Comparative Advertising: Some Positive Results

Gerald J. Gorn, University of British Columbia
Charles B. Weinberg, University of British Columbia
ABSTRACT - Despite increasing use of comparative advertising, tests of its effectiveness have produced few positive results. The present study investigated comparative advertising in a 2x2 experimental design using print ads in a number of product categories. The effects of comparative and non-comparative advertising by a challenger were compared in the presence or absence of advertising by the brand leader. Attitudinal effects were observed as a function of Comparative advertising.
[ to cite ]:
Gerald J. Gorn and Charles B. Weinberg (1983) ,"Comparative Advertising: Some Positive Results", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 377-380.

Advances in Consumer Research Volume 10, 1983      Pages 377-380


Gerald J. Gorn, University of British Columbia

Charles B. Weinberg, University of British Columbia

[The authors thank Cathy Bolrgeault for her help in data collection and analysis and Jane Howell, Mary Burns, and Lee Smith for helpful assistance in the project.]


Despite increasing use of comparative advertising, tests of its effectiveness have produced few positive results. The present study investigated comparative advertising in a 2x2 experimental design using print ads in a number of product categories. The effects of comparative and non-comparative advertising by a challenger were compared in the presence or absence of advertising by the brand leader. Attitudinal effects were observed as a function of Comparative advertising.

In comparative advertising, explicit reference to a competitor's brand name is made in the ad. Since the early 1970's, comparative advertising has been encouraged by the FTC, in the belief that mentioning a competitor by name would help the consumer evaluate a claim of superiority (Wilkie and Faris 1975). Comparative advertising would encourage more competition in markets dominated by a few companies. Market penetration by smaller companies would be facilitated if they could make explicit reference to a larger company's brand in trying to substantiate a claim of superiority. More and, perhaps, better information would be available to consumers and therefore more rational consumer decision making should result.

Although comparative advertising is commonplace today, empirical research in the area casts doubt on its effectiveness. Research using different methodologies and widely different products (from cameras to supermarkets) has generally led to the same conclusion. Comparative advertising is no more effective than traditional advertising when it comes to: believability or credibility of both specific claims and the ad in general (Wilson 1976, Levine 1976, Prasad 1976, Golden 1976, Goodwin and Etgar 1980), attitude towards the brand (Goodwin and Etgar 1980), purchase intentions (Golden 1976, Swinyard 1981), and actual behavior (Swinyard 1981).

Results for some of the variables in these studies revealed that comparative advertising was in fact less effective than traditional advertising(e.g., Swinyard 1981, Goodwin and Etgar 1980, Shimp and Dyer 1978). Swinyard (1981) found that the credibility of a key claim could be enhanced by structuring an ad that was two-sided (contains some negative features of the sponsoring brand in the ad) in addition to making an explicit comparison with a competitor on other dimensions. It is, however, not clear whether the two-sided comparative claim significantly out performed the two-sided noncomparative claim. Etgar and Goodwin (1982) also found that a two-sided comparative ad achieved significantly better results than a one-sided comparative ad, but they did not test noncomparative ads. The general conclusion to be drawn from the literature is that comparative advertising is sometimes equally and sometimes less effective than traditional advertising.

Swinyard utilizes Wright's (1973) cognitive response framework to explain the lack of effectiveness of comparative advertising. In this framework, counterarguing, source derogation, curiosity statements and other positive or negative message related statements are cognitive mediators of message acceptance. Minimizing counterarguing is particularly important in achieving message acceptance. When a brand claims that it is not only good on an attribute, but better or equal on that attribute than another brand, it is considered a fairly strong claim. Direct contrasts of the sponsoring and competing brand require more involvement and therefore more active information processing by the receiver (Wilson and Muderrisoglu 1979). The receiver counterargues if he/she finds the comparative claim is not believable-for one reason-or another. This counterarguing results in decreased message acceptance. Attribution theory should also suggest greater counterarguing with comparative claims (Swinyard 1981). With attribution theory the motivation behind advertising messages would be an important factor influencing message credibility. Counterarguing or a discounting of claims should be expected if consumers suspect that the motivation to present more facts is not for the consumer's benefit. Similarly, one sided formats would be expected to create greater counterarguing than two sided advertisements because the latter may appear to be more honest.

Wilson and Muderrisoglu (1979) found that the incidence of counterarguing was significantly greater for comparative advertising than for noncomparative advertising. Swinyard (1981) not only found a greater incidence of counterarguing with comparative ads, he also found that comparative ads were perceived to be significantly less credible than noncomparative ads.

Counterarguing can explain the results of studies where comparative advertising is less effective than noncomparative advertising. Other factors may explain the majority of the research which has shown comparative and noncomparative advertising to be equally effective. Previous research has often utilized fictitious brands making explicit comparisons to a brand leader (Prasad 1976). -Attitudes towards these brands relative to attitudes toward the brand leader are unlikely to change with a one shot exposure to an ad, whether the ad is comparative or non-comparative. Differences between comparative and non-comparative ads are also likely to be attenuated if the ads are home-made (Goodwin and Etgar 1980) and if the comparative part of the at is not clearly the major aspect of the comparative ad.

Finally, it is important to assess the impact of a comparative ad in a format where post exposure claim believability questions and questions to elicit counterarguing are not asked before questions assessing attitude towards the ad or brand. This hasn't always been the case (Prasad 1976). Asking a respondent whether he/she believes a claim (e.g., -Do you believe that tests show that brand X is very good?") is likely to encourage counterarguing, particularly if the claim is comparative (e.g., "Do you believe that it is a scientific fact that brand X is superior to brand Y?"). Therefore, if claim believability and questions to assess counterarguing are asked, they should be asked following and not prior to questions measuring perceptions of and attitudes toward the ad and brand. Although such an ordering could bias measures of cognitive response (indicating an unfavorable attitude makes me more likely to counterargue), we believe the first priority should be on assessing the impact of comparative advertising on perceptions and attitudes.

The present research focused on responses other than claim believability or counterarguing. It was considered critical to select dependent variables which reflected the aim of comparative advertising. Comparative advertising is usually initiated in a product category by a non-brand leader or challenger. > e goal seems usually to place the challenger in the same league as the brand leader (Prasad 1976). For example, the vice-president of marketing for Wine Spectrum, a division of Coca-Cola Co., stated "Comparative ads are good when you're new, but when you're the standard, it just gives a lot of free publicity to you competitors." (Business Week 1982). Whether or not the challenger claims to be equal or better than the brand leader, a favorable outcome for the comparative ad would be less psychological distance between the challenger and the leader. > is might translate into the challenger being considered together with the leader in the evoked set of brands that came to mind when a consumer is considering a purchase.

Consumers cannot process or store all of the vast amount of information in their environment (Bettman 1979). When a communication is presented, certain cues or central aspects of a message will be attended to (Bettman 1979). With a comparative ad the central feature of the ad would be the two brands and the common dimension(s) on which they are compared. if this is the case, then one outcome of a comparative at would be increased similarity rather than differentiation between the two brands which would be a positive outcome for a challenger when it is taking on a brand leader in its advertising.

The present study was designed to test this notion and its implications for how consumers respond to both the two brands in question and to other brands in this product category and in other product categories.



Since much of the previous research in the area used full-time day students as subjects, in this study an attempt was mate to obtain a broader sample of subjects. One hundred seventy-two people, mostly evening students at a local university, served as subjects; some were full-time students. [Preliminary tests showed that there were no significant differences in response between full and part-time students so all respondents were pooled.] They represented a wide range of occupations (secretaries, housewives, middle level managers and so on).


Three products (toothpaste, low tar cigarettes, and golf balls) were chosen for which a variety of comparative and non-comparative magazine ads were available for selection. [A fourth product category pain relievers was also utilized, but was not analyzed as it was discovered that Bayer aspirin, the leading brand of aspirin, is not the leading brand of pain reliever, which is Tylenol.] In each product category, we selected a brand which did not have the largest market share (hence termed the "challenger") and which in print advertisements explicitly named and compared itself to the brand with the largest market share (the "leader"). These were the two brands studied in each market. The challenger was not always the number two brand in the market, but it was a non-leading brand which used comparative advertising. Although product familiarity is not a directly manipulated variable, it is interesting to note that all respondents were toothpaste users and most were aware of the brand leader. There was less usage and familiarity with the other two product categories.


Photocopies of authentic magazine ads were used in the survey. Comparative and non-comparative advertisements were chosen for each product. In each product class, the comparative ad for the challenging brand specifically mentioned the leading brand. In all cases the claim made in each at was the same for both the comparative and non-comparative copy. Where necessary, the ads were professionally retouched to make sure that the claims in the comparative and non-comparative ad dealt with the sane product attribute. (In a post test, there were no significant differences between treatments on whether the ads were viewed as professional or amateurish. )

The headlines for the toothpaste ads illustrate the treatments:

Challenger nonconparative      "Great news for mothers of cavity-prone children!"

Challenger comparative      "How 2000 children proved that (challenger) could beat (leader) at fighting cavities.

Leader      "Thanks to (leader), millions of kids have gotten less than they expected."

Experimental Design

The following experimental conditions were structured:

1. Non-comparative ad for a brand (referred to as the challenger), which is not the brand leader.

2. Comparative ad for the challenger.

3. Non-comparative ad for challenger plus noncomparative at for leader.

4. Comparative at for challenger plus non-comparative at for leader.

Treatments 3 and 4 were included in the design to simulate the real life situation where the leader in addition to the challenger is advertising (previous research has only incorporated treatments where the challenger is advertising). me design was therefore 2x2, type of at (comparative or non-comparative) X context (presence or absence of advertising by the leader).

Each subject was randomly assigned to a particular condition for all three products. Thus, for example, subjects in Condition 1 were exposed to a challenger's non-comparative at for all three products. The order of exposure to the various ads was randomized within each treatment.

Questionnaire Content and Procedure

The questionnaire was administered during class time. Respondents within each class were randomly assigned to the various treatments.

On a cover sheet they were told that we were doing a study of advertising and showing people many of the ads that have appeared in magazines in recent years. They were then told that they were going to see some of them and we were interested in their reactions.

Following exposure to the ad(s) from the first product category they were asked a series of questions. Wilson and Muderrisoglu (1979), had pointed out that the dependent variables selected for study in earlier research might not have been the most important ones. A critical effect of comparative advertising not previously assessed was the degree to which the challenger and brand leader were seen as similar, that is, whether or not the challenger was "in the same league" as the brand leader. Respondents in this study were therefore asked for their first reaction to how different they thought the challenger and leader were even if they hadn't used those brands. They circled their answer on a nine point scale ranging from identical (1) to completely different (9). Next, they provided their attitude toward the challenger's brand relative to the brand leader on a scale ranging from challenger brand much better (1) to leader much better (9). The third question started with a statement that later in the year there would be two new brands of the product introduced into the market. Respondents were asked for their first reaction to how different they thought the two brands were likely to be from each other on a scale ranging from identical (1) to completely different (9). Respondents also evaluated the ad(s) they saw on a series of bipolar scales: vague (1) precise (5); informative (1) uninformative (5); professional (1) amateurish (5); not persuasive (1) persuasive '5); interesting (1) boring (5); have seen the ad before (1) have not seen the ad before (5), as well as answering a series of questions about the brands and the advertisements.

Subjects then saw ads and answered the sane questions for brands from the second and third product categories. Following exposure to the ads in all three product categories, subjects were then asked to indicate the thoughts they had while looking at the at(s) in the last product category seen by then. The cognitive response data are still to be coded and the data other than the attitude question described in the previous paragraph have not yet been fully analyzed.

Respondents were also asked questions regarding their usage of toothpaste, cigarettes and golf balls. First they indicated whether they smoked cigarettes on a four point scale with categories ranging from "No, I have never really smoked" to "Yes, I smoke a pack or more a week." If they were smokers, they were asked to indicate the brand they smoked most of the time. To obtain perceived brand leader, subjects were then given a list of four low tar cigarettes (including the brand leader and challenger), and asked to circle the brand they thought sold the most low tar cigarettes. For golf balls they market down how often they played (from never (O) to at least once a week in the golf season (4)). They then indicated the brand they used if they played golf and the brand of golf ball they thought was the brand leader on the same scale used for low tar cigarettes. Brand of toothpaste used and perceived brand leader of toothpastes were obtained in a similar fashion. Information about other product categories was also gathered.

In the final section, data on age, sex, occupation, education, and income were collected.


Typically, the goal of comparative advertising is to reduce the challenger's distance from the leader with the two brands and common attribute(s) being featured prominently in the comparative ad. It was therefore predicted that greater perceived challenger-leader brand similarity would be obtained with comparative advertising than with non-comparative advertising. Furthermore, when comparative advertising is used by some brands in a product category then the entire product category may be seen by respondents as containing less differentiated rather than more differentiated brands. This possibility let to the prediction in the present study that exposure to comparative versus noncomparative ads would lead to greater perceived brand similarity for the two hypothetical new brands to be introduced into the market.

Regarding brand preference, it was predicted that the challenger would be more favorably evaluated in the comparative than in the non-comparative condition. The central element of the challenger's comparative ad was a favorable comparison of itself with the leader on a particular attribute. Furthermore, the preference measure was presumably made more sensitive than preference measures of previous studies by asking for a comparative, rather than a separate, evaluation of the challenger and leader.

Regarding the effect of context, it was expected that the presence of an at for the leader would lessen the impact of the challenger's at. Of particular interest would be a context type of ad interaction. Would the potential gains of a comparative ad be completed eliminated when subjects also saw an ad for the leader?


Brand Attitudes

The data were analyzed using ANCOVA.4 There were two treatment factors (type of ad x context) with product usage treated as a covariate for both the golf ball and cigarette responses. Since there was no variation in product usage for toothpaste (everyone brushes their teeth), the particular brand of toothpaste used was treated as a covariate.

The covariate was significant for both toothpaste (F(1,152)=8.32, p<.05) and golf balls (F(1,146)=12.91, p<.001). but not for cigarettes (F(1,157)=.54, p>.05).

The mean brand attitudes from the ANCOVA are contained in Table 1. The results for both toothpaste and cigarettes were in the predicted direction with more favorable attitudes toward the challenger brand with the comparative at (toothpaste F(1,152)=4.72, p<.05; cigarette F(1,157)=8.86, p<.01). There was no main effect for type of at with golf balls (F(1,146)=.57, p>. 05).



While there was no significant type X context interaction, the results for the main effect of context were significant for all three products tested (toothpaste F(1,152)=3.68, p<.05; cigarette F(1,157)=11.73, p<.001; golf balls F(1,146)=12.91, p<.001). As expected, the presence of the leader ad reduced the effectiveness of the challenger's ad, creating less favorable attitudes toward the challenger.


Although the data are only partially analyzed as yet, there seems to be enough evidence to suggest that comparative advertising can be effective for the challenger. The results cast doubt on the conclusions and methodologies employed in previous research which suggested that comparative advertising has little positive impact.

Comparative advertising was successful in creating more favorable attitudes toward the challenger than noncomparative advertising. The effect was significant for both toothpastes and cigarettes, but not for golf balls.

In an attempt to identify some of the factors influencing the relative effectiveness of comparative advertising, the present experiment is different in design from previous studies. While the results of a fair amount of academic research casts doubt on the effectiveness of this advertising strategy, advertisers have been increasing their use of comparative techniques (Jackson et al. 1979). me -business community seems to have greater faith in comparative advertising than the academic community. Their faith is supported by the results obtained in this study. Further analysis of the data in the present study will hopefully yield insights into why their faith has been rewarded.


Bettman, J.R. (1979), An Information Processing Theory of Consumer Choice, Reading, Mass.: Addison-Wesley Publishing Co.

"Creating a Mass Market for Wine," Business Week (March 15, 1982), 108-118.

Etgar, M. and S.A. Goodwin (1982), "One-Sided versus Two-Sided Comparative Message Appeals for New Brand Introductions," Journal of Consumer Research, 8 (March), 460-465.

Golden, L. (1976), "Consumer Reactions to Comparative Advertising," B.B. Anderson, ed., in Advances in Consumer Research, 3, Association for Consumer Research, 63-7.

Goodwin, S. and Etgar, M. (1980), "An Experimental Investigation of Comparative Advertising: Impact of Massage Appeal, Information Load and Utility of Product Class, Journal of Marketing Research (May 1980), 187-202.

Jackson, D., W. Jr., S. W. Brown, and R. R. Harmon, "Comparative Magazine Advertisements," Journal of Advertising Research, 19:6 (December 1979), 21-26.

Levine, P. (1976), "Commercials That Name Competing Brands," Journal of Advertising Research, 16 (December), 7-14.

Prasad, V.K. (1976), "Communications Effectiveness of Cognitive Advertising: A Laboratory Analysis," Journal of Marketing Research, 13 (May), 128-37.

Shimp, T.A. and D.C. Dyer (1978), "The Effects of Comparative Advertising Mediated by Market Position of Sponsoring Brand," Journal of Advertising, 7(Summer), 13-19.

Swinyard, W.R. (1981), "The Interaction Between Comparative Advertising and Copy Claim Variation," Journal of Marketing Research, 18 (May), 175-86.

Wilson R.D. (1976), "An Empirical Evaluation of Comparative Advertising Messages: Subjects' Responses on Perceptual Dimensions," B.B. Anderson, in Advances in Consumer Research, Cincinnati: Association for Consumer Research. 53-7.

Wilson, R.D. and A. Muderrisoglu (1979), "An Analysis of Cognitive Responses to Comparative Advertising," Working Series in Marketing Number 84 (August), Pennsylvania State University.

Wright, P.L. (1973), "The Cognitive Processes Mediating Acceptance of Advertising," Journal of Marketing Research, 10 (February), 53-62.