Lineups and Understanding the Nature of Eyewitness Testimony

It has been my experience over 42 years – that eyewitness testimony – long believed and cherished by juries in determining and resolving the guilt or innocence of the accused – is the weakest kind of evidence. In Colorado, judges routinely deny expert testimony on behalf of the defendant on the fallacies and inherent weaknesses of eyewitness testimony.

Eyewitness expert testimony informs a jury about psychological processes and accuracy related variables in eyewitness testimony.

One Recent Study Favors the Admission of Experts Who Address the Weaknesses of Eyewitness Testimony

Studies of trial dynamics and reactions to eyewitnesses suggest a sizable risk of inordinate eyewitness impact, creating sizable risk of conviction on the basis of mistaken identifications.

Trial simulations examining eyewitness expert testimony indicate it promotes modest, appropriate increases in skepticism about eyewitnesses, even when the expert gives a general overview of research and admits to limitations. The psychological and legal professions should develop responsible guidelines for use of expert testimony in court.

The New Research on the Weaknesses of Eyewitness Testimony – A National Study

Eyewitness identification is used to make cases for the apprehension and prosecution of alleged criminals. Eyewitness evidence is a critical tool for defending against can also be an important tool for exonerating innocent suspects.

Groundbreaking research on eyewitness memory over the past three decades, as well as increasing attention to the problems in the cases of wrongfully convicted individuals, has brought the fallibility of eyewitness memory to the fore.

Eyewitness misidentification is widely recognized as the leading cause of wrongful conviction in the United States, accounting for more wrongful convictions than all other causes combined. Since 1989, DNA evidence has been used to exonerate nearly 200 individuals who were wrongfully convicted. Of those, approximately 75 percent were convicted on evidence that included inaccurate and faulty eyewitness identifications. In some cases, these innocent individuals were misidentified by more than one eyewitness.

In the vast majority of criminal cases, however, DNA or other biological evidence is not available to establish guilt or innocence. Given the persuasive nature of eyewitness evidence, as well as the inherent danger of misidentifications— both in convicting the innocent and allowing the true perpetrator to go free—it becomes imperative that we take stock of the procedures within the control of the criminal justice system that contribute to these problems in order to ensure that the most reliable evidence possible makes it into a courtroom and before a jury.

A number of challenges emerge in pursuit of a more accurate protocol, none more prevalent than an historical lack of communication between scientists and law enforcement. Decades of empirical research have proven that a number of small changes to identification procedures can help improve the accuracy and reliability of eyewitness identifications, and help ensure that the highest quality of eyewitness evidence is collected.

What’s more, when put to the test in numerous jurisdictions throughout the country, these reforms have met with real-life success. Thus, it may seem surprising that these reforms have not been implemented in police districts across the board. While much of the research has been extensively documented and peer-reviewed within the scientific community, and the recommendations for reform are widely accepted by experts in the field, these reforms were initially discussed and developed outside the realm of law enforcement.

Best Practices About Eyewitness Identification Procedures are Known

Starting in the late nineties, however, leading researchers joined with law enforcement and legal practitioners to bridge the gap and comprehensively address eyewitness identification issues at the intersection of the two fields. As a result, guidelines and best practices for law enforcement were developed with the science in mind.

In October 1999, the Department of Justice released a comprehensive guide for law enforcement on procedures for obtaining more accurate eyewitness evidence. However, there is no current national program or federal agency responsible for educating local departments about these reforms—or in assisting with their practical implementation.

Moreover, as reforms are implemented on a jurisdiction- by-jurisdiction basis in some states, there continues to be little opportunity for sharing information and perhaps even less incentive, given the already overloaded criminal caseloads of police, prosecutors and defenders, and the lack of leadership from the courts or legislature on the issue.

This policy review has been designed to facilitate communication among local law enforcement agencies, policymakers, and others regarding the best practices and methods for enhancing the evidentiary value of correct identifications and at the same time reducing the risk of erroneous identifications. By presenting many of the successful methods employed in local jurisdictions, as well as the science behind them, we hope to create a dialogue around recommendations that will enhance the quality of evidence relied upon in criminal trials, as well as confidence in our system of justice.

Introduction A measure of Fairness and Accuracy in the Criminal Justice System

Eyewitness misidentification is widely recognized as the leading cause of wrongful conviction in the U.S., accounting for more wrongful convictions than all other causes combined.

A handful of specific improvements have emerged as pragmatic strategies for minimizing eyewitness error. While modernizing identification procedures to incorporate advances in eyewitness memory science requires retooling long-standing lineup methods engrained in police culture, the substantial benefits of implementing the protocol are leading more jurisdictions to update their procedures to catch up with the science.

Because eyewitness evidence, much like trace evidence, is susceptible to contamination, some eyewitness identification procedures actually increase the risk of false identification. By improving these procedures in subtle ways, the actual quality of eyewitness evidence can be improved.

The following recommendations reflect the consensus in the scientific community — confirmed by successful implementation in numerous jurisdictions — as to the procedural changes that will enable law enforcement to extract the most reliable evidence from eyewitnesses for use in a criminal investigation. These practical changes to the identification process help increase the likelihood of identifying the true culprit while enhancing protections for innocent people accused of crimes.

These reforms are equally effective for photographic lineups and live lineups.

Procedures to Be Used In Lineups

Prior to presenting the lineup members, the eyewitness should be instructed that the perpetrator may or may not be included in the lineup, and that she should not feel compelled to make an identification.

Cautionary instructions respond, in part, to the tendency of witnesses to make a relative judgment by removing some of the pressure on the eyewitness to choose a suspect when the culprit may not be in the lineup.

Effective Use of Fillers

Only one suspect should appear in each lineup. In addition, at least five fillers should be included in a photo lineup, and at least four fillers in a live lineup. The fillers should resemble the witness’ description of the perpetrator, and the suspect should not unduly stand out.

Fillers, if chosen correctly, allow authorities to judge the reliability of an eyewitness. The effective use of fillers is critical to ensuring that an innocent individual is not identified simply because of the composition of the lineup.

Documentation

The identification procedure should be carefully documented. Documentation includes preservation of photos in a photo array or photographs taken of a live lineup, recording all individuals present at the lineup, documentation of the witness’ statements regarding the lineup members during the procedure, and, if an identification is made, documentation of the witness’ degree of confidence in the identification, in the witness’ own words, prior to any feedback from authorities.

Careful documentation of the lineup procedures, including a witness’ level of certainty that she has correctly identified the perpetrator, when taken immediately following the identification, helps the jury to assess the eyewitness evidence appropriately and minimizes the effects of reinforcing feedback that can distort the confidence level of an eyewitness between the time of the identification and the trial.

Double-Blind Administration

The person who administers the lineup should not know the identity of the suspect. This procedure prevents well-intentioned officials from giving inadvertent clues to the witness as to which person in the lineup is the police suspect.

Sequential Presentation

The lineup members should be presented to the witness “sequentially” (one at a time) rather than simultaneously (all at once). Sequential presentation should only occur, however, if the identification procedures comply fully with the double-blind administration recommendation.

Presenting the lineup members one at a time to the witness reduces the tendency for witnesses to engage in “comparison shopping.” Rather, an eyewitness must judge whether each lineup member matches her memory of the perpetrator, as opposed to making a relative judgment.

Recommendations & Solutions Getting it Right the First Time

Reliable eyewitness evidence is critical to criminal investigation and prosecution, and it plays a powerful role within the criminal justice system. The repeated discovery of misidentifications contributing to wrongful convictions, however, has prompted inquiries into the nature of eyewitness evidence used to convict criminal suspects, and the problems that arise in utilizing human memory in criminal investigations.

The scientific community has brought the knowledge built through decades of research and experiments to bear on eyewitness identification procedures.

Important lessons learned in the laboratory, and in the decades of research devoted to eyewitness memory science, are of enormous value in the legal and law enforcement communities. This substantial body of research has revealed that several natural psychological phenomena can undermine the accuracy of eyewitness identification, and that these psychological factors, left unchecked, can lead to unreliable evidence being presented in the courtroom.

Lineups as Experiments

Just as trace physical evidence (such as DNA or fingerprints) can be contaminated if it is not collected precisely and carefully, so too eyewitness evidence can be spoiled if it is gathered in ways that do not properly control for known sources of error.

As some researchers have described, a lineup is essentially an experiment designed to test a hypothesis: whether the suspect matches the witness’ memory of the perpetrator. Like scientific experiments, careful controls must be put in place to ensure accuracy and prevent the witness’ memory from being contaminated or skewed.

Relative Judgment

Relative judgment refers to the natural tendency of a witness to consider lineup participants in comparison with one another, as opposed to a more direct comparison of each lineup member with the witness’ memory of the culprit. A witness viewing a lineup will thus tend to identify the person who looks most like the perpetrator in comparison to the other members in the lineup.

While, at face value, this process seems unproblematic, it can actually lead to inaccurate and unreliable identifications under certain conditions — namely, when the police suspect is innocent.

The purpose of a lineup is to differentiate innocent suspects from those who actually committed crimes using an eyewitness’ memory of an event. Thus, when conducting a lineup, law enforcement officers do not know if a suspect included in a lineup is, in fact, the true perpetrator or simply an innocent person suspected of a crime. If the lineup is full of innocent people (an innocent suspect and a group of innocent fillers), however, relative judgment would mean that an innocent person may be identified, because it is likely that there will always be someone in the lineup who looks more like the person who committed the crime than the other members of the lineup.

Sometimes this person will be a filler, and a witness identification will be dismissed. But sometimes an innocent suspect will be the victim of this tendency toward “comparison shopping,” because the witness is always making a relative judgment — the witness is always picking the person who looks closest to the culprit relative to the other lineup members, even if the lineup is full of innocent people.

Psychological Factors Preventing Unreliable Evidence in the Courtroom

Several natural psychological phenomena can undermine the accuracy of eyewitness identification, and these psychological factors, left unchecked, can lead to unreliable evidence being presented in the courtroom.

Scientific treatments of eyewitness evidence began over 100 years ago, most notably with Harvard Professor Hugo Munsterberg’s 1908 book, On the Witness Stand. While Munsterberg established that eyewitness evidence was much more fallible than previously thought, his research did not show a way forward. Based on the science of the day, the legal system had no capacity for dealing with these mistakes, and the system could not sort the mistakes from the true identifications. The science, at first, only documented the problem, but it could provide no solutions.

In the late 1970’s, however, eyewitness memory scientists began to zero-in on the particular sources of eyewitness error and test revised identification procedures that reduced the risk of mistakes. The guiding principle of this new research was that we must do all we can to ensure good quality evidence on the front end of the process, rather than trying to second guess identifications after the fact. For the research on eyewitness fallibility to be useful, it had to be applied to the criminal justice system in a way that allowed the system to prevent or reduce future mistakes.

Scientists thus focused on the ways that the system of collecting eyewitness evidence could itself cause mistakes, in hopes that these mistakes could be prevented before they occurred.

The past three decades of eyewitness research and discussion have coalesced around this purpose preventing false identifications with research-based improvements to the system. Largely, these improvements focus on controlling the suggestiveness of the lineup procedures themselves.

A discussion of the science behind these improvements follows.

The Science

Demanding changes in eyewitness identification procedures witnesses who saw the same event will often pick someone out of a lineup when the culprit is removed. In other words, regardless of whether a culprit is in a lineup, witnesses tend to pick the person who looks closest to the culprit, even when the culprit is not present.

As leading researchers have noted, “The problem with the relative judgment process, therefore, is that it includes no mechanism for deciding that the culprit is none of the people in the lineup.”

Malleability of Witness Certainty

Traditionally, a witness’ self-reported degree of certainty in an identification was considered a good indicator of accuracy. Unfortunately, a great deal of research in recent decades has proven this intuitive assumption false. The level of certainty a witness expresses in her eyewitness testimony does not necessarily correlate with the level of accuracy of the identification. An eyewitness’ confidence that she has identified the culprit can fluctuate as a result of factors that occur after the identification and have little to do with memory. This is what is referred to as confidence malleability.

For example, experiments have been conducted in which witnesses were shown a staged crime and asked to identify the culprit from a lineup. The lineup they were shown, however, did not contain the culprit. After the witnesses unknowingly made false identifications, they were then asked their level of confidence. Before doing so, however, some of the witnesses were given various types of reinforcing feedback.

Those witnesses who received some confirmation of their false identification, whether the information that a co-witness identified the same individual or some other confirming feedback, were far more confident in their identifications than other witnesses who were given no feedback — despite having given false identifications. These witnesses also distorted and exaggerated certain details, such as how good their view was, how much of an opportunity they had to view the culprit, etc.

Our new and better understanding of the influence feedback plays on a witness’ self-described level of confidence strongly suggests that measures that control for this influence be adopted in our identification procedures.

When the culprit is not, in fact, present in the lineup, this perception, combined with the natural tendency to compare lineup participants and make a relative judgment, may influence an eyewitness to identify an innocent person.

Cautioning an eyewitness that the offender may or may not be in the lineup reminds witnesses that the answer may be “none of the above.”

Research has shown that this extra step, while on its face pointing out a fact that should be obvious, significantly lowers the rate of inaccurate identifications without reducing the number of true identifications.

Effective Use of Fillers

Relative judgment theory means that an eyewitness viewing a simultaneous lineup tends to make a judgment about which individual in the lineup looks most like the perpetrator relative to the other members of the lineup. This is particularly problematic when a lineup only contains innocent people (i.e., a number of fillers and an innocent suspect).

Research has shown, however, that the effective use of fillers when composing a lineup can help combat the tendency for the relative judgment process to lead to the identification of an innocent suspect.

First, ensuring that the suspect in the lineup does not stand out, or that the fillers resemble the witness’ prior description of the culprit at least as much as the suspect does, guards against the eyewitness choosing an innocent suspect simply because the suspect is the only lineup member that resembles the perpetrator.

For example, if the eyewitness describes the perpetrator as an Asian man with a mustache, and there is only one man in the lineup who is Asian and has a mustache, then the lineup is obviously suggestive, and the evidentiary value of any identification is nil. In contrast, if all of the lineup members resemble the prior description of the culprit (or all of the lineup members are Asian men with mustaches), then the eyewitness will have to rely more on comparisons to her own memory of the culprit.

In short, no lineup participant can unduly stand out for a lineup to be effective. This holds true in general, but especially with regard to features of the witness’ description of the culprit. For example, if a witness describes the perpetrator as having a particular feature such as a mustache, the lineup must be composed with all members sharing that feature.

There are certainly cases where selecting fillers is not as clear-cut. For example, if the suspect does not fit the witness’ prior description of the suspect but other evidence creates suspicion of guilt, then it may be appropriate to place that suspect in the lineup, as witness descriptions can sometimes be off the mark. If so, however, the fillers must be chosen to be similar

to the appearance of the suspect. There are methods for dealing with contingencies, but the true test of this rule is whether the suspect stands out relative to the other fillers.

In other words, if a person who is not involved in the case is given a description of the perpetrator, would she be able to pick the suspect out of the lineup? Including only one suspect in a lineup is also a fundamental safeguard against misidentification.

Lineups not only allow police to judge whether a suspect is innocent, they also allow investigators to judge the reliability of an eyewitness. If a lineup contains more than one suspect, however, its ability to test reliability is diminished. This is because it increases the likelihood that a witness would select a suspect based on a guess rather than recognition. The more choices in the lineup test that could be considered “correct” (i.e., suspect identification), the less the lineup can control for witnesses with weak memories or those who guess.

The same considerations underpin the need to include an adequate number of fillers. Doing so also reduces the likelihood of an eyewitness identifying an innocent suspect based on a guess. For example, if there is only one suspect and one filler, the likelihood that an innocent suspect will resemble the culprit more than the other lineup members is 50%. If three fillers and one suspect, the likelihood is 25%; and so on.

While there is no magic number of fillers that should be used, the science has shown that the greater the number of fillers, the greater the reliability of the procedure.

Documentation

Lineup identifications are a critical component of the investigation of criminal cases. Given the overwhelming importance of eyewitness testimony and the weight afforded to it by juries, it is essential to provide sufficient contextual information about an identification in order for factfinders to evaluate its evidentiary weight correctly.

Careful documentation of lineup procedures, where possible, means that a complete and accurate record of the methods used to obtain an identification is preserved for review. Recording the identification and the non-identification results, the dialogue between witnesses and police, and the photos themselves (or photographs of a live lineup), serves as much to protect the police from false claims of influencing a witness as it does to preserve the integrity of the evidence.

Thorough documentation has the power to put an identification beyond reproach. Scientists have shown that a number of procedures within the system actually contribute to misidentifications. Complete documentation allows any suggestiveness in the procedure to be considered by judges and juries in deciding how to weigh the evidence and, when reliable procedures are used, it strengthens the evidentiary value of an identification.

A critical component of appropriate documentation is recording an eyewitness’ statement of confidence (or self-assessment of certainty) immediately after an identification. This guards against the confidence malleability problem — when an eyewitness’ confidence that she has identified the culprit fluctuates as a result of factors that occur after the identification.

To document a witness’ confidence, the witness is asked her level of certainty that the person being identified is the true perpetrator prior to receiving any feedback from authorities or other witnesses. The witness’ confidence level should be recorded in her own words in order to allow judges and juries to evaluate eyewitness testimony in an informed manner.

Studies have shown that information provided to an eyewitness after an identification can influence the witness’ level of confidence, and thus skew a juror’s assessment of the accuracy of the identification. For example, if an eyewitness makes an identification of a suspect, and that same witness later learns that the person identified also has a criminal record, the witness’ confidence level may become artificially inflated. Confirmatory feedback oftentimes occurs without the knowledge or intent of investigators in the case or even the eyewitness, and if a confidence statement is not taken directly after the identification, the window of opportunity for protecting the integrity of the identification evidence as an indicator of accuracy is lost.

Double-Blind Administration

The “double-blind” rule applies the scientific method to lineups, and is rooted in a general strategy for ensuring the objectivity of data collection and interpretation. The purpose of keeping the administrator “blind” as to which person in the lineup is the suspect is to prevent the administrator from unintentionally influencing the results through inadvertent cueing of the witness toward the suspect. A doubleblind protocol also eliminates the problem of interpreting ambiguous witness comments and other behaviors through the lens of the theory that the suspect in a lineup is guilty.

Double-blind protocols are familiar to many in the context of pharmaceutical studies to test a new drug. Not only is the patient unaware of whether she received the drug or the placebo, but the doctor who examines the patient during the study is also “blind” to this fact. If the tester knew that the patient had taken the placebo, the tester might unknowingly skew the examination as a result.

Double-Blind

A critical component of appropriate documentation is recording an eyewitness’ statement of confidence (or self-assessment of certainty) immediately after an identification. protocols are standard practice in such contexts not because we distrust the integrity of the medical and scientific professionals involved, but because we understand the risk of natural human psychological factors that can undermine objectivity. We control for such factors because there is much at stake. Similarly, if a lineup administrator knows which member of the lineup is the suspect, she might unintentionally influence the identification through verbal or non-verbal cues. A cue can be a statement to the witness or even an administrator’s posture or facial expression. Verbal and non-verbal cues are examples of suggestive that can suggest to the witness where a suspect is in the lineup.

Verbal and non-verbal cues can also influence or inflate the certainty of the witness. Given that eyewitness confidence is weighed heavily in the legal system, and given that it has been shown to be highly malleable and particularly susceptible to feedback, it is important to design lineup procedures that eliminate the risk of over-inflating confidence through unintentional suggestion.

Sequential Presentation

Relative judgment theory also serves as the basis for the sequential presentation recommendation. Traditionally, eyewitnesses are shown a lineup or photo array in which the lineup members are presented as a group. This type of presentation actually encourages an eyewitness to “comparison shop” or to judge the lineup members against each other through a process of elimination.

On the other hand, sequential presentation, first articulated by researchers in the 1980s, is a process where the witness is shown the lineup members one by one and asked to decide if the lineup member presented is the perpetrator. By forcing witnesses to consider the lineup members individually, sequential presentation favors a direct and independent assessment of whether each lineup member matches a witness’ actual memory of the perpetrator.

Researchers have shown that the sequential presentation, if implemented in tandem with the doubleblind procedure, results in fewer false identifications. It is important to note that if the administrator is not “blind,” however, the sequential procedure can actually produce higher rates of false identifications, as witnesses may be more susceptible to unintentional feedback from the administrator when considering one lineup member at a time.

While eyewitnesses have been shown to make fewer choices when viewing a sequential lineup, the research suggests that this is due, in part, to fewer guesses on the part of eyewitnesses with a weak memory of the culprit. A comparison of the accurate and mistaken identifications also suggests that a sequential presentation yields a higher probability that a suspect, if identified using this procedure, is in fact the culprit.

In short, the sequential lineup creates a higher threshold for identification by reducing the influence of the tendency to make relative judgments. As a result, the evidentiary value of identifications gained through sequential lineups is much higher, at the cost of some identifications based on weaker witness memory or witness guesses.

Taking this research into the field has shown a generally positive effect. In a pilot project on the sequential procedures in Hennepin County, Minnesota, (Minneapolis) for example, eyewitnesses picked fewer fillers. Such a decrease in known errors confirmed the value of sequential procedures for officials in that jurisdiction. In addition, New Jersey implemented sequential procedures statewide in 2000. A pilot project conducted by the Chicago Police Department in 2006, however, raised concerns that sequential doubleblind lineups were less accurate than conventional methods. Nonetheless, the Chicago study was criticized as having design problems that undermined the study’s ability to yield reliable comparisons.

Researchers are currently pairing with other jurisdictions to add to the credible literature on the topic. While some questions have been raised about the value of sequential presentation, on balance, most experts believe that it has proven to be superior in both experimental research and in the field. The “double-blind” rule applies the scientific method to lineups, and is rooted in a general strategy for ensuring the objectivity of data collection and interpretation.

The benefits of improved eyewitness identification procedures are perhaps best conceived of in terms of the avoided costs. When an eyewitness mistakenly identifies the wrong individual, the costs to public safety are great. Scarce resources in the criminal justice system are misdirected toward investigating and perhaps even trying an innocent person, and not toward convicting and jailing criminals. “The cost to society of inaccurate eyewitness identifications is twofold,” notes psychologist Rod Lindsay of Queens University in Kingston, Ontario. “It’s a double error. Not only are you convicting the innocent — or at least putting them through the process of having to get out of the situation — but the guilty are still out there doing the crimes.”

---