Content-Type: text/html

The Impact of Political Advertising: Differences between Positive Ads and Issue, Image and Mixed Attacks






The Impact of Political Advertising: Differences between Positive Ads and Issue, Image and Mixed Attacks





Sung Wook Shim
Doctoral Student
College of Journalism and Communications
UNIVERSITY OF FLORIDA

Address: 3600 SW 23rd St. #A2
Gainesville, FL 32608
Tel: 352)338-7805
E-mail: <a href="/cgi-bin/wa?LOGON=A3%3Dind0101B%26L%3DAEJMC%26E%3Dquoted-printable%26P%3D8363877%26B%3D--%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D_5852201%253D%253D_%26T%3Dtext%252Fplain%3B%2520charset%3Diso-8859-1" >[log in to unmask]</a>









*Manuscript submitted to the Mass Communication & Society Division for consideration of presentation at the AEJMC annual convention, Phoenix, Arizona, August 2000
 I do no wish to be considered for the Moller Award.


Abstract

The purpose of this study is to identify the impact on the attacking candidate when he/she attacks the attacked candidate with four types of ads: issue, image attacks, both issue and image combined attacks and positive. The study results show that image attack produced a greater negative change than issue attack for evaluation of attacking candidate. The decline was significant between likelihood of voting for attacking candidate in the pretest and likelihood of voting for attacking candidate in the posttest. It is assumed that subjects didn't like negative political advertising. Compared to positive advertising, negative political advertising with issue attack can be treated as almost the same as the positive advertising. Finally, negative political advertising did produce source derogation, causing harmful effects on the character evaluation and the likelihood of voting for attacking candidate.



























Introduction

Campaign advertising in the 1996 presidential race was the most negative in the history of presidential campaigns (Kaid, 1997). Politicians are still going negative. The reason is that negative commercials usually work, even if they might alienate voters (Germond & Witcover, 1996). While the definition of negative political advertising varies, Surlin and Gordon "operationalized the genre as advertising which attacks the other candidate personally, the issues for which the other candidate stands, or the party of the other candidate" (as cited in Garramone, 1984, p. 250). 
Much research shows that negative political advertising is still widely used. In the1992 presidential campaign, fifty-five percent of the Bush campaign ads and sixty-nine percent of the Clinton ads were classified as negative ads attacking the opponent. "The Clinton percentage is the highest percentage of negative ads ever documented in a presidential campaign." (Kaid, 1992,  p. 116). 
The purpose of this study is to identify the impact on the attacking candidate when he/she attacks the attacked candidate with four types of ads: issue, image attacks, both issue and image combined attacks and positive. An experiment with manipulated print ads was conducted to examine the character evaluation, advertising evaluation and the likelihood of voting for the attacking candidate. Other studies have looked at issue vs. image. In this study, differences among three attacks will be examined. In addition to this, the differences between positive ads and three types of negative ads will be found.
Negative political advertising has unintended effects. Some studies about negative political advertising indicate that voters did not like negative ads and found them unnecessary (Garramone, 1984; Merrit, 1984; Pinkleton & Garramone, 1992). It also indicated that a strong attack on a candidate may bring a backlash against the attacker, rather than the attacked (Garramone, 1984). Critics of negative political advertising argue that these ads contribute to voter alienation and apathy, and are harmful to the U.S. democratic process (Pinkleton, 1997). 
In spite of this, negative political advertising continues. Empirical research indicates that negative comparative advertising lowers the attacked candidate's evaluation without lowering the attacking candidate's evaluation (Pinkleton, 1997). Exposure to a spot with technological distortion, which was more likely to occur in negative spots than in positive ones, would fortify the image and vote likelihood for the attacking candidate while lowering the image and vote likelihood for the opponent (Kaid, 1997). 
Because the difference between issue vs. image advertising has been examined, More research needs to be conducted to analyze the effects of  an issue attack, an image attack, and a mixed attack about an attacking candidate and positive advertising. 

Literature Review

Negative political advertising
Negative political advertisements focus on the weaknesses of an opponent rather than on the positive attributes of ad's sponsor. Negative advertising has played an increasingly important part in campaigning, taking on a more important role from the 1964 presidential election to the present (McNair, 1994).  Indeed, 1988 was the year of the best-known negative advertising of all because of numerous negative ads such as the "Revolving Door." This Bush commercial sought to portray Dukakis as soft on crime by claiming the Massachusetts governor had voted against the death penalty and had given weekend furloughs to first degree murderers not eligible for parole (West, 1993). Another example is the negative spot by the Bush's side which contrasted Dukakis' declared "green" policy with his record as governor in Boston, where it was alleged he had allowed the Boston harbor to become polluted (McNair, 1994, p. 94).
As Johnson-Cartee and Copeland (1991) suggest, there are several types of negative ads, various designs used in the transmission of negative issue appeals, and many kinds of stylistic techniques for constructing negative spots. 

Independent variable check
Negative political advertising is attack advertising. It contains an aggressive, one-sided assault, designed to draw attention to an opponent's weaknesses in either character or issue positions (Johnson-Cartee & Copeland, 1989; Merrit 1984; Surlin & Gordon 1977; Pinkleton 1997).
Negative issue attack cites a candidate's position on specific issues or items of public policy. As for the pretest about how important it is that issues are related to financial aid, housing, raising wage level, and health benefit, the mean scores are over 4.20, above the average of 3.5. It is assumed that the issues that were used in this study are appropriate for the students. 
Negative image attack cites a candidate's personal characteristics or traits without addressing specific issue stands. Such advertising may contain information about an opponent's medical history, personal life, religion, sex life, or family members (Johnson-Cartee & Copeland, 1989). As for the pretest about how important it is that images are related to medical history, personal life, religion, sex life, or family members, the mean scores are over 4.00, above the average of 3.5. It is assumed that the images that were used in this study are appropriate for the students. 
Negative mixed attack cites both an issue and an image attack. The traits of both attacks are included in this attack. This format of attack might be typical in negative political advertising.
A one-way ANOVA was used to examine participants' perception of stimulus. The results show significant differences in participants' perceptions of the issue/image/mixed stimuli [F (3,94)=12.75 p<.001.]

The effects of negative political advertising 
Intended effects
The intended effect of negative political advertising is to create negative feelings toward the attacked candidate and positive feelings toward the attacking candidate (Garramone, 1984).  Surlin and Gordon (1977) found direct attack ads were unethical but informative. Comparisons between candidates, even if tilted to one side, are usually rich with information (as cited in Johnson-Cartee & Copeland, 1991).  Negative political advertising may well offer more information to voters than positive ads, and people retain negative information more easily than positive information (Johnson-Cartee & Copeland, 1991). 
In fact, negative political advertising has often been cited as offering greater issue clarity than positive spots, which is especially helpful for people who are not familiar with the campaign or political affairs (Hagstrom & Guskind, 1986). Negative political advertising also increases political debate. It actually opens up the campaign to the issues by making the candidates respond to the charges and countercharges. These negative ads are used strategically to set the terms of the campaign debate by framing the issues (Nugent, 1987). 
Additionally, negative spots can exert a strong impact on memory. People remember negative ads better than positive ads. They also recognize negative ads more accurately and quickly than positive spots (Shapiro & Rieger,  1989 ). There are many reasons for this, like the psychological tendency to attach more weight to negative rather than positive information. In addition to this, the other effects are that the negative spot has a stronger impact if there is an already existing perception, negative spots have a better production value, and negative spots get more news coverage than positive ads (Perloff, 1997).

Unintended effects
Negative political advertising is "often thought to promote alienation because it has been viewed as increasing the negativity of the political process. Political advertising in general is not the most highly regarded form of communication, and negative advertising is seen as something worse" (Johnson-Cartee & Copeland, 1991, p. 276). Garramone (1984) found negative advertising has a "strong negative influence on the viewer's feeling toward the sponsor but only a slight net negative influence on feelings toward the target (p. 256)." Merrit (1984) also investigated negative political advertising and found negative advertising produced negative effects toward both the sponsor and the target.
Theoretical background

Comparative product advertising
Comparative product advertising provides some insights into negative political advertising (Johnson-Cartee & Copeland, 1991). Because of the nature of negative comparative advertisements, they may produce a differential and perhaps more "powerful influence on the receivers' cognitive organization and knowledge structure for the communicated information" (Wilson et al., 1979, p. 566). Because comparative advertising may be used both to make statements about the attacking brand and to make point-by-point contrasts about the attacking and competing brands, this type of advertisement may hold more attention and involvement on the part of receiver than would similar non-comparative advertisements (Wilson et al., 1979). In addition, while the receiver encodes the relationships brought forth in a comparative message, a deeper level of information processing may result, and the receiver may retain a more meaningful perception of the message than would occur under a non-comparative advertising condition (Wilson et al., 1979). 
According to the cognitive response model, thoughts (cognitive response) elicited in various forms of communication can be seen as underlying factors which mediate the overall effectiveness of that communication (Wilson et al., 1979). 
Wright (1973) categorized three cognitive responses to advertising communication as support argument, counterargument, and source derogation. These variables can be summarized as follows:

Support Argument: in relating incoming information to existing beliefs, the receiver may activate responses indicating that congruent associations have been discovered or that the message argument is supported by already entrenched beliefs.

Counterargument: activated when incoming information is compared to the existing belief system and a discrepancy is noted. The spontaneous thought activated is assumed to neutralize or counter message evidence.

Source derogation: an alternate type of resistive response focuses on the source of the information. The source derogating response may serve as a substitute for counterargument and may be used quite frequently in situations where the source is easily viewed as biased (an unfortunate description of mass-media advertising p. 54).

Merrit (1984) showed examples of these variables. Support arguments include negative images of the target ("He is indeed that bad"). The reduction of dissonance through counterargument gives images of the target that balance the image ("He is changed"). Source derogation involves negative images of the sponsor ("He is a mudslinger") or of the message ("It's misrepresentation") in the negative political advertising context (p. 29)
Wright indicated that source derogation is a more frequent response to dissonance than counterargument since the source might be assuredly biased (as cited in Merrit, 1984). Also, it can be expected that negative political advertising is likely to produce source derogation and cause harmful effects toward the attacking candidate when viewers see attacking political advertising. 
In terms of negative political advertising, comparative political advertising is more effective than comparative product advertising in achieving the desired communication goals of recall, image enhancement, and persuasion for a number of reasons (Johnson-Cartee & Copeland, 1991).  The effectiveness of negative advertising can be explained in terms of comparative advertising.

Expectancy theory
Expectancy theory focuses on the relationship between language use and the effectiveness of such language use on persuasion (Burgoon & Miller, 1985). Expectancy theory identifies two violations: positive or negative violation. In terms of the persuasion effects of the two violations, the theory assumes that when messages positively violate people's linguistic expectations, the violation has a positive influence on people's attitudes and evokes persuasive effectiveness (Burgoon & Miller, 1985). Contrary to this, when messages negatively violate people's linguistic expectations, a "boomerang effect occurs, with receivers changing to the position opposite to the one advocated by the communicator" (Burgoon & Miller, 1985, p. 200).  
As one would imagine, expectancy theory can be applied to various fields of study. One application concerns the effectiveness of negative political advertising. Based on the theory, it can be assumed that voters have normative expectations about negative political advertising because such advertising is one of the most common genres in today's politics, and voters have been exposed to various negative ads (Park, 1996). Since most negative ads have been employed to attack the opponent's image or issues, people may expect negative political ads to have standard formats and intense messages. If negative political ads conform to people's normative expectations, expectancy theory defines it as a negative violation and predicts that it arouses negative effects toward the sponsor (Park, 1996). 

Image versus Issue

Not surprisingly, image versus issue is a longstanding theme in politics. As far as West's study is concerned, from 1952 to 1992, domestic performance and specific policy statements have been the object of negative ads more than personal qualities. In 1980, ninety-five percent of ads dealing with domestic matters were negative, as were seventy-three percent of those in 1984 and eighty-three percent in 1988.  In the 1992 primary television spots, fifty-nine percent of these primary ads for all candidates concentrated on candidate images, and twenty-four percent stressed issues (Kaid & Ballotti, 1991). 
In qualifying negative political advertising, Johnson and Copeland (1987) identified 10 negative advertising topics that they classified as either "political" or "personal." Political topics include political record, issue stands, and voting record. Personal topics include personal life, current or past marriage, criminal activities, family members, religion, medical history, and sex life. This political/personal distinction is suggestive of political advertising's longstanding issue/image distinction (Roddy & Garramone, 1988, p. 417). 
While issue appeals are related to specific policies, image appeals are related to personal characteristics of the candidate (Kaid & Sanders, 1978). Tests of the effectiveness of issue versus image political commercials have found that issue ads brought a more favorable candidate evaluation (Kaid & Sanders, 1978) and a greater intention to vote for the candidate (Garramone, 1985) than image ads. Moreover, Johnson-Cartee and Copeland concluded that voters were more likely to tolerate negative commercials that focus on policy than on personality (as cited in West, 1993, p. 52). In the Thorson et al.'s study (1991), issue commercials have a more positive impact on attitudes and voting intent than image commercials but image commercials produce better memory scores than issue commercials. 
Even though political advertising can be divided into issue and image advertising, sometimes the boundary between them might be unclear. Also, because combined issue and image advertising is shown, it is appropriate to examine the mixed advertising in this study.
Based on these studies, it is hypothesized that:



H1: Negative political advertising containing an image attack will produce more negative change in the evaluation of the attacking candidate than advertising containing an issue attack on the attacked candidate.

H2: Negative political advertising containing an image attack will produce more negative change in likelihood of voting for the attacking candidate than advertising containing an issue attack.

H3: Negative political advertising will produce source derogation, causing harmful effects of (a) the character evaluation, and (b) likelihood of voting for the attacking candidate.

Research questions are as follows:
RQ1: What is the difference in the three attacks in the evaluation of the attacking candidate's character when a mixed (issue and image) attack is added? 
RQ2: What is the difference in the three attacks in the likelihood of voting for the attacking candidate when a mixed (issue and image) attack is added?
RQ3: What is the difference in the three attacks in the evaluation of advertising when a mixed (issue and image) attack is added? 
RQ4: What is the difference between positive ads and negative ads in the evaluation of the attacking candidate's character and in the likelihood of voting for the attacking candidate?

Method

A 1(4 pretest-posttest experiment was conducted on November 20, 1998, to test the hypotheses and answer the research questions. All participants enrolled in an advertising course received extra credit for their voluntary participation. Male and female participants were approximately equal in number, and all participants are of legal voting age.
The 100 participants were randomly assigned to one of four treatment conditions:  group 1 (positive ad), group 2 (issue attack), group 3 (image attack), and group 4 (mixed attack). The treatment groups were exposed to different stimuli pertaining to a fictional candidate for a Senate seat in Wyoming. Individual group sizes were 25. 


Stimuli
Two types of stimuli were created for use in the experiment. Pretests of candidate biographic profiles were conducted on dependent variables. These stimuli contained general information about the attacking candidate's education, employment background, volunteer work, and political accomplishment. Advertising stimuli were based on real-world examples of political advertising. Candidates were fictional.
In the initial pretesting, the advertising stimuli were pretested and revised to produce the final stimuli used in the experiment. In terms of execution, stimuli contained the line "Joe Smith  For Senate Vote for Smith!, On October 3rd, Paid by  Joe Smith." The following descriptions were also provided: supports tax relief, supports welfare reform, supports public education, supports Medicare for issue attack, is a parishioner in the church, has a good health record and is able to maintain a political career, is keeping a happy home, and served as infantry squad leader with the U.S. Army's 9th Infantry Division for image attack. Both contents of issue and image attack were combined in mixed attack.
A positive political ad containing public affairs and general information about the attacking candidate was used in place of a stimulus in the first group.

Procedures
At the beginning of the experiment, participants were instructed to read the candidate profiles and complete a series of pretest scales. Next, they were exposed to the advertising stimuli and completed the posttest scales. Participants were debriefed after the instruments were collected. 

Dependent measures
Dependent variables were measured for the attacking candidate (a) evaluation of the candidate's character (b) likelihood of voting for the candidate and (c) evaluation of the candidate's advertising
Character evaluation. Subjects evaluated each candidate on 7-point scales using seven attributes found useful in previous research investigating candidate images (Garramone, 1988). The attributes are intelligent, sincere, believable, honest, persuasive, concerned, and qualified.
Likelihood of voting. Subjects indicated on 5-point scales the likelihood that they would vote for the attacking candidate if they were voting in an election involving the candidates. "1" means not at all likely and "5" means very likely.
Advertising evaluation.  Subjects evaluated each commercial on three 5-point scales according to how informative, believable, and persuasive they felt the commercial was. For each scale, "1" indicated little of the attribute and "5" indicated much of the attribute. 

Results


The average reliability of the scales was assessed using Cronbach's alpha: Candidate evaluation for the pretest was .84, candidate evaluation for the posttest was .93 and advertising evaluation was .85.
A one-way ANOVA was used to examine participants' perception of stimulus. The results show significant differences in participants' perceptions of the issue/image/mixed stimuli [F (3,94)=12.75 p<.001.] An examination of means by Tukey's multiple comparison test revealed that there were significant differences between group 1 (positive ad) and group 3 (image attack) and also significant differences between group 2 (issue attack) and both group 3 (image attack), and group 4 (mixed attack).
To test H1 and H2, the attacking candidate evaluation scores of each treatment group were summed, and the pretest mean was subtracted from the posttest mean to create change-scores. The differences between treatment groups' change-scores were tested by specifying polynomial contrasts after the application of an omnibus F test. The results of analysis of variance indicate significant differences among change-score means: [F (2,71) = 3.47, P<.05.]  Change-score contrasts show that the stimulus containing image attack produced a greater negative change in attacking candidate evaluations than the issue attack stimulus, supporting H1 as shown in Table 2. H2, which states that the image attacks on the attacked candidate will produce a greater negative change in likelihood of voting for attacking candidate than an issue attack on the attacked candidate, is not supported. This is shown in Table 3. 
H3-a suggested that negative political advertising would produce source derogation, causing harmful effects of the character evaluation for the attacking candidate. As can be seen in Table 6, there is a statistically significant difference between the mean character evaluation for the pretest and the posttest [t (1,73)=-7.51, p<.001.]               
H3-b suggested that negative political advertising would produce source derogation, causing harmful effects of the likelihood of voting for the attacking candidate. As shown in Table 6, there is a statistically significant difference between the mean likelihood of voting for the pretest and the posttest [t (1,73)=-2.14, p<.05.]               
Overall, negative political advertising did produce source derogation, causing harmful effects on the character evaluation and the likelihood of voting for the attacking candidate.
The attacking candidate evaluation was tested to answer research question 1 about the impact of three attacks. The results of analysis of variance indicated no significant difference in attacking candidate evaluations [F (2,71)=2.19,  p=.12.]  
The likelihood of voting for the attacking candidate was tested to answer research question 2 about the impact of three attacks. The results of analysis of variance indicated no significant difference in likelihood of voting for attacking candidate [F (2,71)=1.76, p= .18.] As shown in Table 3, differences are small between the pretest in likelihood of voting for attacking candidate. Mixed attack has a higher difference between the pretest and the posttest.
Advertising evaluations were tested to answer research question 3 about the impact of three attacks. The results of analysis of variance showed significant difference in evaluation of advertising [F (2,72)=3.42 p<.05.] An examination of means by Tukey's multiple comparison test revealed that there were significant differences between Group 2 (issue attack) and Group 4 (mixed attack). It is assumed that issue attack is more favorable than mixed attack because of the higher mean score of issue attack than that of mixed attack. As shown in Table 4 and 5, the issue attack had higher mean scores than any other attacks as expected in this study. The mean score of image attack is almost the same as that of mixed attack.
Positive ads and three negative ads were contrasted to answer research question 4-a. There is no statistically significant difference between positive ads and three negative ads for the character evaluation [t=1.71 p=.24.] Importantly, the mean score of positive ad is almost the same as that of issue attack according to Figure 1. It leads to no significant difference between positive ads and three negative ads. To answer research question 4-b, there is no statistically significant difference between the positive ads and three negative ads for the likelihood of voting for attacking candidate [t=.233  p=.816.]

Discussion

The present study sought to determine the relative effectiveness of issue and image negative political advertising and positive and negative ads. The study results show that image attack produced a greater negative change than issue attack for evaluation of attacking candidate. In addition, the issue attack tended to have a higher attacking candidate evaluation even though it was not statistically significant. 
By comparing the means among three attacks, issue attack (M=30.68) is higher than mixed attack (M=28.42) and image attack (M=26.08). These findings are consistent with research results showing an issue attack will be more effective than an image attack (Roddy & Garramone, 1988) and Thorson et al's study (1991) that issue commercials produced greater voting intent and more positive attitudes toward candidates' characters. 
The mean score of character evaluation of attacking candidate decreases after seeing the positive ad (See Table 1). Possibly, this is because these ads don't show subjects the differentiation between candidates because they contain only positive arguments for the candidate.
In terms of voter backlash, sponsorship of negative political advertising has a little impact on the attacking candidate's evaluations and likelihood of voting for the attacking candidate. Attacking candidate evaluations did decline as a result of the use of negative political advertising. In addition, the decline was significant between likelihood of voting for attacking candidate in the pretest and likelihood of voting for attacking candidate in the posttest [t=-2.43, p< .05.] It is assumed that subjects didn't like negative political advertising. 
However, the effects of negative political advertising are complex. Even though the mean scores of candidate evaluation and likelihood of vote declined after the advertising stimuli, the difference is small when comparing mean scores of positive ads (M=30.58) and that of issue attack ads (M=30.68). It is assumed that negative political advertising may have an influence on the attacking candidate, but negative political advertising with issue attack can be treated as almost the same as the positive advertising.
Finally, negative political advertising did produce source derogation, causing harmful effects on the character evaluation and the likelihood of voting for attacking candidate. These findings are consistent with findings that the source derogation is caused by the attacking advertising (Merrit, 1984; Park, 1996). The candidate wanting to attack another candidate may use careful negative political advertising. When the candidate has certain evidence about arguments, which he or she makes in the advertising, he or she can reduce harmful effect on him or her. 
The main limitation of the present study is the use of student participants, who are not fully representative of the general electorate (Pinkleton, 1997). However, research suggested that because their age and education compensate for each other, college students are acceptable as participants in political advertising research (Garramone, 1984 & Pinkleton, 1997). Nonetheless, the study should be considered a single contribution to a body of research in progress. Finally, the study findings have a potential for future voters who are students now.
Another limitation is that the experimental setting was artificial with respect to lack of context. In a true election, voters do not base their decision solely on what they've seen in political commercials. Previous research also has found that individual differences such as age, education, and candidate preference mediate negative advertising effects (Roddy & Garramone, 1988). In this study, however, the fictional stimulus candidates and student sample did not allow for the investigation of such individual differences.
Future research examining the effects of political advertising should attempt to use a more representative sample and natural media exposure environments. Broadcast media should be used to provide information on the role of negative political advertising. 






Reference
Burgoon, M., & Miller, G.R. (1985). An expectancy interpretation of language and persuasion. In Giles, H. and Clair, R. N. St. (Ed.), Recent advantages in language, communication, and social psychology (pp. 199-229). London: Lawrence Erlbaum Associates.

Devlin, L. P. (1989). Contrast in presidential campaign commercials of 1988. American Behavioral Scientist, 32, 389-414. 

Devlin, L. P. (1995). Political commercials in American presidential election. In Kaid, L.L and Holtz-Bacha, C. (Eds.), Political advertising in western democracies: Parties and candidates on television (pp. 198-199). Thousand Oaks, CA: Sage. 

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. Journalism of Broadcasting & Electronic Media, 29(2), 147-159. 

Garramone, G. M., Atkin, C. K., Pinkleton, B. E., & Cole, R. T. (1990). Effects of Negative Political Advertising on the Political Process. Journal of Broadcasting & Electronic Media, 34(3), 299-311.

Germond, J.W., & Witcover, J (1996). As long as Negative Political Advertising Work, It Will be Used. Sun's Washington Bureau.  

Hagstrom, J., & Guskind, R. (1986). Selling the candidates. National Journal, 18, 2619-2626.

Johnson-Cartee, K.S. , & Copeland, G. (1987). Setting the parameters of good taste: Negative political advertising and the 1986 elections. Paper presented at the meeting of the International Communication Association, Montreal, Canada.

Johnson-Cartee, K.S., & Copeland, G.A. (1989). Southern voters' reaction to negative political ads in 1986. Journalism Quarterly, 66(4), 888-893, 986.

Johnson-Cartee, K. S., & Copeland, G. A. (1991). Negative political advertising: coming of age. New Jersey: Lawrence Erlbaum Associates.

Kaid, L.L. (1994). Political Advertising  in the 1992 Campaign. In Denton, R.E. (Ed.), The 1992 presidential campaign :  a communication perspective (pp. 111-127). Westport, CT: Praeger Publishers. 

Kaid L.L. (1997). Effects of the Television Spots on Images of Dole and Clinton. American Behavioral scientist, 40(8), 1085-1094.

Kaid, L. L., & Sanders, K. R. (1978). Political television commercials: An experimental study of type and length. Communication Research, 5, 57-70. 

Kaid, L L., & Boydston, J. (1987). An experimental study of the effectiveness of negative political advertisements. Communication Quarterly, 35(2), 193-201. 

Kaid, L. L., & Ballotti, J. (1991). Television Advertising in Presidential Primaries and Caucuses. Paper presented at the Speech Communication Association Convention.

Lutz, R.J. & Swasy, J.L. (1977). Integrating cognitive structure and cognitive response approaches to monitoring communication effects. Advances in Consumer Research, 6, 367.

McNair, B. (1994). An introduction to political communication. London: Routh Routledge.

Merrit, S. (1984). Negative political advertising: some empirical findings. Journal of Advertising, 13(3), 27-38.

Nugent, J. H. (1987). Positively negative. Campaign & Elections, 7, 47-49. 

Park, J. J. (1996). Effectiveness of negative political advertising. M.A. Thesis. University of Missouri-Columbia. 

Perloff, R. M. (1997). Political communication: politics, press and public in America. New Jersey: Lawrence Erlbaum Associates, 345-377. 

Pinkleton, B. (1997). The Effects of Negative Comparative Political Advertising on Candidate Evaluations and Advertising Evaluation: An exploration. Journal of Advertising, 26, 19-30.

Roddy, B.L., & Garramone, G.M. (1988). Appeals and Strategies of Negative Political Advertising. Journal of Broadcasting & Electronic Media, 32, 415-427.

Shapiro, M. A., & Rieger, R. H. (1989). Comparing positive and negative political advertising. Paper presented at the meeting of the International Communication Association, San Francisco. 

Surlin, S.H. , & Gordon, T.F. (1977). How values affect attitudes toward direct reference political advertising. Journalism Quarterly, 54, 89-98.

Thorson, E., Christ, W. G., & Caywood, C. (1991). Effects of issue image strategies, attack and support appeals, music and visual content in political commercials. Journal of Broadcasting and Electronic Media, 35, 465-486. 

Tinkham, S. F., & Weaver-Larisy, R. A. (1993). A diagnostic approach to assessing the impact of negative political television commercials. Journal of Broadcasting and Electronic Media, 37, 377-399. 

West, D. M. (1993). Air Wars : Television Advertising in Election Campaigns, 1952-1992. Washington, D.C.: Congressional Quarterly Inc, 46-52.

Wilson, R. D , & Muddrrisoglu, Aydin (1979). An analysis of cognitive responses to comparative advertising. Advances in Consumer Research, 7, 566-571.

Wright, P.L. (1973). The cognitive processing mediating acceptance of advertising. Journal of Marketing Research, 10, 53-62. 


































Appendix

Table 1. Change scores for attacking candidate evaluations
Condition
Change Scores
Issue attack
-4.6
Image attack
-10.04
Mixed attack
-6.06
Positive
-5.4
More negative values indicate greater negative change in attacking candidate evaluations from pretest to posttest. 

Table 2. Contrast for attacking candidate evaluation change scores

Difference
T Value
D.F.
T Prob.
Image attack with Issue attack
-5.4400
-2.528
71.0
.014*
Issue attack with mixed attack
1.3167
.606
71.0
.547
Image attack with mixed attack
-4.1233
-1.897
71.0
.062
*p< .05
Table 3. Change Scores for likelihood of voting for an attacking candidate
Condition
Change Scores
Issue attack
0
Image attack
-.28
Mixed attack
-.66
Positive
-.24

Table 4.  Advertising evaluation
Condition
Mean Scores
Issue attack
9.52
Image attack
7.96
Mixed attack
7.72
Positive
9.12

Table 5. Contrast for advertising evaluation
Value
S. Error
T Value
D.F.
T Prob.
Image attack with Issue attack
-1.56
.75
-2.09*
72
.040
Issue attack with mixed attack
1.8
.75
2.41
72
.19
Image attack with mixed attack
.24
.75
.321
72
.749
*p< .05
Table 6. T-test about the pretest and the posttest
T value
df
P
Pretest vs. Posttest
(Evaluation of character)
-7.51**
73
.000
Pretest vs. Posttest
(Likelihood of vote)
-2.14*
73
.036
*p< .05,  **p<.01

Figure 1. Mean of difference between pretest and posttest for character evaluation of candidate