Reducing Pausing at Rich-to-Lean Schedule Transitions: Effects of Variable-Ratio Schedules and Noncontingent Timeouts

Abstract

Transitions from relatively rich to lean conditions of reinforcement often produce extended pausing because this transition is relatively more aversive than other transition types (e.g., rich-rich, lean-lean, and lean-rich). In applied settings, aversive rich-lean transitions may underlie maladaptive aberrant behavior (e.g., self-injury, aggression, and severe stereotypy). Reducing the aversiveness of this critical transition is of basic and applied interest. The length of the pause may be used as an index of the aversive stimulation at rich-lean transitions and as an analog measure of aberrant behavior in clinical settings. Experiment 1 investigated the effects of arranging rich-lean transitions between variable-ratio (VR) schedules or random-ratio (RR) schedules on pausing at rich-lean transitions in four pigeons. Variable- and random-ratio schedules were equally effective at reducing pausing at rich-lean transitions. These results suggest that the aversive stimulation at rich-lean transitions was reduced with variable schedules. The applied implication is that aberrant behavior should be less likely to occur at rich-lean transitions when variable schedules are arranged. In Experiment 2, four transitions were separated by a timeout period imposed between the end of the reinforcer and before the start of the next multiple-schedule component. During the timeout, the response key was darkened and the reinforcement schedule was suspended until the timeout interval had elapsed. Across different timeout durations, ratio sizes, and probes, the timeout produced inconsistent within- and between-subject results. Inconsistent results may be attributable to the timeout reducing the number of transitions completed, performance failing to meet the quantitative and qualitative stability criteria, and insufficient reinforcement in the lean-schedule component. An interesting possibility raised by Experiment 2 is that the multiple schedule may have "chain-like" features (e.g., access to a rich-schedule component), which may maintain responding in the lean component when there is insufficient reinforcement.