The effects of delay of reward on negative contrast effects associated with reductions in reward magnitude

Abstract

The results of three runway experiments in which rats received a reduction in the magnitude of delayed vs. immediate reward are reported. The results indicated that the negative contrast associated with a reduction in reward amount is more pronounced when all rewards are delayed. In the past three decades. considerable theoreticai attention has focused on the behavioral effects of changes in magnitude of reward in instrumental conditioning. Until quite recently. it appeared that \\hile decreases in reward magnitude produced contrast effects. comparable increases in reward did not (d. Black. 1968: Dunham. 1968). Thus. rats shifted from large to small reward typically show negative contrast effects (NCEs). depressed perfor­ mancc to small as compared with that of subjects trained and maintained on small reward. \\hile subjects shifted from small to large reward do not display posItI\e contrast effects (PCEs). facilitation of performance to large reward as compared \\ith that of subjects maintained on large re\\arci. Recent studies. howe\er. ha\'e clearly shown that PCEs attend increases in reward amount if all re\\ards are delayed follo\\'ing the instrumental response (Mellgren. 1972). While it is clear that reward delay interacts with increases in reward magnitude such that PCEs are larger and more reliable under delayed as compared \\ith immediate reward conditions. the theoretical basis of this interactive remains largely unexplored. Most frequently. the interactive has been attributed to the operation of a physiological or psychological effect on the performance of unshifted large reward (contro\) subjects (cf. BO\\er. 1961: Mellgren. 1972; Shanab & Cm·allaro. 1975). According to the ceiling hypothesis, PCEs are not observed under conditions of immediate reward because control subjects have achieved optimal response le\'els: lowering the performance level of control su bjects. through delay of reinforcement. provides a context in which the facilitative effects of increases in re\\ard magnitude may be seen. It should be clear that such a hypothesis does not attempt to