Nonminating Presidential Candidates: The Primary Season Compared to Two Alternatives

Abstract

Study of the impact of the McGovern-Fraser reforms on presidential nominations, though extensive, has focused largely on (1) the effects of reduced control by party leaders, and (2) the nature of inputs including votes, media, and money Despite the unusual practice of holding sepa- rate contests in multiple states to select a single nominee, little research explicitly addresses how such sequential contests function to transmit voters' preferences. I use computer simulations to tackle this seemingly intractable question and compare the results of a primary season to two alternatives, a single national primary and a national primary followed by a runoff. Using first a uniform distribution of voter preferences and then a distribution based on ANES data, I explore the impact different deci- sion rules have on (1) a party's ability to satisfy its own voters' prefer- ences, and (2) its competitiveness in the general election. I find that, on average, the primary season does better both at estimating party voters' preferences and at selecting candidates nearer the general electorate median. A primary followed by a runoff yields results close to those for the primary season, whereas a single primary generates results not dis- similar to those created by random selection. I conclude with a discus- sion of the implications of these findings for the system used to nominate presidential candidates.