A common approach to the treatment of instrumental problem behavior is the introduction of an acceptable alternative source of reinforcement. However, when alternative reinforcement is removed or reduced, the target behavior tends to relapse. The relapse of a target response following the removal of alternative reinforcement has been termed resurgence. Shahan and Sweeney developed a quantitative model of resurgence based on behavioral momentum theory that captures both the disruptive and strengthening effects of alternative reinforcement on the target response. The quantitative model suggests that although higher rates of alternative reinforcement result in faster response elimination, lower rates of alternative reinforcement result in less relapse when removed. The present study was designed to examine the possibility that good target response suppression and less relapse could be achieved by beginning with a higher (rich) rate of alternative reinforcement and gradually thinning it such that a lower (lean) rate of alternative reinforcement is ultimately removed. Furthermore, the data obtained were generated to provide insight into how thinning rates of alternative reinforcement might be incorporated into the quantitative model of resurgence. Results suggest that rich rates of alternative reinforcement were more effective than lean or thinning rates of alternative reinforcement at response suppression during treatment, but when alternative reinforcement was discontinued, the group that experienced rich rates exhibited a substantial increase. Although lean and thinning rates of alternative reinforcement were not as effective at response suppression during treatment as rich rates, they still resulted in substantial decreases in the target response. Furthermore, removal of lean rates of alternative reinforcement did not result in substantial increase in the target response. Advantages and disadvantages of rich, lean, and thinning alternative reinforcement rates are discussed with respect to target response suppression and sensitivity to the end of treatment, and an alternative response rate is discussed. Although a small modification to the quantitative model was able to similarly account for data produced by rich, lean, and thinning alternative reinforcement, as it currently stands the model is unable to account for the finding that alternative reinforcement may not always serve as a disruptor relative to a no alternative reinforcement control.

No description available.