Heuristics for risky choice

Expected utility (EU) theory has been exerting a dominating influence on risky-choice research. The theory is based on the premise that people behave as if they multiply probabilities and value and then maximize. Many experiments, however, challenged EU theory as a descriptive model of human reasoning (e.g. Allais, 1953; Kahneman & Tversky, 1979). To account for these deviations, new theories emerged that modified the theory while retaining the original EU scaffolding. Examples include, among many others, cumulative prospect theory (Tversky & Kahneman, 1992) and the transfer of attention exchange model (Birnbaum & Chavez, 1997). To account for evidence at variance with EU theory, Brandstätter, Gigerenzer, and Hertwig (in press) proposed a sequential heuristic, the priority heuristic, that offers an alternative to the assumption that people always compute trade-offs in the sense of multiplying beliefs and value. Tested on 260 choice problems from four diverse data sets, the priority heuristic predicted choices better than previously proposed heuristics and better than three modifications of EU theory. The overarching goal of the present project is fourfold: to examine (i) the cognitive processes directly by using process tracing procedures, such as mouse-lab) (Payne, Bettman, & Johnson, 1993) and think-aloud protocols (Ericsson & Simon, 1984); (ii) variants of the heuristic that are able to predict interindividual differences; (iii) the cognitive processes that are involved in situations where the priority heuristic has failed to predict; and (iv) the heuristic`s capacity to predict real-world choices. The planned studies are designed to (i) substantiate the heuristic as a process model and (ii) explore both its scope and boundary conditions.

Expected utility (EU) theory has been exerting a dominating influence on risky-choice research. The theory is based on the premise that people behave as if they multiply probabilities and value and then maximize. Many experiments, however, challenged EU theory as a descriptive model of human reasoning (e.g. Allais, 1953; Kahneman & Tversky, 1979). To account for these deviations, new theories emerged that modified the theory while retaining the original EU scaffolding. Examples include, among many others, cumulative prospect theory (Tversky & Kahneman, 1992) and the transfer of attention exchange model (Birnbaum & Chavez, 1997). To account for evidence at variance with EU theory, Brandstätter, Gigerenzer, and Hertwig (in press) proposed a sequential heuristic, the priority heuristic, that offers an alternative to the assumption that people always compute trade-offs in the sense of multiplying beliefs and value. Tested on 260 choice problems from four diverse data sets, the priority heuristic predicted choices better than previously proposed heuristics and better than three modifications of EU theory. The overarching goal of the present project is fourfold: to examine (i) the cognitive processes directly by using process tracing procedures, such as mouse-lab) (Payne, Bettman, & Johnson, 1993) and think-aloud protocols (Ericsson & Simon, 1984); (ii) variants of the heuristic that are able to predict interindividual differences; (iii) the cognitive processes that are involved in situations where the priority heuristic has failed to predict; and (iv) the heuristic`s capacity to predict real-world choices. The planned studies are designed to (i) substantiate the heuristic as a process model and (ii) explore both its scope and boundary conditions.