Contrast effect in the use of punishment.

Abstract

Informal accounts of what has been termed a police riot indicate that more clubbing, prodding, and macing of protestors was done by men who carried a revolver than by those who did not. A contrast hypothesis was suggested whereby a strong but prohibited degree of punishment serves as an anchor in contrast with which lower degrees of permissible punishment are judged to be weaker than the case without the anchor. To compensate for the insufficiency generated when there is such a discrepancy between a strong punishment that cannot be used and a weaker level of punishment that can be used, trainers deliver more of the latter. In an experiment designed to test the contrast hypothesis, the subject trained a rat to press a bar by administering mild and slightly painful shocks during a 12-minute period. In a control condition, only these shock levers were present; in experimental conditions, a moderately painful or extremely painful lever was also present but was strictly forbidden. Results were in complete agreement with the hypothesis: In the first 6 minutes—within which the experimenter said other students had trained their animals—the subjects in the large discrepancy condition delivered significantly higher frequency and proportion of slightly painful shocks than the control subjects. In the second 6 minutes—beyond the normative training time—the difference in absolute frequency is maintained, but the higher proportion of slightly painful shocks was no longer significant. The relative increase in the use of mild shock is attributed to exogenous factors which combine with the contrast effect, namely, a fine tuning effect that occurs toward the completion of shaping or a shift to a normative level implied by the experimenter with a growing sense of ineffectiveness. Additional evidence for the hypothesis is presented, and alternatives are discussed.