抄録
In three experiments, rats were exposed to a concurrent-chains schedule that differed both in terms of overall and relative reinforcement rates. Overall reinforcement rates were specified by high versus low absolute values of choice phase schedules. Relative reinforcement rates were specified by the ratio of choice phase schedules across four conditions (1 : 1, 1 : 2, 1 : 5, and 1 : 9) both in high and low overall reinforcement rate conditions. Experiments 1 and 3 used a forced-choice procedure in which relative and overall reinforcement rates were arranged by a single variable-interval schedule. Experiment 2 used a free-choice procedure in which relative and overall reinforcement rates were arranged by two independent variable-interval schedules. The fixed-interval schedules in the outcome phase did not differ between alternatives in Experiments 1 and 2, whereas the fixed-interval schedules differed in Experiment 3. When log response ratios were plotted against log reinforcement ratios, sensitivity to relative reinforcement rate under the forced-choice procedure varied with changes in overall reinforcement rate when the fixed-interval schedules differed between alternatives (Experiment 3), but did not vary when the fixed-interval schedules did not differ (Experiment 1). However, sensitivity to relative reinforcement rate obtained under the free-choice procedure varied even when the fixed-interval schedules did not differ (Experiment 2). These results suggest that the effects of overall reinforcement rate, the choice procedure as well as the length of the choice and outcome phases should be incorporated in modeling choices between alternatives associated with different reinforcement rates
