We investigate the amount of negative campaigning in the 1998 senatorial primaries and the ramifications of negative campaigning on primary turnout and general election outcomes. A large literature has developed to show whether primary divisiveness has significant consequences for electoral outcomes, though we do not have much knowledge about what primary divisiveness entails (Bernstein 1977: 540). We employ a holistic measure of campaign negativity measured by coding newspaper articles three months prior to the primary to uncover how much negativity exists in senatorial primaries, which campaigns turn negative, and the relationship between primary negativity and several campaign factors. We find that primary divisiveness is strongly related to campaign negativity, negativity boosts primary turnout, while divisiveness depletes a nominee's general election fortunes.
The adoption and spread of the nominating primary election at the end of the Progressive era spawned an industry of investigation into its effects on the political party, the voter, and governance. The classic work on state government by V. O. Key, Jr. (1956) presented convincing evidence that primaries erode the party organization, change the face of representation, relocate the locus of competition from interparty to intraparty, and perhaps handicap those emerging from divisive primaries. Research thereafter has built upon those themes, with a rather extensive literature exploring, among other facets, the impact of contested primaries on a nominee's fate in the general election (Hacker 1965; Bernstein 1977; Kenney and Rice 1987; Abramowitz and Segal 1992).
In another related branch of electoral research, scholars over time increasingly have taken up the topic of negative campaigning, mostly through the medium of negative advertising (Ansolabehere and Iyengar 1995). The questions have focused on whether negative advertising demobilizes the electorate (Ansolabehere et al. 1994, 1999, including the exchange in the APSR), degrades campaign discourse (Jamieson 1992), and affects candidate choice (West 1997).
We integrate these two literatures to the advantage of both. Through the use of aggregate election returns in the divisive primary literature, some of the connective tissue of primary division to nominee success is lost. We have little knowledge of what divisiveness means (Bernstein 1977: 540). While we would assume that close primary contests are more negative than barely contested nominations, that relationship is untested. Few references are substantiated concerning what valuable resource division might deplete, which might include trampled voter party loyalties, emptied campaign warchests, personal and policy credibility, favorability, and electability (though see West 1994, 1997). Several pieces have shown, however, how divisiveness alienates intraparty challenger activists (Johnson and Gibson 1974; Buell 1986; Miller, Jewell, and Sigelman 1988). To begin to fill some of these gaps, we attempt to document the relationship between primary divisiveness, campaign negativity, and a few of these campaign resources: general election support and primary turnout.
Why should we expect the mere closeness of the primary outcome to influence behavior in the general election? The effect that the primary campaign experience might have on the general election stems mostly from the content of the primary campaign. Thus, we believe that a measure of campaign negativity, as we refer to it, derived from the substance of the campaign environment provides a more detailed picture of the primary campaign than traditional aggregate outcome measures.
Finally, this work provides new leverage on the relationship between campaign negativity and turnout by testing theories on a new set of elections: primaries. Ansolabehere et al. (1994) and Ansolabehere and Iyenger (1995) suggest that voters are demobilized because the negative campaign denigrates the political process. …