Why Crossover Voters Are Not "Mischievous Voters":The Segmented Partisanship Hypothesis
Abstract
Although extensive crossover voting in open primaries has been documented, scholars have never found empirical support for the allegation that considerable mischievous crossover voting occurs in such circumstances. Formal theorists have argued that such voting is not consistent with the voting decision logic that most voters practice, others have inferred that most voters are not equal to the sophistication of the calculations requisite to strategic voting, and still others have hypothesized that voters are induced to cross over by candidates' personalities or issues. This exploratory study presents and examines the hypothesis that mischievous crossover voting is limited because most crossover voters are segmented-partisan identifiers who at some level of the federal polity actually identify with the party whose primary they invade. Segmented partisans do cast most crossover votes, and are more likely than consistent partisans to evaluate the opposing party favorably on thermometer scales, to evaluate favorably its past governing performance, and to defect to its ticket during general elections. The hostility toward the partisan out-group, posited in The American Voter and supporting studies, thus is not characteristic of most of those who cast crossover votes.