When Primary Campaigns Go Negative: The Determinants of Campaign Negativity

• Author: David A. M. Peterson,Paul A. Djupe
• Published date: 01 March 2005
• Date: 01 March 2005
• DOI: http://doi.org/10.1177/106591290505800104
• Subject Matter: Articles

45

Negative campaign tactics have received tremendous

amounts of attention from scholars. This attention

has focused largely on two questions: historical

trends in negativity (e.g., Jamieson, Waldman and Sherr

1998; Kaid and Johnston 1991; West 1997) and the impact

negativity has on turnout (Ansolabehere, Iyengar, and

Simon 1999; Ansolabehere, Iyengar, Simon and Valentino

1994; Kahn and Kenney 1999; Lau, Sigelman, Heldman,

and Babbitt 1999; Wattenberg and Brians 1999). The find-

ings of the former literature are clear—negativity has

increased in all types of elections in recent years (though the

magnitude of the increase is an unresolved debate). In the

latter literature, the linkage is still an open question—in

some studies negativity boosts turnout, in others it

depresses turnout.

A largely separate, but equally important, literature fur-

thers our understanding of how the primary contest actors

behave (Abramowitz 1989; Abramson, Aldrich, Paolino,

and Rohde 1992; Bartels 1988; Cooper and Munger 2000;

Geer 1989; Gerber and Morton 1999; Kenney and Rice

1987; Mutz 1995; Popkin 1991; Stone, Rapoport, and Atke-

son 1995). The dominant questions in this literature focus

on the vote choice of individuals, and the forces of momen-

tum for candidate evaluations and campaign contributions.

Largely absent from both of these literatures is much more

than conjecture (Skaperdas and Grofman 1995) and anec-

dotal evidence about patterns of negativity within primary

campaigns—why do primary campaigns turn negative

when they do (though see Haynes and Rhine 1998 for pres-

idential campaigns)? Given the importance of negativity in

both primary and general election campaigns, we find this

oversight surprising.

We examine which races are likely to go negative and

when during the campaign candidates will attack their

opponents, with an eye toward understanding some of the

structural and dynamic elements driving campaign choices.

That is, the content of campaigns is predictable based on the

general dynamics of campaigns and the types of challenges

and opportunities candidates face. We test our hypotheses

on every contested Senate primary from 1998.

This article proceeds as follows. In the next section, we

lay out our theoretical expectations for when we expect to

see candidates attack their opponents. We then briefly

describe the data, a content analysis of media coverage, and

the method, a binary time series cross sectional model.

Finally, we present our results and conclude with a brief

statement about what we can learn about candidate deci-

sionmaking in campaign context.

PREDICTING NEGATIVITY

We begin with the assumption that the campaign choices

candidates make are driven by the desire to attain elective

office. The standard advice given to candidates by one

prominent political consultant includes: “Go negative early,

often, and right through Election Day. . . . Define your

opponent, before he or she can define him/herself or you.

. . . If attacked hit back even harder. . .” (Kamber 1997: 46-

7). Put simply, the advice most candidates get is to attack

from the start and never let up.

The literature on negativity in general elections suggests

that candidates take this advice. Candidates use negative

campaigning as a tool to drive down the positives of their

opponent and demobilize voters of the other party

(Ansolabehere and Iyengar 1995). This strategy, however,

has the potential to backfire as voters may view the attacker

more negatively as well. Of course, this strategy is a more

likely choice in some races than others. Because negative

campaigning is a dangerous choice, it is more likely to be

used in competitive races (Damore 2002; Hale, Fox, and

Farmer 1996; Kahn and Kenney 1999; and Theilmann and

Wilhite 1998). Additionally, once the negativity begins the

tone of the race can spiral downward—early negative cam-

paigning the tone for the rest of the race (Kahn and Kenney

1999; though see Damore 2002).

The electoral history of the candidates in the race matters

as well. Because they have a clear record, usually greater name

When Primary Campaigns Go Negative:

The Determinants of Campaign Negativity

DAVID A. M. PETERSON, TEXAS A&M UNIVERSITY

PAUL A. DJUPE, DENISON UNIVERSITY

Standard investigations of both campaign negativity and primary elections focus on either the electoral insti-

tutions or the primary voters. In this article, we begin to explore the factors affecting the content of the infor-

mation environment voters face by examining the effects of timing and electoral context on which primary

races are likely to become negative and when. Using a content analysis of newspaper coverage of every con-

tested Senate primary in 1998, and binary time-series cross-sectional methods, we demonstrate that negativ-

ity is an interdependent function of the timing of the race, the status of the Senate seat, and the number and

quality of the challengers in the primary.

Political Research Quarterly, Vol. 58, No. 1 (March 2005): 45-54

PRQ_March05_II 3/24/05 9:18 AM Page 45