| Abstract: | In choosing between small, immediate and large, delayed reward, an organism behaves impulsively if it chooses the small reward and shows impulse control if it chooses the large reward. Work with nonhumans suggests that impulsivity and impulse control may be derived from gradients of delayed reinforcement. A model developed by Ainslie and by Rachlin suggests that preference for the rewards should be a function of when the choice is made: small reward with no delay may be preferred to large reward with delay X, but adding delay T to both alternatives should shift preference to the large reward. Three experiments investigated this preference reversal in humans, using termination of 90 dba white noise as the reinforcing event. Experiment 1 showed that under some instructional conditions 90-sec noise off with no delay was preferred over 120-sec noise off after a 60-sec delay, but that preference shifted to the large reward when a 15-sec delay (T) was added to both alternatives. Experiment 2 replicated this preference reversal under two conditions of large, delayed reward, and with three rather than two values of T. Experiment 3 confirmed this effect of T and showed that some humans committed themselves to the large reward when commitment could be made some time before presentation of the reward alternatives. These data support the Ainslie-Rachlin model and extend it to human choice behavior. |
