Resistance to extinction after varying amounts of discriminative or nondiscriminative instrumental training.

Abstract

This experiment investigated the hypothesis that acquisition level and resistance to extinction would be monotonically related for a simple (nondiscriminative) instrumental response (bar-pressing in a Skinner box), but the function would be nonmonotonic for a comparable discriminative response (successive brightness discrimination). Four groups of Ss were trained on the simple instrumental response and allowed 200, 400, 800, or 1,600 reinforced responses, 50 per day. The same numbers of reinforced responses were given to four corresponding groups of discriminatively trained Ss, the procedure employed with the latter differing only in the insertion of occasional S delta periods. All groups were exposed to one 10-min. extinction period (in the former S delta) on each of 5 successive days. Trend analyses of the data supported the initiating hypothesis; and as expected, the discriminatively trained Ss were, as a group, more resistant to extinction than the Ss trained on the simple instrumental response.