William N. Swain, Gary A. Copeland and Karen S. Johnson-Cartee
[WJMCR 3:4 September 2000]
An experiment tested the effects of the race and gender of the sponsor of a humorous negative political commercial on perceptions of the sponsoring candidate, the commercial itself, and a white male political opponent. Five groups of participants read a brief biography of the candidate sponsoring the commercial, four of the biographies with a different pictures of the fictitious candidate-black female, black male, white female, white male-and the fifth with no picture at all with the biography. Participants were then shown a humorous negative political commercial and asked to express perceptions on semantic differential scales. The race and gender of the sponsor of the humorous negative political commercial had no apparent effect on perceptions of the candidates or the commercial.
The ability of negative political advertising in several effective forms-direct attack, direct comparison, and implied comparison-has been found, in research of the 1990s, to be a powerful force in defining the political campaign battleground and lowering a targeted candidate’s evaluations more than those of the sponsoring candidate, in spite of the low esteem viewers ascribe to negative ads.1 For example, Bruce Pinkleton, in an experimental investigation, found that comparative political advertising with negative content appears to reduce voting preferences for targeted candidates without engendering political disaffection for or backlash against sponsoring candidates, and may, in some instances, enhance the situational political involvement of voters.2
Arguing that voters in latter-day elections have become more likely to vote against rather than for a candidate, Karen S. Johnson-Cartee and Gary A. Copeland suggest that “it has become increasingly important to provide prospective voters with reasons for not voting for the opposition.”3 Negative advertising,they add, has been the dominant form of political advertising since the 1980s.
Noting President Bill Clinton’s ability, in his 1996 campaign, “to run the most negative advertising campaign in the history of presidential elections and yet be perceived by voters as less negative than Bob Dole,” Judith S. Trent and Robert V. Friedenberg credit Clinton’s use of attack commercials, a strategy which they suggest can contribute to a perception of a candidate as strong and decisive.4
Trent and Friedenberg also have noted a strategy in which some candidates “have fastened labels on their opponents and then defined those labels negatively.”5Consistent with that strategy is the suggestion by David Okeowo and William N. Swain that a political entity seeking the approval of an audience may benefit by first positioning the opponent as worse on some standard of acceptability, thereby claiming a comparatively more positive valence perception than the opponent, and may then successfully employ communication to achieve increased visibility for the comparative valences.6
As the concept of negative political advertising has became better operationalized in research, its various forms and its powers to shape campaigns have become better understood and accepted.7 Research on the topic has thus increasingly incorporated investigation of the factors and circumstances that may determine the effectiveness of negative political advertising.
Research has revealed much more about the influence of gender on negative political advertising8 than it has about the influence of race9; and, there is virtually no literature on the dual influences of race and gender in negative political advertising.
Studies on race stereotype and gender stereotype also do not explore the potential impact of one factor upon the other.10 Stereotype is important to consider in negative political advertising because it can influence our evaluations of individuals, sometimes in conflicting ways.11
The less familiar a candidate is to a voter, the more gender stereotype has been found to be a factor in voter support.12 However, even in a national election, in which candidates tend to be more familiar to voters, gender stereotype has tended to work in favor of male candidates.13 Racial stereotype also serves as a cue in low-information elections, where voters tend to label black candidates as more liberal, more concerned about minority rights and the plight of the poor.14
There also is evidence that racially intolerant white voters simply will not support a black candidate-regardless of his personal characteristics, experience, or stands on issues-while racially tolerant white voters, sensitive to the social consequences of expressing racial sentiments, may express support for a black candidate, but then may behave differently in the voting booth.15 Lee Sigelman and Susan Welch found, in a secondary analysis of social survey data, that a quarter of respondents said either that they would not vote for a presidential candidate who was female or that they would not vote for a candidate who was black.16
Wayne Parent argues that racial political polarization continues in part because of racial partisanship. “After all,” Parent added, “elections are run by consultants who base strategies on existing evidence. It is therefore understandable that campaign messages�especially voter mobilization messages-involve some appeal to racial consciousness.”17
Finally, in the single study we found that explored race and gender together, it was found that, while many white men and women differ in their positions on some political issues, most black men and women hold similar viewpoints in ideologies, party affiliations, women’s issues, military support, and social programs.18
Turning to the role of gender and race in negative political advertising, it is evident that candidates construct their messages either to reinforce or to violate expectations based on gender stereotype.19 While radio, television, and video have awakened male candidates to the advantages of adopting a more feminine style20, women have tended to assume more masculine traits, portraying themselves as aggressive and career-oriented, while avoiding traditional feminine traits such as nurture, cooperation, or supportiveness.21
In attack ads, women tend to attack issues more than character.22 They offset images of competence and strength with feminine images to accommodate “gender-based stereotypes of femininity even while using a campaign tool that is inherently aggressive, confrontational, argumentative, or, in other words, stereotypically masculine”.23 In essence, a female candidate cannot portray herself as too masculine or she will alienate voters who expect her to be feminine, and she cannot appear too feminine or she will alienate voters who expect her to display masculine traits of strength and leadership.24
On the other hand, because attack ads are more in keeping with perceived masculine traits, research has suggested that a male candidate does not have the balance issues that a female has25, and that, because he is already assumed to possess competence (unless proven otherwise), a male candidate can employ feminine traits (e.g., nurture or compassion) to enhance his image.26
However, a study of advertising for statewide elections in 1990 found little evidence that gender affects negative advertising communication styles; researchers reported that women’s campaigns construct negative ads just as men’s campaigns do: they attack an opponent’s orientation or ethics via an unseen narrator and then present undocumented evidence to support the claims.27
While scholars are exploring the impact of gender on negative political advertising and making great strides in understanding what makes it effective, there remains a dearth of information about the influence of race on negative political advertising. An extensive search of the published literature revealed only one such study in which blacks�especially blacks with group-oriented values�perceive “direct reference” (or attack) ads as providing more positive information and as creating a more favorable attitude toward the sponsor of the ad than whites do.28
A paper presented at the April, 2000, convention of the Southern States Communication Association by Hong Sik Yu, Meg Lamme, Kelly Lewis, Karen S. Johnson-Cartee and Gary A. Copeland reported on a study of the effects of gender and race of the sponsor of a humorous negative political commercial on the commercial, the sponsor, and a white female opponent. A companion to the research reported here, the experimental study employed a similar design but with a white female rather than a white male opponent. The convention presentation reported one significant finding: that a black male candidate whose commercial attacks the record of a white female opponent is perceived less positively than such a candidate who is a white female, a white male, a black female, or a candidate whose race and gender are unidentified.29
This scarcity of information could be due to a lack of artifact; that is, there just may not be many negative political ads sponsored by black candidates. But even if this were true, the increasing numbers of black men and women running for office30, combined with the efficacy and ubiquity of negative political ads in campaigns at all levels31, sets the stage for the study and understanding of another dimension in negative political advertising, the combined influences of race and gender.
Hypotheses for this study were based on findings of prior studies suggesting that negative or attack advertisements by women candidates for political office are typically structured differently from those of political candidates who are men32, that negative political advertising can be effective for black candidates33, and that stereotypes work in favor of dominant demographic groups.34 Two hypotheses each were constructed to test the effect of ethnicity and gender on the three subjects of evaluation in the study: the candidate sponsoring the negative advertisement, the negative advertisement itself, and the target of the negative advertisement, the political candidate’s opponent.
Hypotheses one and two relate to evaluation of the target of the political advertisement, the opposing candidate:
H1: The perceived ethnicity of the sponsor of a humorous negative political advertisement will significantly affect the evaluation of the target of the ad.
H2: The perceived sex of the sponsor of a humorous negative political advertisement will significantly affect the evaluation of the target of the ad.
Hypotheses three and four related to evaluation of the political candidate, the fictitious Pat Adams, who is the sponsor of the ad:
H3: The perceived ethnicity of the sponsor of a humorous negative political advertisement will significantly affect the evaluation of the sponsor of the ad.
H4: The perceived sex of the sponsor of a humorous negative political advertisement will significantly affect the evaluation of the sponsor of the ad.
The fifth and sixth hypotheses related to evaluation of the negative political advertisement itself:
H5: The perceived ethnicity of the sponsor of a humorous negative political advertisement will significantly affect the evaluation of the ad.
H6: The perceived sex of the sponsor of a humorous negative political advertisement will significantly affect the evaluation of the ad.
An experiment was conducted at a small Southern university to investigate the effect of race and gender of the sponsor of a negative political television commercial. The study evaluated the impact of five versions of a fictitious candidate named “Pat Adams” upon the ad sponsor, the ad target, and the ad, itself. Each of five groups of subjects was presented with one of five biographical sheets, four of which were printed with one of four photos: a black man, a black woman, a white man, and a white woman. The fifth biography did not include a photo nor did it identify the gender or race of Pat Adams.
Student volunteers (n=150) were recruited from classes in communication, psychology, and English, and randomly assigned to one of five treatments (n=30 each treatment). Participants read and signed an informed consent form before participating in the experiment, and were later briefed on the purpose of the study, following their participation.
Women comprised 60 percent of the group, and men 39 percent; 1 percent did not specify their gender. By ethnicity, 60 percent of participants were Caucasian, and 31 percent African American; 9 percent were of Asian, Hispanic, Native American or of unspecified heritage. By age, they ranged from 18 to 52, but 79 percent were 22 or below, and 93 percent were 27 or below. In political party affiliation, 40 percent associated themselves with Democrats, 26 percent with Republicans, 18 percent Independent, and 14 percent other.
Participants were asked to read a biography of a fictitious political candidate, Pat Adams. Each of the five treatment groups saw a different version of the page containing the biography. Four of the five biographies were accompanied by a picture of the candidate, one of a black female, one of a black male, one of a white female, one of a white male. The fifth version of the biography was presented with no picture, as a control. The pictures used were each the most favorable to the race and gender category of a set of pictures pretested before the experiment.
After reading the biography, participants were asked to evaluate the candidate on 22 semantic differential scales. The 7-level bi-polar scales assessed the candidate’s competence, strength, intelligence, qualification, thrift, trustworthiness (worded two ways), sympathy, reliability, safety, faithfulness, effectiveness, confidentiality, benevolence, candor, deceitfulness, straightforwardness, respectfulness, considerateness, honesty, sincerity, and carefulness. Those scales had been successfully used in previous studies of the effects of political advertising on political candidates.
Participants were then shown a television commercial, one from an actual campaign, in which the legislative and personal spending record of Pat Adams’ presumed opponent, State Legislator Buzz Andrewzeski, is attacked through ridicule. The commercial, filmed against a neutral background, had a lab-coated clipboard-holding announcer enter and position himself between two clear vats approximately four feet in diameter by five feet high, each with a chain-suspended handle over it. The announcer then read from the clipboard a series of alleged facts about Buzz Andrewzeski’s spending record, together with some side comments about not attending legislative sessions and other shortcomings. After each claim of excessive spending, the announcer pulled the chain above the vat labeled “spending,” and a large splash of water descended from above the frame. When the announcer finished reading the list, the “spending” vat was full to the brim. Then the announcer looked for Buzz Andrewzeski’s legislative accomplishments on the clipboard, finding only one bill, for the renaming of a highway. He then pulled the chain above the vat labeled “accomplishments,” and a single drop of water descended from above into the vat, complete with sound effects.
The negative ad used was the same for each of the five treatment conditions. The commercial used disparaging humor to attack a male opponent with a last name of Polish derivation and thus white by implication, though not by specification. Although such humor “belittles, debases, demeans, humiliates, or otherwise victimizes” others, and people tend to resent humor that ridicules someone whom they respect35, the humor in the commercial shown was identical for each of the five groups in our study and thus constant as a variable. Participants viewed the commercial three times, the number of viewer exposures recommended as optimal for a television commercial.36
After viewing the commercial, participants were asked to complete three more 7-level questionnaires, one a set of Likert scale questions on the participants’ perceptions of the commercial itself, then semantic differential scales for the opponent, Buzz Andrewzeski, and for the candidate, Pat Adams, similar to that used to assess Pat Adams before the commercial was shown.
Participants were then asked to assess the likelihood of their voting for Pat Adams, of voting for Buzz Andrewzeski, and of voting in an election between the two, each on a 10-point Likert scale. Finally, participants were asked to provide demographic information: age, gender, ethnicity, class standing, citizenship, political party affiliation.
A principal components factor analysis identified factor clusters of variables for the commercial, for political opponent Buzz Andrewzeski, and for the pre-commercial reaction to the political candidate, Pat Adams. Component correlations in excess of .35 for most comparisons justified the use of direct oblimin rotation in the factor analysis procedure.
For Pat Adams, the candidate sponsoring the commercial, five variables�not deceitful, saver, qualified, strong, and competent�were related as a factor to be called performance. Eight variables�not deceptive, can be trusted, candid, straightforward, safe, respectful, honest, and sincere�were identified as a factor to be called personality. Seven variables�careful, saver, qualified, competent, sympathetic, effective, and sincere�clustered as a factor to be called character.
For Buzz Andrewzeski, Pat Adams’ opponent, criticized in the commercial, three factors emerged. The variables smart, qualified, effective, can be trusted, sympathetic, and safe clustered as a factor labeled competence. A factor labeled politic included the four variables confidential, considerate, candid, and not deceitful. Twelve variables clustered in a factor to be called nature: benevolent, honest, sincere, careful, candid, not deceitful, faithful, can be trusted, sympathetic, reliable, safe, and trustworthy.
Three factors also emerged for the commercial. The seven variables annoying, offensive, exaggerated, hilarious, enjoyable, irritating, and funny comprised a variable to be called qualities. The seven variables believable, convincing, informative, offensive, exaggerated, accurate, and tasteless clustered as a factor to be called perception. Eight variables formed a factor to be called reaction: believable, convincing, tasteless, informative, hilarious, enjoyable, entertaining, and funny.
To test the hypotheses, one way ANOVA statistical tests were run for each factor, incorporating each component variable of the factor. Delta scores were calculated for Pat Adams, the candidate sponsoring the commercial by subtracting pretest scores from post test scores, and the one way ANOVA was run on the data representing the difference.
Hypotheses One and Two: Evaluation of the Target of the Commercial
Hypotheses one and two anticipated that the evaluation of the target of a humorous negative political commercial would be significantly affected by the perceived ethnicity (H1) and sex (H2) of the sponsor of the commercial. Those hypotheses were tested by one-way ANOVA tests on the component variables of three factors identified by factor analysis of the evaluations of the target of the Pat Adams commercial, the opponent, Buzz Andrewzeski. The results of the Buzz Andrewzeski factor analysis and ANOVA tests are contained in Table 1.
No statistically significant (p<.05) differences were found in any component variables of any factors identified in the evaluations of the target of the commercial, the opponent Buzz Andrewzeski, based on the ethnicity or sex of the sponsor of the humorous negative ad, Pat Adams. Thus, neither hypothesis one nor hypothesis two is supported by the data. The ethnicity and sex of the sponsor of the commercial had no effect on evaluation of the target of the humorous negative commercial, the political opponent of the sponsoring candidate.
Hypotheses Three and Four: Evaluation of the Sponsor of the Commercial
Hypotheses three and four anticipated that the evaluation of the sponsor of a humorous negative political commercial would be significantly affected by the perceived ethnicity (H3) and sex (H4) of the sponsor of the commercial. Those hypotheses were tested by one-way ANOVA tests on the component variables of three factors identified by factor analysis of the evaluations of the sponsor of the commercial, political candidate Pat Adams. Evaluations of Pat Adams, the candidate, were performed by study participants both before and after exposure to the humorous negative commercial, and therefore, both before and after, participants were aware that Pat Adams was the sponsor of the humorous negative commercial attacking the opponent. Factor analysis was performed on delta scores calculated for Pat Adams by subtracting pretest scores from post test scores. The results of the Pat Adams factor analysis and ANOVA tests are contained in Table 2.
For the evaluations of the sponsor of the humorous negative commercial, candidate Pat Adams, none of the component variables in the three identified factors yielded a statistically significant difference based on the ethnicity and/or sex of the sponsor of the humorous negative ad. Thus, neither hypothesis two nor hypothesis three is supported by the data. The ethnicity and sex of the sponsor of the commercial had no effect on evaluation of the sponsor of the humorous negative commercial.
Hypotheses Five and Six: Evaluation of a Humorous Negative Commercial
Hypotheses five and six anticipated that the evaluation of a humorous negative political commercial would be significantly affected by the perceived ethnicity (H5) and sex (H6) of the sponsor of the commercial. Those hypotheses were tested by one-way ANOVA tests on the component variables of three factors identified by factor analysis of the evaluations of the Pat Adams commercial. The results of the commercial’s factor analysis and ANOVA tests are contained in Table 3.
No statistically significant (p<.05) differences were found in any component variables of any factors identified in the evaluations of the humorous negative commercial based on the ethnicity or sex of the sponsor of the humorous negative ad, Pat Adams. Thus, neither hypothesis five nor hypothesis six is supported by the data. The ethnicity and sex of the sponsor of the commercial had no effect on evaluation of the humorous negative commercial itself.
In summary, no statistically significant differences among the ethnicity and gender identification treatments at the p<.05 level were found in the ANOVA tests on any of the factors or variables for candidate Pat Adams, the candidate’s opponent, Buzz Andrewzeski, or the commercial. Thus, none of the hypotheses are supported by the data. It would appear that the perceived ethnicity and the perceived sex of the sponsor of a humorous negative political advertisement made little or no difference to the evaluation of the sponsor, the advertisement, or the opponent who is the target of the humorous negative political commercial. Those findings are very nearly identical to the Yu et al. study37 that is a companion to this study, and the similarity of the findings suggests an acceptable degree of reliability for both studies.
The findings of this study are at variance with the findings of earlier studies that formed the basis for the hypotheses in this study38, but may be consistent with the findings of studies suggesting that female and male political candidates have moved closer to one another in advertising style and portrayal of their traits and issues.39 No evidence was found of the gender stereotype effect on political candidacies found in prior research,40 either in the gender or in the race of the candidate.
The findings of this study are not at odds with the assertion of Johnson-Cartee and Copeland that documented evidence in support of an attack on character or ethics enhances the strength of the ad.41 The commercial used in the experiment offered certain presumed facts, read from a clipboard, but offered no documentation in support of the accusations.
The findings may also suggest that young, college-capable citizens of the type who participated in this study may be prepared to evaluate candidates for state office on the merits of their qualifications, rather than on their race or gender. Another possible interpretation is that, in the absence of compelling evidence of qualifications and documentation of charges, they simply may not choose to exercise their selective judgment; and in that case the role of race and gender in forming perceptions may be moot.
Means for the variables evaluating the humorous negative political advertisement were all well below the midpoint of the scales, indicating that the audience did not view the commercial’s characteristics of appeal positively. By extension, the low regard for the commercial, coupled with the fact that the commercial had no dramatic impact on the perception of its sponsor, Pat Adams, may suggest that participants in the study evaluated the commercial without transferring their low evaluation of the candidate’s advertisement to the candidate himself/herself in any systematic way. Such an interpretation would be consistent with the findings of prior research in several studies.42
The use of college students as participants in this study may be regarded as a limitation of the study; however, the use of college students in experimental communication research has been defended by Pinkleton and by Johnson-Cartee and Copeland on the grounds that college students’ education and the limited life experience associated with their ages compensate for each other, and that citizens who complete college are found to vote much more often than non-college-educated citizens and so, by extension, may represent a major portion of the future electorate.43
Future research might profitably focus on race and gender differences in perceptions when the advertisement or commercial presents documentation in support of negative claims, when the advertisement is presented in other media, and when, before the experimental treatment, the participants can be identified as favorable, unfavorable, and neutral towards the issue raised in the advertisement. Further, Yu, et al. assert that the increasing representation of Latino and Asian-American candidates in American politics justifies research on the effects of those ethnic backgrounds on campaign communication.44
The effects of ethnicity and gender in circumstances of negative political advertising when the advertisement adopts a tone other than humorous may also be an avenue for further research. Trent and Friedenberg note that, in addition to humor and ridicule, candidates have attempted, in campaign advertising, to apply guilt by association, negative labeling, fear appeals, suspicion, anxiety, non-specific implications and other negative techniques.45
Finally, the effects of race and gender in positive political advertising46 and in negative comparison advertising might also bear scrutiny. Johnson-Cartee and Copeland suggest a classification of negative political advertising that includes direct attack, direct comparison, and implied comparison advertising.47 Pinkleton has focused on comparison advertising, but not on the effects of race and gender together in comparison advertising.48
About the Authors:
William N. Swain is Associate Professor in the Department of Communication at University of Louisiana at Lafayette. Gary A. Copeland is Professor of Telecommunication and Film, and Communication Studies in the College of Communication and Information Sciences at The University of Alabama. Karen S. Johnson-Cartee is Professor of Advertising and Public Relations, and Communication Studies in the College of Communication and Information Sciences at The University of Alabama. Please send communication regarding this article to William N. Swain, Department of Communication, P.O. Box 43650, University of Louisiana at Lafayette, Lafayette, LA 70504, or call 337/482-6358 (office), 337/993-7869 (home). Send faxes to 337/984-3670, and emails to firstname.lastname@example.org and email@example.com.