Police Identifications Are Occasionally More Accurate, but a General Conclusion About Their Performance Is Premature

Abstract

Reisberg and Pezdek raise concerns about our study, in which police trainees outperformed civilians on an identification task. They argue that we did not report our results in a sufficiently disaggregated form and that our study does not overturn their earlier conclusion that the superiority of police witnesses is a myth. We maintain that police trainees' identification decisions in our experiment were, on aggregate, more accurate than civilians'. However, we do not claim that the literature generally supports the conclusion that police witnesses are more accurate than civilians when making identifications. The existing literature includes only seven relevant experiments, and two use only target-present lineups. Moderators and confounds may play a role in several studies. We conclude with a call for a systematic review and meta-analysis and encourage future research to consider under-represented jurisdictions and sampling plans to reflect the heterogeneity of the police body.