Experiments on the Core: Some Disconcerting Results for Majority Rule Voting Games

In the context of spatial majority voting games, considerable experimental support exists for the core as a solution hypothesis when it is not empty (Berl et al., 1976; Fiorina and Plott, 1978; Isaacs and Plott, 1978). Specifically, these experiments show that if a simple majority voting game possesses a Core point-a point that cannot be defeated by a majority vote-subjects choose outcomes at or near this point. In this article, however, we present some additional experimental evidence to suggest that committee choice in simple majority rule games is not dictated solely by whether or not the core is nonempty. We conclude that, in the experimental context of open and free discussion, the performance of the core may be affected by the structure of the alternative space, and by factors that seem irrelevant to the theory, such as the structure of the dominance relation underlying the core in the social ordering.