Presentation Title

Individual Differences in Face Recognition Ability and the Disguise Replication Effect

Faculty Mentor

Iris Blandón-Gitlin

Start Date

23-11-2019 9:00 AM

End Date

23-11-2019 9:15 AM

Location

Markstein 306

Session

oral 1

Type of Presentation

Oral Talk

Subject Area

behavioral_social_sciences

Abstract

If an eyewitness perceives a perpetrator in disguise, what lineup test would lead to better accuracy in identifying the culprit? A lineup test with undisguised individuals (real-world police practice) or a lineup test that replicates the disguise in each individual (an experimental procedure)? Recent research from our lab tested and confirmed the hypothesis that replicating the disguise in the lineup members (disguised lineup) yields better accuracy than not replicating the disguise (undisguised lineup). The results further showed the benefit of the disguise replication to be greater when encoding times were long than short, and the lineup test immediate rather than delayed. Thus, the replication effect appears to be related to good observation/testing conditions.

In an attempt to understand the cognitive mechanisms underlying the disguise replication effect, the current study investigates the role of individual differences in face recognition ability, using the widely used Cambridge Face Memory Test (CFMT). In our previous research, we obtained individual CFMT scores (N = 566), which are the focus of the current study. The study assesses whether face recognition ability is associated with the disguise replication effect and how this varies as a function of estimator variables (encoding time/test delay).

We analyzed the relationship between hits (correct identification) and false alarms (innocent suspect identification) with CFMT scores as a function of lineup type (disguised/undisguised) in both the best (long encoding/short delay) and worst (short encoding/long delay) memory conditions. There was an overall positive correlation between CFMT score and hit rate in good, but not poor, conditions, r = .203, p = .008, 95% CIs [.34, .05], with disguise serving as a moderator, F(1) = 8.009, p < .01.

The results provide an initial suggestion that replication is most useful with strong face recognizers exposed to good encoding/testing conditions. Theoretical and forensic implications will be discussed.

This document is currently not available here.

Share

COinS
 
Nov 23rd, 9:00 AM Nov 23rd, 9:15 AM

Individual Differences in Face Recognition Ability and the Disguise Replication Effect

Markstein 306

If an eyewitness perceives a perpetrator in disguise, what lineup test would lead to better accuracy in identifying the culprit? A lineup test with undisguised individuals (real-world police practice) or a lineup test that replicates the disguise in each individual (an experimental procedure)? Recent research from our lab tested and confirmed the hypothesis that replicating the disguise in the lineup members (disguised lineup) yields better accuracy than not replicating the disguise (undisguised lineup). The results further showed the benefit of the disguise replication to be greater when encoding times were long than short, and the lineup test immediate rather than delayed. Thus, the replication effect appears to be related to good observation/testing conditions.

In an attempt to understand the cognitive mechanisms underlying the disguise replication effect, the current study investigates the role of individual differences in face recognition ability, using the widely used Cambridge Face Memory Test (CFMT). In our previous research, we obtained individual CFMT scores (N = 566), which are the focus of the current study. The study assesses whether face recognition ability is associated with the disguise replication effect and how this varies as a function of estimator variables (encoding time/test delay).

We analyzed the relationship between hits (correct identification) and false alarms (innocent suspect identification) with CFMT scores as a function of lineup type (disguised/undisguised) in both the best (long encoding/short delay) and worst (short encoding/long delay) memory conditions. There was an overall positive correlation between CFMT score and hit rate in good, but not poor, conditions, r = .203, p = .008, 95% CIs [.34, .05], with disguise serving as a moderator, F(1) = 8.009, p < .01.

The results provide an initial suggestion that replication is most useful with strong face recognizers exposed to good encoding/testing conditions. Theoretical and forensic implications will be discussed.