Fitting Human Data with Fast, Frugal, and Computable Models of Decision-Making

Abstract

There is a long-standing debate regarding the extent to which human decision-making processes are “rational” in the formal sense of the word. One approach to evaluating and contrasting the psychological validity of multiple theories of decision-making is to statistically calculate the predictive validity of the algorithms and equations that instantiate those theories against actual human performance data. Previously, we presented the predictive validity results of linear regression and fuzzy logic models of decision-making data generated by participants in a military-like task (Campbell, Buff, Rhodenizer, & Dorsey, 1999). In this paper, we evaluate the predictive validity of four “fast and frugal” algorithms against these same data. While no single approach will resolve the debate regarding the nature of decision-making processes, each adds a piece to the puzzle. A better understanding of decision-making processes could improve our ability to train and provide performance support for all types of decision-makers.