Eyewitness Memory

In some respects, the story of eyewitness memory has remained unchanged for decades. For example, because memory is malleable, eyewitness memory evidence (like all types of forensic evidence) can be contaminated. Although the field of psychology has appreciated the malleability of memory for decades, the legal system’s failure to appreciate that fact has led to many wrongful convictions. In other respects, the story of eyewitness memory – particularly eyewitness identification from a lineup – has changed in recent years. We focus on that story
here. For decades, the field mistakenly believed that sequential lineups were diagnostically superior to simultaneous lineups, that confidence was, at best, only weakly predictive of accuracy, and that suboptimal estimator variables (e.g., short exposure, weapon focus, etc.) reduced the reliability of eyewitness identifications. But none of that is true, which raises a question: What went wrong? The chief problem seems to be that, long ago, applied psychology became divorced from basic experimental psychology, particularly with respect to recognition memory. As a result, eyewitness identification research did not make use of theories from basic science that have proven their ability to accurately guide the interpretation of data, particularly signal detection theory. In recent years, signal detection theory has been applied to a variety of eyewitness identification issues. As it turns out, simultaneous lineups are superior to sequential lineups, initial confidence (before contamination) is highly predictive of accuracy, and estimator variables are largely irrelevant once initial confidence is taken into account.