|Page 2 of 2 <|
Negative Ads Turnoff Voters, Enthrall News Media
Not all voters reacted uniformly to campaign ads. An axiom of political campaign strategists is that advertising fires up the "base;" Democrats and Republicans both react enthusiastically to "in-party" advertising and with distaste to "out-party" advertising. As shown in Figure 1, the average dial rating for Democratic and Republican positive and negative ads revealed precisely this pattern. The ads clearly polarized strong Democrats and Republicans, for whom the gap in the dial rating was some 40 points. Attack ads (especially Republican attacks) tended to evoke less enthusiastic ratings across the board, thus muting the level of polarization.
The level of polarization generally increased during the playing of the ads. Viewers needed some time to register whether the ad was sponsored by Democrats or Republicans. The trend in the dial scores suggests that many viewers were initially uncertain over the source of the ad, but began to respond in a partisan direction as the ad played. The sponsoring candidate typically appears towards the end of the ad, thus providing partisans with a clear cue concerning the origins of the ad. This "learning" effect is captured by assessing the level of polarization in the dial scores over the last 15 seconds of the ads. As shown in Figure 2, the partisan divide in the dial ratings was heightened during the closing stages of the ad.
The level of polarization in the dial ratings is a measure of an ad's ability to mobilize the base. Figures 1 and 2 suggest that both sides generally appealed to their respective partisans effectively.
There were two striking instances, however, of ads that failed to rally strong partisans. In Virginia, Sen. George Allen's (R) campaign attacked James Webb for using an endorsement from former President Reagan over the objection of Mrs. Reagan. The dial results for this ad (see Figure 3) were distinctive: Republicans, Democrats and independents rated the content equally unfavorably. The charge against Webb was too complicated for most viewers to grasp. In their haste to attack Webb, the Allen campaign concocted an ad that was virtually impossible to decipher, even by staunch Republicans.
A similar failure occurred on the Democratic side in the Maryland Senate race where the Democratic Senate committee aired an ad linking Republican candidate Michael Steele with President George W. Bush. With the exception of the strong Democrats (who were only slightly favorable), everybody else rated this ad below the neutral point. Linking Steele with Bush did not seem to work.
From a candidate's perspective, the million dollar question is not whether people like or dislike the ad, but whether the dial ratings translate into voting preferences and more positive feelings about the ad's sponsor. We asked participants to rate each candidate on a 0-100 feeling thermometer scale with higher numbers indicating warmer, more favorable evaluations of the candidate in question. For each race, we subtracted the Democratic candidate's thermometer score from the Republican's score. Positive "net" scores indicate a preference for the Republican, while negative scores show a preference for the Democrat. These net scores can be used as proxies for vote choice because they are almost perfectly correlated with the voting intention question.
Figure 4 shows a set of "impact coefficients" gauging the effects of a one-unit shift in the dial ratings on the net thermometer rating. A coefficient of .5, for example, would indicate that for every increase of ten points in the dial rating, the net thermometer score rises by 5. Because the net thermometer scores were derived by subtracting the Democratic candidate's score from the Republican's, the coefficients for the Democratic dial scores will necessarily be negative (suggesting that those who reacted favorably to Democratic ads were more in favor of the Democratic candidate) while the coefficients corresponding to the Republican dial scores will be positive.
Figure 4 reveals two clear patterns. First, the effects of the dial ratings on evaluations of the candidates were typically larger for independents. Not having a strong prior preference either way, independents are more susceptible to ad campaigns. When independents feel good about an ad, they also feel good about the ad's sponsor. Partisans, on the other hand, are already inclined to support "their" candidate; how they feel about a particular ad is somewhat redundant.
Second, reactions to the negative ads left a lesser mark on evaluations of the candidates. This effect was foreshadowed in Figures 1-2 which showed that strong partisans reacted more favorably to their own side's positive ads. The relative advantage gained by going positive was more pronounced in the case of Democratic ads; voters who felt good about a Democratic positive ad ended up with a net thermometer score that favored the Democratic candidate to a greater degree than did people who felt good about the Democratic negative ad. The increased effectiveness of positive ads was also apparent on the Republican side, but to a lesser degree.
Going Negative Is the Only Option?
Campaign consultants typically claim their clients have no choice but to go negative, either to head off anticipated attacks or rebut those already in circulation. But what is the social cost of continuous negativity? Previous research suggests that exposure to negative rather than positive campaign advertising encourages a sense of cynicism among voters, especially those with no partisan attachments. Independents, for instance, are more apt to believe that their opinions do not count and that ordinary citizens have little influence over policy outcomes.
This study replicates these findings -- among voters who classified themselves as independents or "independent-leaners," exposure to negative campaigning strengthened cynical beliefs about politics and politicians (see Figure 5 ). On the other hand, the views of partisans -- who were significantly less cynical to begin with -- showed no such pattern. Their feelings about voting, the sincerity of candidates, and the responsiveness of political institutions were entirely unaffected by the tone of the ad campaign.
The dials data suggest that candidates who seek votes are in fact better off using positive ads. Why, then, is going negative such a dominant strategy? As noted above, the reactive nature of campaigning often forces both sides into attack-counter-attack mode. But there is an even stronger incentive to use negative ads: the news media. Controversial attacks are much more newsworthy than ads promoting a candidate's credentials or performance record.
Candidates go negative because negativity increases a candidate's ability to set the campaign agenda by getting the media to focus on an issue of the candidate's choosing. In 1988, the "Willie Horton" ad prompted an avalanche of news reports on prison furlough programs. This proved advantageous to Vice President George H.W. Bush, who was perceived as being tougher on crime than Massachusetts Gov. Michael Dukakis (D).
In 2004, the "Swift Boat" ad -- which attracted more than 500 news reports in the week after it aired -- raised serious questions about Senator Kerry's integrity. And in 2006, the "Playboy ad" used against Harold Ford Jr., although hardly a hit among Tennessee Republican voters, was a smash hit among journalists. The discussion in the press of the appropriateness of the "Call me, Harold" tag line directed voters' attention to the character issue and distracted attention from Iraq and the state of the nation.
If the news media devoted less attention to ad campaigns, and more attention to a candidate's track record and positions on substantive issues, rational candidates might begin to consider the persuasive effects of ad content on voter choice, rather than shaping the ads mainly designed to attract the attention of journalists. The results from this study suggest that if the press -- and hence the candidates -- actually focused on education and persuasion concerning substantive issues, ad campaigns would not necessarily be more negative than positive.
Shanto Iyengar is Professor of Communication and Political Science and Director of the Political Communication Lab at Stanford University. His research on negative campaigning is described in Media Politics: A Citizen's Guide, to be published by W. W. Norton in November 2006.