A model of probabilistic choice satisfying first-order stochastic dominance

This paper presents a new model of probabilistic binary choice under risk. In this model, a decision maker always satisfies first-order stochastic dominance. If neither lottery stochastically dominates the other alternative, a decision maker chooses in a probabilistic manner. The proposed model is derived from four standard axioms (completeness, weak stochastic transitivity, continuity, and common consequence independence) and two relatively new axioms. The proposed model provides a better fit to experimental data than do existing models. The baseline model can be extended to other domains such as modeling variable consumer demand. Key words: probabilistic choice; first-order stochastic dominance; random utility; strong utility History: Received March 8, 2010; accepted October 19, 2010, by Peter Wakker, decision analysis. Published online in Articles in Advance January 28, 2011.
1. Introduction Empirical studies show that decisions under risk are often contradictory (e.g., Hey and Orme 1994). Such decisions are best captured through a model of probabilistic choice. Popular models [...]