Students in a large state university were surveyed to determine their beliefs about penalty levels appropriate for persons found guilty of various acts of academic dishonesty. There were two samples: one was random, and the other consisted of students who were strongly believed to have cheated by copying answers on one or more multiple-choice tests the preceding academic quarter. (These latter subjects were identified by a powerful and thoroughly documented statistical procedure that evaluated the probability of observed answer correspondences.) It was hypothesized that the suspected cheaters would be more lenient generally, and this was borne out. However, differences for the most part were small in terms of practical effect. Only on questionnaire items concerning tests taken from a classroom contrary to an instructor's rule were suspected cheaters substantially more lenient. (Author)