Can a toolbox of simple heuristic rules help explain lottery choices relative to expected utility theory (EUT)? While a mixture model of EUT plus heuristic rules will obviously fit data better than EUT only, given the small sample sizes, there is a danger of overfitting. Therefore, instead of goodness-of-fit measures, we focus on forecasting performance. Using two data sets of binary lottery choices and reasonable holdout subsets for testing forecasting performance, we find that the EUT-only model forecasts better than the toolbox mixture model with EUT. Even when the toolbox model with EUT fits the data significantly better, EUT-only forecasts better.

There is much interest in ambiguity-averse behavior under uncertainty, and many theories have been advanced to explain this. Empirical analyses of choices involving ambiguous options have typically used a representative agent model. We address the question of whether representative agent models are accurate approximations of reality or whether there is substantial heterogeneity in ambiguity preferences. In contrast to the representative agent model, we find that the vast majority of participants are not significantly ambiguity averse and that a significant proportion of participants are consistent with expected utility theory. This finding has important implications for the application of behavioral economics.

Experimental data have consistently shown diversity in beliefs as well as in actions among experimental subjects. This paper presents and compares alternative behavioral econometric models for the characterization of player heterogeneity, both between and within subpopulations of players. In particular, two econometric models of diversity within sub-populations of players are investigated, one using a model of computational errors and the other allowing for diversity in prior beliefs around a modal prior for the subpopulation.