No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
How many faces can we remember? Why this matters when assessing eyewitnesses
Abstract
Most research on eyewitness memory has focused on single-perpetrator crimes. However, crimes to which eyewitnesses may bear testimony are often committed by groups of perpetrators. A consequence of researching only single-perpetrator crimes is that we know very little about how set size (i.e., the number of faces) at encoding impacts recognition performance. We do not know much more about this question in the face recognition literature either but the small extant literature does appear to converge on one conclusion, namely that recognition performance is worse for larger set sizes. In the case of eyewitness memory, the presence of multiple perpetrators poses an additional unique question: Eyewitnesses not only need to identify perpetrators, but also need to testify to the perpetrators' actions. Few researchers have investigated this second aspect. In this chapter, we review literature in the areas of face recognition and eyewitness memory to shed light on these questions, and present two laboratory studies that test the effects of set size on face and person recognition. Results show that recognition performance decreases as a function of set size, but that this is differentially true for faces and roles, and is in fact dramatically reduced when faces and roles are paired. There are serious applied implications for this latter finding in particular.
... Wells and Turtle (1986) showed that from a statistical perspective single-suspect lineups will always be superior and recommended against multiple-suspect lineups (their argument is quite technical, so we won't consider it further here). There are of course considerations other than those researchers have the freedom to consider, and Nortje, Tredoux, and Vredeveldt (2017) report that multiple-suspect lineups are very frequently used by police in multiple-perpetrator cases, mostly for practical reasons. It seems important to us to study ways of improving multiple-suspect lineups, since single-suspect lineups are often unaffordable in jurisdictions that are still required to conduct live lineups. ...
... Multiple-perpetrator crimes deserve attention from eyewitness researchers, as they appear to differ from single-perpetrator crimes in one very important respect: Whereas the actions carried out in a single-perpetrator crime are unique and usually accurately recalled by eyewitnesses, this is not the case in multiple-perpetrator crimes, where memory for actions and who carried them out are often dissociated. Although few studies have investigated the reliability of eyewitness memory in multiple-perpetrator crimes, it appears that as the number of perpetrators increases, overall eyewitness identification accuracy declines (Clifford & Hollin, 1981;Megreya & Bindemann, 2011;Nortje et al., 2017). Kask and Bull (2009) considered the interesting case where multiple perpetrators of a crime are of different ethnicities to each other and concluded that it was better to test recognition in such a crime with separate single-ethnicity lineups. ...
The efficiency of visual search for one (single-target) and either of two (dual-target) unfamiliar faces was explored to understand the manifestations of capacity and guidance limitations in face search. The visual similarity of distractor faces to target faces was manipulated using morphing (Experiments 1 and 2) and multidimensional scaling (Experiment 3). A dual-target cost was found in all experiments, evidenced by slower and less accurate search in dual- than single-target conditions. The dual-target cost was unequal across the targets, with performance being maintained on one target and reduced on the other, which we label “preferred” and “non-preferred” respectively. We calculated the capacity for each target face and show reduced capacity for representing the non-preferred target face. However, results show that the capacity for the non-preferred target can be increased when the dual-target condition is conducted after participants complete the single-target conditions. Analyses of eye movements revealed evidence for weak guidance of fixations in single-target search, and when searching for the preferred target in dual-target search. Overall, the experiments show dual-target search for faces is capacity- and guidance-limited, leading to superior search for 1 face over the other in dual-target search. However, learning faces individually may improve capacity with the second face.
The vagaries of eyewitness identification are well known; the annals of criminal law are rife with instances of mistaken identification.
In a simulated crime situation, 3 groups of eyewitnesses viewed 2 target persons through a one-way mirror and were tested for accuracy of identification after delay periods of 2, 21, or 56 days. Subjects made identifications from either a live “show-up” or by looking at photographs. In both methods of testing only 1 of the 2 targets was actually present in the 5-man array. Results indicated that delay affects number of false alarms, test method affects number of hits. Photographs produced less accurate performance than live show-up. More than 60% of the subjects erroneously selected one of the 4 distractor persons, a finding relevant to the use of eyewitness identification in criminal investigations. Only 28% of the subjects made no errors of identification. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Rape prevention efforts are hampered by a chronic lack of adequate epidemiological and surveillance data. Information on identifying factors such as the who, when, where and how of rape is needed to inform the design of effective intervention programmes. Results from a demonstration Rape Surveillance Project show that records of rape cases presenting at three medico-legal clinics provide a valuable source for the epidemiological surveillance of rape. From January 1996 to December 1998 a surveillance questionnaire was completed for rape victims presenting at the Hillbrow, Lenasia South, and Chris Hani Baragwanath Medico-Legal Clinics in Gauteng. Analysis of the data suggests which women are most at risk for being raped, by whom they are raped, the areas where attacks most often occur, and the day and time when rapes are mostly committed. Despite the limited database and difficulties with generalizing findings beyond the three clinics, it is apparent that surveillance procedures have enormous import for sexual violence prevention and intervention. Implications for prevention strategies, aftercare, policy formulation, and future research are discussed. Methodological issues and institutional constraints are also discussed with a view to strengthening and developing such information management systems.
The accuracy of eyewitness identification in a two-culprit crime was examined using two traditional lineup procedures – simultaneous and sequential – and a newly developed procedure called the two-person serial lineup. Target-present and –absent lineups with each procedure were used. In addition, the ability of witnesses to describe and match culprits to their actions was examined. Participants (n=150) watched a videotaped, staged, theft involving two culprits that had different roles in the crime (assailant vs accomplice). Witnesses were significantly more likely to correctly identify the accomplice than the assailant. There was no evidence of role transference. When the culprits were present in the lineup, identification accuracy did not vary as a function of lineup procedure. When the culprits were absent from the lineup, there was a trend for the two-person serial lineup to produce higher correct rejections than the simultaneous or sequential procedures. Witnesses were significantly more likely to describe the appearance of the assailant, however, they were more accurate with their appearance descriptions of the accomplice. Witnesses were more accurate when describing what the assailant did than the accomplice. A relation between description length and identification accuracy was not found.
Investigated 60 undergraduates' accuracy in testimony and identification following either a violent (mugging) or a nonviolent (direction-seeking) videotaped incident in which 1, 3, or 5 perpetrators participated. The relationship between Ss' objective accuracy and their subjective feeling of certainty concerning correctness was also examined. Testimony was less accurate following the witnessing of the violent incident, and the decrease in accuracy was a function of the increase in the number of perpetrators seen, especially under the violent condition. The accuracy of identification was very low, with only 27% of Ss making a correct identification; chance performance was observed with 5 perpetrators. A positive accuracy–confidence relationship held for identification under the nonviolent condition but not under the violent condition. (25 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Presents a meta-analysis of 128 eyewitness identification and facial recognition studies involving 960 experimental conditions and 16,950 Ss. The meta-analysis was designed to determine what knowledge has been accumulated on factors that influence facial identification performance and what areas of facial identification research would benefit from further research. Two techniques were used: an effect size analysis, which integrates the effect sizes of independent variables across studies, and a study-characteristics analysis, which integrates the influence of study characteristics on performance. A number of variables operating at the encoding and retrieval stages yielded large effects on performance. These variables included context reinstatement, transformations in the appearance of faces, depth of processing strategies, target distinctiveness, and elaboration at encoding. Additional variables yielding strong effects on recognition performance were exposure time, cross-racial identification, and retention interval. There was little correspondence between a variable's impact on hit rate and false-alarm rate. A list of the studies used in the meta-analysis is appended. (49 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Tested 392 high school graduates for memory of names and portraits of classmates selected from yearbooks. The retention interval since graduation varied from 2 wks to 57 yrs. Performance was adjusted by multiple regression procedures to control the effects of various conditions on original learning and rehearsal of the material. Results show that identification and matching of names and faces remained approximately 90% correct for at least 15 yrs, even for members of very large classes. Free recall was independent of class size and recognition memory, and declined with negative acceleration by 60% during 48 yrs. This very slow forgetting process is attributed to distribution of practice and overlearning effects far more extreme than those observed in laboratory research. The decline of free recall is associated with the forgetting of the nature of the interpersonal relationship which serves as a retrieval mediator. (33 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
IN AN EARLIER EXPERIMENT SS EXAMINED AN EXTENDED SEQUENCE OF PHOTOGRAPHS ATTEMPTING TO IDENTIFY THOSE PHOTOS THAT WERE OCCURRING FOR THE 2ND TIME WITHIN THE SEQUENCE. THIS NOTE DESCRIBES THE RESULTS OBTAINED WHEN, AFTER PERIODS RANGING FROM A DAY TO A YEAR, THE SAME SS WERE ASKED TO DISTINGUISH BETWEEN NEW PHOTOS AND PHOTOS THAT THEY HAD SEEN WHILE PARTICIPATING IN THE ORIGINAL EXPERIMENT. RESULTS INDICATE FAIRLY SUBSTANTIAL RETENTION OF BRIEFLY PRESENTED PICTORIAL INFORMATION OVER LONG DURATIONS. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Studies an aspect of short-term memory somewhat neglected in the past, namely, the ability to remember, i.e., to recognize, complex meaningful visual configurations. Ss' task was to inspect an extended sequence of photographs of assorted content and to identify those which were occurring for the 2nd time within the sequence. The probability of recognizing the recurrence of a photo as such was very high even with as many as 200 items intervening between its 1st and 2nd occurrence. (French summary)
In the present study, we investigated age-related decline in face recognition memory and whether this decline is moderated by the age of the target faces and by the number of faces that the participant must learn (memory load). Thirty-two participants in each of three age groups (18-39 years, 60-75 years, and 76-96 years) completed a face recognition task. Signal detection analyses confirmed that face recognition accuracy declined with age. However, this finding was qualified by an interaction between participant age and target age, which revealed that the age-related decline in face recognition accuracy occurred only for young target faces. Increased memory load was associated with comparable performance decrements across all age groups. However, memory load appears not to be the cause of these decrements. Instead, they appear to be a product of recognition load (the number of stimuli presented in the recognition phase).
Following a question raised by members of a police force in England and Wales regarding the accuracy with which an eyewitness makes an identification of multiple perpetrators, this study explores presenting the line-ups in a novel way to eyewitnesses. Currently, there is little previous research into the accuracy of eyewitness identifications when witnesses have viewed multiple perpetrators for one crime event. With reference to theories of source monitoring, the novel procedure examines whether presenting all video line-up parades one after another before making any identification decisions increases accuracy compared with current police procedures in the UK. Furthermore, this study investigated whether instructing witnesses, prior to viewing the line-up parades, to identify the role that each perpetrator played in the event increased identification performance. Results reveal that overall total accuracy was affected by presentation style and, when instructed to, witnesses were more accurate at identifying the role that each perpetrator played in the crime. Future research needs to explore these techniques in more depth both in a laboratory and operational setting.
There is an important mismatch between empirical research on the accuracy of eyewitness identification and the real world of criminal investigation. Most research models single-perpetrator crimes, but in the real world, most crimes involve multiple perpetrators. This study examined how the number of perpetrators affects eyewitness identification by manipulating the gender of accomplices. Observers viewed a video of a staged crime. The crime was committed by a male or female perpetrator, who was presented alone or with an accomplice of the same or the opposite gender. The observers were then asked to identify the perpetrators from target-present or target-absent line-ups. The results revealed a double-perpetrator disadvantage, which was manifested in reduced identification accuracy on target-present line-ups. Importantly, however, the gender of the perpetrator or the accomplice had no effect on this disadvantage. This double-perpetrator disadvantage is attributed to the need to divide attention between two concurrent people, compared to single-culprit crimes, rather than an impairment of visual encoding processes. The implications of these findings are discussed.
Detection Theory is an introduction to one of the most important tools for analysis of data where choices must be made and performance is not perfect. Originally developed for evaluation of electronic detection, detection theory was adopted by psychologists as a way to understand sensory decision making, then embraced by students of human memory. It has since been utilized in areas as diverse as animal behavior and X-ray diagnosis. This book covers the basic principles of detection theory, with separate initial chapters on measuring detection and evaluating decision criteria. Some other features include: complete tools for application, including flowcharts, tables, pointers, and software;. student-friendly language;. complete coverage of content area, including both one-dimensional and multidimensional models;. separate, systematic coverage of sensitivity and response bias measurement;. integrated treatment of threshold and nonparametric approaches;. an organized, tutorial level introduction to multidimensional detection theory;. popular discrimination paradigms presented as applications of multidimensional detection theory; and. a new chapter on ideal observers and an updated chapter on adaptive threshold measurement. This up-to-date summary of signal detection theory is both a self-contained reference work for users and a readable text for graduate students and other researchers learning the material either in courses or on their own. © 2005 by Lawrence Erlbaum Associates, Inc. All rights reserved.
This experiment examined the effects of memory load over three periods of delay. Following presentation of 20, 35, or 50 targets, subjects were required to select these from an equal number of distractors 10 min, 1 week, or 2 weeks later. Increased target load (independent of increases in recognition load) decreased accuracy mainly by decreasing hit rate. Increasing delay decreased accuracy largely as a result of an increased false alarm rate. Most individual ROC curves, plotted on z-coordinates, had slopes < 1. Implications of this finding for the use of d' are discussed.
Two frequent types of violence by young men in groups are rapes of women and assaults on perceived gay men. This article draws
on scholarship from multiple disciplines to propose a theoretical linkage between these two seemingly disparate forms of violence.
It proposes that the two forms of violence serve near-identical functions for certain young men. These functions include social
bonding, the celebration of power, and the public display of heterosexual masculinity through the subordination of women and
men perceived as feminine. In other words, group rape of women and violence against homosexuals are parallel forms of cultural
theater, with the victims serving as interchangeable dramatic props. This argument is advanced through an examination of a
2003 case in which athletes from Mepham High School in Long Island, New York, sexually assaulted their younger teammates during
a week-long summer training camp.
There is a large body of work investigating face identification, but most of this addresses recognition of a single person. Here, we examine how recognition is affected by the presence of a second face. In Experiments 1 and 2, we demonstrate that memory for an unfamiliar face is severely reduced if it is seen alongside a second person. Sequential presentation of two target faces further reduces accuracy. In Experiments 3 and 4, we demonstrate the same disadvantage for two-face targets in a matching task, where subjects have no time limits or memory requirement. In matching, the damaging effect of a second face is greatest when the targets are placed close together. Furthermore, there is a general advantage for faces presented to the left. We suggest that it may not be possible to extrapolate results from single-face studies to experiments (or realistic situations) involving more than one person. Copyright (c) 2006 John Wiley & Sons, Ltd.
The Ss looked through a series of about 600 stimuli selected at random from an initially larger population. They were then tested for their ability to recognize these “old” stimuli in pairs in which the alternative was always a “new” stimulus selected at random from the stimuli remaining in the original population. Depending upon whether this original population consisted solely of words, sentences, or pictures, median Ss were able correctly to recognize the “old” stimulus in 90, 88, or 98% of the test pairs, respectively. Estimated lower bounds on the informational capacity of human memory considerably exceed previously published estimates.
This study took place over a three month period, from April-July 2003 and had two broad aims: 1. To examine the extent and nature of sexual assault in Southwark, with a particular focus on sexual offences and young people, rape and indecent assault 2. To identify the availability and effectiveness of current service provision for victims and perpetrators in Southwark. Report commissioned by the Serious and Violent Crime Sub-Group of the Safer Southwark Partnership, London Borough of Southwark.
This study examined effects of an event and dynamic mug shots on facial recognition. Witnesses saw either a film of a liquor store holdup staged by six perpetrators, slides of the same targets, or freeze-framed views from the filmed event. Slides of different facial views of targets among 12 foils, or dynamic 180 degrees mug shots of the same men, or multiple static views were administered 24 hr later in yes-no recognition tests. Recognition (d') was poorer for the same men seen in the event than in slides, but better for dynamic than for static mug shots. Nondynamic event characteristics were responsible for poor recognition of people seen in events, whereas dynamic aspects of facial transformations, not multiple views, seemed responsible for the superiority of dynamic mug shots. The implications of these results for issues of ecological validity of traditional face recognition methodology and for static feature models vs. dynamic transformation models of face recognition are discussed.
Sexual assault in Australia: A statistical overview
Australian Bureau of Statistics. (2004). Sexual assault in Australia: A statistical
overview.4523.0. Retrieved from http://www.abs.gov.au/AUSSTATS/ abs@.nsf/
DetailsPage/4523.02004.
Costs and benefits of eyewitness identification reform: Psychological science and public policy
- N Brewer
- R Potter
- R P Fisher
- N Bond
- M A Luszcz
Brewer, N., Potter, R., Fisher, R. P., Bond, N., & Luszcz, M. A. (1999). Beliefs and data
on the relationship between consistency and accuracy of eyewitness testimony.
Applied Cognitive Psychology, 13, 297-313. Clark, S. E. (2012). Costs and benefits
of eyewitness identification reform: Psychological science and public policy.
Perspectives on Psychological Science, 7, 238-259.
The feasibility of using multiple-suspect sequential lineups. (Unpublished Honour's thesis)
- P Jacob
Jacob, P. (1994). The feasibility of using multiple-suspect sequential lineups.
(Unpublished Honour's thesis). Queen's University, Canada.
Contextual effects on lineup identification of multiple perpetrators. (Unpublished Honour's thesis)
- S Laldin
Laldin, S. (1997). Contextual effects on lineup identification of multiple perpetrators.
(Unpublished Honour's thesis). Queen's University, Canada.
The psychological impact of rape trauma: A longitudinal study of adult female survivors in the Western Cape, South Africa
- S Maw
Maw, S. (2012). The psychological impact of rape trauma: A longitudinal study of adult
female survivors in the Western Cape, South Africa. (Unpublished doctoral thesis).
University of Cape Town, Cape Town.
Stimulus load and age effects in face recognition: A comparison of children and adults
- M M Metzger
Metzger, M. M. (2002). Stimulus load and age effects in face recognition: A comparison
of children and adults. North American Journal of Psychology, 4, 51-62.
Remembering multiple faces is harder than you think! The effect of set size on face recognition. Paper presented at the American Psychology-Law Society conference
- A Nortje
- C G Tredoux
- A Vredeveldt
Nortje, A., Tredoux, C.G., & Vredeveldt, A. (2015, March). Remembering multiple faces
is harder than you think! The effect of set size on face recognition. Paper presented at
the American Psychology-Law Society conference, San Diego, USA.
Role and offender identification in multiple-perpetrator crimes. Presentation at the Annual Face Science Seminar
- A Nortje
- M Deglon
- C Tredoux
- A Vredeveldt
Nortje, A, Deglon, M., Tredoux, C., & Vredeveldt, A (2016, November). Role and
offender identification in multiple-perpetrator crimes. Presentation at the Annual
Face Science Seminar, Cape Town, South Africa.
Estimated percent distribution of violent victimizations by multiple offenders by type of crime and perceived race of offenders
Sourcebook of Criminal Justice Statistics. (2008). Estimated percent distribution of
violent victimizations by multiple offenders by type of crime and perceived race of
offenders, Retrieved from http://www.albany.edu/sourcebook/pdf/t3312008.pdf.
Standing, L. (1973). Learning 10000 pictures. Quarterly Journal of Experimental
Psychology, 25, 207-222.
The effects of the sequential presentation of lineups with multiple perpetrators on eyewitness identification. (Unpublished Honour's thesis)
- A Vanderwal
Vanderwal, A. (1996). The effects of the sequential presentation of lineups with multiple
perpetrators on eyewitness identification. (Unpublished Honour's thesis). Queen's
University, Canada.