A test of the response deprivation hypothesis in a multiple-response context.

Reinforcement contingencies defined by the response deprivation hypothesis were evaluated in two multiple-response contexts. In one context a high probability response was freely available, and in the other context a low probability response was freely available. In the presence of the low probability, freely available response, a consistent reinforcement effect was evident for all 3 moderately mentally retarded participants; however, in the presence of the high probability response, the reinforcement effect was attenuated. Results were qualitatively consistent with several recent mathematical descriptions of behavior emphasizing the importance of the context in which reinforcement contingencies are introduced.