Maximization of reinforcement by two autistic students with accurate and inaccurate instructions

The present study examines maximization of reinforcement by two autistic individuals under conditions of no instructions, accurate instructions, and inaccurate instructions. Accuracy of instructions and magnitude of reinforcement for differential responding in a choice paradigm were systematically varied across phases. Subject one maximized reinforcement across all three conditions in seven experimental phases. Subject two maximized across these same seven phases, but also experienced three additional phases. In two of the additional phases, subject two maximized reinforcement. In a ninth phase, when reinforcement was intermittent rather than continuous, he failed to maximize reinforcement. Implications of the results for the controversies surrounding the concept of rule-governed behavior are discussed.