Beginning with the works of Anthony Downs (1957) and Duncan Black (1958), a large literature on voting in mass elections and voting in committees has developed that uses the mathematical tools of game theory to enhance our understanding of politics. As these political events are often conducted by secret ballot, it is not surprising that this work on spatial models of voting has taken as given that voters cast their ballots simultaneously. With a secret ballot, the order in which votes are cast is irrelevant because the last voter has no more information about the other ballots than the first voter. However, several important types of elections are sequential in nature; later voters can observe the votes or election results of earlier voters. This proposal investigates the effect that these informational differences have on the behavior of voters who are late in the voting sequence. This research develops a formal model in which voters start with private information about the quality of the candidates. Voters can supplement this information with inferences drawn from earlier votes in the sequence. Several extensions of this model are developed that capture some of the unique features of the presidential nominating process.
