Twenty schizophrenics, 20 manics, and 20 depressives were given two sets of multiple choice questions, one testing the subject's social knowledge of how people tend to act in a social situation and the other tapping their knowledge of events or objects which are relatively free of a social component. Schizophrenics were significantly impaired on the former set of questions relative to manic, and were significantly worse on both than depressives. It is suggested that these results represent an objective measure of the social naïvete of schizophrenics. The significant difference from manics indicates that the results are not merely the general effects of psychosis, particularly because the manics performed worse on an attentional test than the schizophrenics.