Assessing U.S. Senators' Response to a Competitive Primary Challenge with Increased Partisan Roll Call Voting

Much of the political punditry in the United States discusses the notion that facing a primary election results in legislators voting in a more partisan fashion in the legislature. A common refrain of this analysis is that facing the primary election constituency (Fenno 1978) or even the threat of facing the primary election constituency causes the senator to vote with the ideological extremes of the party in following sessions of congress. The literature on congressional elections has examined this area of research as it applies to the U.S. House, but few studies fully examine the impact of primary elections on roll call voting in the Senate. This study examines Senate primary elections to see how facing a primary, specifically a competitive primary, influences how a senator votes in the legislature in the first term following the election. This study specifically asks if senators who face a competitive primary challenge and win reelection vote with their party more often in subsequent congresses than senators who do not face a competitive primary challenge. Using OLS regressions and a number of control variables shown in the literature to impact roll call voting patterns, I examine the percentage of the vote that a senator received in her primary election compared to her party unity score in the Senate after the election. While my models demonstrate that facing a competitive primary correlates with a senator having a higher party unity score than senators who do not face a competitive primary, in all of my models the coefficient on this variable is not statistically significant. However, serving in the majority party and being elected in certain election years did have a statistically significant impact on a senator's partisan voting behavior. When testing an interaction effect between facing a competitive primary and serving in the majority party after the election, I also did not find a significant relationship between the interaction and a senator's change in party unity score. These findings add to our understanding of congressional elections by exploring an under-studied aspect of elections in the United States, and future research that adapts and refines the methodology of this study could further develop these results.