Lineup Members Research Articles

Parallelism in eyewitness and mock witness identifications

The fairness of eyewitness identification lineups is assessed through the use of a series of fairness measures. All the measures proposed to date are based on the lineup choices of ‘mock witnesses’. Mock witnesses are persons who have not previously seen the offender but who have been given information about him, usually a verbal description. In a fair lineup mock witness choices should be distributed approximately equally across all lineup members. Simultaneous presentation of lineup members is the most common form of identification procedure, although recent evidence suggests that sequential presentation decreases the proportion of false identifications while maintaining the proportion of correct identifications. The present study used lineups from a recent court case to explore the limits of the mock witness paradigm using different photographs of the defendant, variations in the verbal description, and two presentation modes. Results indicated that suspect choices and the distribution of foil choices were sensitive to changes in the suspect photograph and alternate forms of the verbal description. The patterns of mock witness choices were similar across presentation modes. Our findings raise questions concerning the limits of the validity of the mock witness paradigm. Specifically, they illuminate the cognitive process differences between eyewitnesses and mock witnesses, and the information available for the identification decision. Copyright © 1999 John Wiley & Sons, Ltd.

Measuring lineup fairness

SUMMARY The fairness of line-ups and photospreads is a traditional concern of research and policy development in the area of eyewitness identification. Quantification of fairness, the construction of fairness indices, and the development of evaluation procedures started in the 1970s and continues to this day. This paper reviews the historical development of the field as an introduction to the articles that follow. The entire set of articles addresses current questions and raises new issues of measuring the fairness of identification procedures. Copyright # 1999 John Wiley & Sons, Ltd. The literature on the measurement of line-up fairness began with the report by Doob and Kirschenbaum (1973) of their attempt to measure the bias contained in an identification process brought to their attention by a defence attorney. The verbal description of the oAender was extremely vague. She stated only that the criminal was ‘rather good looking’. The possibility existed that the witness had simply chosen the best-looking person from the line-up, whether he was the oAender or not. To test this conjecture, Doob and Kirschenbaum gave the description of the criminal and a photograph of the line-up to a set of persons who had never seen the criminal. They were asked to choose the suspect from the line-up. If these ‘mock witnesses’ (Wells et al., 1979) were able to identify the suspect, never having seen him but solely on the basis of the verbal description, the line-up could be considered biased in the sense that the alternative line-up members were not full (eAective, functional) alternatives to the suspect. The underlying assumption of Doob and Kirschenbaum’s method was that the alternatives (distractors, fillers, or foils) in a line-up ought to protect innocent suspects from identifications based solely on their match to the witness’s description of the criminal (recall). Instead, the witness should have to rely on the presumably more informative data provided by recognition of the face as a match of the suspect’s appearance to the witness’ memory image derived from having witnessed the original event (Wells, 1993). Thus line-up foils ought to be plausible choice alternatives: the suspect should not stand out in comparison to them. The reductio ad absurdum would be to place a suspect described as a European‐American in a line-up with an African‐American, an Asian‐American, an elephant, and a refrigerator. Doob and Kirschenbaum’s (1973) procedure oAered an attractive paradigm for post hoc evaluation of fairness in line-ups and photospreads, through the use of the

Measures of lineup fairness: do they postdict identification accuracy?

Various measures have been proposed to test the fairness of lineups (Doob & Kirshenbaum, 1973; Malpass, 1981; Wells, Leippe and Ostrom, 1979). In three experiments, we examined the relationship of identification accuracy to measures of lineup bias (proportion of mock witness choices, defendant bias, functional size) and lineup size (effective size, number of acceptable lineup members). Lineup bias measures successfully postdict false positive choices, but only for some lineup procedures. Lineup size measures rarely postdict false positive decisions. Lineup fairness measures do not postdict false identifications from sequential lineups. Implications for measuring and reporting lineup fairness in court are discussed. Copyright © 1999 John Wiley & Sons, Ltd.

Elimination lineups: An improved identification procedure for child eyewitnesses.

Elimination lineup procedures were proposed that required the witness to eliminate all but I lineup member before being asked if the remaining lineup member was the criminal. Elimination lineups were designed and tested with the aim of reducing false-positive choices by child eyewitnesses (n = 587 children. 10-14 years, M = 12 years; n = 185 adults). Elimination lineups decreased false-positive responding in children without significantly reducing correct identifications. Fast elimination lineups with modified instructions emphasizing the negative consequences of identifying an innocent person and explaining how to make an absolute judgment significantly decreased children's false-positive rate to a level comparable with adults shown a simultaneous lineup. The potential benefits of elimination lineup procedures for child witnesses are discussed.

The effect of authority and social influence on eyewitness suggestibility and person recognition

This study investigated the influence of an authority figure on eyewitness identification. Participants watched a staged crime and then were administered a photo lineup by either an authority (policeman) or non-authority figure (civilian). Participants in the authority condition were more likely to choose a lineup member than those in the non-authority condition. There was no effect of authority, however, on the accuracy of the identification decisions. The lack of a deleterious effect suggests that the presence of a police officer during identification procedures does not create an unduly suggestive situation or have undesirable effects on eyewitness identification decisions. Although witnesses' choosing behaviors did not increase the rate of identification errors, the effect of the administrator's authority on choosing was reduced when unbiased instructions (vs. biased instructions) were presented to the witnesses. Thus, support was found for the use of neutral instructions during eyewitness identification procedures.

Maximizing the Utility of Eyewitness Identification Evidence

affect eyewitness accuracy and are under the control of the criminal jus tice system have been labeled sys tem variables. Examples of system variables include instructions to eye witnesses, question formats, and the lineup procedure. Eyewitness researchers use slides, videotapes, or live staged events to create eyewitness experiences for unsuspecting people. These eyewit nesses can then be tested under con ditions controlling for system vari ables to determine the best ways to maximize eyewitness accuracy. The discovery of system variables that af fect eyewitness accuracy can yield a technology that can be applied to actual criminal cases for the purpose of minimizing eyewitness error.1 Our emphasis in this review is on police lineups, with particular con cern for the system variables that contribute to accurate and false identifications. Conducting a lineup is analogous to conducting a psychology experi ment.2 In a psychology experiment, the researcher begins with a hypoth esis, selects or creates an appropri ate set of stimuli, creates a design, places in motion a carefully crafted set of procedures, records the sub ject's responses, and reinterprets the hypothesis in light of the obtained data. A police investigator is like a researcher who has a hypothesis (that the suspect is the actual cul prit), selects an appropriate set of stimuli (selects lineup members), creates a design (e.g., places the sus pect in Position 4), initiates a proce dure (e.g., gives instructions to the eyewitness), records the subject's re sponses (e.g., identification of the suspect or not), and then reconsiders the original hypothesis in light of the

How Police Officers Construct Lineups: A National Survey

In criminal investigations, considerable weight is given to eyewitness identification evidence. In some cases, like assault and robbery, this evidence may be the only kind available. Over the last two decades, considerable research has been aimed at the factors that cause identification errors. However, virtually all of this work has involved undergraduates and naive lay persons as subjects. There is little known on what police officers do in the course of their work. The present research investigates the procedures that police investigators employ when they construct live and photographic lineups. Surveys were sent to 500 U.S. police jurisdictions along with a cover letter requesting that it be completed by the person most experienced in constructing lineups; 220 were returned. On average, the respondents had 12 years experience as police officers and had constructed a mean of 329 lineups (89 live, 240 photographic). For many items, the results were consistent with those of previous laboratory research. For example, the police officers reported giving more attention to upper face features (e.g., hair) in selecting non-suspect lineup members (foils) which is consistent with research on feature saliency. However, other results indicate that the police use different procedures than those recommended in the research literature. For example, the police officers report using similarity as the major basis for selecting the nonsuspects (foils), whereas, research shows that selection of foils based exclusively on suspect appearance can produce biased lineups. In addition, there is considerable research showing that sequential lineups are more fair than simultaneous lineups, yet only 40% of the police investigators reported having used the sequential technique. These and other findings provide direction and implications for research opportunities. It is concluded that systematic investigation of actual police procedures is a more direct approach of studying ways to decrease identification errors.

Eyewitness identification and the selection of distracters for lineups.

Previous treatments of eyewitness lineups have focused exclusively on the importance of homogeneity (similarity of common features) in the physical characteristics of lineup members. This has led to some confusion about the proper way to select distracters. We argue that distracters should not be selected for their similarity to the suspect but rather for their similarity to the witness's description of the culprit. The similarity-to-suspect strategy fails to define limits to the number, type, and degree of featureal matching and falls prey to the logical extension that a good lineup is composed of the suspect's clones. Accordingly, the similarity-to-suspect strategy has no supportive logic in recognition memory theory and gives no credit to the importance of hit rates. The similarity-to-witness's-description-of-culprit criterion, on the other hand, specifies a finite and manageable set of feature requirements for distracters, articulates a role for heterogeneous features, meets all functional requirements for fairness to the suspect, has a supportive logic in recognition memory theory, preserves hit rates, and is not subject to the clone argument.

Standards for Evaluating the Fairness of Photograph Lineups

In theory a lineup is fair, from the perspective of the suspect, to the extent that the suspect is not distinctive in appearance from the other lineup members (foils) and all of the foils are reasonably similar in general appearance to the suspect. Lineup fairness can be divided into two dimensions-size and bias. Several proposed methods for empirically evaluating lineup fairness were assessed in terms of their sensitivity to absolute standards of lineup fairness and their discriminability, the extent that they differentiated between lineups designed to be fair and those constructed via less rigorous selection procedures. One of three lineup size measures and both measures of lineup bias showed significant discriminability. The sensitivity of the measures depended on the criteria arbitrarily chosen or, when using statistical significance, on sample size. Race of mock witness had some impact on evaluations of Black lineups but not on evaluations of White lineups. Results indicated that a sample size of 18 ...

Enhancing lineup identification accuracy: two codes are better than one.

Ways of improving identification accuracy were explored by comparing the conventional visual lineup with an auditory/visual lineup, one that paired color photographs with voice recordings. This bimodal lineup necessitated sequential presentation of lineup members; Experiment 1 showed that performance in sequential lineups was better than performance in traditional simultaneous lineups. In Experiments 2A and 2B unimodal and bimodal lineups were compared by using a multiple-lineup paradigm: Ss viewed 3 videotaped episodes depicting standard police procedures and were tested in 4 sequential lineups. Bimodal lineups were more diagnostic than either visual or auditory lineups alone. The bimodal lineup led to a 126% improvement in number of correct identifications over the conventional visual lineup, with no concomitant increase in number of false identifications. These results imply strongly that bimodal procedures should be adopted in real-world lineups. The nature of memorial processes underlying this bimodal advantage is discussed.

Bias in Photospreads of Faces: A Comparison of Two Lineup Construction Methods

Recent research suggests that the current method of lineup construction produces biased or suggestive lineups. Earlier studies used face composite stimuli to assemple the lineups. The present study uses more realistic materials, actual face photographs. Ten pairs of subjects constructed photospread lineups using the traditional method of selecting lineup members who are similar in appearance to the suspect. Another ten pairs of subjects constructed lineups using an alternative construction method. The lineups were then given to a separate group of subjects who had never seen the photographs before and were asked to try to select the face that was the basis for each lineup. The results showed that traditional lineup constuction method produced bias towards the target/suspect. The alternative construction method produced less bias, but not significantly less than the traditional method. These results have implications for law enforcement personnel concerned with the construction and presentation of lineups.

Empirical assessment of lineups: Getting down to cases.

Empirical assessment has been proposed as a process for evaluating the fairness of lineups in forensic settings. The mock witness paradigm, which asks a nonwitnesses to choose lineup members best fitting a given description, was used in a 2×2 ANOVA design (N=165) to evaluate two photospreads (same defendant, different foils) from an actual rape case. Two groups (context 1) were asked to pick the best matching photo from the two photospreads after hearing the real witnesses' description and to estimate the features of each foil. Two groups (context 2) were asked for feature estimates only. A significantcontext effect (p<.01) on height and weight estimates shows that hearing a description polarized the absolute values (higher) but not the relative distribution, which saw the defendant rated as the heaviest and tallest member. The defendant was chosen by 58% (p<.001), in one photospread rated as biased. Complications in the second photospread prompted critique of recommended fairness indicators.

Most Similar is Different: Response Bias in Lineups

Previous research has shown that distractor faces made up of features of previously seen target faces are more likely to be selected as targets than the targets themselves. The present experiments sought to determine whether there is a tendency to pick a face out of a lineup if the face is more similar to the other lineup members than they are to each other. In Experiment 1, 25 subjects ranked members of lineups on similarity to target faces. The important comparisons were between rankings of the faces similar to the distractors (the critical faces), target faces, and other distractors. When the target faces were present in the lineup, no significant difference between rankings of the targets and the critical faces was found. When no targets were present in the lineup the critical faces were judged to be more like the targets than were the other distractors. This latter finding was replicated in Experiment 2 using visual stimuli other than faces. Implications for choosing police lineups are discussed.

The Psychology of Lineup Identifications1

A review is made of issues and data on eyewitness identifications, and a relative‐judgment conceptualization is proposed. It is argued that eyewitnesses are prone to choose the lineup member who most resembles the perpetrator relative to other lineup members as evidenced by studies that manipulated similarity of lineup members. The relative‐judgment strategy is fallacious because of the unpredictable occurrence of target‐absent lineups and is not corrected fully by instructions to eyewitnesses. An extension of the relative‐judgment conceptualization proposes an inverse relationship between the goodness of witnesses' memories (quality and quantity of relevant information available in memory) and witnesses' tendencies to rely on relative judgments. This extended conceptualization was used to derive expectations regarding an experiment (N= 192 eyewitnesses) that used a blank lineup prior to presenting eyewitnesses with the actual lineup. The data indicated that a blank lineup can yield a diagnostic split of eyewitnesses; those who made no identification when presented with a blank lineup were less likely to make false identifications on the actual lineup than either the witnesses who identified someone from the blank lineup or the witnesses who were not presented with a blank lineup. The blank lineup did not produce a significant loss in accurate identifications. The practical implications of using blank lineups and the theoretical utility of the relative‐judgment conceptualization are discussed.

Does Knowledge that a Crime Was Staged Affect Eyewitness Performance?1

The Staged crime has generally been considered a valid paradigm for assessing eyewitness identification accuracy. Yet the typical procedure informs the witness before the lineup task that the event was staged. The current study manipulated whether or not witnesses were informed that the witnessed crime was staged (information variable). Two other variables, the presence or absence of the perpetrator in the photo‐lineup and the physical similarity of lineup members, were manipulated to asses any possible interactions with the information variable. A theft was staged individually for 184 unsuspecting witnesses who were then randomly assigned to be either informed that the theft was staged or not informed. Next, all witnesses were met by a uniformed security officer who showed them a photo‐lineup of six individuals. Half of the witnesses viewed a perpetrator‐present lineup with either high or low physical similarity between the foils and the perpatrator and half viewed a perpetrator‐absent lineup with either high or low physical similarity. Three confidence measures, one cooperation measure, and an Embedded Figures Test score were obtained. The results showed that (a) the information variable did not affect the likelihood that a witness would attempt an identification; (b) the information variable did not have a main effect on the particular choices made by the witnesses nor did it interact with lineup similarity, but it did interact with the presence‐absence variable; and (c) the information variable did not affect witness confidence either as a main effect or interaction effect, but, was relatively consistent in affecting the correlation between confidence and accuracy such that the correlation was lower for witnesses who were not informed. The interaction between the information variable and the presence‐absence variable on accuracy was due to the uninformed witnesses being less accurate than the informcd witnesses in the perpetrator‐present condition. The results suggest that accuracy among actual eyewitnesses may be lower than obtained in the typical staged crime procedure and that accuracy‐confidence correlations may be overestimated by the typical staged crime.

Perspectives on the impact of lineup composition, race, and witness confidence on identification accuracy.

This paper addresses a number of issues suggested by the Lindsay and Wells (1980) study on lineup composition and identification accuracy. The interaction between bias in instructions to the witness, presence or absence of the criminal from the lineup, and similarity between the suspect and the lineup foils are discussed. Although Lindsay and Wells suggest that witness confidence has little or no relationship to witness accuracy, it is pointed out that a recent field study found a substantial accuracy-confidence relationship when criminal-present photo lineups were used. There are not yet clearcut findings on the accuracy-confidence relationship in criminal-absent lineups, partially because of ambiguity in the definition of confidence in this situation. Although there is much research on the impact of the race of suspect and witness on identification accuracy, little attention has been paid to the race of the person who constructs the photo or corporeal lineup. Recent research results lead to the prediction of an interaction between all three of these factors on identification accuracy, with greatest accuracy when the lineup constructor and lineup members are of the same race and the witness is of a different race. Issues in the applicability of the results of Bayesian analyses to the judicial system are briefly discussed. Potential issues include the tendency to see researchers solely as advocates for the defense, and the tendency of people to disregard statistical summaries such as base rate data and research results when making individual decisions. An additional issue concerns Bayesian diagnosticity ratios (derived from rates of correct and false identifications) which can be logically and statistically equivalent to one another but differ considerably in their legal applications and value relevance.

Accuracy, confidence, and juror perceptions in eyewitness identification.

Subjects of both sexes individually witnessed the staged theft of a calculator. The 127 witnesses were then given the opportunity to identify the thief from a six-person picture array; from this sample, 24 accurate-iden tification witnesses and 18 inaccurate-id entification witnesses were cross-examined with either leading or nonleading questions. Jurors were unable to distinguish accurate from inaccurate witnesses across the 42 cross-examina tion sessions (d' = .02). However, jurors in the leading-questions conditions were significantly more likely to believe accurate than inaccurate witnesses (<f = .35), whereas the reverse effect held for nonleading questions (<f = — .39). Jurors' attributions of witness confidence were unrelated to witness accuracy, even though these attributions accounted for 50% of the variance in jurors' decisions to believe witnesses. The poor accuracy-confidence relationship among witnesses is discussed in relation to the research on probability calibration. Eyewitness identification of criminal suspects is heavily relied on in the criminal justice system. (See Woocher, 1977.) Yet, staged-crime research indicates that eyewitness identifications are often unreliable. Indeed, a recent study indicated that following a staged crime, 34% of the witnesses made no identification from a six-person photo spread, 31% made accurate identifications, and 35% made false identifications (Leippe, Wells, & Ostrom, 1978). Although it is not clear to what extent current research may be presenting an exaggerated picture of the unreliable eyewitness (see Wells, 1978), there is little doubt that the rate of false identifications (i.e., identifying an innocent lineup member) in staged crimes is unacceptably high. The current investigation was designed to more fully analyze the implications of false iden

Guidelines for empirically assessing the fairness of a lineup.

Issues regarding the fairness of lineups used for criminal identification are discussed in the context of a distinction between nominal size and functional size. Nominal size (the number of persons in the lineup) is less important for determining the fairness of a lineup than is functional size (the number of lineup members resembling the criminal). Functional size decreases to the extent that the nonsuspect members of the lineup are easily ruled out as not being suspected by the police. The extent to which the identification of the suspect can be considered an independently derived piece of incriminating evidence is positively related to functional size. Empirical estimates of functional size can be obtained through pictures of the corporal lineup from which mock witnesses make guesses of whom they believe the police suspect. A distinction is made between a functional size approach and hypothesis testing approaches. Uses of functional size notions in the court, by police, and in research are discussed.