Confidence and Response Time as Indicators of Eyewitness Identification Accuracy in the Lab and in the Real World

The criminal justice system should consider the confidence an eyewitness expresses when making an identification at the time the initial lineup procedure is conducted. High confidence expressed at this time typically indicates high accuracy in the identification. Because the suspect identification-not filler identifications or no identifications - matters most in the court of law, confidence-accuracy characteristic (CAC) analysis provides information most relevant to stakeholders. However, just as high confidence identifications indicate high accuracy, fast identifications may also indicate high accuracy. We tested whether a new technique that is similar to CAC analysis, called response time-accuracy characteristic (RAC) analysis, could inform stakeholders about the likely accuracy of an identification while usefully summarizing response time data. We argue this is the case in the lab and in the real world. Furthermore, CAC and RAC results are not completely redundant so both, considered together, are useful to the criminal justice system. General Audience Summary During a police investigation, an eyewitness may be presented with a lineup. If the suspect is identified, and if that suspect becomes the defendant in a court of law, judges and jurors should know the likelihood that an eyewitness identification of the defendant is accurate. One indicator of accuracy is confidence. That is, if an eyewitness expresses high confidence in the identification of the suspect during the initial identification procedure (as opposed to a later time, such as in a courtroom during the trial proceedings), then the identification is more likely to be accurate than if low confidence is expressed. Another indicator of accuracy may be the time it takes for an eyewitness to identify the suspect during the initial identification procedure. That is, if an eyewitness makes an identification of the suspect quickly, is that identification more likely to be accurate than if an identification is made slowly? Whether this is the case and whether confidence and response time could each provide useful independent information for judges and jurors were the main questions that we investigated in this research. For both eyewitness participants in the lab and real eyewitnesses in the field, confidence and response time both provide information about the accuracy of the identifications. Moreover, they provide partially non-overlapping information. Thus, we suggest that, where possible, judges