Mock jurors were given information about a stabbing incident which varied in incrimination value (high or low) and a psychological profile of the perpetrator which varied in the degree of indicated insanity (high, medium, or low) in four replications of a 2 × 3 design. In one replication, the information was presented in the context of a sanity hearing where guilt was not an issue. In three other replications, the context was a criminal trial in which the defendant pleaded not guilty on the facts, not guilty by reason of insanity (NGI), or both. Judgments in the insanity plea replication were significantly affected by incrimination information even though sanity is legally the only issue. Insanity information predictably affected judgments in the sanity hearing but had only weak effects in the insanity plea context. It was concluded that the judgment dimension which equates insanity with not guilty results in a conflict which subjects resolved by viewing the dimension as more of a guilt dimension.