Robert Richards and Dr. Jay Goodliffe, Political Science Department
Introduction
In recent presidential elections, we have seen many long, drawn out nomination battles in party primary elections. The divisive primary hypothesis states that divisive presidential primaries hurt the party’s performance in the general election (Kenney and Rice 1987). Research on this claim has generally focused on the party’s performance in terms of vote shares (for example, see Henderson, Hillygus, and Tompson 2010), but there are problems with this approach. Aggregate vote data only allows analysis of two points in time (the primary and the general election), and does not permit an analysis of individual behavior. Panel survey data can give an individual-level picture of support for the party nominee as the campaign progresses, but panel data is very difficult to collect and suffers from the limitations of survey research. Do donors to losing primary candidates tend to rally around the nominee even after a divisive primary? Analyzing campaign in this way provides a different perspective that the other methods cannot give.
There are a couple of important assumptions underlying my research. First, it may be difficult to precisely define a primary as divisive or not. This project does not purport to identify primaries as divisive or not, but rather makes the assumption that the 2008 Democratic and 2012 GOP primaries were divisive. This assumption limits the conclusions of this research.
The second important assumption is that donors to presidential primary campaigns are representative of a candidates’ supporters generally. Donors do tend to be more politically active than non-donors (Magleby et al, forthcoming). However, the people most likely to be affected by divisive primaries are those who are politically active and politically well-connected. Thus, donors provide a powerful test of the effects of divisive primaries.
Methods and Procedures
This project used campaign finance data on the 2008 and 2012 presidential elections provided by the FEC. I analyze only the 2008 Democratic and 2012 GOP primaries. The data comes as donation data, which was then manipulated using a computer program to determine which donations came from the same donor1. This allows analysis at the level of the donor, which is actually a very new innovation in the study of campaign finance. Prior studies relied on the donation-level data, since the methods of probabilistic record matching had not yet been applied to this research (for example, see Brown, Powell, and Wilcox 1995, 149-151). Using a Cox survival model, I calculated the likelihood that an average donor would give to the party’s nominee given who he/ she donated to in the primary election.
Results, Discussion
The 2008 results show that Obama’s primary donors were indeed more likely to give to Obama in the general election than donors to other primary candidates. Donors to Hillary Clinton, the heir-apparent to the nomination until Obama’s surge in the beginning of 2008, were the least likely to rally behind Obama. This lends strong support to the divisive primary hypothesis.
A different story appears in the 2012 model. Donors to Mitt Romney’s primary campaign were not much more likely2 to give to him in the general election than were donors to his primary opponents. Thus, the 2012 case fails to support the divisive primary hypothesis. However, there is some debate about whether the 2012 primary divided the GOP in the same way that 2008 divided the Democrats. The recent Republican primary was certainly loud and had a very crowded field, but the frontrunner ended up winning (not the case in 2008 for the Democrats). Also, Barack Obama’s candidacy probably provided a strong incentive for the Republican establishment to rally around their nominee. There seems to have been a strong “anyone but Obama” sentiment among Republicans in 2012.
These results suggest that the divisive primary hypothesis may hold for campaign donors and politically aware individuals generally, but in a way that depends heavily on the broader electoral context. Factors such as incumbency, different campaign styles, dynamics of the electoral horserace, and “crowdedness” of the primary field probably play a key role. Further research on these aspects is necessary to make more definitive conclusions.
References
- Brown Jr., Clifford W., Lynda W. Powell, and Clyde Wilcox. 1995. Serious Money: Fundraising and Contributing in Presidential Nominating Contests. New York: Cambridge University Press.
- Giraud-Carrier, Christophe, Jay Goodliffe, and Bradley Jones. 2010. “Improving the study of campaign contributors with record linkage.” Unpubliushed working paper. Available at goodliffe.byu.edu/papers/linkage.pdf.
- Henderson, Michael, D. Sunshine Hillygus, and Trevor Tompson. 2010. “Sour grapes” or rational voting? Voter decision making among thwarted primary voters in 2008. Public Opinion Quarterly, 499.
- Kenney, Patrick J. and Tom W. Rice. 1987. The relationship between divisive primaries and general election outcomes. American Journal of Political Science 31, no. 1: pp. 31-44.
- Magleby, David B., Jay Goodliffe, and Joseph Olsen. Forthcoming. Message, Messenger, and Medium: Understanding the 2008 Presidential Donor Surge. Book manuscript.