Learned helplessness and the detection of contingency: A direct test

Two studies tested a basic hypothesis of the learned helplessness model: That performance deficits associated with exposure to uncontrollable outcomes are directly mediated by an individual's perception of response‐outcome independence. In the first experiment 48 subjects were exposed to noise bursts. For one experimental group, the termination of the noise was response‐contingent. For five other groups, noise‐burst termination was independent of subjects' responses. These five groups varied in the number of trials on which they received positive feedback: As predicted, subjects overestimated the amount of control they had over noise termination as a positive linear function of the amount of noncontingent positive feedback they received. Although subjects exposed to either noncontingent positive or negative feedback showed subsequent performance deficits on an anagrams task, the expected relation between perceived control and subsequent performance failed to emerge. These findings were replicated in a second experiment. In addition, subjects' locus, stability, and globality attributions failed to predict subsequent performance. These results call into question the central premises of helplessness theory: That perceived uncontrollability and causal attributions mediate learned helplessness.