Empirical Tests of Equilibrium Selection Based on Player Heterogeneity

We develop and test a model of initial play for symmetric normal-form games with multiple Nash equilibria. First, we specify an encompassing model that incorporates both equilibrium selection principles and boundedly rational behavioral models. We then design experimental games that can identify a variety of equilibrium selection principles. Model comparisons and hypothesis tests indicate that: (1) boundedly rational behavior is prevalent in initial-period play, (2) homogeneous population models can be strongly rejected in favor of heterogeneous population models, and (3) deductive selection principles add no statistically significant contribution to explaining the data. J.E.L. Classification Nos.: B41, C14, C51, C72, C90