Two tests of the Sheffield hypothesis concerning resistance to extinction, partial reinforcement, and distribution of practice.

Abstract

Two experiments were performed to test Sheffield''s interpretation of resistance to extinction following partial reinforcement. Sheffield''s design and procedure were duplicated and the experiments differed from hers only in the motivational-reward conditions employed. In Expt. 1 hungry Ss were rewarded with dry food, while in Expt. 2 thirsty Ss were rewarded with water. Sheffield had employed a wet mash as reinforcement. The results were not in agreement with Sheffield''s, and did not support her interpretation. Partial (50%) reinforcement groups were more resistant to extinction than continuous (100%) groups regardless of intertrial interval in acquisition or kind of reinforcement. These differences were significant only in an analysis of covariance which adjusted the groups for differences in acquisition levels of performance. An associative mechanism was postulated to account for the results. Although, in Expt. 1, massing of trials increased resistance to extinction, the effect of distribution of extinction trials on resistance to extinction depended also upon the distribution of acquisition trials: the switched groups were less resistant to extinction than the unswitched ones. In Expt. 2, massed extinction groups were more resistant to extinction than spaced ones after both massed and spaced training and after both partial and continuous reinforcement, although here too generalization decrement seemed to be a factor.