Every Exit Is An Entry Somewhere Else: An Experimental Investigation of Group Membership Adjustments and Policy Shifts


The extent to which party members vote together -- party unity -- is an important feature of the political landscape. Congressional scholars have extensively studied the impact of political parties on voting behavior. However, most prior studies have taken the existence of legislative parties as given and have ignored the question of why and how parties matter. This study begins with the presumption that individual group members have heterogeneous policy preferences, so that some are more distant from the party center than others. They are often confronted with situations in which they lack complete information about other members’ policy preferences and whether their party will represent their preferred policies or not. In situations where intra-party rules restrict the entry and exit of members, collective action problem can motivate tolerant individuals whose ideal points are far from the party center to band together in subgroups and ensure that they have more to gain by remaining in the party than leaving it. When decisions to make contribution to the collective policy outcome of the party are purely a function of the level of diversity in initial party composition and the level of tolerance for policy among party members, we expect to see that more diversity of preferences and tolerance for policies among members lead to increasingly volatile party policy decisions. The theoretical argument is fairly well developed, but testing it empirically requires data that are difficult or impossible to gather in the real world. Based on initial theory building and a pilot experiment conducted at EITM in summer 2011, we employ economics laboratory experiments to examine the ability of individuals to make contributions to the collective policy outcome of their group when decisions to contribute involve complex interaction between diverse preferences and policy tolerance of group members. This study challenges the conventional wisdom that largely treats political parties as unitary actors and party policy positions as static. Therefore, it is an important step to improve our understanding of party cohesion and tracing the lines between individual preferences and party policy decisions.