Decisions, error and heterogeneity

Probabilistic choices or 'choice errors' may play an important supporting role in shaping observed risky choices, and a variety of restrictions on such choice probabilities play the role of identifying restrictions in empirical decision research. We test six restrictions in the context of common ratio effect data for a general class of theories of risky choice. Two restrictions, one for aggregate and one for individual data, are supported. But none of the heterogeneity restrictions that make the aggregate data model powerful survive scrutiny; nor do estimable parametric versions of the individual data model. There is room for new stochastic identifying restrictions in empirical decision research.