| In contrast to the progress psychology has made in developing descriptive models of decision under uncertainty, little progress has been made in describing decision over time. The experimental work in psychology has largely been limited to testing the normative model of decision over time found in the economic literature, the Discounted Utility model. I review the experimental evidence that bears on the appropriateness of employing the Discounted Utility model, or its variants, as a descriptive model of intertemporal choice. These models assume that all future outcomes are discounted according to a stable discount function and that the value of each outcome in a sequence is independent of the value of all other outcomes. I present evidence that neither of these assumptions is met in actual choices and, therefore, I conclude that these models fail as descriptions of choice over time. I then examine what information a successful descriptive model of choice must consider. I argue that a model that incorporates outcome information alone will fail to describe actual intertemporal choice because people's choices are often influenced by the situations that produce the outcomes, and, particularly, by the meanings that those situations carry. In support of this claim, I present evidence that people make different choices between the same objective outcomes depending upon the situations that produce those outcomes. Specifically, I show that sequences of outcomes produced by situations that (a) carry connotations of status, (b) are related to one another, or (c) are distinctive from other situations in one's life are more likely to be preferred in an increasing order than are the same sequences when those features are absent. Finally, I discuss some limitations of this approach and directions for further work. |
