Abstract
In this experiment, using a lever-pulling response in humans, 8, 12, and 16 reinforcements in acquisition were combined with 67% reinforcement administered either in a pattern, which was random, random with the last 4 correct responses reinforced or random with the last 6 correct responses reinforced. On the last correct response reinforcement was terminated and S allowed to continue responding as long as he wished. The total number of lever pulls and the number of correct responses made in this period were the dependent measures. It was found that the pattern of reinforcement did not significantly affect resistance to extinction as reflected by either of these measures, whereas Ss who had received 8 reinforcements were more resistant to extinction than Ss who had received either 12 or 16 reinforcements.