Testing sincere versus strategic split-ticket voting at the aggregate level: Evidence from split house–president outcomes, 1900–2004

We compare aggregate level predictions of sincere and strategic (balancing) models of voting behavior using data on the number of districts exhibiting split outcomes of elections for President and for representatives in U.S. House over the period 1900–2004. Our prime focus is on the mean ideological differences between Republican and Democratic House members (DW-NOMINATE scores). If voting is sincere then we argue that this variable will be negatively linked to the proportion of split outcomes. In contrast, balancing models of ticket splitting [Fiorina, M.P., 1996. Divided Government. second ed. Allyn and Bacon, New York] suggest that the further apart are the parties, the greater will be the degree of ticket splitting among individual voters, since, ceteris paribus, there will be more voters close to neither party who need to balance off House and Presidential candidates to come closer to their own more centrist position. In testing these differing predictions we control for other factors that affect the proportion of split outcomes for these two offices at the congressional district level: (1) margin of presidential victory; (2) magnitude of incumbency advantage; (3) magnitudes (standard deviations) of ideological differences among House members of the same party; and (4) mean ideological differences between Republican and Democratic presidential candidates. In our multivariate analyses, we find more support for sincere than for strategic voting as a determinant of aggregate levels of split-ticket outcomes.