Critical Tests of Decision Models for Eyewitness Identification

University Of California-Riverside, Riverside CA

Investigators

Abstract

It is well-established that eyewitnesses make identification errors. If the eyewitness is unable to make a correct identification, a criminal may escape justice and continue to present a threat to society. If the witness makes a false identification, an innocent person may be prosecuted and convicted of a crime he did not commit. The healthy functioning of the criminal justice system requires that such errors occur as infrequently as possible. In order to minimize eyewitness errors, we must understand why witnesses make errors. The present research seeks to understand the causes of eyewitness identification errors by developing, testing, and refining a psychological theory of the eyewitness. The theory, called the Witness model, describes the memory and decision processes that underlie eyewitness identification decisions, and makes quantitative predictions that can be compared to data. The data come from eyewitness identification experiments, conducted under controlled conditions. The participants in the experiments become witnesses to a staged crime, and are later shown either a perpetrator-present or a perpetrator-absent lineup. The perpetrator-present lineup simulates those real-world cases in which the police suspect is guilty, whereas the perpetrator-absent lineup simulates those real world cases in which the police suspect is innocent. The key data are the correct identification rates of the guilty and the false identification rates of the innocent. The probative value of identification evidence is maximized when the correct identification rate is high and the false identification rate is low. The first series of experiments examines how the probative value of identification evidence varies as a function of the way police create lineups and the decision processes that witnesses use to make identification decisions. The second set of experiments examines how a witness?s decision processes distribute accuracy and error across the members of a lineup.