Behavioral compliance and devaluation of the forbidden object as a function of probability of detection and severity of threat.

Abstract

This experiment was devised to test the prediction derived from cognitive dissonance theory that mild threats are more effective than severe threats in obtaining internalization of prohibitions. Previous research had revealed more devaluation of the forbidden object under mild threat, but obtained 100% behavioral compliance both under mild and severe threat. The following hypotheses were tested: (1) Mild threats lead to higher behavioral compliance both under high and low probability of detection, whereas severe threats lead to high compliance only under high probability of detection. (2) Highest devaluation of the forbidden object occurs under low probability of detection and mild threat, lowest devaluation occurs under high probability of detection and severe threat, and intermediate devaluation occurs under the other two conditions. The results were opposite to those predicted, but are of interest as a cross-cultural disconfirmation of hypotheses derived from cognitive dissonance theory. Both common sense and learning theory suggest that large rewards should be more effective than smaller rewards, and that severe punishment should be more effective than mild punishment. However, dissonance theory has considerably shaken this belief. Aronson and Carlsmith (1963), for instance, have shown that children show higher devaluation of a forbidden toy under conditions of mild threat than under severe threat. These results are explained by dissonance theory as follows: The cognition that one has refrained from playing with a toy is dissonant with the cognition that the toy is very attractive. Under severe threat of punishment the child has sufficient justification for not playing with the toy, but under mild threat the justification is insufficient; therefore, the child reduces dissonance by deval1 This study was supported by a grant from the American Psychological Foundation to the first author.