadvertently be introduced by the police during their first contact with a witness, such as a 911 call. Non-leading, open-ended questions; a thorough effort to obtain complete information; and careful record-keeping (ideally audio or video records, whenever possible) are among the suggestions made by research psychologists for minimizing the influence of the police on witness memory and for later accurate reporting of the witness’s memory as it existed at the time of the contact with the police.
B. The Lineup
The main goals for improving lineup accuracy are reducing potential sources of suggestion and the influence of relative judgment processes. Research suggests that a substantial amount of guessing goes on by eyewitnesses in lineups. Sometimes guessing results in accidentally identifying a guilty party, sometimes in fingering the innocent, as defined in simulation and archival studies.21 The potentially pernicious influence of guessing must, therefore, also be controlled. Here are some of the various techniques for accomplishing these goals:
1. Sequential Lineups: The usual lineup procedure is to present all suspects to the witness simultaneously in a line.22 However, this process encourages relative judgments, that is, choosing the person who, among those in the line, looks most like the perpetrator.23 With sequential lineups, the witness views one lineup participant at a time and is not told how many he will see.24 As each participant is presented, the eyewitness states whether or not it is the perpetrator.25 The witness is thus encouraged to compare the individual participant’s face to the witness’ recollection of the perpetrator’s face rather than also comparing the participants’ faces to one another in a quest for the “best match.” Once an identification is made in a sequential procedure, the procedure stops.26
There is near uniform agreement in all the published literature that the sequential procedure “produces a lower rate of mistaken identifications when the perpetrator is absent….”27 This conclusion was reaffirmed in a recent “meta-analysis” of studies conducted around the world using a variety of methodologies.28 This meta-analysis concluded that false identifications were twice as likely in target-absent arrays using simultaneous presentation than when using sequential presentation.
The vast majority of researchers also conclude that sequential methods result in “little loss of accuracy when the perpetrator is present.”29 However, what constitutes a “little loss” is debatable. Thus the same “meta-analysis” noted above reported a 15 percent loss in correct identifications by foregoing simultaneous methods in favor of sequential ones.30 Professor Steven Penrod suggests that this loss may be attributable, however, largely to former “lucky guessers” now making no choice whatsoever, thus constituting no real loss at all. Sequential methods arguably function best, however, only in conjunction with the “blind lineup” procedure, in which no one involved in administering the lineup knows who is the suspect, a procedure about which there is no scientific dispute and is also of critical importance in administering accurate simultaneous lineups. These methods may face resistance in the field because they differ so much from the old ones and are not self-evidently superior based on officers’ everyday commonsense.31 Nevertheless, some police departments in the United States are already making tentative efforts toward adopting sequential methods.32
Justice Robert Kreindler also ordered a sequential lineup in a recent case in which he concluded that the scientific community was “unanimous in finding that sequential lineups are fairer and result in a more accurate identification.”33 Justice Kreindler further noted that he found not “a single scientific article criticizing the sequential lineup or criticizing the scientific method used by psychologists in their experiment.”34
Justice Kreindler was not, however, entirely correct. There is a growing dissenting view among some very well-respected social scientists that the research has not proceeded far enough to determine under what conditions, if any, a sequential lineup is to be preferred to a simultaneous lineup.35 Moreover, say some researchers, it may be that there are factors other than the simple order of presentation that are the cause of better outcomes for sequential lineups in many experiments. Additionally, field studies have not been done to determine the practicability of sequential methods, though new technologies entering the marketplace now may substantially reduce the time and out-of-pocket costs involved. These dissenters do not argue that simultaneous lineups are the preferred method, and some seem to believe that sequential lineups will eventually be proven superior in many circumstances. Nevertheless, their current view, if accepted, suggests that the scientific evidence is insufficient to choose one method over another; therefore, either might do. To add to the knowledge base and to test the practicability of the sequential method, Illinois has by statute mandated that some police districts in cities of varying sizes uses sequential methods, while others use simultaneous methods, with careful tracking of the results and problems by social scientists or by others working under their guidance. Although the dissenters are thus far few in number, this Committee finds their critique persuasive and the Illinois approach most consistent with an effort to improve the long run accuracy of lineups and thus the chances of convicting the guilty while acquitting the innocent.
2. Lineup Size: Lineups in the United States typically involve five or six participants.36 Given the substantial evidence of eyewitness guessing, larger lineups should reduce the chances of a false positive - - of a guesser selecting the (in fact innocent) suspect focused on by the police. The math is straightforward: there is 1 in 6 chance of selecting the suspect by entirely random guessing (if no other forces are at work) in a 6 person lineup but only a 1 in 12 chance of doing so in a 12 person lineup.37
There is no magic correct number. Britain, for example, uses arrays of 9.38 The point is simply that any increase in size will help to reduce the false positive rate.39 But many researchers believe that 6 person lineups create an unacceptably high risk of error, one study concluding, for example, that in real-world 6 person lineups the likely risk of a false positive would be 10% even if most of the other recommendations to improve lineup accuracy were followed.40 This report therefore urges larger size lineups than is currently the case whenever practicable. However, given debate over the necessary lineup size, this report does not mandate a specific minimum number of foils, leaving that to the judgment of local jurisdictions in light of the teachings of science and the resources available to local departments. It is useful to note, nevertheless, that computerized databases should make it easier to have more foils in photo arrays than in live lineups so that there need not necessarily be the same required minimum number of foils in both sorts of procedures.
3. Foil Selection: Foils should be selected so that they fit the witness’s description of the culprit rather than that the foils and the suspect look like one another.41 If all foils fit the suspect description, then a witness cannot guess based on who comes closest to that witness’s description - - a relative judgment process and a reasoned guess.42 On the other hand, if every effort is made to select foils because they all look so much like the suspect rather than because they fit the suspect description, then, at some point, “the lineup would be composed of clones,” unduly interfering with recognition of a guilty suspect.43 Furthermore, there are a small number of special circumstances in which alternative foil-selection strategies make more sense.44
At the same time, the lineup must be designed to avoid the suspect’s standing out unduly from the foils. For example, if the suspect is the only one wearing clothes similar to those worn by the perpetrator during the crime,45 that would draw undue attention to that suspect.46
4. Avoiding Instruction Bias: The instructions given by the lineup administrator can significantly raise the risk of false identification, even where the biases are subtle.47 Eyewitnesses must be told that the perpetrator may not be in the lineup, that they should not therefore feel that they must make an identification, and that the person administering the lineup does not himself know which person is the suspect.48 Non-verbal cues must also be avoided by the lineup administrator. 49 Of course, where there are multiple eyewitnesses, each lineup must be conducted with one witness at a time and out of the sight of other witnesses.
5. Collecting Confidence Judgments: “A clear statement should be taken from the eyewitness at the time of the identification and before any feedback as to whether he or she identified the accurate culprit.”50 This accurately preserves the witness’s confidence level at the time the identification was made and before other influences can taint or alter the witness’s memory of how confident he was in his choice.51 Ideally, the witness should never be told whether he selected the “right man” so that his confidence is not artificially inflated by the time of trial.
6. Accurately Record the Lineup, Including Videotaping: Accurate records of a lineup procedure can help to improve later judicial and jury assessment of the quality of the lineup and the accuracy of the identification.52 Videotaping would especially enable accurate recreation of lineup circumstances.53 However, multiple cameras are likely necessary to achieve this goal most effectively; the procedure does not itself improve lineup accuracy; and videotaping can be costly in terms of time, money, and equipment. 54 Nevertheless, on balance, videotaping or digital video recording of lineups seems highly desirable, where practicable.
Show-ups involve showing a single suspect to an eyewitness and asking him to identify or reject the suspect as the perpetrator.55 There is clear evidence that show-ups are more likely to yield false identifications than properly constructed lineups.56 Show-ups hint to the witness that the police believe that “this is the man,” a highly suggestive message.57 Moreover, given no other options, it is often hard independently to judge the accuracy of the witness’s choice.58 On the other hand, there is some research suggesting that “a show-up is preferable to a poorly constructed lineup,”59 though well-constructed lineups are unquestionably the best choice.60 Furthermore, show-ups can enable the quick release of innocent persons at the crucial early stages of an investigation. 61 Many representatives of law enforcement at the recent American Judicature Society Conference on Wrongful Convictions described show-ups as common and as essential to effective law enforcement, contrary to the constitutional mandate that they be used only when “necessary.” Given these competing concerns, it is difficult, absent further research, to craft a general rule concerning when even prompt show-ups should or should not be permissible, so this Report postpones any recommendation on this subject.
Photo arrays are governed by substantially the same principles as for lineups. Thus blind and sequential spreads of adequate size, with foils selected to match eyewitness descriptions, with efforts made to avoid the suspect’s standing out, and with proper instructions from the lineup administrator, are generally advised by researchers.62 One study comparing subject responses to photos of lineups versus videotaped lineups maintains that a photo of a lineup and a photo array (a collection of photos of individuals) are very different things.63
Photo arrays are probably becoming increasingly important. As a Washington Post investigative staff writer recently explained:
Like woolen uniforms, wooden batons and six-shot revolvers, the old-fashioned lineup is a vanishing part of police work. The DC police department is the only one in the Washington area that still uses it regularly, and only a decade ago it conducted 300 lineups a year.
Police departments today are far more apt to ask victims or witnesses to identify photographs of suspects instead of the suspects themselves. Detectives can use computer programs to comb through photo databases and can quickly create an array of pictures from which a suspect can be identified at any time or place.
A live lineup is “a big hassle, compared to what we can do with what’s already on the computer,” said Capt. John Fitzgerald of the Montgomery County police.64
Some in law enforcement continue to be lineup advocates, however, noting that “lineups display a suspect’s profile, posture and other features that a simple mug shot cannot capture, all of which can aid the victim or witness in making an identification.”65 Lineups also add the dimension of voice that is missing from photospreads.66 As former United States Attorney for the District of Columbia, Joseph E. diGenova, explains: “They [the witnesses] didn’t look at a photo when the crime was committed. They looked at a person.”67
If photospread use is indeed rising relative to lineup frequency, that merely underscores the importance of using the same principles for sound identification procedures, whether done by lineup or photospread.
Caution in administering photospreads and show-ups is especially important because flawed ones can easily taint later lineup and at-trial identifications. 68
IV. Improving the Processing Stage
Once an identification has been made at a lineup, show-up or photo array, a new set of concerns must be addressed: How, if at all, can we improve factfinders’ abilities properly to evaluate the fairness and accuracy of lineup identifications. A variety of options have been suggested. Here we discuss just two that we found most promising: use of experts and revamped jury instructions.
A. Expert Testimony
There is substantial psychological research establishing that eyewitness identification and memory processes are not common knowledge69 and correspondingly not within the knowledge of most jurors.70 Jurors are likely unaware of such phenomena as weapons focus, retention intervals, and instruction bias.71 Wells and colleagues summarize matters thus:
Taken together, the survey, post diction and mock-juror experiments, and the confidence-accuracy studies converge on a worrisome set of conclusions: Jurors appear to overestimate the accuracy of identifications, fail to differentiate accurate from inaccurate eyewitnesses - - because they rely so heavily on witness confidence, which is relatively nondiagnostic - - and are generally insensitive to other factors that influence identification accuracy. Furthermore, this picture is even gloomier when one considers that eyewitness confidence proves to be highly malleable.72
Expert testimony is thus needed to educate jurors. Moreover, much of such expert testimony, if properly presented by a qualified witness, should logically survive scrutiny under Daubert and other potential hurdles to admissibility. Nevertheless, the courts are divided on the question. Some state and federal courts have found such expert testimony inadmissible because it concerns knowledge within jurors’ everyday understanding73 or because cross-examination is deemed adequate to reveal deficiencies in eyewitness testimony.74 Other courts leave it within the discretion of the trial judge to admit or exclude expert testimony, such trial-judge-deference apparently being the predominate view among both federal and state courts.75 Still other jurisdictions allow expert testimony only under specific circumstances.76
In a recent study, researchers found that expert testimony enhances jurors’ sensitivity to the factors that influence identification accuracy without overly increasing juror skepticism of the witness’s identification.77 These conclusions are largely consistent with numerous earlier trial simulation studies concluding that expert testimony does indeed increase juror awareness of factors affecting eyewitness accuracy, assists them in evaluating eyewitness testimony effectiveness, and reduces conviction rates.78
D. Jury Instructions
Some courts give special instructions about eyewitness testimony, often as a replacement for expert testimony. The earliest and most well-known of these is the Telfaire instruction.79 The Telfaire instruction, however, omits many important factors and can be misleading, for example, by suggesting that witness confidence is a good predictor of eyewitness accuracy when the research shows otherwise.80 The “Revised Telfaire Instruction,” proposed by Professor Edith Greene, has been found in her own jury research to be more effective than the original Telfaire instruction.81 This revision is simpler, more accurately conveys the lessons of the research, and explains the processes by which various factors affect eyewitness memory.82 But, in the view of one of the leading experts in the area, California’s Wright instruction does an even better more thorough job.83 The Wright instruction can readily be updated with new research and easily tailored to the factors important to a particular case.84
In most jurisdictions, the question whether a jury instruction is proper is left to the trial court’s discretion. Some courts find an instruction necessary where the evidence raises serious doubts as to the accuracy of an identification.85 Omission of an instruction is usually found to be an abuse of discretion only where identity is the central issue, there is no corroborating evidence, and the circumstances raise doubts about the reliability of the defendant’s identification.86 Some jurisdictions, by contrast, hold as a general proposition that special instructions are unnecessary.87 Still others consider it adequate to use only general instructions about judging the credibility of any witnesses,88 or special instructions to be unnecessary where identification testimony has been corroborated by other evidence.89 Jury instructions in other areas generally have not, however, had a good record of sufficiently altering jury reasoning processes as intended so that the efficacy of a more specific instruction, at least absent other reforms, such as use of expert testimony on the subject and improvement of the quality of identification procedures themselves, is in doubt.90 Jury instructions about eyewitness identification accuracy tend to instruct jurors on general principles, such as “unconscious transference,” that are relevant to the facts involved in a particular case but do not more specifically instruct the jury about how those principles apply to the case at hand.91 The instructions are necessarily general because even experts cannot reliably opine after the fact that a particular identification was reliable. Generality also avoids “usurping the jury’s role” as factfinder. 92 Nevertheless, some prosecutors object to specific instructions precisely because, in their view, generalities tell the jury nothing about the particular case.93 Some judges might also hesitate to give instructions not supported by expert testimony at the particular trial.94 On the other hand, jury instructions on the areas in which there is widespread scientific consensus can save time, much in the way that operation of the doctrine of judicial notice does.95 Moreover, there is reason to believe that well-crafted jury instructions in this area can at least have some positive impact, however modest, on creating a more-informed jury better able to reach a rational decision.96
V. Other Efforts at Reform and Conclusions
Well-known efforts at reform have been undertaken by the National Institute for Justice (NIJ), the New Jersey State Police, Former Illinois Governor Ryan’s Commission on Capital Punishment, and North Carolina’s Actual Innocence Commission.97 Here we offer the briefest summaries of those approaches and a comparison among them.
Three of these four organization’s reports mandate double-blind lineups, with the fourth (DOJ) acknowledging that double-blind is the best practice. These same three reports mandate sequential lineups, with DOJ acknowledging their likely advantages but questioning their practicability absent field studies. The Illinois state legislature, as noted above, rejected the Ryan Commission’s mandating of sequential procedures, instead adopting a pilot study requiring three police districts, each in police departments in municipalities of various sizes, to use sequential procedures and to evaluate their effectiveness and practicability using mechanisms “consistent with the most objective scientific research methodology.” This Council recommends a conservative approach similar to that adopted by Illinois. All these reports mandate a specific minimum number of lineup or photospread foils, but these numbers, while minima and not maxima, are still smaller than the best practices suggested by the science. However, rather than specify a precise number, and given resource concerns, this Council has simply recommended embracing the principle that there should, where practicable, be a sufficient number of foils to minimize the risk of error by guessing, an approach that makes larger numbers of foils aspirational, but not mandatory, and that allows for local variation and change as the teachings of science improve. Most of the remaining recommendations by this Council concerning lineup and photospread procedure are largely inspired by similar variations adopted in one or another of these reports. The best practices recommended in this Report – including blind lineups, experimental use of sequential methods, enhanced number of foils, expert testimony, and special jury instructions -- are fully supported by the scientific data and will go far toward improving identification procedure accuracy.
Norman Maleng, Chair, Criminal Justice Section, August 2004