Download files
Abstract
Many of the decisions people face involve outcomes that are both probabilistic (risky) and delayed in time (inter-temporal). This thesis seeks to increase our understanding of these risky inter-temporal choices by testing the assumptions of existing models of risky inter-temporal choice.
Existing models of risky inter-temporal choice assume that risky choices and inter-temporal choices are special cases of risky inter-temporal choice and that all three types of choice are made by a utility-comparison process. Chapter 1 tests these two assumptions by forcing participants to choose between a risky option and a delayed option that they had previously indicated were worth the same amount. According to a utility-comparison process participants should have no preference between these two options; instead they tended to prefer the delayed option.
In Chapter 2 a common utility-comparison process underlying risky and inter-temporal choice was directly tested by fitting a common model to both choice types. This common model was found, using both BIC and Bayesian model comparisons with informed priors, to perform better than fitting each choice type with a separate model.
Chapter 3 directly compared the performance of several models of risky inter-temporal choice. Hierarchical versions of each model were fit to a baseline dataset using Bayesian parameter estimation. The parameterizations obtained were then used to assess each models ability to predict behaviour under three manipulations- magnitude, immediacy and certainty. The Risky-Inter-Temporal-Choice-Heuristic (RITCH) model, an attribute-comparison model, performed best under all manipulations.
Finally Chapter 4 focused only on inter-temporal choices and attempted to explain the sign, magnitude and immediacy effects by taking into account differences between objective measures and subjective perceptions of time. Time perception was found to play a role in immediacy, but not sign or magnitude, effects.
Taken together these Chapters provide a range of behaviours any model of risky inter-temporal choice needs to explain. Furthermore Chapter 2 suggests a common model could underlie both risky and inter-temporal choices, while Chapters 1 and 3 suggest this model needs to involve attribute-comparison, rather than utility-comparison mechanisms. The General Discussion offers insights into how these results could be reconciled within a single theoretical framework.