02 WW lllllllllllllllllllllllllllllllllllllllHllllllllllllllll ‘ 3 1293 00604 8320 ”BRA“ 1 Michigan State L I.lnivel'sit:_J This is to certify that the thesis entitled NEGATIVE POLITICAL TELEVISION COMMERCIALS: ‘ AN EXPERIMENTAL STUDY OF APPEAL TYPES AND RESPONSE STRATEGIES presented by Brian L. Roddy has been accepted towards fulfillment of the requirements for MASTERS OF ARTS degree in TELECOMMUNICATIONS é zajor professor Date “/30/€S’ ‘ 0-7639 MS U i: an Afi'mnative Action/Equal Opportunity Institution W , i ’7' - >77 fl" - ,, - ,, 7" PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. =— DATE DUE DATE DUE DATE DUE MSU Is An Affirmative ActioNEquel Opportunity Institution ' NEGATIVE POLITICAL TELEVISION COMMERCIALS: AN EXPERIMENTAL STUDY OF APPEAL TYPES AND RESPONSE STRATEGIES BY Brian L. Roddy A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Telecommunication 1988 $04233? ABSTRACT NEGATIVE POLITICAL TELEVISION COMMERCIALS: AN EXPERIMENTAL STUDY OF APPEAL TYPES AND RESPONSE STRATEGIES BY Brian L. Roddy The purpose of this study is to evaluate the effectiveness of negative political advertising in the form of 30-second TV commercials. Appeal types, issue versus image, and response strategy, direct rebuttal versus positive message, are compared to study differences in impact on viewers. It is hypothesized that viewers will have a greater tolerance for attacks concerning issues, rather than attacks on candidates' character (or image). The second hypothesis states that a direct rebuttal should be a more effective response strategy than ignoring the attack - a direct rebuttal may create a greater "backlash effect", resulting in a negative reaction toward the sponsor of negative commercials. An experimental procedure was implemented with a total of 274 college students participating, divided into four treatment groups. Six 30-second Tv spots were produced and inserted into a 15-minute news program. One group was exposed to a negative-issue commercial by one candidate and a rebuttal-issue commercial by the 'opponent'. The second group saw the same negative-issue ad coupled with a positive-issue ad. The third and fourth groups are similarly set-up, but feature image- related ads. Both candidates were fictional, and no mention was made of political party. The results showed that attack ads were more effective when they addressed issues rather than candidate image. Regarding the second hypothesis, the rebuttal response strategy proved to be only partially effective. The rebuttal strategy did not produce as strong a backlash effect as expected. Copyright by BRIAN LEO RODDY 1988 ACKNOWLEDGMENTS The people I'd like to thank for their assistance in this project are too numerous to mention individually. I greatly appreciate the contributions made by those involved in the production of the television commercials. I would especially like to thank my thesis committee members, Bella Mody in the Department of Telecommunication, and Gina Garramone in the Department of Advertising, for their patience and guidance. Additionally, I thank Gina for her great efforts in seeing this project submitted to the 1988 International Communication Association conference in New Orleans, and for publication in the Journal of Broadcasting and Electronic Media. TABLE OF CONTENTS INTRODUCTION.OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0.00.00.000 Characteristics of Television-Centered Campaigning PREVALENCE OF NEGATIVE POLITICAL ADVERTISING..................... Modern Political Campaigning and Negative Advertising Negative Campaigning: 1984 Negative campaigning: 1986 Analysis of Negative Advertising's Prevalence PERCEHIONS OF NEGATIVE ”VERTISING.S VALUEOOOOOOOOO0.00000.00... Negative Aspects Positive Aspects EFFECTIVENESS OP NEGATIVE POLITICAL ADVERTISING.................. Intended versus Backlash Effects Attitudes Toward Negative Political Appeals TYPES OP NEGATIVE ADVERTISING APPEALS............................ Issue-oriented versus Image-oriented Political Ads Research into Issue and Image Types of Political Appeals Testing Differences between Issue and Image Appeals RESPONSE STRATEGIES FOR NEGATIVE POLITICAL COMMERCIALS........... Political Advertising Strategies: Responding to Negative Ads Testing Rebuttal and Positive Advertising Strategies Interaction between Appeal Type and Response Strategy WDOOOOOOOOOOOOOOOOOO000......0.00.00.00.00.IOOOOOOOOOOOOOOOOO Treatment Conditions and Subjects The Stbmulus Commercials The Experimental Procedure Description of the Variables ”SULTSOO0.0.0......0.0.0....00......O...OOOOOOOOOOOOOOOOOOOOOOOC ‘vi 13 13 15 18 18 19 21 10. 11. 12. 13. DISCUSSIOROOOOOOOOOOOOOOOO00.00....0.00IOIOOOOOOOOOOOO00.0.0.0... ”CLUSIONeeeeeeeeeeeeeeeeoeeeeeeeeeeeeoeeeeeeeeeeeeeeeeoeeeeoeee ”mexh-m1m000000.000.0000IOOOOOOOOO ”szxa-QUESTIOWAIEOOOO...00......OOOOOOOOOOIOOOOOOOOOOOOO LIST orWGSOCOOOCOC.OOOOOOOCOOOOOOOOOO 38 66 47 53 55 Table 1 Table 2 Table 3 LIST OF TABLES Total Mean Scores, Standard Deviations, and Alpha CoeffiCIents................................... ..... 34 Mean Scores on Dependent Variables by Appeal Types and Response StratngOOOOOOI...00.00.000.00.00.0.0000... 36 Responses to 'Agree-Disagree' Statements Concerning Attitudes toward Negative Advertising............... 41 viii LIST OF FIGURES Figure l Attitudes toward Using Negative Advertising to Criticize: Candidate Record vs. Character (Percentages) ......... 42 Figure 2 Attitudes toward Using Negative Advertsing: As Often As They Like vs. Never (Percentages) ................. 43 ix INTRODUCTION In an election year, television viewers are exposed to a multitude of messages by candidates running for public office. Commercial appeals ' range in length from short lS-second spots to 30-minute programs. In recent years, perhaps the most common advertising vehicle for federal, state, and local political candidates has been the 30-second spot. Longer messages — five minutes, or the rare 30-minute program - are usually delivered on behalf of major office candidates with high media budgets. The 30-second spot has been the most frequently employed message length for product advertising as well, although there has been a trend toward even shorter commercials in recent years, particularly the 15- second spot. Political communicators, however, have not yet adopted that message length. As it is, there has been much criticism leveled by analysts and participants in the political process concerning this usage of short TV ads by candidates to convey their ideas. The importance of political TV spots is evidenced by the preponderance of 30—second ads for even the lower profile offices. It is now common to see candidates for such offices as state representative and judicial posts make use of the medium of television to promote their campaigns. Increasingly, this emphasis on television comes at the expense of traditional party politics. Candidates today are more likely to bypass party leaders and, instead, make direct appeals to voters via the mass media. Characteristics of Television-Centered Campaigning In order to compete for public office in the 1980's and beyond, candidates need to be knowledgeable about the kind of image they project on television. A candidate with an attractive appearance and an ability to look at ease in front of a camera, will generally have the advantage over an unattractive opponent who is uncomfortable with the medium. Consultants for the National Republican Congressional Committee unanimously agreed that, "'Candidates should be pictured in a television spot only when they are physically attractive'" (cited in Sabato, 1981, p. 145). The 1960 Kennedy-Nixon debate is often mentioned as an example of utilizing television effectively (or conversely, of having physical flaws exposed). Nixon was undermined by a poor make-up job and an unshaven, tired appearance, while Kennedy seemed more at ease and presidential-looking. Voters listening to the debate on the radio were more impressed with Nixon than those watching on television. (Nixon's successful comeback was due in large part to his advisors' ability to better control the image he projected in the media.) In the 1988 presidential primary campaign, two candidates with virtually no political backgrounds have demonstrated significant public support. Both Jesse Jackson on the Democratic side and Pat Robertson as a Republican, are experienced, charismatic public speakers and are comfortable on-camera. Neither has much experience in government, yet both made their presences known in the presidential race, particularly Jackson. In comparison, Alexander Haig and Bruce Babbitt have, between them, spent decades in the public sector, but public support for each was minimal, and both were forced to drop out of the contest early. Media consultants - with television and film skills, and experience in advertising or public relations - have taken on vital roles in modern campaigns for national and state offices. Political professionals today apply social research methods, such as telephone interviewing and focus groups, to measure public opinion of candidates and advertisement effectiveness. Strategies are more than ever, media-centered: aimed at increasing a candidate's standings in polls, or at decreasing public support for the opponent. A significant trend in political campaigning has been the more frequent use of negative advertising strategies. Office-seekers attack their opponents with a variety of accusations and innuendos concerning issues and personal character. In the not-so-distant past, political communication strategists viewed negative advertising as being very risky, fearing voter disapproval of such "mudslinging" tactics. While it is still viewed with disdain among some politicians and consultants, inhibitions against using negative advertising have diminished considerably. With television now the dominant mode of mass communication in U.S. society, it is apparent that further research into the effects of the medium is needed. The impact of negative political television commercials on viewer perceptions of candidates is one area requiring additional investigation. This project presents an experimentally-designed study comparing different types of commercial messages and response strategies. The focus is on exploring the relative effectiveness of different types of negative political television advertising appeals, and the relative effectiveness of strategies for responding to them. PREVALENCE OF NEGATIVE POLITICAL ADVERTISING Modern Political Campaigning and Negative Advertising To be accurate, it must be pointed out that negative political advertising is not a recent phenomenon. Washington, Jefferson, Lincoln, and all past presidents, were subjected to partisan and personal attacks by opponents. A Republican media consultant, Roger Ailes, was interviewed on NBC News the night before election day, 1986. He stated, "I don't think there's any more of a trend toward negative commercials today than back in the days when people did it with posters, and whispered about candidates' illegitimate children" (Nov. 3 1986). Many others would dispute that assessment. Larry Sabato, professor of government at the University of Virginia and author of The Rise of Political Consultants wrote: If there is a single trend obvious to most American media consultants, it is the increasing preportion of negative political advertising. Going on the offensive, "attack politics", is becoming more popular because, while vicious, it has gained a reputation for effectiveness among professionals. (pp. 165-166) The principal difference between today's mudslinging and that of the past is the medium. Television is a far more pervasive and affective means of transmitting messages than are handbills or newspaper advertisements. Virtually all homes in the U.S. have at least one television set, allowing candidates to directly communicate specific visual images and audio messages to nearly every potential voter. The medium has proven to be the most efficient means of reaching the broadest possible range of people. For better or worse, television has become the dominant mode of communication in political campaigns, and it is likely to remain so for a long time. Negative Campaigning: 1984 The 1984 national election campaign was characterized by a number of races in which candidates used the TV spot as a point of attack against opponents. The campaign for Senate in North Carolina, for example, featured a particularly bitter contest between incumbent Jesse Helms and Governor James Hunt. The candidates spent millions of dollars on TV advertisements attacking each other and defending themselves. The commercials became a central part of the campaign. Although it was the Helms' campaign which relied more heavily on negative advertising, Hunt was probably hurt more by a "backlash" effect. One TV spot directed against Helms featured still photos of dead bodies, victims of Salvadoran "Death Squads". Senator Helms was portrayed as a friend of the Salvadoran leader allegedly responsible for atrocities against civilians. This controversial ad drew national attention because of its graphic nature, and backfired on the Hunt campaign. Several months prior to the election, Hunt had held a substantial lead in the polls: eventually, he lost. Certainly, television advertising by itself could not account for the final result, but many observers believed it was the deciding factor in that race (Louden, 1987). After the 1984 election, a bill entitled The Clean Campaign Act of 1985 was introduced in the U.S. Senate. Its purpose was to combat the increasing number of negative political spots in campaigns. The legislation (which did not pass) would have required candidates who attack their opponents to appear in person in the ad, and required broadcasters to provide free response time for the attacked candidate. Criticism of the bill focused on the question of regulating broadcast content, and the First Amendment rights of broadcasters. Also, the measure was perceived by some as favoring incumbent office- holders, as they are more likely to be targets of negative advertising (Broadcasting, Sept. 16 1985). Negative campaigning: 1986 From a strategic perspective, the 1986 election campaign was characterized by an even greater use of attack ads than the 1984 campaign, considered to be a landmark year for negative advertising. In the judgement of many political observers and commentators, 1986 was marked by a higher amount of negative TV commercials, and a nastier tone, than usual. Columnist Tom Wicker of the New York Times described the elections as "surely the nastiest, least relevant, most fraudulent campaigns - by statesmen of both parties - in the history of a long-suffering nation" (Nov. 14 1986, p. 27). Syndicated columnist Gary Wills agreed with the assessment that the tone was worse than usual. "Trashing politicians is an honored American practice . . . Yet even veteran pol-bashers have been dismayed by the reactions of disgust over negative ads during this campaign" (Detroit Free Press, Nov. 10 1986, p. 9A). John Anderson, a former congressman and a presidential candidate in 1980, wrote about how TV ads have changed the complexion of politics over the past thirty years. "Is negative campaigning on the rise because of political television ads? It is the rare person who would not agree” (USA Today, Oct. 16 1986, p. 8A). Another new development arising from the 1986 campaign season was the gégly use of negative ads. Normally, campaign spots get more negative as the election approaches, when the targeted candidate has less time to recover from an attack. In 1986, candidates were attacking much earlier. Democratic political consultant Robert Squier explained the prevalence of negative ads by saying, "We are in an era where there are fewer solid voters on either side - the campaigns are more contested so the ads are more contested". A consultant for several Republican candidates, Larry McCarthy, suggested that it might not be the closeness of races, but that viewers have become accustomed to tough spots, as long as they're not blatantly false or cheap shots (New York Times, August 27 1986, p. A18). Analysis of Negative Advertising's Prevalence Political communication analysts and researchers have studied the practice of negative advertising, and concluded that, undoubtedly, there has been a rise in the amount of attack ads broadcast. According to Sabato (1981), at least one third of all TV spots in recent campaigns have been negative. That estimate may be conservative, as it does not include the 1984 and 1986 election campaigns. Negative strategies have since filtered down from presidential and other high-level office races, to races for state and local offices. (Simultaneously, there's been a trend over the past several years in product advertising toward the use of comparative strategies, whereby one company directly refers to a competitor. There is a strong relationship between product and political advertising strategies, as many media advisors bring their private, commercial experiences to campaigns.) Research conducted for Wake Forest University included a content analysis of the Helms-Hunt 1984 Senate race (Louden, 1987). The analysis compared "level of candidate presence" and "level of opponent presence", controlling for positive and negative spots. A total of 60 negative spots were identified with the Hunt campaign producing 24 of those (along with 17 positive spots), and the Helms campaign 36 (accompanied by 18 positive). The total amount of ad time analyzed for the challenger Jim Hunt was 1,360 seconds, and for the incumbent Jesse Helms, 1,380 seconds. The tabulated results showed that Hunt devoted 730 seconds (53.7%) to negative advertising, compared with 860 seconds (62.3%) by Helms. Both of these amounts of time for negative advertising are very high, and possibly, atypical. Nevertheless, the analysis does represent what goes on in some campaigns. One conclusive piece of evidence gathered from a variety of sources relates to the ability of media users to recall negative advertisements. If nothing else, negative ads are memorable. For instance, Garramone (1984) reported results from a telephone survey of Michigan voters. Her study found that over 75% of the respondents recalled seeing a TV commercial in which one candidate criticized another: Similarly, Johnson and Capeland (1987) analyzed results from a survey of nearly 2,000 respondents in six Southern states during the 1986 elections. They found that 66% remembered seeing a negative ad. PERCEPTIONS OF NEGATIVE ADVERTISING'S VALUE Political pollsters have extensively examined the role of the media in the campaign process, discovering that, over the years, the media has come to displace what used to be the role traditionally served by political party members and officials. Although people increasingly rely more on the media for information, they also seem skeptical of political advertising in general. For example, a poll conducted by the Detroit News (Nov. 3 1986) found that 72% of Michigan's likely voters considered the news media to be their main source of political information. Twelve percent said they got most of their information from political ads. However, only 49% consider political advertisements to be trustworthy and 46% consider them untrustworthy. Some experts have analyzed and reported on negative political advertising's harm. Others have pointed out potential beneficial aspects of negative campaigning. Negative Aspects Debating the Clean Campaign Act of 1985, one of the measure's co- sponsors, John Danforth, argued before the Senate Commerce Committee that negative 30-second commercials were "corrupting the political process.” Former Senate Majority leader Howard Baker supported the bill, stating that, Not only is the negative ad the sleaziest new element in politics, it may also be the most dangerous. The first victim is the person under attack, but the greater victim is the integrity of the political system itself. (Broadcasting, Sept. 16 1985, p. 64) 10 11 Curtis B. Gans, director of the Committee for the Study of the American Electorate - formed to examine the problem of declining turnout - testified that negative commercials affect voter turnout: That type of ad turns people off. They are tuning out, turning off, and staying home .‘. . the impact of television as an institution is to make the American public spectators and consumers rather than active participants. (p. 66) Researchers have studied public reactions to negative ads, measuring attitudes toward ads themselves, and toward both the attacked and the attacking candidates. Garramone (1984) found that, overall, there is a strong audience resistance to negative advertising, measured in terms of approval or disapproval. Almost 75% of 211 respondents expressed disapproval, compared with 19% showing approval. In the aforementioned Johnson and Copeland study (1987), about 65% responded that they disliked negative political ads, with about 43% expressing a strong dislike. Only about 23% reported having some degree of acceptance for the ads. Surlin and Gordon (1977) found that people from lower socio-economic status (SES) groups believed negative ads to be more unethical, than did middle SES respondents, and both groups judged such ads to be "slightly unethical" on an ethics scale used by the authors. Positive Aspects While the prevalence of negative advertising is clearly evident, the yalgg of it is at least somewhat open for debate. Not everyone has demonstrated that negative advertising is detrimental to the political process. Surlin and Gordon (1977) reported that, overall, both lower and middle socioeconomic status (SES) respondents said that direct 12 reference (or "attack") ads were slightly informative. Lower SES respondents believed the ads to be more informative than middle SE5 respondents. Although the subjects perceived the negative ads as unethical, the authors found that as "low-energy information seekers", lower SES respondents were "favorably oriented toward attacking . . . ads" (p. 89). Their results may support the claims of some who insist that attack ads are not as bad as they're made out to be. Media consultant Larry McCarthy contended, "The idea that commercials are somehow evil because they're negative is a real mistake. Campaigns are about comparing candidate A and candidate B" (New York Times, Aug. 27 1986, p. A18). EFFECTIVENESS OF NEGATIVE POLITICAL ADVERTISING The question of whether negative political advertising undermines the public interest or is a natural, perhaps even beneficial, result of an evolving political system which emphasizes competitiveness and public discourse, is debatable. Another matter is whether negative advertising is effective. Researchers have sought answers to the questions of how effective political advertising is in general, and how effective negative tactics are in particular. Analyzing the 1980 elections, Robinson (1981) downplayed the significance of "paid media" in presidential campaigns, but stated that: Once ad campaigns move below the rarefied atmosphere of presidential politics, they can work relative wonders . . . [and] when ad campaigns are not challenged by the other side - especially at the congressional level - they have considerably greater chance of success. (p. 186) Intended versus Backlash Effects The basic intent of a negative ad is to degrade an opponent, with the expectation that support for the attacking candidate will increase. Research indicates that negative political ads do, indeed, accomplish that goal. An experimental study by Raid and Boydson (1987), for instance, concluded that, "Negative advertising reduces the image evaluation of the targeted politician” (p. 193). That is the intended effect. Previous research has also unveiled an unintended effect. Generally, the unintended effect is referred to as a backlash or boomerang effect. The phenomenon of the backlash effect has been studied in psychological, as well as political, 13 l4 contexts. Replacing the term backlash with reactance, the effect is clearly defined by Miller (1976), who refers to J. Brehm's A Theory of Psychological Reactance (1966): Reactance, as described by Brehm, is a boomerang effect which results in negative evaluation as a reaction to a perceived threat to freedom. Thus, when individuals believe that a persuasive manipulation infringes upon their right to decide for themselves, they often react in a manner opposite to the persuader's intention. (p. 230) A prime example of the backlash effect is represented in a study of the 1980 elections, the year in which the National Conservative Political Action Committee (NCPAC) became a noticeable force in political campaigns. Robinson (1981) reported results of their controversial, but effective, media campaign. NCPAC was one of the most powerful political action committees. They spent $1.2 million in a campaign targeting six liberal Senate incumbents. They also conducted an evaluation of their campaign's effectiveness in four of the six races. One evaluation involved a negative campaign against Thomas Eagleton in Missouri. NCPAC used a field experiment in the city of Springfield, Missouri to measure public attitudes toward Eagleton both before and after a radio and TV campaign during the summer. The results showed the media campaign to be a huge success - at least in the short run. The ads focused on the Senator's record and elicited drastic changes in public opinions toward him on issues such as national defense, energy, taxes, and economics. Support for Eagleton (people who thought he was doing a good job) on those issues dropped from an average rating of about one-third to about one- sixth. Eventually, however, Eagleton won re-election. In the fall campaign, NCPAC's ads became an issue, not only in the Missouri race, 15 but in the other targeted states as well. Targeted senators criticized NCPAC and the negative advertisements, and a backlash effect was created. NCPAC acknowledged the occurrence of a backlash effect, but believed that their ads helped conservative candidates more than they hurt. Among Robinson's conclusions regarding political advertising: "What is clear . . . is that paid media work well when public images of the candidate are vague, [and] when the Opponents let the paid media go unanswered” (p. 190). Garramone (1984) examined the variable "perceived truthfulness" of advertisements, and feelings toward candidates. Results showed that the impact of negative commercials on attitude toward the attacking candidate was "distinctly negative". Attitudes toward the target was also negative, but the study found that the overall effect of negative advertising was a backlash, as negative feelings toward the attacker "was not compensated for by increased negative feelings toward the target" (p. 253). Merritt (1984) also studied voter perceptions of candidates involved in negative campaigns. A survey involving a state assembly election in California showed that negative advertising evokes negative feelings toward both the attacked and attacking candidates. The author also suggested that the overall effect of negative advertising may be an increase in voter cynicism and a decrease in political participation. Attitudes Toward Negative Political Appeals Analysis of past research into negative advertising's effectiveness indicates that, while overall effects are likely to be either intended 16 or backlash, this may not be true for all types of appeals. Some appeals will evoke strong unfavorable reactions toward the gpggggg if the attack is perceived as unwarranted, irrelevant, or unfair. If an attack goes too far (and it is difficult to say what is "too far"), the "free media” may focus extra attention on it, which may increase the backlash effect. For example, in a 1986 Congressional election primary in a conservative district of Michigan, the incumbent urged fundamentalist ministers to help him "break the back of Satan" by supporting his candidacy (Ann Arbor News, Aug. 6 1986, p. A9). The implication was that a vote for his opponent was a vote for Satan. The publicity generated by the remarks forced the congressman to defend himself against protests by political and religious officials and the media, and resulted in an unexpected defeat. Voters are not necessarily opposed to all negative political ads, per se. Backlash effects do occur, but intended effects are also achieved with a significant amount of success. The trick is to determine the conditions under which a negative strategy will work. Johnson and Copeland (1987) identified several topics which people felt were legitimate areas of attack. These topics, and the percentage of respondents in their study who believed each to be fair for attack, are; candidate stands on issues (93%), pglitical activities (83.4%). votingirecord (83.5%), and alleged criminal activities (88%). Demographic differences among voters are also important in determining reactions to political advertising. Garramone's 1984 study of Michigan voters reported that age and education were influential factors. Older voters, and those with higher education, were more favorably influenced toward the target than younger and less educated l7 voters. Thus, negative commercials are viewed with varying degrees of acceptance among segments of the population. Another relevant factor is level of involvement. Voters highly involved in politics were more likely to be negatively influenced regarding the commercial's Eggggg than less involved voters. This was primarily true of those who favored neither candidate. The author reasoned that voters who are interested, or involved, in politics and are undecided about the candidates are seeking information and thus are more receptive to advertisements. TYPES OF NEGATIVE ADVERTISING APPEALS Issue-oriented versus Image-oriented Political Ads One topic of great debate in political campaign circles is the strategy of using imggg or igggg types of appeals. Image and issue ads may sometimes be referred to as "personal" and "political", "style” and "substance", or perhaps something else. Most often, researchers and professional consultants call them "image" and "issue", and the relative merits of each are debated by political strategists. Issue appeals are those which, either in a general or a specific way, focus on matters of public concern - such as crime, taxes, and national defense - in terms of how they relate to the candidates. An image appeal describes personal attributes or qualities of candidates; for example, honesty, leadership, intelligence, or family life. Some campaign strategists prefer to use television as a means of creating an image for their candidates, while others emphasize getting candidate messages across. Two of the most influential political consultants are Robert Goodman, who handles Republican candidates, and Robert Squier, for the Democrats. Goodman has a reputation for producing TV spots aimed at creating or enhancing a candidate's image. His purpose is to produce an emotional impact, to make voters feel good about the candidate. A typical Goodman spot may consist of patriotic flag-waving and uplifting themes. Squier tends to produce more issue- oriented spots, although candidate positions on issues are usually not clearly defined (Detroit News, Oct. 21 1986). For all political TV spots, imagery is important. Democratic pollster Patrick Caddell commented that, "Too many good people have been 18 19 defeated because they tried to substitute substance for style" (Sabato, 1981, p. 144). Most candidates favor image over issue advertising because they feel voters judge them more on feelings about personalities than on issues. According to Sabato, "Voters are presumed to want to know what kind of person the candidate is, whether he is likeable and of good character, whether he is honest and sincere, and whether he appears effective, authoritative, and knowledgeable . . . " (p. 145). Despite his issue-oriented strategy, Robert Squier concurs. "Issues are important, but character, in the end, is always the deciding factor" (Detroit News, Nov. 1 1987, p. 25A). Research into Issue and Image Types of Political Appeals Although television is generally considered by political consultants and candidates to be the medium for conveying images, media research indicates that issues are also addressed in political advertising spots. Patterson and McClure (1975) found that political ads were heavy in imagery, but also provided a good deal of information; in fact, they found that TV news coverage of the 1972 presidential campaign contained less information about candidates than did the TV ads. A number of issues (12) received five times more air time via advertising as they did via the average weeknight network newscast. People who were heavily exposed to political spots became more informed about the candidates' issue positions. Josyln (1980, 1984) conducted a study of commercial types. A content analysis of political spots revealed the existence of four types of message appeals; partisan, demographic group identities, candidate 20 attribute, and issue-related. Partisan appeals involve identifying a candidate's party and other members of that party, with the expectation that such identification will attract loyal party voters. Demographic group appeals feature the candidate addressing a particular segment of the population, presenting him- or herself as the one who shares their concerns. Candidate attribute (or image) appeals focus on personal characteristics of the candidates - such as leadership, integrity, intelligence - and describe what qualities a candidate has or what the opponent lacks. Finally, issue-related appeals deal with matters of public policy. Josyln divides this last group into three categories. One involves the candidate's issue concerns - those issues which are salient to the candidate. This type of appeal may or may not explain Egg; the candidate intends to do about a problem, only that it is of concern. The second type features a candidate's policy preference, stated only in a vague manner. The third type of issue appeal articulates a more specific position on an issue. Josyln analyzed 156 television spots from a variety of campaigns. Among the findings most relevant to issue and image types of appeals: approximately 47% included mentions of candidate attributes, while about 58% contained issue-related appeals. Most of the issue spots presented vague policy positions of candidates. To study the effectiveness of political commercials, Kaid and Sanders (1978) conducted an experiment in which a comparison of television commercial type (issue and image) was made, measuring subjects' evaluations of candidates and recall of the commercial content. They found that igggg commercials resulted in a higher candidate evaluation while image commercials resulted in greater recall 21 of content. All scores were low, indicating that voters would not base their voting decision simply on viewing advertisements. Regarding negative political advertising and issue versus image appeals, Johnson and Copeland (1987) reported results from a political communication project involving scholars from several Southeastern states. Their intent was to examine sources of political information used by voters and non-voters during the 1986 election. In a pretest, respondents were asked about negative political attacks. Open-ended responses to questions about fair and unfair attacks resulted in an identification of ten types of negative ads: political record, personal life, issue stands, current or past marriage, criminal activities, family members, voting record, religion, medical history, and sex life. The ten types were included on the final survey instrument. From the ten, two categories of ads were defined, political issue and personal issue. The former includes political record, issue stands, and voting record, while personal issues consisted of past or present marriage, criminal activities, sex life, religion, medical history, and family members. This second category corresponds to what others call image ads. The authors' results point to a conclusion that while a majority of people oppose negative advertising in general, people more readily accept attacks involving issues, rather than attacks on candidates' personal attributes. Testing Differences between Issue and Image Appeals Having identified two main types of political television spots, and briefly summarizing some research findings related to these types, the first of two hypotheses to be tested can be stated. The goal will be to 22 investigate possible differences in effectiveness between issue and image negative appeals. It is hypothesized that viewers of negative political TV spots will evaluate the attacking candidate more favorably when the attack involves issues than when the attack involves the opponent's character, or ig§g_. From the perspective of the targeted candidate, he should fare better when there is an attack on his imggg, in comparison with an attack on issues; it is anticipated that viewers will react more favorably toward the target of an attack when the attack is made on personal character. To state the hypothesis formally: Hla: Exposure to issue-oriented negative political commercials will result in more favorable evaluations of the attacking candidate, in comparison with exposure to image-oriented negative commercials. Hlb: Exposure to issue-oriented negative political commercials will result in less favorable evaluations of the targeted candidate, in comparison with exposure to image-oriented negative commercials. Reactions from viewers will be measured in terms of: (a) likelihood of voting; (b) an overall feeling toward the candidate; (c) commercial evaluation; (d) candidate character evaluation; and (e) a total index of viewer perceptions of each candidate. Negative-issue commercials should produce a more favorable reaction toward the attacking candidate than negative-image commercials. The intended effect - to decrease support for the opposing candidate - will offset any possible backlash effect, when the attack is made on issues. The targeted candidate, then, should benefit more from a backlash effect when the attack is made on image, rather than issues. RESPONSE STRATEGIES FOR NEGATIVE POLITICAL COMMERCIALS Political Advertising Strategies: Responding to Negative Ads According to Lau (1982), the rational politician attempts to avoid controversy: Being well thought of is not as important to a candidate as not being badly thought of. There is always the fear of being labelled a mudslinger. The backlash effect is constantly in the minds of media strategists. "The trick," writes Lau, "is being fortunate enough to have many negative things known about your opponent without your having to bring those negative facts to light". His findings indicate "that when voters are forming impressions of candidates, the negative things they know are more important than the positive things” (p. 373). Political strategists and consultants today are very much aware of possible backlash effects when negative advertising is used. Sometimes a strategy of attacking an opponent backfires. If the public perceives the attack as unfair or unjustified, sympathy may be aroused for the targeted candidate, or anger at the attacker. Among other decisions media advisors and other campaign officials must make, are what types of commercials and what kinds of advertising strategies are best suited for their candidate's campaign. A typical campaign consists of a mixture of image and issue types of commercials, along with varying amounts of positive and negative strategies. Positive commercials focus on the qualities of the sponsoring candidate, putting him or her in a most favorable light. The purpose of the negative commercial is to drive voters gygy from an opponent. The frequency and intensity of each advertising type and strategy varies, depending upon the stage of the campaign, candidate standings in 23 24 polls, and the personal preferences of different media consultants, campaign officials, and candidates. Clearly, a popular candidate running against a weak opponent has little need for negative advertising. On the other hand, a challenger to an incumbent may be trailing in the polls and therefore rely more heavily on an attack strategy (Diamond and Bates, 1984). In addition to positive and negative strategies, there is another. A rebuttal is a defensive strategy consisting of a direct response by the target of an attack. When a negative commercial is aired, the attacked candidate may choose to ignore it or to directly respond to it. There may be times when an attack is made which results in a significant shift of public support away from the targeted candidate. When that happens, there is little choice but to rebut the charges. If the attack is ineffective, it can be ignored. Sometimes it's not possible to measure the effectiveness of an attack strategy until Egggg an election. Even then, it is difficult to assess the full impact of an advertising strategy. If the attacking candidate wins, after trailing in pre- election polls, it is often assumed that the negative ads were effective. Often, candidates cannot afford the risk of pg; responding to an attack, and so they produce a rebuttal. A rebuttal may or may not criticize the attacking candidate directly. It may indirectly refer to an attack by an opponent without specifically identifying the attacker by name. For the purposes of this study, the term "rebuttal" describes a gigggg response by the targeted candidate. Comparisons will be made between positive and rebuttal response strategies to negative commercials. 25 Testing Rebuttal and Positive Advertising Strategies Contrasting rebuttal and positive responses to negative advertising, it is anticipated that the rebuttal should produce a stronger backlash effect. The intended effect of the negative commercial should be weaker when it is directly rebutted. Garramone (1985) showed that a rebuttal by the targeted candidate increases a backlash effect, but failed to influence perceptions of the target; in other words, both the sponsor and target of a negative ad are adversely affected. In that study, comparisons were made between a rebuttal condition and a non-rebuttal condition. In this study, the "non- rebuttal" condition is represented by a positive commercial. The second hypothesis can now be presented. Stated formally: H2a: Exposure to rebuttal strategy commercials will result in less favorable evaluations of the attacking candidate, in comparison with positive strategy commercials. H2b: Exposure to rebuttal strategy commercials will result in more favorable evaluations of the targeted candidate, in comparison with positive strategy commercials. Directly rebutted commercials should result in a more negative reaction toward the attacking candidate than the positive (non-rebutted) commercials, in terms of: (a) likelihood of voting; (b) an overall feeling toward the candidate; (c) commercial evaluation; (d) candidate character evaluation; and (e) a total index measuring viewer perceptions of candidates. Interaction between Appeal Type and Response Strategy Comparisons between appeal types and response strategies involving negative political TV spots shall be investigated. Additionally, a test 26 for an interaction effect will be conducted. Possibly, viewer reactions to negative image or issue appeals will depend on which response strategy is used. For example, it may be that the rebuttal will be more effective when an image appeal is employed. No formal hypothesis will be explicitly presented, however. Instead, it will be assumed - until proven otherwise - that the two treatment variables, appeal type and response strategy, are independent of each other. To summarize, the purpose of this project is to study the effectiveness of negative political advertising. Appeal types - issue and image - are investigated, to determine whether they differ in impact on viewers. Another aim is to simulate a campaign in which one candidate responds directly to a negative commercial versus a scenario whereby a candidate ignores the attack. There will be no 'non-rebuttal' situation, per se; rather, there will be a positive message by the target of an attack. The question is how effectively a candidate can respond to a negative ad - by directly refuting it or by ignoring it. METHOD To test the hypotheses that negative-issue commercials would show a greater intended effect than negative-image ads, and that a rebuttal strategy would generate a greater backlash effect than a positive strategy, an experimental procedure was implemented. The first step was to acquire the proper stimuli. There were several options considered when looking for commercials. One alternative was to record off-the-air broadcast TV spots from the on- going 1986 elections involving local and state candidates in Michigan. A second option was to purchase or borrow ads from previous campaigns. These could be obtained from political communication archives. The third choice was to produce the necessary commercials using Michigan State University's TV production facilities. Each alternative had advantages and drawbacks. The easiest solution would have been to record current ads, saving time and money. However, there would have been no way to control the amount of exposure. Subjects who watched TV would have already seen the ads in varying amounts prior to the experiment. Also, many other factors would have come into play when subjects were questioned about the candidates (e.g., political party affiliation, news coverage of candidates, candidate personal appearances). The main problems with the second alternative was the cost of acquiring copies of previous commercials, and having to select appropriate ads (issue and image, rebuttal and positive). The best solution from the standpoint of controlling ad content was to actually produce the spots. Although this took some time to accomplish, the 27 28 result was a group of ads that were realistic (although not highly sophisticated), and guaranteed candidates unknown to the subjects. Treatment Conditions and Subjects Two of the four conditions involved issue-related ads, and the others were image-oriented. The first condition consisted of exposure to one negative-issue commercial by the attacking candidate (”Carl Martin"), and one rebuttal-issue ad in which the targeted candidate ("Phil Jackson") directly responds to the attack. Condition 2 compares the same negative spot with a positive-issue ad by the target. Conditions 3 and 4 are identically set-up, but with image-oriented content. The four conditions are as follows: Condition 1: Negative-issue vs. Rebuttal-issue Condition 2: Negative-issue vs. Positive-issue Condition 3: Negative-image vs. Rebuttal-image Condition 4: Negative-image vs. Positive—image The subjects were recruited from Michigan State University's College of Communication Arts and Sciences undergraduate classes (mean age = 21), and were awarded extra credit points by their instructors for participating in the experiment. The total number of subjects was 274, approximately equally divided among the four condition groups (Group 1 = 68, Group 2 = 70, Group 3 = 67, Group 4 = 69). For the analysis, Groups 1 and 2 are combined to form an igggg treatment for comparison with Groups 3 and 4 (imggg). Also, a rebuttal treatment consists of Groups 1 and 3, for comparison with the pgsitive treatment (Groups 2 and 4). 29 The Stimulus Commercials The experiment consisted of four conditions involving ads for two candidates for a congressional office. Each candidate was fictional, and two photographs were used. The target of the negative advertising was pictured in both his own and the attacker's ads (using similar but not identical photos). The photos are of a businessman, taken out of a magazine. In order to minimize the effect of personal appearance, the selected man is neither very attractive nor unattractive. There were a total of six commercials produced, each being thirty seconds in length. Only two negative spots were needed, one issue and one image. Each of these were paired with both one rebuttal and one positive. The negative-issue spot focused on the targeted candidate's positions on crime and the environment, claiming that his record as a state senator showed that he was lenient toward criminals and against environmental protection. Visually, it shows a photograph of him with the superimposed words "released criminals" and "opposed environmental protection” reinforcing the audio message. The rebuttal ad depicts another photo of the targeted candidate and a voice-over claiming that his opponent, the attacker, is ”distorting the picture” (as the camera becomes unfocused) and is engaging in negative campaigning. Also, the voice argues that he (the target) has ”fought against crime and for a cleaner environment." The positive spot ignores the attack, and describes the targeted candidate's efforts to reduce pollution and strengthen law enforcement, and is accompanied by still images of the environment and a prison. 30 In the negative-image spot the target is accused of being "indecisive and inconsistent", of missing votes as a state senator, and also implies that he unethically accepted a trip to Hawaii. Visually, an image of several arrows pointing in different directions is first shown, along with a photo of Hawaii, and a photo of the candidate. The image-rebuttal ad is visually identical to the issue-rebuttal, but directly refers to the attack on the target's honesty and claims that his attacker is relying on negative and inaccurate messages. The final spot presents a series of positive words against a black background which accompany the audio portion. The words describe the targeted candidate as a "good leader", "tough fighter", ”honest", "responsible", and "hard working". In each positive commercial, no mention of the attacking candidate is made. Each negative spot concluded with the visual identification tag, "Paid for by Friends for Carl Martin", and each of the other ads stated, "Paid for by Jackson for Congress Committee.” Neither candidate was identified as being the incumbent or challenger, nor was there any mention of their parties. (See Appendix A for complete scripts.) The Experimental Procedure The experiment was carried out over a two-week period of time, in late March and early April, 1987. The subjects were instructed to report to a classroom at a particular time. They were told that they would be seeing a news broadcast, about 15 minutes in length, and would be asked to fill out a questionnaire afterward. The number of subjects at each showing ranged from about 5 to 50, with an average group size of about 25. 31 The subjects were shown a tape-recorded newscast (1/2" video) from a TV station in Montana. The stimulus ads were placed within the broadcast alongside regular product commercials. The product ads were low budget productions for such things as used cars, a mail-order calculator/watch, and a local restaurant. The newscast itself was of the low budget variety found in small TV markets. This was of benefit to the experiment, as the political commercials were also quite simple, without the high-tech look of national ads. The experiment was pretested using one of the four conditions, with a group size of 40. The pretest showed that the subjects accepted the commercials as being real ads for a congressional campaign in Montana. However, there was some criticism of the ads' quality by a couple of subjects. To improve the quality, two experienced public speakers were enlisted as narrative voices in place of the novices who assisted in the original spots. Each commercial was presented twice within the program, with the negative ad shown first, followed later by the rebuttal (or positive). then the negative again, and lastly, the other. Each ad was paired with a product commercial, and the news program separated each showing. Description of the Variables Immediately following the program the questionnaire was administered. Subjects were asked several questions about the candidates and the commercials. The experiment consisted of the two treatment variables, appeal type (issue, image) and response strategy (rebuttal, positive), along with five dependent variables for each candidate (for a total of ten). 32 The first dependent variable consisted of a five-point scale indicating how likely the subject would be to vote for the attacking or targeted candidate, the scale ranging from Very likely (l) to Not at all lifigiy (S). Secondly, a rating on a scale of l to 100 measured how subjects felt about each candidate (overall feeling). (To reinforce the idea that the candidates were real, subjects were also asked about their feelings toward actual political figures, such as Ronald Reagan, George Bush, Jimmy Carter, and others.) The third dependent variable described commercial evaluation and consisted of an index comprised of another five-point scale evaluating the commercials according to how informative, believable, and persuasive the subjects felt they were. This scale ranged from Not at all (1) to Eggy (5). The fourth variable described evaluations of the candidates' characters, and was based on subjects' perceptions of candidates using a seven-point scale for the following items: intelligent, sincere, mean spirited, believable, honest, arrogant, persuasive, concerned, and qualified. This scale ranged from Eggy (1) to Not at all (7). A factor analysis produced two distinct variables describing candidate character. The first factor consisted of the mean spirited and arrogance items, and the second included the remaining seven items. Testing for reliability of the two factors (that is, whether the items could be combined to form appropriate measures of character evaluation), it was found that only the second group of items produced a sufficiently high alpha for each candidate (.88 for the attacking candidate and .89 for the target). For the other two items, the alphas were .52 and .57 respectively. The variable candidate character, then, is an index of seven items which correlated highly and describe subjects' perceptions of the candidates. 33 The fifth variable is a total index, or a summation of standard scores of the other four variables. Means, standard deviations, and reliability coefficients for the constructed indices are presented in Table l. The reliability coefficients show that items were appropriately combined to form indices which describe commercial and candidate evaluations. One-tailed g tests were employed to test: (1) the five dependent variables for the targeted candidate by appeal type (issue, image) and response strategy (rebuttal, non-rebuttal); (2) the attacking candidate's dependent variables by appeal type and response strategy. Analysis of variance was used to test for a possible interaction effect between the two treatment variables, appeal type and response strategy. An alpha level of .05 was required for significance. Additionally, respondents were directly asked several questions regarding their attitudes toward negative political advertising. These were intended to measure the subjects' feelings about how frequently ("as often as they like”, "never") candidates should use attack ads, along with what gypgg ("performance record”, "personal character”) of attacks were acceptable. A group of statements was presented and the subjects indicated their level of agreement or disagreement on a four- point scale, ranging from Strongly agree (1) to Strongly disagree (4). TABLE 1 34 Total Mean Scores, Standard Deviations, and Alpha Coefficients. TARGET Commercial Evaluation Character Evaluation Likely to Vote for Overall Feeling Total Index ATTACKER Commercial Evaluation Character Evaluation Likely to Vote for Overall Peeling Total Index Mean 8.91 28.26 3.05 43.49 .0040 7.72 25.99 2.53 35.70 -.0003 Standard Deviation 2.60 6.37 .99 20.44 3.59 2.73 6.45 .92 20.94 3.45 Alpha .787 .889 .764 .789 .883 .726 NOTE: Asterisk indicates variable is not constructed. RESULTS The first treatment involved effects of appeal type. It was anticipated that the issue treatment would produce a highg; evaluation for the attacking candidate than the image treatment. Conversely, the targeted candidate was expected to do better in the image treatment. Results showed that four of the ten dependent variables were significant, or nearly significant at a .05 level. (Table 2) Of the attacking candidate's five variables, three showed significant differences in the appeal type treatment. With regard to commercial evaluation, the subjects' evaluations of the attacking candidate's issue commercials were significantly higher than evaluations of the image commercials (8.38 vs. 7.02). Evaluations of the attacker's candidate character evaluation is significantly higher for the issue commercials (26.8 vs. 25.15). For the total index variable, which is an overall candidate/commercial measure, there is a significantly higher evaluation of the attacking candidate's issue commercials relative to the image commercials (.48 vs -.49). Regarding the targeted candidate, the likelihood of voting variable approaches significance. In this case, the subjects were more likely to vote for the target after exposure to the image commercials, compared to the issue commercials (3.17 vs. 2.94), supporting the hypothesis. The second treatment involved comparing effects of response strategy. The expectation was that the direct rebuttal would increase the backlash effect. Therefore, scores for the attacking candidate in the rebuttal condition should be lower than scores in the other 35 36 TABLE 2 Mean Scores on Dependent Variables by Appeal Type and Response Strategy. APPEAL TYPE RESPONSE STRATEGY TARGET Issue Image Rebuttal Positive Commercial Evaluation 9.15 8.68 .068 8.50 9.32 .005* Character Evaluation 28.36 28.16 .401 27.95 28.56 .213 Likely to Vote for 2.94 3.17 .029* 3.03 3.08 .340 Overall Feeling 43.35 43.63 .454 41.87 45.06 .099 Total Index -0.02 0.02 .466 —0.31 0.30 .051* ATTACKER Commercial . Evaluation 8.38 7.05 .000* 7.72 7.72 .499 Character Evaluation 26.80 25.15 .017* 25.79 26.18 .306 Likely to Vote for 2.55 2.50 .337 2.41 2.64’ .019* Overall Feeling 37.47 33.91 .080 34.67 36.71 .204 Total Index 0.48 -0.49 .001* -0.21 0.20 .141 NOTE: The asterisks indicate significance, or near significance, at alpha=.05 level, using one-tailed 5 test. 37 condition. Scores for the targeted candidate should be gighgg in the rebuttal condition, compared with the other. Here it was found, to some surprise, that the non-rebuttal condition resulted in significantly higher scores on three of the dependent variables. The first of these describes the subjects' likelihood of voting for the attacking candidate. In this case, as expected, the rebuttal produced a backlash effect. The attacker is significantly less likely to be voted for when there is a rebuttal to the negative ad (2.41 vs. 2.64). The remaining variables which show significant differences are the targeted candidate's commercial evaluation and the total index. The results show lower scores for the target on these variables in the rebuttal condition relative to the non- rebuttal (8.50 vs. 9.32, and -.31 vs. .30, respectively). The rebuttal, it seems, was ineffective in reducing the intended effect of the negative advertising. Each of the scores for the attacker was higher in the pgsitive (non-rebuttal) condition, in comparison with rebuttal, although only one (likelihood of voting) was significant. On the other hand, each score for the target was also higher in the non-rebuttal condition, contradicting what was predicted. Regarding possible interaction effects between the two treatment variables (appeal type, response strategy), there were no significant effects detected. It appears that the effects of issue versus image types of ads are independent of the presence of a direct rebuttal, and vice versa. DISCUSSION It was hypothesized that viewers may have a higher tolerance for negative-issue commercials than for negative-image. Viewers may object more to attacks on candidate personal attributes than to attacks on their stands on issues. An unintended, or backlash, effect might result for the negative-image commercials. Also, the backlash effect may be more evident when the negative spot is paired with a rebuttal than when it is paired with a positive commercial. Table 1 showed clear differences between evaluations of the attacking and the targeted candidates. Mean scores for the sponsoring candidate are lower for each of the dependent variables, indicating a general rejection of the negative strategy. Neither candidate is rated very highly on any variable. The scores indicate no better than lukewarm evaluations for either, which is not surprising given that the subjects could only use the commercials as a guideline for evaluations. The results supported the first hypothesis regarding issue versus image appeal types, but contained an unexpected finding concerning the response strategy conditions. The first hypothesis predicted that negative commercials would be more effective when they consisted of issue content rather than image. It was reasoned that voters might be more critical of an attack on a candidate's personal character, and therefore react unfavorably against the sponsor of the ad. A greater backlash, or unintended effect was expected for the image condition. The results showed that the sponsor of the attack ad was rated lower in the image condition than in the issue. The subjects apparently saw the 38 39 attack on the target candidate's positions on crime and the environment as more acceptable than the attack on his character. The second hypothesis was concerned with the effect of rebuttal. It was expected that the target of a negative commercial could enhance a backlash effect by directly rebutting the attack; in other words, the intended effect of the negative advertising strategy would be diminished or possibly produce an unintended, sympathetic reaction in favor of the targeted candidate. The target should be perceived more favorably in terms of likelihood of voting, overall feeling toward, commercial evaluation, and character evaluation. There was some reason to doubt this hypothesis, as it could be speculated that viewers may prefer candidates to 'take the high road' (i.e., ignore the attacks and not indulge in mudslinging tactics). Rebuttal ads usually contain a counterattack, whereby the target of the original attack criticizes his opponent. (The rebuttal ads used in this experiment refuted the original charges, claiming that the attacker was "distorting the picture” and "resorting to negative tactics".) If this were true, scores for the rebuttal condition may not show a backlash effect. Instead, the backlash could be offset by the negative tone of the rebuttal. This appears to have happened, at least to some extent. The results showed that the gggggg candidate received significantly lower scores for the commercial evaluation and for the total index variables in the rebuttal condition, contradicting the stated hypothesis. It seems the subjects accepted the charges against the candidate and did not find the rebuttal effective in refuting those charges. 40 Overall, the scores for the positive condition are higher, for both candidates. From the perspective of the negative ad's sponsoring candidate, there is a significantly less likelihood of voting for the sponsor when there is a rebuttal. For the other variables, there are no significant differences, although scores are slightly higher for the positive ad. From the targeted candidate's point of view, there is a significant difference on the total index and the commercial evaluation variables, with viewers of the ads perceiving the non-rebuttal situation more favorably. These results suggest a positive response strategy, one which ignores a negative attack, is more appropriate. Perhaps this is because neither of the candidates is well known, and opinions are based only on the commercials. In an actual campaign other factors would come into play. Table 3 and Figures 1 and 2 present a summary of subjects' responses to questions about negative ads. These results lend some support to the other measurements. It appears that the subjects have a general tolerance toward the use of negative commercials. However, there is a strong opposition to attacks on candidates' personal character. About 58% believe that it is okay to criticize an opponent's record, but 86% believe that it is g9; okay to attack a candidate's character. Also, subjects disagree with the statement that candidates should use negative ads as often as they like (85%), but on the other hand, they also disagree with the belief that candidates should never use negative commercials (57%). In short, the experiment's participants show a general dislike of negative ads, but the results suggest that an attack strategy can be effective when it focuses on candidate performance or issues. TABLE 3 41 Responses to 'Agree-Disagree' Statements Concerning Attitudes toward Negative Advertising. (Percentages) (1) Candidates should use negative commercials as often as (2) (3) (4) they like. Strongly Agree Agree No Opinion Disagree Strongly Disagree Candidates should Strongly Agree Agree no opinion Disagree Strongly Disagree (1.5) Total Agree=10.6% (9.1) (4.7) (41.2) (43.4) Total Disagree=84.6% never use negative commercials. (17.5) Total Agree=37.9% (20.4) (4.7) (46.7) (10.6) Total Disagree=57.3% It's okay to criticize an opponent's performance record in a commercial. Strongly Agree Agree no opinion Disagree Strongly Disagree (5.5) Total Agree=58% (52.5) (6.2) (22.6) (13.5) Total Disagree=36.1% It's okay to criticize an opponent's personal character in a commercial. Strongly Agree Agree no opinion Disagree Strongly Disagree (0.7) Total Agree=ll.3% (10.6) (2.9) (38.3) (47.4) Total Disagree=85.7% Percentage (%) 42 Figure l Attitudes toward Using Negative Advertising to Criticize: Candidate Record vs. Character (Percentages) 100 . I AGE NOOPNION 80 I DISAGREE so .h C N O record character Percentage (%) 43 Figure 2 Attitudes toward Using Negative Advertising: As Often As They Like vs. Never (Percentages) 100 _ I AGES I NOOPINION 80 I! DSNNEE 60 4O 20 ' 5 0‘ Z%%%%/ 2%fifl3/ oflen never CONCLUSION As with any experimental procedure, there are certain precautions to be taken in conclusion-drawing. Some problems have already been mentioned. Certainly, most viewers of political advertisements do not base their voting decision solely on what they've seen in TV spots. Factors such as party identification, particular campaign issues, news coverage of candidates, debates, name recognition, are considered. Most media consultants live by the general rule that you don't do negative ads until after you've established an identity first. Once that is accomplished, the use of negative ads are determined largely by the candidate's standings in the polls. Furthermore, the negative strategy is normally just one phase of a political advertising campaign (Diamond and Bates, 1984). This study deliberately eliminated several other factors in order to focus on the political commercials themselves. Another constraint of this study concerns the selection of subjects for the study. The attitudes and behaviors of college students are undoubtedly different from those of other demographic groups. It would be difficult to apply the results of this study to the general population with a large degree of certainty. Still, candidates do often target their messages for young voters, so these results can be applied to an important part of the electorate. The 1984 Garramone study indicated that negative ads are more effective for younger voters. This effectiveness is demonstrated to some degree in this project as well. However, more educated voters tend to react against the sponsor of negative ads, so the full impact of the attack strategy could have been diminished by this factor. 44 45 An additional source of concern might be production differences between the commercials. The overall quality was about equal for each ad. The principal difference consisted of the narrators' voices. The voice for the negative spots belonged to an older, more experienced speaker. The other voice was that of an student actor. Both were practiced at public speaking, but the subjects may have preferred one over the other. Subjects may have evaluated both candidates lower due to the relatively unsophisticated productions, but they probably would not favor one over the other for that reason. In real campaigns, many, if not most, commercials combine issue and image appeals. Patterson and McClure (1976) reported that few spots are either pure issue or pure image commercials. The issue ads used in this experiment addressed themselves to crime and the environment, two issues which are very common ones, and the image ads were mainly aimed at questioning the target candidate's honesty, work habits, and decisiveness. The results of this study suggest a greater tolerance of negative commercials which address an opponent's positions on issues, compared with an attack on character. The study also suggests that it may be more prudent, at times, to ignore an attack, rather than to respond in kind. This is especially true if voters know little about either candidate. A positive response to an attack is better under these conditions. A rebuttal commercial must be designed carefully, so that it is perceived not as an attack, but a defense. A mitigating factor might be the sharpness or severity of the attack. The negative ads in this experiment are moderately harsh. Real campaign commercials vary in 46 the degree to which a candidate is criticized. What may be needed is a comparison of different levels of attacks, ranging from mild to severe. Whether one uses a controlled experimental design, a survey, or a field experiment, it is difficult to gauge exact voter behaviors regarding negative advertising. Many researchers, politicians, and commentators have voiced the belief that by indulging in mudslinging with the aid of the TV spot, candidates are creating a more cynical electorate which accounts for a declining turnout. This idea has not yet been extensively investigated, but deserves attention. APPENDIX A — SCRIPTS SCRIPT 1 VIDEO (Fade in ) Photo of target candidate (Super) "Released criminals" "Opposed environmental protection" Photo (Black - Super) "Carl Martin - U.S. Congress" "Paid for by Friends for Carl martin" (Fade out) 47 NEGATIVE; ISSUE AUDIO (Narrator) Phil Jackson claims to be tough on crime. But look at his record. As a state senator he spon- sored a bill which would have released hundreds of convicted criminals from state prisons. Phil Jackson claims to be for a clean environment. But 3 times he voted against environmental protection bills. Now he wants your vote for Congress. We don't need a Congressman with a record like that. Vote for Carl Martin - for Congress. SCRIPT 2 VIDEO (Fade in) Photo of candidate Photo - out of focus Photo - coming into focus (Super) "Phil Jackson - U.S. Congress" "Paid for by Jackson for Congress Committee" 48 REBUTTAL: ISSUE AUDIO (Narrator) Phil Jackson has been around politics long enough to know that when you run for office, you have to be able to take criticism. But his opponent is carrying things too far. Carl Martin is giving the voters a distorted picture. The truth is, Phil Jackson has consistently fought against crime and £3; a cleaner environment. You can reject the politics of negativism. Elect Phil Jackson to Congress. SCRIPT 3 — VIDEO (Fade in) (Super) "Phil Jackson" Photo of toxic dump site (Super) "Eliminate dumping" Photo - wilderness "Protect environment" Photo - prison "Improve criminal justice system" Photo - candidate "Vote for Phil Jackson - U.S. Congress" "Paid for by Jackson for Congress Committee 49 POSITIVE, ISSUE AUDIO (Narrator) As a state senator, Phil Jackson led the fight against toxic waste dumping in our state. He's been an outspoken advocate of conserving our wilderness areas. Phil Jackson has worked to improve our state's criminal justice system, and has consistently voted for measures which would strengthen law enforcement. Vote for Phil Jackson - for the future. SCRIPT 4 - NEGATIVE: VIDEO (Fade in) Photo of arrows pointing in several directions (Super) "Inconsistent" "Indecisive" Photo of candidate "Missed votes" Photo - Hawaii "Carl Martin for Congress" "Paid for by Friends for Carl Martin" 50 IMAGE AUDIO (Narrator) Indecisive and Inconsistent. Two words that describe Phil Jackson. - You don't know which way he's coming from. But that's not the worst of it. He claims to be honest and hard working. But what about all those votes he missed? And how about that trip he took to Hawaii? Who really paid for that? Not voting is one thing, Mr. Jackson - But not being honest with us is another. Carl Martin will be straight with us. SCRIPT 5 VIDEO (Fade in) Photo - candidate Photo - out of focus Photo - back into focus (Black - Super) "Phil Jackson — U.S. Congress" "Paid for by Jackson for Congress Committee" 51 — REBUTTAL, IMAGE AUDIO (Narrator) Some people will say anything to get elected. Carl Martin has reached a new low in his attempt to discredit Phil Jackson. He's giving the voters a distorted picture. His latest attack questions Phil Jackson's Honesty. He must be desperate if he has tc rely on negative and inaccurate messages. Come on, Mr. Martin. The people deserve better than that. Vote for Phil Jackson — for Congress. SCRIPT 6 VIDEO (Fade in) Photo of candidate (Black - Super) "Leader" "Responsible" "Honest" "Hardworking" Photo — candidate "Phil Jackson - U.S. Congress" "Paid for by Jackson for Congress Committee 52 POSITIVE, IMAGE AUDIO (Narrator) Phil Jackson. A tough fighter, and good leader for the people of our state. A responsible man, who cares about people and wants to help. Congress needs someone who will get things done — someone honest and hard- working. Congress needs - Phil Jackson Vote for Phil Jackson - for the future. APPENDIX B - QUESTIONNAIRE 53 Q1 The news ogram you've ust seen contained several political commercia . Based on t ose commercials, if you were voting in an election involving these two candidates, PNIL JACKSON and CARL MARTIN, about how likely would on be to vote for - PNIL JACKSON? (Please check the appropr ate number.) [1] Very likely [2) Somewhat likely [3] Possibly [4] Not very likely [51 Not at all likely About how likely would you be to vote for - CARL MARTIN? [1] Very likely [2] Somewhat likely [3] Possibly [4] Not very likely [51 Not at all likely Q2 Based on the comercials, please rate candidate PHIL JACKSGI on the following characteristics. Circle the appropriate I‘. not at all INTELLIGENT 1 2 3 4 S 6 not at all SINCERE 1 2 3 4 5 6 not at all MEAN SPIRITED 1 2 3 4 5 6 not at all BELIEVASLE 1 2 3 4 S 6 not at all HONEST 1 2 3 4 S 6 not at all ARROGANT 1 2 3 4 S 6 not at all PERSUASIVE 1 2 3 4 5 6 not at all CONCERNED 1 2 3 4 5 6 not at all QUALIFIED 1 2 3 4 5 6 Q3 Again based on the commercials, please on the following characteristics. not at all INTELLIGENT 1 2 3 4 5 6 not at all SINCERE 1 2 3 4 5 6 not at all MEAN SPIRITED I 2 3 4 S 6 not at all RELIEVARLE 1 2 3 4 5 6 not at all NONEST 1 2 3 4 5 6 not at all ARROGANT 1 2 3 4 S 6 not at all PERSUASIVE 1 2 3 4 S 6 not at all CDNCERNED 1 2 3 4 5 6 not at all QUALIFIED 1 2 3 4 5 6 very 7 INTELLIGENT very 7 SINCERE VGPY 7 MEAN SPIRITED VOVY 7 BELIEVABLE very 7 NONEST very 7 ARROGANT very 7 PERSUASIVE very 7 CDNCERNED VOPY 7 QUALIFIED rate candidate CARL MARTIN VPFY 7 INTELLIGENT very 7 SINCERE very 7 MEAN SPIRITED VGVY 7 IELIEVAILE VPVY 7 HONEST very 7 ARROGANT very 7 PERSUASIVE very 7 CONCERNED very 7 QUALIFIED Q4 Please rate your feelings toward the following political figures on a scale from 0 to 100, with 0 meaning VERY UNFAVORARLE, 100 meaning VERY FAVORARLE, and 50 meaning NEUTRAL. _Ronald Reagan _.li'y Carter ___Nalter Mondale ___George Bush Phil Jackson ___Gary Nart Carl Martin ___Richard Nixon 54 05 Now we would like you to evaluate each candidate's commercial. On a scale of 1 to 5, please rate the advertisements on the following factors. IST Commercial (for CARL MARTIN) NOT AT ALL VERY informative 1 2 3 4 5 believable 1 2 3 4 5 negative 1 2 3 4 5 persuasive 1 2 3 4 5 2ND Commercial (for PHIL JACKSON) NOT AT ALL VERY informative 1 2 3 4 S believable 1 2 3 4 5 negative 1 2 3 4 5 persuasive 1 2 3 4 5 Q6 Now we would like to find out how you feel about negative political TV advertisments, that is, commercials wh ch candidates use to criticize their opponents. For each of the following statements, please indicate whether you strongly agree (SA), agree (A), disagree (D), strongly disagree (SD), or have no opinion (N). a. Candidates should use negative commercials as often as they like. SA A D SD N b. Candidates should only occasionally use negative commercials. SA A D so N c. Candidates should never use negative commercials. SA A D SD N d. If a candidate is going to criticize an t in a commercial the candidate oulddoso mum mi. SA A n so u e. It's okay to criticize an opponent's performance record inaco-ercial. SA A D SD N f. It's okay to criticize an opponent's personal character in a commercial. SA A D SD N Q7 Do you consider yourself a Republican, a Democrat, an Independent, or what? I I Republican I I Democrat I I Independent I I Other QB Did you vote in the election this past November? I IYes I INo Q9 About how interested were you in this past election? I I Not at all interested I I Not very interested I I Somewhat interested I I Very interested 010-12 Uhat is your - AGE? SEX? I I Male I I Female RACE? I I Nhite I I Black I I Asian I I Hispanic I I Other Q13 Nhat is your home town/city and state? 55 LIST or REFERENCES Anderson, J. (1986, October 16). TV's political ads: A battle of the banal. USA Today, p. 8A. Ann Arbor News (1986, August 6). 'Satan' tape proved smoking gun for Siljander, p. A9. Berke, R. L. (1986, August 26). Mud: There seems to be more of it this year. New York Times, p. A18. Much ado about ads knocking candidates. (1985, September 16). Broadcasting, pp. 64-72. Diamond, E. & Bates, S. (1984). The Spot: The Rise of Political Advertising on Television. Cambridge, MA: MIT Press. Firestone, J. (1986, November 3). Most voters skeptical of political ads. Detroit News, pp. 18, 6B. Garramone, G. M. (1984). Voter Responses to Negative Political Ads. Journalism Quarterly, 61, 250-259. Garramone, G. M. (1985). Effects of Negative Political Advertising: The Role of Sponsor and Rebuttal. Journal of Broadcasting & Electronic Media, 12, 147-159. Johnson, K. S. a Copeland, G. (1987). Setting the Parameters of Good Taste: Negative Political Advertising and the 1986 Elections. Paper presented at International Communication Association conference. Joslyn, R. A. (1980). The Content of Political Spot Ads. Journalism Quarterly, gl, 92-98. Joslyn, R. A. (1984). Mass Media and Elections. Addison-Wesley. Kaid, L. L., & Boydson, J. (1987). An Experimental Study of the Effectiveness of Negative Political Advertisements. Communication Quarterly, gs, 193-201. Raid, L. L., s Sanders, K. R. (1978). Political Television Commercials: An Experimental Study of Type and Length . Communication Research, é, 57-70 a Katz, D. (1987, November 1). Candidates face unblinking stare. Detroit News, p. 25A. Lau, R. R. (1982). Negativity in Political Perception. Political Behavior, 5, 353-377. Louden, A. (February, 1987). Lecture given for Department of Communication, Michigan State University. East Lansing, Michigan. 56 Martin, R. (1986, October 21). Consultants shape TV appeals. Detroit News, pp 18; 3B. Merritt, S. (1984). Negative Political Advertising: Some Experimental Findings. Journal of Advertising, 1;, 27-38. Miller, R. L. (1976). Mere Exposure, Psychological Reactance and Attitude Change. Public Opinion Quarterly, 29, 229-233. NBC Evening News. (1986, November 3). Patterson, T. E., & McClure, R. D. (1975). The Unseeing Eye: The Myth of Television Power in National Politics. New York: G. P. Putnam‘s Sons. Robinson, M. J. (1981). The Media in 1980: Was the Media the Message? in The American Elections of 1980. Ed. by Austin Ranney. Washington, D.C.: American Enterprise Institute for Public Policy Research. Sabato, L. J. (1981). The Rise of Political Consultants. New York: Basic Books. Surlin, S. H., S Gordon, T. F. (1977). How Values Affect Attitudes Toward Direct Reference Political Advertsing. Journalism Quarterly, g2, 89-98. Wicker, T. (1986, November 14). Big Losers of 1986? New York Times, p. 27. Wills, G. (1986, November 10). Mud-slinging: You asked for it. Detroit Free Press , p. 9A. "I111mmES