The impact of discredited evidence.

How do people revise their beliefs when evidence is discredited? In three studies, mock jurors read simplified criminal cases and then judged the probability that a suspect was guilty on the basis of sequentially presented evidence. Study 1 showed an extension effect: When two items of incriminating evidence were presented, a subsequent discrediting of the second item also lessened belief in the first item, irrespective of whether it was directly related to the discredited item. Study 2 showed that this effect depended on the order of evidence presentation: When the discrediting evidence was presented early, rather than late, in the sequence, there was no extension to unrelated items. Study 3 showed that the extension effect held only when items of evidence were both incriminating or both exonerating, but not when they were mixed. To explain these findings, we draw on coherence-based models of juror reasoning and propose that people group evidence according to its direction with respect to the guilt hypothesis.