Reinforcement: A test of Premack's differential probability rules

Abstract

According to Premack's (1959) basic assumptions, reinforcement will occur only in a contingency relationship where the independent probability of the contingent response is higher than the independent probability of the instrumental response. To test this assumption, free-operant baselevels of running alone (phase I), drinking alone (phase II), and running and drinking simultaneously (phase III) were established for six rats. Independent response probabilities were calculated from these data. A run to drink contingency was then instituted which required each subject to increase the probability of the instrumental response above baselevel in order to maintain the contingent response at baselevel. In all cases reinforcement was observed regardless of the prior independent probabilities of the two responses.