Diagnosing eyewitness accuracy

Eyewitnesses frequently mistake innocent people for the perpetrator of an observed crime. Such misidentifications have led to the wrongful convictions of many people. Despite this, no reliable method yet exists to determine eyewitness accuracy. This thesis explored two new experimental methods for this purpose. Chapter 2 investigated whether repetition priming can measure prior exposure to a target and compared this with observers’ explicit eyewitness accuracy. Across three experiments slower responses to target faces were consistently observed irrespective of eyewitness accuracy in a lineup task. This indicates that repetition priming can provide a covert index of eyewitness accuracy. However this method could not reliably assess the accuracy of individual eyewitnesses. Chapter 3 therefore explored an alternative test of eyewitness accuracy which was based on a multiple lineup procedure for faces. The characteristics of this method were assessed over five experiments which showed that only some eyewitnesses can actually identify a perpetrator repeatedly. Chapter 4 then showed that such repeat-identifications can provide a direct index of eyewitness accuracy in a field study. Over two experiments, the success of this method was such that eyewitnesses who consistently acted on the same identity over six lineups were always accurate eyewitnesses. These results demonstrate that multiple lineups of faces could provide a useful method for assessing eyewitness accuracy. The implications of these findings, both for further study and for forensic application, are discussed.