Strategic Voting in Plurality Elections

This paper extends the Calculus of Voting of McKelvey and Ordeshook, providing the first direct derivation of the conditions under which voters will vote strategically: choose their second-most preferred candidate in order to prevent their least-preferred candidate from winning. Addressing this theoretical problem is important, as nearly all empirical research on strategic voting either implicitly or explicitly tests hypotheses which originate from this seminal model. The formal result allows us to isolate the subset of voters to which strategic voting hypotheses properly apply and in turn motivates a critical reevaluation of past empirical work. In making this argument, we develop a unified and parsimonious framework for understanding competing models of tactical voter choice. The typology helps to elucidate the methodological difficulties in studying tactical behavior when faced with heterogeneous explanatory models and suggests the need for both theoretical caution and more precise data instruments in future empirical work.