Effectiveness of Negative Political Advertising
[WJMCR 2:1 December 1998]
Throughout history politicians have used various methods, such as whistle-stop speeches, political advertising and political rallies, to achieve their "primary goal, the winning of votes."1 However, over the years, politicians have found that it is most advantageous to use political advertising to persuade voters. According to Ansolabehere, Behr and Iyengar (1991), campaign spending in the United States grew fifteen-fold since 1952 from $140 million to nearly $2 billion in 1988.2 They say less than 5 percent of campaign expenditures in 1952 were devoted to radio and television time. By 1972, 15 percent of campaign expenditures were for broadcasting. By1988, about 20 percent of nearly $2 billion went to purchase airtime. If you add the salaries of media consultants and advertising production costs, at least 40 percent of campaign dollars are spent on media expenditures in 1988.3 Dwight Eisenhower's campaign in 1956 was the first presidential campaign which relied heavily on political television commercials.4 After the election, "Truman, referring to the effects of political advertising . . . commented that it was the first time in 148 years that a president had been elected without carrying a Congress with him."5
Most remarkable in the 1996 presidential campaign was an enhanced dependence by the candidate on television advertising. This was apparent in the primaries. Advertising expenditures in New Hamshire were up from $851,000 in 1992 to $2,696,500 in 1996, an increase of 317%.6 Without a competitor, Clinton spent $12 million of his $30.9 million limit on television commercial in the middle of the primaries, $42.4 million during postprimary/preconvention, and $44 million on television ads in the general election. A total of $98.4 million was used on television commercials. On the other hand, Dole spent $47.1 million on television advertising and Perot spent $22 million which was over 75% of his money on television advertising.7
Because political advertising, unlike product advertising, must get results in a short period of time, political practitioners use several kinds of political advertising: image, issue and negative advertising. Although there exists no rule in using political advertising, prior to the 1980s candidates usually used issue or image ads at the beginning of a campaign to establish their positive image and then used negative ads at the end of the campaign to attack the opponent. However, those strategies were abandoned in the 1980s. A significant trend in today's political advertising is the increasing use of negative political advertising. In today's political campaign, candidates, either challengers or incumbents, use negative ads from the beginning of a political campaign. After examining more than 1,100 political commercials, Sabato asserted that:
Even when television is used to communicate political truth (at least from one candidate's perspective), the truth can be negatively packaged—attacking the opponent's character and record rather than supporting one's own. If there is a single trend obvious to most American consultants, it is the increasing proportion of negative political advertising.... At least a third of all spot commercials in recent campaigns have been negative, and in a minority of campaigns half or more of the spots are negative in tone or substance.8The increasing use of negative political advertising has been promoted by two unrelated legal touchstones.9 First, the Communication Act of 1934 made an important distinction between candidate ads and product or service ads. It stated that broadcasters could refuse all deceptive advertising except for political commercials. Second, the 1976 amendment to the Federal Election Campaign Act allowed private individuals and political action committees or PACs, to spend unlimited amounts on behalf of candidates. According to L. Sandy Maisel, the 1976 amendment to the Federal Election Commission allowed PACs and individuals to make unlimited independent political expenditures on behalf of a candidate, i.e. they can spend as much as they want to independently support a particular candidate, including making their own commercials. These expenditures are different from contributions to a specific candidate's official campaign, which the act limits to $1000 for an individual and $5000 for a PAC. The amendment was prompted by a U.S. Supreme Court decision, Buckley v. Valeo (424 U.S. 1, 1976). As Maisel puts it, the court "ruled that limitations on such independent expenditures was an abridgment of the freedom on speech."10 Although these two groups cannot make unlimited donations to a political campaign, they may cover certain expenses, such as advertising, without any limit. Because advertising sponsored by individuals or PACs is usually more aggressive than candidate-sponsored commercials, these ads are often negative.
Why do political practitioners increasingly use negative political advertising? Do they think negative ads are the most effective way to persuade voters in a short period of time in order to win an election? Many researchers have conducted studies, but the results are inconsistent. According to politicians, because attack ads work, they increasingly use such ads.
Up to now, the increasing use of negative advertising in political campaigns has created a considerable controversy between researchers and political practitioners. Thus, more research needs to be conducted in a variety of ways to measure the effectiveness of negative political advertising.
Various theories have been developed for the study of political advertising. However, the present study adopts two theories concerning the effectiveness of negative political advertising.
Expectancy theory. Expectancy theory focuses on the relationship between language use and the effectiveness of such language use on persuasion.11 Expectancy theory assumes that "since language is a rule-governed system, people develop norms and expectations concerning appropriate usage in given situations."12
Most cultures and societies shape their own patterns of language and determine normative or non-normative patterns of language use. When messages conform to people's norms and expectations, "the norms and expectations are strengthened, but the messages exert minimal impact on attitudes."13
On the other hand, when "communicators, intentionally or accidentally, violate norms governing appropriate language usage," they violate "the expectations of message receivers, and, in turn, affect their receptivity."14 If messages violate people's norms and expectations, they can have more or less persuasive effects, depending on the circumstances.15
Expectancy theory identifies two violations: positive or negative violation. In regard to the persuasion effects of the two violations, the theory assumes that when messages positively violate people's linguistic expectations, the violation has a positive impact toward people's attitudes and evokes persuasive effectiveness.16
In contrast, when messages negatively violate people's linguistic expectations, "a boomerang effect occurs, with receivers changing to the position opposite to the one advocated by the communicator."17
Expectancy theory can be applied to many fields of study, one being negative political advertising. Based on the theory, it can be assumed that voters have normative expectations about negative political advertising because such advertising is one of the most common genres in today's politics, and, as a result, voters have been exposed to numerous negative ads.
Because most negative ads have been employed to attack the opponent's image or issues, people may expect negative political ads to have standard formats and intense messages. If negative political ads conform to people's normative expectations, expectancy theory defines it as a negative violation and predicts that it evokes negative effects toward the sponsor.
Many studies support the prediction.18 It should be noted, however, that the studies have the limitation that research was not conducted in actual elections. "In a true election, factors such as news coverage of the campaign, and especially news coverage of candidate advertising, may influence reactions to particular political advertising strategies."19
Ansolabehere and Iyengar say that news coverage of ads create a "ripple effect." Campaign advertisements have become so important, that they are now a subject of news coverage and public debate in and of themselves. This ripple effect in the news is a significant incentive for campaigns to devote event more resources to advertising."20
Ansolabehere and Iyengar say media coverage of such ads further turn off independent voters and reinforce partisan voters (see below), amplifying the effects of negative ads. They argue that "advertisement tends to reinforce partisan inclinations, the ad-watches simply amplified its effects, thus playing into the hands of the candidates and their handlers."21
Cognitive response model. The cognitive response model was developed to monitor the effects of persuasive communications and was widely employed in many fields of study: "fear appeals, distraction, source credibility, and advertising effectiveness."22
The response model assumes that when people are exposed to a message, they have psychological processes that determine the impact of the message.23 Wright asserted that three distinct responses are identified in the psychological process of a message acceptance: counterargument, source derogation and support argument.24
In the negative political advertising context, counterargument occurs when voters attempt to defend the target. When counterargument is activated, receivers would neutralize the message ("he [the target] is changed") to reduce the discrepancy.25 Source derogation involves "negative images of the sponsor ('he is a mudslinger') or of the message ('it's misrepresentation')."26 Support argument involves "negative images of the target ('he is indeed that bad')."27
Wilson and Muderrisoglu conducted "an initial study on the analysis of cognitive responses generated by comparative and non-comparative ads."28 Through the study, the authors found that comparative ads produced significantly more counterarguments, source derogations, and negative ad-related statements, but fewer support arguments and positive ad-related statements."29
Merritt cocluded that "since negative advertising is a variant of comparative advertising, [the] findings of product-related effects of comparative advertising suggest parallel effects for negative advertising."30
Research31 on the impact of negative political advertising also supported the results of Wilson and Muderrisoglu. For example, Garramone found that negative political advertising evokes source derogations.32 Wright also asserted that "source derogation is a more frequent response to dissonance than counterargument in situations where the source might be assumed biased."33 Therefore, it can be expected that negative political advertising would be likely to produce source derogations, in turn, causing harmful effects toward the sponsor.
Effectiveness of negative political advertising. The growth of negative political advertising has drawn the attention of numerous researchers. Most research has tried to find its effectiveness on "cognitive, affective, and conative components of voters responses,"34 but the results of the research are inconsistent. While some research supports the use of negative advertising, others assert that attack politics evoke a boomerang effect.
According to Basil et. al., the counterproductive aspects of negative political advertising "may arise from the fact that negative advertisements are rated as 'effective' because the message itself remembered, but 'ineffective' because the candidate sponsoring the ad is harmed."35
Ansolabehere and Iyengar argue that campaign advertising is effective. Using experiments primarily, their results show that political advertising is persuasive but not manipulative.36 Specifically, they conclude that the ads "inform voters about the candidates' positions"37 and "allows voters to develop differentiated images of the candidates, images that play an important role in shaping voting choices."38 Although they say that more involved and attentive voters learn more from competitive political advertising, they say people who do not pay close attention to campaigns, such as low-income voters, can benefit from the information in the ads.
Intended effects. Garramone defined the intended effects of negative advertising as "creating negative feelings toward the targeted candidate and positive feelings toward the sponsoring candidate."39 Kaid and Boydston in an "experimental study of negative newspaper and television advertising by an independent sponsor," found that "negative advertising reduces the image evaluation of the targeted politician."40
Garramone et. al. asserted that negative ads are very effective when differentiating or discriminating candidates' images, and they pointed out that:
By providing concrete substantive information, a negative political ad may allow voters to distinguish candidate qualities, positions, and performance more readily than would other types of political information that provide less explicit information. Also, the greater perceived differences between candidates may lead voters to greater attitude polarization regarding the candidates. That is, by discerning clear differences between candidates, voters may be more likely to strongly like one candidate while strongly disliking the other.41
Garramone asserted that one tactic to obtain the intended effect is to make the opponent untruthful, and she suggested that the perceived truthfulness of negative political advertising may determine its impact.42
Persuasion research indicates that the more credible a source, the more persuasive the message. Thus, the more truthful negative political advertising is perceived, the greater should be its impact. Sabato indicated that although academic researchers have found that negative political ads cause a backlash effect, political consultants believe otherwise.43
Political consultants assert that since "academic researchers usually look at negative advertising divorced from the strategy that propels it," they can't exactly prove its effectiveness.44 Janet Mullins, the 1988 Bush campaign's media director, claimed that "everybody hates negative ads; then they rate them most effective in terms of decision making. There isn't any long term effect . . . It is kind of like birth pains. Two days later, you forget how much it hurt. The same is true for negative advertising . . ."45
Political consultants generally suggest that negative advertising is more effective in terms of information than positive advertising. Lau in a study of negativity in political perception, agreed with political consultants and indicated that "the tendency for negative information to have more weight than equally extreme or equally likely positive information appears in a variety of cognitive processing tasks."46
Tinkham and Weaver-Lariscy's study of the impact of negative political TV commercials also supported "the negativity effect"47 asserting that:
Because negative messages derive their impact from a broader range of intervening variables than do positive messages, candidates whose message strategies are exclusively positive limit their potential voters to fewer criteria for judgment, and at the same time, limit themselves to fewer avenues of potential impact. Effective opponent denigration, in contrast, adds another broad dimension for judgment and influence.48
In addition, Nugent suggested that since "people are more apt to vote 'against' than 'for' something"49, negative advertising is effective in real political campaigns. "Political consultants' reliance on the use of negative political advertising in recent years is strongly grounded in the evidence provided by cognitive research."50
Unintended effects. Unintended effect usually refers to a boomerang effect. Garramone defined a boomerang effect as one which "may create more negative feelings toward the sponsor, rather than toward the target."51According to Garramone, "voters response varies with content theme, but backlash or boomerang may be the most common effect of negative political advertising."52
In a 1985 experimental study53, Garramone assumed that a boomerang effect of negative advertising resulted from two reasons:
First, many viewers disapprove of advertising that attacks a candidate54 and such viewers may develop negative feelings toward the sponsor of the advertising. Second, viewers may perceive the negative advertising as an infringement upon their right to decide for themselves. Such a perception may result in reactance, a boomerang effect in which the individual reacts in a manner opposite to the persuader's intention.55
Stewart supported Garramone's assertion in his study of voter perception of mud-slinging in political communication. Stewart found that "the majority of respondents view the mud-slinger as an untrustworthy, dishonest person who will do anything to win an election."56 Through a telephone survey of voters in Atlanta and Philadelphia, Surlin and Gordon also found that voters consider negative ads to be unethical.57
Garramone conducted a telephone survey of mid-Michigan voters to investigate effects of negative advertising in terms of "perceived truthfulness" and "feelings toward both sponsor and target."58 The results of the study proved that negative advertising has "a strong negative influence on the viewer's feeling toward the sponsor but only a slight net negative influence on feelings toward the target."59 Garramone also suggested that 75 percent of respondents disapproved the use of negative advertising.60
Merritt's study of voters' responses to negative advertising in California Assembly District found that "the negative political advertising evokes negative affect toward both the targeted opponent and the sponsor."61
The hierarchy of effects model. Lavidge and Steiner created a model to measure advertising effectiveness based on a classic psychological model that classifies behavior into three categories:
H1. Negative political advertising will cause respondents to believe they are better informed.
The primary purpose of the study is to investigate the effectiveness of negative political advertising. Survey method was used to collect desired data, and four graduate students of the University of Missouri were employed as interviewers. Before conducting an actual survey, the interviewers were trained to understand the purpose of the present study and the contents of the questionnaire.
During two weeks of October 1996, 297 interviews with randomly selected residents of Columbia Missouri were completed.
Hypothesis 1 suggested that negative political advertising will cause respondents to believe they are better informed. As can be seen from Table 1, only 21.9 percent of the total respondents evaluated negative political advertising as informative while 47.5 percent disagreed or strongly disagreed that negative political advertising is informative, and 30.6 percent said that they are neutral.
Interestingly, only 2 percent strongly agreed while 14.1 percent strongly disagreed that negative political advertising is informative. The overall mean score was 3.38. Thus, it can be inferred that the overwhelming response to informativeness of negative political advertising was negative. Hypothesis 1, therefore, is not supported.
Hypothesis 2 suggested that negative political advertising might have a negative impact on believability (truthfulness). As can be seen from Table 2, almost half of the total respondents evaluated negative political advertising as unbelievable while only 13.5 percent agreed that such advertising is believable. 36.4 percent were neutral, and the overall mean score was 3.48. Thus, hypothesis 2 can be supported.
To test hypothesis 3 that negative political advertising will produce a negative impact on attitudes toward both the sponsor and the target, descriptive analyses were conducted.
As Tables 3 and 4 indicate, more than half of the respondents expressed negative attitudes toward both the sponsor and the target, while less than 30 percent expressed positive attitudes toward both candidates. Of respondents, 14.1 percent had neutral -attitudes toward the sponsor, and 23.3 percent had neutral attitudes toward the target. As predicted, hypothesis 3, therefore, is supported.
Research question 1 suggested that perceived believability of negative political advertising will have a positive relationship with a favorable attitude toward the sponsor and a negative relationship with a favorable attitude toward the target. To find out correlations among dependent variables, a Pearson correlation test was conducted.
As Table 5 indicates, believability has a high positive correlation with favorable attitudes toward the sponsor and has a negative correlation with favorable attitudes toward the target. In other words, those who are likely to believe negative political advertising tend to have positive attitudes toward the sponsor and negative attitudes toward the target.
In addition, believability is positively correlated with approval. That is, those who are likely to believe negative political advertising tend to approve of that candidates use of such advertising. Informativeness is positively correlated with believability, approval and favorable attitudes toward the sponsor, although it is negatively related to favorable attitudes toward the target.
Approval of negative political advertising has a high positive correlation with favorable attitudes toward the sponsor and a negative correlation with favorable attitudes toward the target.
Research question 2 is "Does negative political advertising have different effects in terms of informativeness, believability and attitudes toward the sponsor and the target depending on gender?" One-way analysis of variance (ANOVA) was employed to examine research question 2. Tables 6, 7, 8 and 9 contain the results of the tests.
There is no significant difference in evaluating informativeness of negative political advertising between male and female respondents. The means are males, 3.35, and females, 3.39. There also is no significant difference in evaluating the believability of negative political advertising between male and female respondents. Means are males, 3.45, and females, 3.51. There is a significant difference (F = 4.44, p < .05) between male and female respondents on attitudes toward the sponsor. Means are males, 22.92, and females, 24.17.
There is a highly significant difference (F = 18.67, p < .001) between male and female respondents on attitudes toward the target. Scheffe's test also supported the result. Mean scores are males, 22.34, and females, 24.61.
Research question 3 is "Does negative political advertising have different effects in terms of informativeness, believability and attitudes toward the sponsor and the target depending on age?" One-way analysis of variance (ANOVA) was used to examine research question 6.
There is no significant difference in informativeness of negative political advertising depending on age.
There is a significant difference (F = 3.07, p < .01) in evaluating believability of negative political advertising depending on age (Table 6). Scheffe's test identified that there is a significant difference between group 1 and group 4. That is, it can be assumed that younger people are more likely to believe negative political advertising than older people.
There is a highly significant difference (F = 7.61, p < .001) on attitudes toward the sponsor depending on age. Based on mean scores of each group, Scheffe's test identified that there is a significant difference between group 1 and group 5, between group 2 and group 5, between group 1 and group 4 and between group 2 and group 4. That is, it can be inferred that younger people, in general, have more positive attitudes toward the sponsor than older people. There is no significant difference on attitudes toward the target depending on age.
Research question 7 is "Does negative political advertising have different effects in terms of informativeness, believability and attitudes toward the sponsor and the target depending on income?" One-way analysis of variance (ANOVA) was used to examine research question 7.
There is a significant difference (F = 3.38, p < .01) in evaluating informativeness of negative political advertising depending on income (Table 7). Scheffe's test revealed that there is a significant difference between group 1 and group 5 and between group 1 and group 4. Mean scores of each group are reported in Table 7.
There is no significant difference on attitudes toward the target depending on income. Scheffe's test also supported the result. However, according to mean scores of each group (group 1 = 24.88, group 2 = 24.24, group 3 = 23.39, group 4 = 23.48 and group 5 = 22.66), group 1 has the most negative attitudes toward the target.
The present study investigated the effectiveness of negative political advertising in terms of informativeness, believability and attitudes toward both the sponsor and the target. Before discussing the results of the present study, it should be noted that "it is difficult to determine the effectiveness of political [advertising]."66 As Devlin asserted, because many things happen simultaneously in a real election, "it is difficult to isolate the impact of political advertising."67 The present study, however, revealed many interesting effects of negative political advertising.
The findings suggest that negative political advertising was perceived as uninformative. However, this seems contrary to the previous research68, which measured information levels and political practitioners' claims concerning negative information.
As expected, negative political advertising was perceived as untruthful, and perceived truthfulness was positively related to favorable attitudes toward the sponsor and negatively related to favorable attitudes toward the target.
Although the perceived truthfulness of negative political advertising was as expected, a minority of the respondents perceived such ads to be true. Overall, negative political advertising produced negative evaluations of both the sponsor and the target. Those effects are consistent with the findings of the previous research.69 As Garramone noted, such a negative effect might be related to the respondents' overall attitudes toward negative political advertising, revealing that more than half of them expressed disapproval of using such ads.70
As mentioned above, although decisive conclusions are not drawn to determine the impact of negative political advertising, the findings of the present study have implications for politicians. political consultants and advertising agencies. The findings raise doubts about using negative political advertising in a political campaign. As Merritt noted, respondents' negative attitudes toward both candidates and their overall disapproval of negative political advertising resulted from the increasing use of negative political advertising during the 80's, producing cynicism toward politics and declining political participation.71
Because negative political advertising that identifies the sponsor and the target hurts both candidates, when a candidate uses such advertising, it would be better not to identify the sponsor. However, it is normally required that by law the sponsor be identified. Garramone also pointed this out and suggested that "independent political action committees sponsoring negative advertising offer the candidates they help this anonymity advantage. Independent sponsors may contribute the additional benefit of greater credibility."72
Surlin and Gordon asserted that African Americans are more likely to believe that negative political advertising is informative and have more positive attitudes toward the sponsor than whites.73 But, those findings were not supported in the present study. The findings of the present study indicated that there was no significant difference in evaluating informativeness, believability and attitudes toward the sponsor and the target between African Americans and whites. That is, both African Americans and whites, in general, considered negative political advertising as uninformative and unbelievable and had negative attitudes toward both the sponsor and the target.
There were no significant gender differences in evaluating informativeness and believability of negative political advertising. But, interestingly, in terms of attitudes toward candidates, women were more negative toward both the sponsor and the target than were men.
The findings suggested that although both younger and older people agreed that negative political advertising is not informative, older people consider negative political advertising as less believable and have more negative attitudes toward the sponsor than younger people.
Negative political advertising would be effective with lower income level people. They perceived negative political advertising as more informative and more believable and had more positive attitudes toward the sponsor than higher income level people.
Won Ho Chang is Professor and Director of the Stephenonson Research Center at the University of Missouri's School of Journalism.
He can be reached by e-mail at Won_Chang@jmail.jour.missouri.edu.
Jae-Jin Park is a Public Relation Speicalist with the LG Corporation in Seoul, Korea.
Sung Wook Shim is a graduate student in the School of Journalism at the University of Missouri.