How do citizens respond to campaign events? We explore this question with a unique repeated measures survey design, fielded during the 2000 presidential campaign. We model transitions in support for the major party candidates following the party conventions and presidential debates. In the aggregate, Gore support increases following the conventions (but not the debates), while Bush support increases with the debates (but not the conventions). But there is considerable microlevel variation in the data: responsiveness to campaign events is greatest among Independents, undecided voters, and ?mismatched partisans,? but exactly how these groups respond differs for each event. Moreover, attitudes toward then President Clinton mediate the effect of the campaign events on voter preferences. Two primary conclusions follow: (1) rich data sets are required to observe the effects of campaign events; (2) the influence of campaign events on vote choice is conditional on previous preferences, partisan dispositions, and political context.