Sobering up after "Partisan Intoxication or Policy Voting?"

Abstract

“Partisan Intoxication or Policy Voting?” raises questions central to understanding the extent to which individuals vote their partisanship and brings important attention to the potential observational equivalence between partisan and policy voting. In this response, I affirm some of Fowler’s arguments but also build upon existing studies to highlight that tests of the policy voting hypothesis need to seriously consider both the direct and indirect effects of partisanship to understand the relative role of policy versus partisanship. Such consideration is particularly significant as partisanship’s indirect effects can have troubling implications for democracy. I also reexamine the southern realignment and voters’ responses to hypothetical candidate policy positions, and when accounting for elite decision-making and complex information environments, I find voters respond less to candidate ideology and policy positions than suggested by Fowler’s original analyses. Together, my findings underscore the point that “policy voting and partisan intoxication are neither exhaustive nor mutually exclusive explanations” of voter behavior (Fowler 2019, 5). 1 I thank Michael Barber, Andrew Englehardt, Tyler King, Nolan McCarty, Seth McKee, and John Sides for their helpful feedback and assistance for this response. (2) “Partisan Intoxication or Policy Voting?” raises questions central to understanding the extent to which individuals vote their partisanship and casts doubt that “partisanship is a hell of a drug.” I encourage readers to seriously consider Fowler’s challenges and critiques, which shed important light on what we know about partisanship’s and policy’s role in voter decision making. In this response, I affirm some of Fowler’s assertions but also provide nuance to Fowler’s arguments to bring greater attention to an underemphasized point: “policy voting and partisan intoxication are neither exhaustive nor mutually exclusive explanations” of voter behavior (Fowler 2019, 5) a point that can be at times lost in this intoxicating debate. I respond to many of Fowler’s points in the order originally made. I first broaden Fowler’s challenge to voter behavior scholars and argue that we not only need to find measures of partisanship independent of policy preferences but also continue to understand partisanship’s indirect effects on voting via policy opinions. Second, I build on Fowler’s study of the southern realignment to show that even though older and younger voters experienced the same realignment, those who came of political age prior to the civil rights movement exhibit more intoxicated voting behavior, which would be puzzling if we were in a purely policy driven world. I further highlight that the southern realignment is a story of both voter and elite electoral behavior, and once we account for elite behavior, voters respond less to candidate ideology than suggested by Fowler’s original analyses. Third, I reexamine the survey experiment considered by Fowler and provide new evidence to support the policy voting hypothesis. I also discover survey respondents are less likely to exhibit policy motivated voting behavior in more complex information environments and respond to non-policy characteristics, such as candidates’ age, race, and religion, consistent with explanations of voting behavior rooted in social groups. I conclude by encouraging political scientists from each side of this debate to avoid becoming tribal ourselves, as we may miss important contributions from members of the other “tribe” in the partisanship and policy debate. The New Coke Challenge Fowler poses a challenge to voter behavior scholars, analogous to Krehbiel’s (1993) challenge to those who study Congress. Members of Congress likely join political parties based on their policy preferences, leading to observed correlations between party membership and roll-call behavior. Voters similarly may adopt