Performance under differential outcomes: Contributions of Reward-Specific Expectancies

Both acquisition and performance across delays are enhanced when each correct stimulus-response sequence in a conditional discrimination task is paired with a different reinforcing outcome, a procedure called differential outcomes (DO). It has been suggested that in the DO procedure, each discriminative stimulus comes to evoke an expectancy of a specific reward and this expectancy exerts stimulus control over choice behavior. Furthermore, expectancy control may reduce control by the discriminative stimulus itself under DO. In the study reported in this article, subjects under DO showed reduced ability to match based on the sample alone when expectancy could not be used to determine specific choices. The results generally support the contention that expectancy control is strong and sample stimulus control is weak under DO.