Previous- and following-component contrast effects using a three-component multiple schedule

Behavioral contrast can be defined as an inverse relation between the rate of behavior in one component of a multiple schedule of reinforcement and the conditions of reinforcement in another component. In the present experiment, four pigeons pecked a key for mixed-grain reinforcers delivered by a three-component multiple schedule. During baseline conditions, the same variable-interval schedule was in effect during each component. In contrast conditions, the rate of reinforcement in the second component was increased or decreased. Behavioral contrast was usually observed in both the first and third components. Across-component contrast effects were neither larger nor more reliable when the change in reinforcement occurred in the following component than when it occurred in the previous component. Furthermore, in both the first and third components, the size of contrast was almost always largest early in the component and decreased as the component progressed. Both across- and within-component data were analyzed using the model proposed by Williams and Wixted (1986). The results question the adequacy of the model. The results also pose difficulties for several existing theories of contrast.