Abstract
Study objective: This present study was designed to determine the effects of a differential reinforcement procedure on response variability. The procedure reinforced sequences that differed from those emitted during previous trials. The sequences were either 2 or 4 responses long. Design: A between-subjects, yoked design was used in which the reinforcement rate during the yoked condition was equal to the reinforcement rate during the differential reinforcement condition. Setting: The participants, seated in an experimental room pressed a 2-button computer mouse. Participants: The participants were 28 college students, 15 men and 13 women, randomly assigned to 1 of 4 groups. Independent variables: (1)A differential reinforcement procedure was used, in which points were earned whenever the current sequence differed from all sequences emitted during the preceding N trials. N was changed systematically. (2)The length of the response sequences was the second independent variable. Measure: The relative frequency of all possible response sequences and the periodicity of the generated response sequences were used to measure response variability. Results: Response variability was higher in the differential reinforcement condition than in the yoked condition. When the variability requirement(N)increased in the 4-response condition, the periodicity of the sequences decreased; the participants seemed to emit sequences randomly. Conclusion: Response variability increases with the reinforcement of variability and with sequence length.
