This paper studies games when agents have second-order expected utility. It examines the theoretical predictions of the model and compares its performance in explaining behavior in experimental data on games with that of quantal response equilibrium.
