Partisanship and voting behavior reconsidered in the age of polarization

• Published in: Electoral studies 2024-04, Vol.88, p.102752, Article 102752
• Main Authors: Stapleton, Carey E., Langehennig, Stefani R.
• Format: Article
• Language: English
• Subjects: Elections; Partisanship; Polarization; Vote choice
• ISSN: 0261-3794; 1873-6890
• DOI: 10.1016/j.electstud.2024.102752

Given the changes in partisan dynamics in the 21st century, is partisanship still the most important driver of voting decisions in American presidential elections as posited by Bartels’s (2000) pivotal work? To answer this, we extend Bartels’s original analysis on the changing importance of partisanship in presidential and congressional voting through the 2020 presidential election. We also provide a more fully specified model of vote choice to ensure any changes in the role of partisanship are not driven by an omitted predictor variable. In the bivariate model, we show that partisanship remains relatively flat since Bartels’s analysis ended in 1996, with a spike during the 2012 election. In the model including additional covariates, we show that the impact of partisanship on presidential vote choice is higher now than at any point since 1972, although the change in importance of partisanship since 1984 is negligible. This is driven largely by including political ideology in the vote choice model. We conclude that partisanship remains the most important driver of vote choice in presidential elections, even if its power has been slightly overstated in the past.