Expert testimony concerning the reliability of eyewitness identifications is generally restricted to factors believed to affect a significant percentage of witnesses. This study explored evidence for one such factor that has been labeled unconscious transference. This phenomenon occurs when an eyewitness identifies a person as having perpetrated a crime when, in fact, the witness encountered the person as a bystander or under circumstances unrelated to the crime event. For the study, 650 undergraduate students viewed a videotaped, simulated robbery. Participants were significantly more likely to select a bystander from a photoarray than the actual perpetrator, and they were more confident in their misidentifications. Participants who were shown a photoarray without the bystander present were over six times more likely to select the perpetrator than observers who were shown an array that included the bystander. The significant misidentifications of the bystander were eliminated when the event was restaged to show both bystander and perpetrator for a few seconds in the same frames of the video. Tables, references