WikiSummary, the Social Science Summary Database

Ansolabehere and Iyengar: Riding the wave and claiming ownership over issues

From WikiSummary, the Free Social Science Summary Database


Ansolabehere and Iyengar. 1994. Riding the wave and claiming ownership over issues. Public Opinion Quarterly 58:335-57.


The authors examine two competing hypotheses, "issue ownership" and "riding the wave," as well as a possible interaction between them. They find evidence only for the "issue ownership" hypothesis (without interaction).

Compare the "issue ownership" hypothesis with Conover and Feldman's (1981) argument.


H1: Issue Ownership

Republicans and Democrats are associated with certain issues. For example, Republicans "own" national security and crime-fighting, whereas Democrats "own" women's rights and civil rights. Two-part hypothesis: When candidates run ads on an issue their party "owns," they will see the largest gains in voter support. Similarly, when news broadcasts emphasize certain problems, the candidate who "owns" these problems will gain support.

This hypothesis rests on two psychological premises. First, the "confirmatory bias" says that people are more receptive to messages that confirm their existing stereotypes. Second, messages are more persuasive if they are more credible, and candidates are more credible if their party "owns" an issue.

H2: Riding the Wave

Candidates are better off if news messages and candidate advertisements are consistent and mutually reinforcing. Thus, by "riding the wave"--i.e., by advertising on issues that are currently in the news--candidates can establish their credentials on these issues and show that they care about them. Thus, campaign advertising is most persuasive when it concerns an issue currently dominating the news.

Interaction: H1 * H2

A three-way interaction between news coverage, advertising, and candidate partisanship: If you see news coverage, and you see an advertisement, and the advertisement is from a candidate who "owns" an issue, then you will be more likely to support the advertiser.


Study 1: Crime and unemployment

An experiment. The authors show a newscast that includes stories about crime or unemployment. Inserted into the newscast (during a commercial break) is a campaign advertisement. All ads show the same video and have the same text, but differ in whether they ultimately endorse the Republican or Democratic candidate for California Senate in 1992 (actual candidate names are used during an ongoing campaign season).

The variables:

  • X1: Whether there was news coverage of the issue being studied (either crime or unemployment)
  • X2: Whether there was a relevant campaign advertisement inserted into the broadcast
  • Y: Perceptions of the two candidates
  • Controls: partisanship, income, age, gender, etc of respondents.

Study 2: Sexual harassment

Similar design, but this time it's not comparing Republicans against Democrats. To try to separate gender bias from party bias (since the two Democrats running for Senate were both women and the two Republicans were both men), the second study compares Feinstein (woman), Boxer (woman), and Bill Clinton. The issue was sexual harassment (the Anita Hill accusations).


Only the "issue ownership" was supported. Neither H2 nor the interactive hypothesis was supported. Additional news coverage of crime helps Republicans; additional coverage of unemployment helps Democrats. Moreover, Republican advertisements on crime had an effect, but their advertisements on unemployment did not (ditto for Democrats and unemployment). With sexual harassment, Boxer and Feinstein could exploit the issue better than Bill Clinton.

Comments and Criticism

Perhaps the authors are naive to assume that a single news broadcast within their experiment can actually affect whether the candidate ads are "riding the wave." It could be that participants already know what "the wave" is, based on their life outside the experiment, and that this one additional news broadcast was insufficient to test the "riding the wave" hypothesis.