Previous research claims that the number of parties affects the representation of social cleavages in voting behavior, election turnout, patterns of political conflict, and other party system effects. This article argues that research typically counts the quantity of parties and that often the more important property is the quality of party competition-the polarization of political parties within a party system. The author first discusses why polarization is important to study. Second, the author provides a new measurement of party system polarization based on voter perceptions of party positions in the Comparative Study of Electoral Systems, which includes more than 50 separate elections from established and developing democracies. Third, the author compares party polarization and party fractionalization as influences on cleavage-based and ideological voting and as predictors of turnout levels. The finding is that party polarization is empirically more important in explaining these outcomes.
Keywords: party polarization; party fractionalization; turnout; votingO ne of the most widely examined properties of party systems is the counting of the number of parties. A large body of research examines the merits of a two-party system versus a multiparty system (or a range of parties) and links the number of parties to the representation of social cleavages in voting behavior, turnout in elections, representation, and levels of political conflict (e.g., Blais & Dobrzynska, 1998;Kim, Powell, & Fording, 2006;Norris, 2004;Powell, 1982). Douglas Rae (1971) and others have examined how the number of parties influences strategic voting and other aspects of electoral behavior. At a systemic level, Lijphart (1999) includes the number of parties as a prime indicator for the development of consociational democracy and links a range of effects to the consociational model.