Contamination of Eyewitness Self-Reports and the Mistaken-Identification Problem

Abstract

Mistaken identification testimony by highly confident eyewitnesses has been involved in approximately 72% of the cases in which innocent people have been convicted and later exonerated by DNA testing. Lab studies of eyewitness identification, however, show that mistaken eyewitnesses are usually not highly confident and that there is a useful confidence-accuracy relation that can help distinguish accurate from mistaken eyewitnesses. We describe research on important variables that can cause mistaken eyewitnesses to give inflated self-reports about their confidence and other testimony-relevant judgments. These testimony-bolstering variables tend to be controlled in pristine experiments, thereby permitting good confidence-accuracy relations. In the real world, however, methods of obtaining identifications and supporting testimony permit testimony-bolstering variables to contaminate witness self-reports in ways that make it difficult to distinguish between accurate and mistaken eyewitness identification testimony. We describe how the justice system can dramatically reduce the chances of such contamination.