Comparing the diagnostic accuracy of suspect identifications made by actual eyewitnesses from simultaneous and sequential lineups in a randomized field trial

Abstract

Eyewitness misidentifications have been implicated in many of the DNA exoneration cases that have come to light in recent years. One reform designed to address this problem involves switching from simultaneous lineups to sequential lineups, and our goal was to test the diagnostic accuracy of these two procedures using actual eyewitnesses. In a recent randomized field trial comparing the performance of simultaneous and sequential lineups in the real world, suspect ID rates were found to be similar for the two procedures. Filler ID rates were found to be slightly (but, in the key test, nonsignificantly) higher for simultaneous than sequential lineups, but fillers will not be prosecuted even if identified. Moreover, filler IDs may not provide reliable information about innocent suspect IDs. Here, we use two different proxy measures for ground truth of guilt versus innocence for suspects identified from simultaneous or sequential lineups in that same field study. The results indicate that innocent suspects are, if anything, less likely to be mistakenly identified—and guilty suspects are more likely to be correctly identified—from simultaneous lineups compared to sequential lineups. Filler identifications are not necessarily predictive of the more consequential error of misidentifying an innocent suspect. With regard to actual suspect identifications, simultaneous lineups are diagnostically superior to sequential lineups. These findings are consistent with recent laboratory-based studies using receiver operating characteristic analysis suggesting that simultaneous lineups make it easier for eyewitnesses to tell the difference between innocent and guilty suspects.