PARTIAL AND CORRELATED REWARD IN ESCAPE LEARNING1

1959) have demonstrated that mag-nitude of reinforcement operates as a variable in instrumental escape learning. In those experiments rats were run in a straight alley in which the startbox and runway provided a continuous and intense electric shock, and the goalbox provided a continuous but weaker shock. It was found that running speed was faster the greater the reduction in shock intensity at the goal. More-over, these performance effects were reversible; when the amount of shock reduction was changed, performance shifted to the level appropriate to the new amount of shock reduction. Such studies have involved condi-tions in which S receives the same reward on every trial; however, of equal interest are those conditions in which the reward varies from trial to trial. Thus, a large reward (i.e., large amount of shock reduction) may be given on some trials and a small reward on others; or reward may be given immediately on some trials and delayed on others. Experiment I is an investigation of partial rein-forcement which may be considered as a limiting case of variable magni-tude of reward. Of further interest are those conditions in which varia-tions in the reward are contingent 1 This experiment was conducted during the writer's tenure on a predoctoral fellow-ship, MF-8S8S, granted by the Nationa