PHOTO 15.2 In the London Dungeon study (Valentine & Mesout, 2009) only 18% of visitors who reported feeling anxious in the dark, disorienting environment of the labyrinth, identified the actor from an array.
Source: © coka/Shutterstock
Valentine and Mesout (2009) tested the ability of visitors to the London Dungeon to identify an actor they had met in the Horror Labyrinth. Only 18% of visitors who reported feeling most anxious in the dark, disorienting environment of the labyrinth identified the actor from a nine-person photograph lineup. In contrast, 75% of visitors who reported experiencing less anxiety in the labyrinth were able to identify the actor.
Deffenbacher, Borstein, Penrod, and McGorty (2004) reported a meta-analytic review of studies that successfully manipulated stress, demonstrated by measures taken as soon as possible after encoding the target person. They found that heightened stress had a moderate negative effect on identification and on recall of a target person. The effect of stress on identification was restricted to the number of correct identifications made when the target person was included in the lineup; there was no effect of stress on the rate of correctly rejecting the lineup when the target person was not present (see also Chapter 6 ).
15.5.2.3 Ethnicity
Witnesses tend to be less accurate in recognising a perpetrator of an ethnic origin different from their own. The effect size is moderate and may depend upon the experience of the witness. Chiroro and Valentine (1995) found that experience of people of a different ethnic origin in daily life may reduce or eliminate any effect of ethnicity, but does not necessarily do so. The quality of the social contact may be an important mediating factor. The effect of ethnicity on face recognition can be interpreted within a framework in which individual faces are recognised by their distinctive qualities in relation to the population of faces experienced in one’s lifetime (Valentine, 1991; Valentine & Endo, 1992; see Meissner & Brigham, 2001 for a review).
PHOTO 15.3 Witnesses tend to be less accurate in recognising a perpetrator of an ethnic origin different from their own.
Source: © g-stockstudio/Shutterstock
15.5.2.4 Witness age
Laboratory studies have found that older people make fewer correct responses in tests of face recognition (e.g. Bartlett & Fulton, 1991; O’Rourke, Penrod, & Cutler, 1989). O’Rourke et al. (1989) found that identification accuracy declined at around age 50. The effect of witness age has been found in terms of older witnesses making both fewer correct identifications and more mistaken identifications (Searcy, Bartlett, & Memon, 1999, 2000; Searcy, Bartlett, Memon, & Swanson, 2001).
15.5.2.5 Confidence
A confident eyewitness may provide compelling evidence and be highly influential on a jury or judge. It has been appreciated for a long time that a confident witness may be mistaken. Many studies of eyewitness identification have suggested that the relationship between confidence and accuracy is low or negligible, leading psychologists to conclude that witness confidence is an unreliable means to assess accuracy. In recent years our understanding of the confidence-accuracy relationship has become more sophisticated. The relationship is stronger when a wide range of viewing conditions is considered (Lindsay, Read, & Sharma, 1998). One factor that may have restricted the relationship in experimental studies is that participants usually view a live or video mock crime under identical conditions. Furthermore, the correlation is stronger (typically in the region of r = .5) if only witnesses who identify somebody from a lineup are considered; it is lower amongst witness who reject the lineup (Sporer, Penrod, Read, & Cutler, 1995). However, a correlation of this order will still mean that a confident but mistaken eyewitness will be encountered fairly frequently (see Brewer, 2006, for a review).
15.5.3 Archival Studies of Estimator Variables
At live lineups conducted in England to investigate real criminal cases, approximately 40% of witnesses identified the suspect, approximately 40% of witnesses did not make any identification, and 20% of witness made a mistaken identification of an innocent foil (Valentine, Pickering, & Darling, 2003; Wright & McDaid, 1996). A more recent study of the new video lineups procedure (described in more detail in section 15.6.2) found similar results: 39% of witnesses identified the police suspect, 35% made no identification and 26% identified an innocent foil (Horry, Memon, Wright, & Milne, 2012). Note that in archival studies it is not known how many lineups contained the actual perpetrator. The known mistaken identifications were made despite the witness having been cautioned that the person they saw may or may not be present in the lineup. An archival analysis of 58 live lineups conducted in US criminal cases found that the suspect was identified in 50% of cases, a foil was identified in 24% of cases and the witness was unable to make an identification or rejected the lineup in 26% of cases (Behrman & Davey, 2001).
Valentine, Pickering, & Darling (2003) examined the effect of estimator variables on the outcome of identification attempts made by approximately 600 witnesses who viewed over 300 live lineups organised by the London Metropolitan Police. The suspect was more likely to be identified if: the witness was younger than 30; the suspect was a white European (rather than African-Caribbean); the witness gave a detailed description; viewed the culprit at the scene for over a minute; and made a fast decision at the lineup. Sixty-five percent of witnesses identified the suspect from lineups held up to seven days later, while only 38% of witnesses identified the suspect in lineups held eight days or more after the incident. There were no independent, statistically reliable effects of the use of a weapon during the incident, or a suspect of a different ethnicity from the witness.
Behrman and Davey’s (2001) archival analysis of eyewitness identification in American criminal cases included an analysis of the outcome of 289 photographic lineups. The typical format is to present the witness with six photographs simultaneously arranged in two rows of three images. They found that 48% of witnesses identified the suspect. In common with the British studies cited above, there was no effect of the presence of a weapon in the crime on the likelihood of the suspect being identified. Lineups held within seven days of the incident produced a higher rate of identifications of the suspect (64%) than lineups held after eight days or more (33%). However, in contrast to the British data, Behrman and Davey did find an effect of ethnicity. Sixty percent of witnesses of the same ethnicity as the suspect identified the suspect, compared to 45% of witnesses of different ethnicity.
15.6 SYSTEM VARIABLES
The criminal justice system can exert influence over some aspects of identification procedures (system variables). For example, a lineup may be presented in photographs, video or live; the method of selecting the “foils” for a lineup and the instructions given to the witness are specified in official guidance.
15.6.1 Presentation Mode
The effect of presenting lineups in different media (photographs, video, live) and manipulating the richness of cues available (e.g. stills, moving images, people walking) is surprisingly small. Cutler, Berman, Penrod, and Fisher (1994) concluded: “the conservative conclusion is that, based on available research, there is no reason to believe that live lineups, videotaped lineups or photo arrays produce substantial differences in identification performance” (p. 181). A possible reason is that faces can be sufficiently well perceived from a good-quality still photograph. Therefore, relying on cues such as gait, build or colour images adds little extra benefit. However, a more recent analysis found a small effect of lineup format (Clark, Moreland, & Rush, 2015). Responses were more accurate for procedures that included more information (e.g. various formats that included moving video) than for those with less information (e.g. static photographs).
15.6.2 Fairness of Video Identification
From 2003 video technology began to replace live identity parades in England and Wales. In 2011 video identification procedures became the preferred method for formal identification. Video lineups consist of 15-second clips of head and shoulders shots of each lineup member. First, they are looking at
the camera and they then rotate their head to show both profiles before looking back at the camera. The images are captured under standardised conditions. Each lineup member is shown sequentially one at a time, with a digit in the top left corner of the screen that can be used to identify each individual. The benefits of video identification compared to live lineups are listed in Box 15.3.
BOX 15.3 THE BENEFITS OF VIDEO IDENTIFICATION PROCEDURES
Video can dramatically reduce the delay before an identification can be organised. Live lineups have been subject to long delays to enable a selection of appropriate foils to be available to stand on a lineup (typically of one–three months, see Valentine, Harris, Colom Piera, & Darling, 2003). In contrast, a video lineup can be produced within two hours of request.
Approximately 50% of live lineups were cancelled, for example because a bailed suspect failed to attend. With video identification, the cancellation rate has fallen to around 5% (Pike, Kemp, Brace, Allen, & Rowlands, 2000).
A large database of video clips (approximately 23,000) is available, providing more foils for selection. This helps to make lineups fairer (see text for further details).
Video is less threatening to victims, who no longer have to attend an identification suite where, for example, their attacker may be physically present. Intimidation can result in a witness feeling too threatened to make a positive identification of a police suspect at a lineup. Use of video does not prevent witness intimidation, but any means of reducing the perceived level of threat at an identification procedure is beneficial.
Video equipment can be taken to a witness who is unable to attend the police station. In 2005, Abigail Witchalls, a victim of an attack who was left paralysed, was able to view a video lineup from her hospital bed. As a result a suspect was eliminated from the enquiry.
Research has shown that video lineups from real criminal cases are fairer to the suspects than live lineups (Valentine & Heaton, 1999; Valentine, Harris, Colom Piera, & Darling, 2003). In these studies, participants (known as mock witnesses) were given the first description of the offender provided by the original witness and required to guess which lineup member was the suspect. The mock witness had not seen the perpetrator, so the suspect should not have stood out in any way. In a perfectly fair lineup the suspect should not be chosen more often than any other lineup member. If the lineup contained a suspect and eight foils, the suspect should not be chosen by more than one in nine (11%) of mock witnesses. Valentine and Heaton (1999) found that the mock witnesses identified the suspect in live lineups more frequently (25%) than by chance but were not able to select the suspect from video lineups significantly more often than chance (15%). Valentine et al., (2003) found video lineups of Afro–Caribbeans and of white Europeans were equally fair.
15.6.3 Instructions Given to Witnesses
Witnesses may assume that they would not have been invited to attempt an identification if the police did not have good reason to believe that their suspect was guilty, and that it will help the police if they identify the suspect. It is important that the instructions to the witness should emphasis the possibility that making no identification may be the right decision because the suspect may not be guilty. Most commonly, this point is made by including an instruction that the person who the witness saw “may or may not be present”. Instructions that do not point out that the culprit may not be in the lineup are regarded as “biased” instructions (e.g. “Look at these photographs. Can you identify the man who assaulted you?”). A meta-analysis of 18 studies showed that when biased instructions are given, witnesses are more likely to make an identification, whether it is correct or incorrect (Steblay, 1997). Although biased instructions produce more correct identifications from culprit-present lineups, they increase the likelihood of an innocent suspect being identified from culprit-absent lineups (Clark, 2012).
15.6.4 Blind Administration of Lineups
‘Blind’ is used in the sense of meaning that the person who administers the lineup procedure to the witness does not know (i.e. is blind to) the identity of the suspect in a lineup. The procedure is often referred to as “double-blind”, meaning that both the witness and the lineup administrator are blind to the identity of the suspect.
The expectations of the experimenter can influence the outcome of behavioural research (Harris & Rosenthal, 1985). The lineup administrator should be blind to the identity of the suspect in order to prevent any inadvertent influence on the witness. Such influence can be very subtle and entirely unconscious. For example, the administrator may look at the witness when the suspect’s image is being viewed, or be more likely to accept a tentative identification if it is of the suspect. Double-blind administration of identification procedures removes all possibility of leading the witness. Therefore, the integrity of identification evidence is enhanced and any potential claim of bias can be rebutted. Clark’s (2012) analysis of the literature found that double-blind lineup administration led to fewer mistaken identifications from culprit-absent lineups but also fewer correct identifications of the perpetrator when present in the lineup.
15.6.5 Repeated Identification Procedures
If the police have not identified a suspect they may show the witness photographs (mugshots) of people previously convicted of a similar offence, in the expectation that the witness may be able to identify the perpetrator. In this procedure all of the people are suspects. Therefore any identification will lead to that person being investigated. Later in the investigation the police may want to collect formal identification evidence from a lineup. Would a subsequent lineup be biased against the suspect if the witness has previously seen their photograph in a mugshot album?
Deffenbacher, Bornstein, and Penrod (2006) provide a systematic review of the effects of mugshot exposure. They found that prior viewing of a photograph of somebody who subsequently appears in a lineup increases the probability of a mistaken identification from the lineup. The effect is due to transference of familiarity from the photograph, which is mistakenly attributed to having seen the person at the crime scene. The effect is strong if the person was mistakenly identified as the perpetrator from the mugshot photographs. This is known as an effect of commitment to the earlier identification. There was no ill-effect of showing photographs if none of the people seen appeared in the subsequent lineup. Deffenbacher et al. (2006) point out that transference of familiarity can occur for a bystander in the original mock crime who is included in the lineup. Experiments using this bystander design showed a significant effect of increased mistaken identification, but the effect was stronger when the face had been seen in a mugshot rather than as a bystander. The increased risk of a mistaken identification when a bystander, but not the perpetrator, is included in a lineup is very relevant to cases in which the suspect admits presence at the scene but denies involvement in the offence (for example, a bystander at a fight).
A street identification (show-up) is the identification most frequently used by British police to investigate street robberies (Davis, Valentine, Memon, & Roberts, 2015). A common method is for the police to drive a victim of street robbery around the area in a police car to point out the perpetrator, if seen. An identification procedure was attempted in 31% of almost 700 robbery cases analysed; a street identification was attempted in 23% of cases. A positive street identification strongly influenced whether a suspect would be charged. Only 7% of suspects were charged if there was no identification evidence; 52% identified in a street identification alone were charged; and 100% of suspects identified in both a street identification and a subsequent video identification procedure were charged, although there were only five cases in this last group. Experimental studies designed to replicate the use of a street identification and a subsequent video identification procedure found a strong commitment effect, which occurred even if the initial street identification was mistaken (Valentine, Davis, Memon, & Roberts, 2012). Over 80% of suspects mistakenly identified in a simulated street identification were also mistakenly identified
when the same suspect was seen in a police video lineup by the same witness between a week and a month later.
15.6.6 Selection of Foils
The code of practice in England and Wales specifies that the foils for lineups must be selected to “resemble the suspect”. This is known as a suspect-resemblance strategy. Luus and Wells (1991) argued that a better strategy is to select foils who match the witness’s description of the culprit. It is reasonable to assume that the witness can remember the description that he or she gave to the police, and may expect to identify somebody who matches their description. Therefore, the witness may be inclined to disregard any foils that do not match their description, or conversely pay special attention to anybody who is a better match to their description than the rest. To be fair, all lineup members should match the witness’s description of the culprit.
Luus and Wells (1991) suggested it does not introduce a bias against the suspect if lineup members differ on some feature that was not mentioned in the original description. Diversity amongst features not mentioned in the description will help a witness with a reliable memory to distinguish the culprit from the foils. If the suspect is not the culprit, he or she is no more likely to be mistakenly identified by some feature not mentioned in the description, because the witness has not seen the suspect before. A lineup that consists of a number of people chosen because they closely resemble the suspect in all aspects of their appearance will make it difficult even for a reliable witness to identify the culprit, if present.
Forensic Psychology Page 70