Electronic Theses and Dissertations UC San Diego

Abstract: Behavior analytic discussions of self-control have focused on temporal discounting as the primary index of selfcontrol behavior. In this measure, choice between discrete, mutually exclusive, delayed outcomes is observed. The outcome of this self-control measure is well described by hyperbolic models of intertemporal choice. In the last ten years, a second measure of self-control has been developed, the single-player iterated prisoner's dilemma. This measure is based on a series of choices between two discrete outcomes, one of which is larger in the immediate trial, but leads to a lower overall payout while the smaller immediate outcome leads to a higher overall payout. Recent theoretical and empirical reports suggest that while the two self-control measures are unique, the same process modeled by hyperbolic discounting may also influence performance in the single-player prisoner's dilemma. To test this possibility, six experiments compared the correlation between performance on temporal discounting tasks with performance in iterated prisoner's dilemma games. Results from these experiments suggest that : (1) previous reports of a correlation between hyperbolic discounting and performance on the iterated prisoner's dilemma are replicable, but (2) this correlation is not a universal characteristic of these tasks. Rather, the relationship between the two measures is dependent on specific experimental parameters. Further explorations of the necessary parameters found that the relationship between discounting and performance on the iterated prisoner's dilemma is influenced by the type of IPD game, type of discounting task, and the probability of reciprocation used in the IPD game