An Empirical Evaluation of Comparative Advertising Messages: Subjects' Responses on Perceptual Dimensions

R. Dale Wilson, (student), The University of Iowa
ABSTRACT - This paper attempts to evaluate comparative advertising messages which provide little or no factual information to consumers. Subjects' evaluations of simulated comparative advertisements, when compared to other subjects' responses to simulated single-product ads, generally supported the expectation that the comparative messages would be evaluated less favorably than the single-product messages. The results of this study provide preliminary evidence that advertisers should avoid subjective, non-factual comparisons with named competitors.
[ to cite ]:
R. Dale Wilson (1976) ,"An Empirical Evaluation of Comparative Advertising Messages: Subjects' Responses on Perceptual Dimensions", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 53-57.

Advances in Consumer Research Volume 3, 1976      Pages 53-57

AN EMPIRICAL EVALUATION OF COMPARATIVE ADVERTISING MESSAGES: SUBJECTS' RESPONSES ON PERCEPTUAL DIMENSIONS

R. Dale Wilson (student), The University of Iowa

[This paper has benefited from the guidance and comments of Professor E. John Kottman (University of Iowa). The author also wishes to thank Gilbert A. Frisbie, Jr. (Bowling Green State University) for his assistance with data collection.]

[R. Dale Wilson is a Ph.D. candidate in marketing at The University of Iowa.]

ABSTRACT -

This paper attempts to evaluate comparative advertising messages which provide little or no factual information to consumers. Subjects' evaluations of simulated comparative advertisements, when compared to other subjects' responses to simulated single-product ads, generally supported the expectation that the comparative messages would be evaluated less favorably than the single-product messages. The results of this study provide preliminary evidence that advertisers should avoid subjective, non-factual comparisons with named competitors.

INTRODUCTION

Of the recent innovations in marketing, perhaps one of the most noticeable is comparative advertising. Contrary to the prohibitive rules of thumb and ethical standards that have existed in the past, some advertisers are now willing to use direct confrontation in their advertising campaigns. By calling their competitor(s) by name and by using side-by-side comparisons, these advertisers believe that they are better able to "prove" the problem-solving utility of their products to existing and potential customers. The popularity of comparative advertising can readily be noted from casual inspection of the print and broadcast media; readers and viewers have recently been exposed to comparative ads for personal care products (e.g., Dial Very Dry antiperspirant vs. Arrid, Sure, and Right Guard), wearing apparel (e.g., No Nonsense pantyhose vs. L'eggs), major appliances (e.g., Magnavox Videomatic TV vs. Zenith and RCA), and automobile warranties (e.g., American Motors vs. Ford, Chrysler, and Chevrolet) among others.

BACKGROUND

The door to comparative advertising was formally opened in 1971 when the Federal Trade Commission (FTC) encouraged the major television networks to accept advertisements that included comparisons with (named) competitors' brands. Ulanoff (1975) has indicated that direct comparisons were encouraged by the FTC to provide an acceptable alternative to indirect comparisons using "Brand X" or electronic "beep" sound techniques. Presumably, the FTC's action was motivated by the intuition that comparative advertising would

". . .encourage the provision of more factual product information to the consumer and simultaneously to discourage deception by eliminating comparison by innuendo" (Oliver, 1975).

While the FTC actively supported head-to-head competition using comparative advertising, the criteria for determining fairness, deception and misrepresentation, and substantiation of claims was, for all practical purposes, left for the media themselves to decide. Subsequently, the broadcast media (Christopher, 1974a; 1974b) and the advertising trade associations (Broadcasting, 1974a; American Association of Advertising Agencies, 1974) established a series of self-regulations governing what could, and what could not, be said and done in a comparative ad. These guidelines, coming just before a substantial surge in comparative advertising, were constructed in response to a ruling by the National Advertising Review Board (NARB) that the Schick Flexamatic shaver's advertised claims of superiority over Norelco, Remington, and Sunbeam shavers were in some respects false and were misleading overall (Broadcasting, 1974b).

Despite the good intentions of the media and the self-regulatory boards, the information content of many advertisements using the comparative strategy remains tainted by trivial comparisons that neglect competitive advertising's role as an aid to consumer decision-making. Comparative ads, when they merely compare one brand with another without providing factual information, simply are not useful to the consumer (Chevins, 1974). Evidence that many comparative advertisements consist of claims that are of questionable worth is provided by the persistent warnings of advertising authorities that abuse of this type of advertising must be halted. If action is not taken within the industry itself, the image and credibility of advertising will surely be further damaged, and ultimately the privilege of 3comparative advertising might be withdrawn by the FTC. [See comments by Chevins (1974), McMahan (1974), Light (1975), Schneider (1975), and Tyler (1975).]

Further evidence of the subjectivity of some comparative ads is the large number of complaints and challenges from those companies who feel that they have been unfairly treated in their competitors' advertisements. Indeed, the volume of protests is so great that one major television network recently issued a set of procedural guidelines specifically designed to expedite the handling of these objections (Advertising Age, 1975). Many of these comparative ads are so obviously subjective (if not false and/or misleading) that their claims defy substantiation. [See Cohen (1974, p. 11) for an enlightening discussion of the legal criteria with regard to unsubstantiated claims and the determination of fairness in advertising.]

The managerial and public policy questions stemming from the type of comparative advertising described above are many and varied. Putting aside the obviously important issues regarding the regulation of marketing practices, let us turn to an evaluation of the managerial ramifications of using comparative ads that provide consumers with little or no factual information on which they can base their purchasing decisions. Is this type of advertising not likely to contribute to the negative image of advertising? Will this type of advertising have a negative impact upon consumers' evaluations of products that are advertised in this way? Will advertisements that use subjective and non-factual comparisons lead to a decreased level of credibility and trustworthiness for those companies who advertise their products in this way? Based upon the beliefs of several members of the advertising community [See comments by Chevins (1974), McMahan (1974), Light (1975), Schneider (1975), and Tyler (1975).] it was expected that these questions could be answered in the affirmative. Consequently, a study was designed to explore the differences in consumers' perceptual and attitudinal responses towards comparative and single-product advertising messages.

METHODOLOGY

In order to assess the impact of subjective claims in comparative advertising, 80 male and female undergraduates enrolled in introductory marketing classes at the University of Iowa were asked to evaluate eight different advertising messages on seven different response dimensions. Each subject was presented with a booklet containing either eight randomly ordered single-product messages or eight randomly ordered comparative messages.

Each simulated advertisement, which consisted of typed headlines and copy, contained a message for a well-known brand in each of the following product classes: automobiles, bath soap, beer, cat food, credit cards, deodorant, mouthwash, and toothpaste. [Product classes and brands were chosen on the basis of a product usage and brand familiarity survey administered to a similar group of undergraduates three months prior to the main study.] In the comparative ads a competitor's brand name was mentioned while the single-product ad was modified by deleting the competitor's brand name or by replacing the competing brand name with a phrase such as "any other brand" or "other brands." For example, the comparative advertisement for toothpaste was as follows:

ONLY A DENTIST CAN GIVE CHILDREN A BETTER FLUORIDE TREATMENT THAN COLGATE.

A kid in the cavity-prone years can wind up with a mouthful of cavities. Eleven is average between five and fifteen. But a Colgate fluoride treatment every day can help prevent some of those cavities. The cavity fighters in Colgate make it more effective than any other toothpaste--even Crest.

Colgate with MFP fluoride helps strengthen teeth against decay. Remember, only your dentist can give teeth a better fluoride treatment.

In the single-product ad, the phrase "even Crest" was merely deleted from the copy. The length of the comparative ads ranged from 49 to 117 words and averaged 74.1 words while the single-product ads averaged 1.4 words longer.

The subjects were told that the advertisements had appeared or were scheduled to appear in the print and/or broadcast media. In reality, only four of the messages (for a brand of automobile, cat food, credit card, and toothpaste) were modified from actual printed advertisements; the remaining ads were constructed specially for the purpose of the experiment. Subjects who received single-product messages were not aware that other subjects were evaluating comparative messages, and vice versa.

Following each simulated advertisement was a page of 7-point scales. Subjects were asked to rate each advertisement on four dimensions (i.e., extremely informative--not at all informative; extremely believable--not at all believable; extremely offensive--not at all offensive; and extremely interesting--not at all interesting), each advertised product on two dimensions (i.e., definitely would change my view towards the product--definitely would not change my view towards the product and extremely high quality--extremely low quality), and each sponsoring company on one dimension (i.e., extremely trustworthy--not at all trustworthy). Thus, seven response measures were collected on each of eight advertisements for every participating subject.

The methodology discussed above, in which each subject evaluates either comparative or single-product advertisements for eight different products forms an 8 x 2 two-factor design with repeated measures on one factor. [This type of "mixed" design is known in the experimental design literature under a variety of names. It is discussed in detail by Lindquist (1953, pp. 267-73) and Wirier (1971, pp. 518-39).] Briefly, this design is a two-factor (A X B) design in which all subjects are exposed to each of a group of A treatments in combination with any one B treatment, but each group of Ss is exposed to a different B treatment. In our case, one group of Ss was exposed to a series of eight single-product ads, and the other group was subjected to a series of eight comparative ads. This mixture of the A X B factorial design and the treatments X subjects (A X S) design contains treatment comparisons which are mixed--some are inter-subject (between) comparisons while others are intra-subject (within) comparisons. While the two-factor design with repeated measures on one factor is not widely used in consumer research, it offers an important advantage over alternative designs in that a limited number of subjects can be more effectively utilized. Even though carry-over effects are a potential problem for most behavioral applications, this problem can usually be minimized by randomizing the A treatments.

RESULTS AND DISCUSSION

Each of the seven response variables was considered independently for the purposes of analysis. However, it should be pointed out that inspection of the response variables' correlation matrices shows a relatively high degree of correlation among the seven response variables. But because the correlation coefficients vary widely depending upon the product and type of advertisement being considered, this intercorrelation problem is difficult to handle. Factor analysis of the seven response variables does not provide conclusive evidence as to which, if any, of the response variables should be eliminated from further study. The factor loadings, as well as the number of emerging factors, vary depending upon the product and type of advertisement being analyzed. Consequently, caution must be exercised when interpreting the results of this study.

The results of the analysis of variance stemming from the two-factor design with repeated measures on one factor are presented in Table 1. Each analysis of variance was computed for 75 subjects (35 Ss in the single-product ad treatment and 40 Ss in the comparative advertisement treatment). Five questionnaires were dropped from the analysis because they were improperly completed. The measurement scales were numbered so that 7 indicates a positive response (e.g., extremely informative) and 1 indicates a negative response (e.g., not at all informative).

TABLE 1

SUMMARY OF ANALYSIS OF VARIANCE

As can be seen from the table, the sums of squares for the type of simulated advertisements (i.e., single-product vs. comparative ads) is a "between" subjects effect and the sums of squares for products and the products/ ads interaction are "within" subjects effects. The relevant F-ratios-are formed by-dividing the mean squares by the appropriate error term--the type of ad F-ratio is obtained by using the "between" error mean square while the products and products/ads interaction F-ratios are determined by using the "within" error mean square as the denominator.

The results of the ANOVA tables indicate a highly significant (p<.001 in all cases) main effect for the product treatment. Because of the diversity of the products under consideration, this was to be expected. Of primary interest, however, is the main effect of the type of ads used in the study. Table 1 indicates that for response variables believability and offensiveness, the main effect is significant (p<.10 and p<.005 respectively). The significant F-ratios are indicative that the treatment groups differentially responded to the believability and offensiveness dimensions based upon whether they read single-product messages or comparative messages. For the remaining five response means, the main effects of the type of ad were not statistically significant (p>.10).

For the believability and offensiveness response dimensions, the differences between the type of ad treatment means were evaluated via Scheffe's (1953) test for multiple comparisons. On the believability dimension, significant differences were found for bath soap (p<.05), deodorant (p<.10), and mouthwash (p<.10). For offensiveness, significant differences between the type of advertisement treatment means for the brands of automobiles (p<.05), bath soap (p<.01), beer (p<.10), cat food (p<.10), credit cards (p<.O1), and deodorant (p<.01) were found to exist.

The nature of the differences between response variable means for each product can be more clearly examined in Table 2, which presents differences between means for the single-product and comparative advertisement groups. Of the 56 response differences, 39 (or 69.6%) are in the expected direction. Of these 39 in which the single-product ad is rated more positively than the comparative advertisement, 16 (or 41.0%) are significant at p<.10 or less.

The results of the analysis of variance and the t-tests between means lends support to the notion that the respondents generally evaluated subjective comparative advertising messages more negatively than they did those ads which did not mention a competitor's product by name. This finding must be qualified, however, by stating that the results of this study seem to be product (or advertisement) specific. In addition, the differences in response means across all products on the ability of the ads to change the respondents' views towards the product, the quality of the product, and the trustworthiness of the sponsor appear to be more random than systematic. From Table 2, however, it is clear that for all advertising messages except mouthwash and toothpaste, respondents differentially evaluated the types of ads on the informative value, believability, offensiveness, and/or interest produced dimensions.

CONCLUSION

The results of the study presented in this paper seem to indicate that advertisers may be using comparative ads in a less-than-desirable way. Current advertising campaigns serve as evidence that some advertisers believe that brand comparison, in and of itself, will lead to a more favorable image in the consumer's eye. However, the empirical study discussed earlier suggests that in cases where brand comparisons provide little or no factual information, comparative advertising offers no advantage. In fact, on the dimensions of believability and offensiveness (and to a lesser extent for informative value and interest produced), the two-product comparative ads are evaluated more negatively than single-product advertisements. When one brand does not offer a clear-cut, substantial advantage over another brand, it seems that consumers should be encouraged to compare products for themselves instead of having nonfactual comparisons prepared for them.

TABLE 2

DIFFERENCES BETWEEN MEAN RESPONSES FOR SINGLE-PRODUCT AND COMPARATIVE ADVERTISING MESSAGES

The study discussed in this paper adds credence to the suggestion that comparative ads may further damage the institution of advertising. While recent studies have documented the negative image of advertising, [See, for example, the results of a study by Warwick and Legler, Inc., summarized in Business Week (1972).] comparative advertising may contribute even more to that image. This would seem to be especially true if this type of advertising continues to use comparisons that do not assist consumers in their decision-making processes.

At the present time, an overall evaluation of comparative advertising suffers from insufficient theoretical development [Theoretically, comparative advertising may be considered to be a special case of two-sided advertising (see Karp, 1971). If so, McGuire's (1964) inoculation theory may prove useful as a basis for a theoretical evaluation of comparative advertising. In addition, Wilkie and Farris (1974) have provided a theoretical framework for comparative advertising based on Lavidge and Steiner's (1961) "hierarchy of effects" model.] and empirical research. Although this paper summarizes one attempt at an empirical evaluation of the content of comparative ads, its findings must be interpreted with caution. The study was conducted on a limited budget, in one geographical area, and used a limited sample size and a small number of advertisements. Additional research must be conducted in order to determine generalized patterns of consumer response to this unique form of advertising.

REFERENCES

Advertising Age, "NBC Sets Guidelines for Comparative Ad Plaints," 46(January 6, 1975),8.

American Association of Advertising Agencies, Policy Statement and Guidelines for Comparative Advertising (New York: AAAA, April, 1974).

Broadcasting, "NBC Puts on Paper its Rules for Comparative Ads," 86(February 4, 1974a),41.

Broadcasting, "Code Review Board Adds More Teeth to TV Rules for Comparison Ads," 87(October 7, 1974b),61.

Business Week, "The Public is Wary of Ads, Too," (January 29, 1972),69.

Anthony Chevins, quoted in "Care in Comparison Urged...", Advertising Age, 45(August 5, 1974),31-2.

Maurine Christopher, "NBC Spells Out New Formal Guides for Comparative Spots," Advertising Age, 45(January 28, 1974a),1+.

Maurine Christopher, "ABC Comparative Rules Stress Test Procedures," Advertising Age, 45(March 18, 1974b),1+.

Dorothy Cohen, "The Concept of Unfairness as It Relates to Advertising Legislation," Journal of Marketing, 38(July, 1974),8-13.

Robert E. Karp, "Creative Advertising Strategies: The Two-Sided Approach," Business and Society, 12(February, 1971),18-25.

Robert J. Lavidge and Gary A. Steiner, "A Model for Predictive Measurements of Advertising Effectiveness," Journal of Marketing, 25(October, 1961),59-62.

Lawrence Light, quoted in "Underdog Advertiser Wins in 'Naming Names': BBDO," Advertising Age, 46(March 10, 1975),56.

E. F. Lindquist, Design and Analysis of Experiments in Psychology and Education (Boston: Houghton Mifflin, 1953).

William J. McGuire, "Inducing Resistance to Persuasion: Some Contemporary Approaches," in Leonard Berkowitz (ed.), Advances in Experimental Social Psychology (New York: Academic Press, 1964),191-229.

Harry Wayne McMahan, quoted in "Advertisers' Name-calling Could Hurt, McMahan Says," Advertising Age, 45(February 18, 1974),72.

Robert E. Oliver, quoted in "Open Way for Comparative Ads, Canada Meet Told," Advertising Age, 46(May 12, 1975),6.

Henry Scheffe, "A Method for Judging All Contrasts in the Analysis of Variance," Biometrika, 40(June, 1953), 87-104.

Alfred Schneider, quoted in "ABC Censor Raps Trend to Naming Rivals in Ads," Advertising Age, 46(March 31, 1975),98.

William D. Tyler, "Comparative Advertising: A Powerful Selling Tool When It is Not Abused," Advertising Age, 46(April 21, 1975),58+.

Stanley M. Ulanoff, "Comparison Advertising: An Historical Retrospective," working paper (Cambridge, Mass.: Marketing Science Institute, February, 1975).

William L. Wilkie and Paul Farris, "Comparison Advertising: Issues and Prospects," working paper (Cambridge, Mass.: Marketing Science Institute, August, 1974).

B. J. Winer, Statistical Principles in Experimental Design, Second Ed. (New York: McGraw-Hill, 1971).

----------------------------------------