Changes in response strength with changes in the amount of reinforcement.

Abstract

A study to determine whether, as one aspect of Hull''s theory would indicate, the slopes of performance curves produced by changes in the amount of reinforcement would be the same, whether the change was an increase, or a decrease even to zero (experimental extinction). The subjects (Ss) in the experiment were 40 albino rats. They were trained to run down an 18-in. alley to obtain a 10% sucrose solution at one of five goal cups. After 20 such trials Ss were divided into 4 experimental groups which received a 5%, 10%, or 20% solution or nothing. The different groups behaved differently during this phase of the experiment on both of the measures available in the experiment. Because of an unsatisfactory degree of irregularity in the percentage of response measure, however, the analysis is largely confined to the speed-of-running data. With 1 exception, these data fit very well predictions drawn from Hull''s theory. As the theory predicts, shifting the amount of reinforcement alters the asymptotes approached by the learning curves. Moreover, the functions change at the same rate except for the group in which the shift was from 10% to 20% sucrose. In this case the increase in performance was much greater than the theory would predict. This finding suggests that the influence may be on performance rather than on learning as assumed in this paper. It does not, however, alter the major conclusion that extinction may be treated as the limiting case of a reduction in the amount of reinforcement.