Negative Campaign Advertising: Demobilizer or Mobilizer

Martin Wattenberg, Craig Leonard Brians

Introduction

In the mid-1990s, a prevailing narrative in political science suggested that the increasing prevalence of negative "attack" advertising was poisoning the American democratic process. This "demobilization hypothesis," most notably advanced by Stephen Ansolabehere and Shanto Iyengar, posited that negative ads breed cynicism and alienate non-partisan voters, ultimately driving them away from the polls. In their 1996 study, Martin P. Wattenberg and Craig Leonard Brians challenge this consensus, arguing that negative advertising may actually serve as a powerful "mobilizer" by providing critical information and highlighting substantive differences between candidates.

The Demobilization Hypothesis vs. Real-World Behavior

The core of the demobilization argument rested on experimental findings where subjects were exposed to specific ads in controlled settings. Wattenberg and Brians question the generalizability of these lab results to the "noisy" environment of an actual election cycle. They argue that in the real world, voters are not passive recipients; they choose what to pay attention to. Furthermore, from a strategic standpoint, campaigns have a greater incentive to "convert" an opponent's supporter (a two-vote swing) than to simply "demobilize" them (a one-vote swing).

Methodology

To test their counter-hypothesis, the authors employed a two-pronged empirical approach:

  1. Survey Analysis (1992 NES): Using the 1992 National Election Study (NES), the authors analyzed open-ended responses from approximately 2,500 citizens. They identified respondents who recalled specific negative ads or expressed concerns about "mudslinging."
  2. Multivariate Modeling: To isolate the effect of ad recall, they used logistic regression to control for confounding variables such as education, age, party identification, and general media exposure.
  3. Aggregate Data Replication: The authors attempted to replicate Ansolabehere et al.’s analysis of 1992 Senate races. They compared the original researchers' dataset against official Federal Election Commission (FEC) returns to identify potential discrepancies in turnout and "rolloff" calculations.

Key Findings

The Information and Efficacy Benefit

Contrary to the idea that negative ads breed apathy, the data revealed that respondents who recalled negative ads actually exhibited higher levels of political efficacy. Wattenberg and Brians suggest that when an ad addresses a substantive issue a voter cares about, it signals that the politician is engaged with contemporary concerns, thereby encouraging participation. Negative ads often draw sharper distinctions on policy than "feel-good" positive ads, providing "peripheral voters"—those less engaged with politics—with the information necessary to justify voting.

Statistical Evidence for Mobilization

The 1992 NES data provided strong evidence for the mobilization effect:

  • Turnout Rates: Reported turnout for those who complained about negative ads was 82%, compared to 76% for those who did not.
  • Specific Recall: Among those who could recall a specific "attack" ad, the turnout rate rose to 84%.
  • Significance: Logistic regression confirmed that recalling negative advertising was a statistically significant predictor of voting (p = .05), even after controlling for education and campaign interest.

Correcting the Record

A pivotal moment in the research was the discovery of significant flaws in the aggregate data used to support the demobilization hypothesis. By comparing the previous researchers' dataset with official FEC returns, Wattenberg and Brians found that errors—such as the exclusion of absentee ballots and failure to count minor-party votes—had skewed "rolloff" rates. Once these errors were corrected, the supposed link between negative ad tone and lower turnout disappeared (p > .15).

Conclusion

Wattenberg and Brians conclude that the dangers of attack advertising have been greatly exaggerated. Rather than suppressing the vote, negative ads appear to provide a turnout advantage by highlighting the stakes of the election. While pundits may decry the "poisonous" tone of modern campaigns, the authors suggest that the volume of communication is a more significant driver of mobilization than the tone itself. Their work serves as a reminder of the importance of validating experimental findings with rigorous real-world data.

Summary updated

Abstract

With political campaigns becoming increasingly adversarial, scholars have recently given some much-needed attention to the impact of negative advertising on turnout.In a widely recognized Review article and subsequent book, Ansolabehere and his colleagues (1994, 1995) contend that attack advertising drives potential voters away from the polls. We dispute the generalizability of these claims outside of the experimental setting. Using NES survey data as well aggregate sources, we subject this previous research to rigorous real-world testing. The survey data directly contradict Ansolabehere et al.'s findings, yielding evidence of a turnout advantage for those recollecting negative presidential campaign advertising. In attempting to replicate Ansolabehere et al’s earlier aggregate results we uncover quite significant discrepancies and inconsistencies in their dataset. This analysis leads to the conclusion that their aggregate study is hopelessly flawed. We must conclude that attack advertising’s demobilization dangers are greatly exaggerated by Ansolabehere et al., while they completely miss negative political advertising’s turnout benefits -- at least in voters’ own context.
BibTeX
@inproceedings{wattenberg1996negative,
  author = {Martin Wattenberg and Craig Leonard Brians},
  title = {{Negative Campaign Advertising: Demobilizer or Mobilizer}},
  booktitle = {eScholarship},
  year = {1996},
}
Figure