How do triers of fact infer the accuracy of eyewitness identifications? Using memory for peripheral detail can be misleading.

Eyewitnesses (n = 107) to a staged theft made identifications from a photo spread and then responded to 11 questions that measured their memory for peripheral details (e.g., how many pictures were in the room where the theft occurred?). Results indicated that witnesses who accurately identified the thief (n = 57) averaged fewer correct answers on the peripheral details test than did eyewitnesses who identified an innocent person (n = 32). The remaining witnesses (n = 18) made no identification. Cross-examinations of willing eyewitnesses who made an identification (47 accurate and 24 inaccurate) were conducted using a method that either scrutinized and documented the witnesses' memory for the peripheral details or did not. Subjects in the role of jurors (n = 96) viewed cross-examinations and indicated their belief that the witness had properly identified the thief versus an innocent person. The cross-examination that scrutinized and documented the witnesses' memory for trivial details lowered subjects' belief of the eyewitnesses' identification accuracy (from 72.9% to 47.9%). This discrediting effect was stronger for accurate eyewitnesses (75.0%-37.5%), however, than it was for inaccurate eyewitnesses (70.8%-58.3%). It is argued that subjects assumed a positive correlation between accuracy in identifying the thief and memory for peripheral details, which led to more discrediting of the witnesses who should have been believed than it did of the witnesses who should have not been believed.