Abstract
The effects on behavioral variability of differential reinforcement of response sequences that differed from preceding response sequences were investigated in college students. Two-response sequences were used as response units in Experiment 1 and three-response sequences were used in Experiment 2. Development of variability was compared across the following two settings : (a) systematic changes in the number of preceding responses used as the reference for differential reinforcement, and (b) two kinds of yoked conditions, within -subject yoking and between-subject yoking, in which response sequences were not differentially reinforced but distribution and frequency of reinforcement were equal to the differential reinforcement condition. Both experiments showed high behavioral variability and more frequent responses that differed from preceding responses under the differential reinforcement condition, but less variability and less frequent responses under the yoked condition. The effect was smaller when the number of preceding responses used as a reference for differential reinforcement was one than when greater than one. These results suggest that the increase in behavioral variability was produced by the differential reinforcement of responses that differed from preceding responses.
