The effects of contextual bias on face recognition decisions.

Abstract

Contemporary research has demonstrated the effects of bias on, even expert, forensic decision making. The paper aimed to test if forensically relevant face recognition decisions could be influenced by biasing information. A 3 (Bias (within-subjects): positive bias vs. negative bias vs. control) × 2 (evidence strength (between-subjects): weak video evidence (N = 97) vs. strong video evidence (N = 98)) × 2 (target presence (within-subjects): absent vs. present) mixed-design was utilized. Confidence, accuracy, and decision time were measured. In total, 195 participants were recruited. The Cambridge face memory test+ was used to measure face recognition ability. Participants saw 36 videos emulating Closed Circuit Television (CCTV) footage of a person walking down the corridor. Participants were randomly allocated to either the strong or weak evidence condition. Participants were shown a statement for each video that contained either a positive bias (target face matched the face in the video), a negative bias (target face did not match the face in the video), or control (no statement provided). Participants were then presented with a target face and asked if it matched the face seen in the previous video. There was a significant interaction between the bias and the target presence factors, with accuracy and confidence increasing and decision times decreasing when a positive bias statement was used when the target was present. Face recognition abilities did not act as a covariate. Bias may influence facial recognition decisions, and superior face recognition abilities do not undermine the influence of bias. Recommendations/implications, such as linear sequential unmasking, were discussed.