Effect of choosing an incorrect photograph on a later identification by an eyewitness.

Abstract

Possible biases in police practices of obtaining identifications of suspects from eyewitnesses were investigated in a test of the hypothesis that picking out an incorrect photograph from an array would affect witness performance on some later identification task. A psychology class was interrupted by a confederate. Twenty-five minutes later, half of the class (the control group) was dismissed, and the remaining subjects (the prior choice group) were asked to pick out the face of the confederate from an array of 12 photographs, none of which was actually that of the confederate. Four to 6 days later, all subjects were asked to pick out the face of the confederate from an array of 6 photographs, 1 of which was of the confederate and 1 of which, for participants in the prior choice group, was the incorrect face selected by that participant in the first identification task. Control subjects picked out the face of the confederate with better than chance accuracy (39% of the time,/? < .05), whereas prior choice subjects picked out the actual confederate no more often than chance (22% of the time) and picked out the face they had previously chosen from the array of 12 photographs significantly more often than chance (44% of the time, p < .005).