Eyewitness identification discriminability: ROC analysis versus logistic regression

To reach conclusions regarding the respective accuracy of two conditions, eyewitness researchers evaluate correct and false identification rates computed across participants. Two approaches typically are employed. One approach relies on ratio-based probative value measures; but Wixted and Mickes (2012) and Gronlund, Wixted, and Mickes (2014) showed that these measures fail to disentangle an assessment of accuracy (i.e., discriminability between guilty and innocent suspects) from response bias (i.e., a willingness to make a response). Our focus is on a second approach, logistic regression analyses of the correct and of the false identification rates. Logistic regression also fails to disentangle discriminability from bias. Therefore, it only can denote the most accurate condition in limited circumstances. The best approach for reaching the proper conclusion regarding which condition is most accurate is to use receiver operator characteristic (ROC) analysis. Simulated ROC data illustrate the problem with a reliance on logistic regression to assess accuracy.     

Lineup administrator influences on eyewitness identification and eyewitness confidence

Lineup administrators were trained to respond to witnesses in such a way as to redirect them from making non-identifications or foil identification responses toward making identifications of the suspect. Compared to a no-influence control condition, suspect identification rates in the influence condition increased substantially and proportionally for guilty and innocent suspects. Administrators steered witnesses more specifically toward the suspect when the suspect was guilty than when the suspect was innocent. Post-identification confidence for correct identifications of the guilty suspect did not differ significantly across the influence and no-influence groups. However, post-identification confidence for false identifications of the innocent suspect was significantly lower for the influence group than for the no-influence group because witnesses who were influenced to make false identifications tended to be those who were less confident prior to the lineup, and also because those witnesses became less confident from pre- to post-identification.     

Gender Differences in Moral Influences on Adolescents’ Eyewitness Identification

Abstract

In this study, 232 (89 11- to 12-year-olds, 71 13- to 14-year-olds; 72 15- to 16-year-olds) students recruited from grades 6th–11th in an urban public high school participated in a study of eyewitness identification. The focus of this study was on the effects of age, gender and moral orientation on decisional bias and, as a secondary outcome, on accuracy (using signal detection analysis).The primary purpose of this and previous studies in this series is to uncover implicit moral decision-making in decisional bias. In this study the perpetrator, the bystanders and the foil were all females. Prior to completing the eyewitness identification task, participants were given instructions that emphasized either (a) fairness and crime prevention, or (b) neither. These instructions had no discernible effect on accuracy but, as in past studies, younger participants (below the age of 13) had lower decisional criteria, resulting in a higher rate of false alarms/positives. Further, those who judged the transgression as worse had a lower decisional criterion, indicating more false alarms. Females were more accurate than the males in identifying the female perpetrator and scored significantly higher on how bad they would feel if they were the victim than did the males.     

Age differences in the effects of social influence on children's eyewitness performance and their metacognitive monitoring

The current work investigated the effects of social influence on children's recall accuracy and metacognitive monitoring. Two studies were conducted in which 8- and 10-year-olds were confronted with postevent information in an interview situation. An interviewer (Study 1) or a confederate (Study 2) provided postevent information with two levels of assertiveness, inducing (a) a variation of conformity pressure and (b) a variation of information credibility. Afterwards, children's confidence judgments were assessed. The results revealed significant age differences in children's ability to adequately cope with variations of social influence. Although conformity pressure was especially important for the 8-year-olds, effects of informative social influence were independent of age. However, 10-year-olds were also able to act appropriately on low credibility, thereby demonstrating a more sophisticated consideration of social influence sources. Moreover, varying assertiveness also affected the quality of children's confidence judgments by improving their metacognitive differentiation skills.