Who Wins Nominations and Why? an Updated Forecast of the Presidential Primary Vote
Steger, Wayne P., Political Research Quarterly
This article builds on previous forecasts of the presidential primary vote by incorporating elite party endorsements and estimates of candidate electability. Elite endorsements are significant predictors of both parties' primary vote, but cash reserves are significant only for Democrats while Gallup poll results are significant only for Republicans. The model updated by the New Hampshire primary results indicates that the primary affects both parties' remaining primary vote, especially for Democrats. Republican nominations are largely predictable and determined mainly by effects occurring prior to the primaries, while Democratic nominations are relatively unpredictable before the primaries and are susceptible to momentum during the primaries.
Keywords: presidency; nominations; elections forecast; candidate endorsement
Political parties and presidential nominations are inseparably linked. The stakes are high for aspirants and party constituencies alike. The presidential nominee of a political party becomes the foremost spokesperson of the party (Tedesco 2001), the personified image of the party (Miller and Gronbeck 1994), the main selector of issues and policies for the party's general election campaign (Riker 1986), and a major factor in defining the ideological direction of a political party (e.g., Herrera 1995). Understanding why certain candidates get nominated is important for understanding several of the most important institutions in American politics. Why certain candidates get elected also matters for evaluating how democratic presidential nominations are. For example, how democratic are nominations if campaign contributors and party elites play bigger roles than does the mass party membership?
Previous models forecasting the presidential primary vote find that the main predictors are candidates' standing in preprimary national polls (Mayer 1996b, 2003) and cash reserves at the beginning of the primary season (Adkins and Dowdle 2000, 2001). These models, however, do not address two theoretically relevant factors that could improve the prediction of the presidential primary vote. One, forecasting models ignore the resurgence of the political party establishments as factors affecting presidential nominations. Steger and Davis (2000) and Cohen et al. (2002, 2003) argued that candidate endorsements by party elites affect primary voters' decisions by providing information about candidates' desirability. Second, forecasting models do not account for the effects of electability. At least some primary voters cast ballots strategically, taking into account their expectations for candidate performance in the general election (Brady and Johnston 1987; Abramowitz 1989; Abramson et al. 1992).
This study incorporates variables representing the central elements of these approaches into a forecasting model of the presidential primary vote. The study also compares a forecasting model using information from the preprimary period with a "momentum" model that incorporates preprimary year information updated by the results of the Iowa and New Hampshire nominating elections. The study is both explanatory and predictive. Given the importance of presidential nominations for the options of voters in the general election and the implications for governance after the election, it is worth figuring out who gets nominated and why. This study builds a forecasting model to shed light on these questions and implicitly offers a test of competing hypotheses through models that incorporate information derived from the various strands of theory in the presidential nomination literature.
The dominant perspective on the contemporary presidential nomination process holds that the primaries are the critical stage of the nomination (e.g., Aldrich 1980). The earliest models of presidential primary voting focused on patterns of candidate spending in each state (Goldstein 1978). Others focused on voter preferences for candidate characteristics and policies (e. …