Consumer Reactions to Comparative Advertising

ABSTRACT - This study investigates the relative influence of comparative and non-comparative advertising upon purchase intentions. The advertiser's competitive position, claim substantiation and theme are considered, while brand loyalty is controlled. The results of the factorial experiment, employing covariance analysis, indicate that comparative advertisements appear to be no more effective than non-comparative advertisements, except when theme is considered.

Citation:

Linda L. Golden (1976) ,"Consumer Reactions to Comparative Advertising", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 63-67.

Advances in Consumer Research Volume 3, 1976      Pages 63-67

CONSUMER REACTIONS TO COMPARATIVE ADVERTISING

Linda L. Golden, University of Texas at Austin

ABSTRACT -

This study investigates the relative influence of comparative and non-comparative advertising upon purchase intentions. The advertiser's competitive position, claim substantiation and theme are considered, while brand loyalty is controlled. The results of the factorial experiment, employing covariance analysis, indicate that comparative advertisements appear to be no more effective than non-comparative advertisements, except when theme is considered.

INTRODUCTION

Comparative advertising is a recently popularized technique which has generated considerable attention and controversy. The purpose of this paper is to briefly describe the background and controversies surrounding comparative advertising and to discuss an empirical investigation of the impact of comparative advertising upon the consumer. For purposes of this investigation, comparative advertising may be defined as any form of paid promotion which:

1. compares the advertised brand to one or more specifically named or recognizably referred to brands within its generic product class;

2. with respect to one or more specific and/or over-all product characteristics.

The first criteria specifies that brands must either be specifically named or referred to in a manner which would make the brands recognizable to the consumer. This criteria excludes any advertising which does not make clear to the audience the brands against which the comparison is made. The second criteria in the definition allows inclusion of advertisements which either compare specific product characteristics or state that the advertised brand is "better" over-all than certain other identifiable brands.

BACKGROUND AND CONTROVERSY SURROUNDING COMPARATIVE ADVERTISING

Until recently the primary form of brand comparisons used in the consumer advertising field were Brand X euphemisms. Although comparisons with named competitors have been a long-established personal selling technique for consumer products, it was not until 1968 when American Motors compared the Javelin to the Mustang that direct brand comparisons began frequenting the print media. The use of named competitors has been slow to evolve in electronic media presentations of consumer products. However, due to the relaxation of traditional advertising taboos, the frequency of direct brand comparisons in consumer advertising has increased dramatically in the last two years.

This increase may be largely attributed to endorsements of comparative advertisements by the Federal Trade Commission and special interest groups. In 1972, the FTC issued a statement calling for advertisers to name competing brands in their advertisements as an alternative to the Brand X euphemism (Bradshaw, 1974). In conjunction with their request to advertisers, the FTC asked major television networks to change their policies prohibiting explicit mention of competitors in advertisements and to accept these otherwise suitable commercials on a trial basis as a means of informing the consumer about who competes against whom and how (Christopher, 1974).

Other groups, including consumerism advocates, have also cited benefits of comparative advertising in their endorsements. Joan Bernstein, Assistant Director of the Bureau of Consumer Protection, has stressed the potential of comparative advertising for delivering information not previously available to consumers (Grant, 1973). And although the American Association of Advertising Agencies has previously publicly discouraged the use of comparative advertising, a recent policy statement recognizes that comparative advertising provides the consumer with needed and useful information when used truthfully and fairly (Tannenbaum, 1974).

In spite of the benefits acclaimed to comparative advertising and the endorsements, some groups have begun to evaluate comparative advertising for negative side-effects. Jack Roberts, of Ogilvy and Mather, expressed concern that with comparative advertising the environment could become one of claims, counterclaims and contradictions resulting in increased opportunities for misleading or deceptive claims (Roberts, 1973). Both the Television Code and the Radio Code imply recognition of the potential of comparative advertising to be deceptive and misleading. The Television Code states that, "Advertising should offer a product or service on its positive merits and refrain by identification or other means from discrediting, disparaging or unfairly attacking competitors, competing products, other industries, professions or institutions" (Szybillo and Heslin, 1973). And the Radio Code states that "any identification or comparison of a competitive product or service, by name, or other means, should be confined to specific facts rather than generalized statements or conclusions unless such statements or conclusions are not derogatory in nature" (Ulanoff, 1972).

The American Broadcasting Company and the National Broadcasting Company have now developed formal guidelines for comparative commercials to help avoid the possibility of deceptive and misleading information. Although the guidelines of both networks are very similar, ABC places stress on the test procedures for comparative claims. Thus, the controversy over the deceptive nature of comparative advertising has led to more concern over an already scrutinized advertising technique, claim substantiations.

With regard to comparative advertising, several things seem apparent. First, public policy makers, media representatives, and other special interest groups have taken an increased interest in the use and impact of comparative advertising. The result is that many groups have endorsed comparative advertising on the assumption that it will provide the consumer with more and better information than will traditional advertising. Other groups, however, have expressed concern over the ability of comparative advertising to convey deceptive and misleading information to the consumer. Second, consumers have experienced and will probably continue to experience an increase in the use of comparative advertising strategies. And finally, since the use of comparative advertising has only recently been accepted, advertisers and public policy makers have virtually no information concerning its impact on the consumer. The endorsements and criticisms of comparative advertisements are a result of hypotheses which have neither-been tested nor validated, and no one has asked if comparative advertising is effective for the advertiser.

As advertisers are increasingly taking advantage of the fact that they can now employ comparative advertising, it is important that the relative impact of comparative and noncomparative advertising upon the consumer be empirically investigated. Since the ultimate objective of a firm's advertising is to elicit a purchase response from the consumer, the general purpose of this research is to investigate the relative effects of comparative and non-comparative advertising upon purchase intentions. Further, this investigation studies the effects of the communications with respect to other variables which may affect consumer reactions to comparative advertisements. Specifically, the effects of the consumer's degree of brand loyalty toward the advertised brand, degree of brand loyalty toward the competing brand in a comparative advertisement, the competitive position of the advertised brand and claim substantiation are considered in this investigation. The influence of three different but parallel copy themes is also investigated in order to eliminate copy testing of one theme.

METHODOLOGY

A four part questionnaire was administered to 594 university juniors and seniors. The first section of the questionnaire obtained information concerning the respondent's degree of brand loyalty toward alternative brands of deodorants. The respondent was asked to report what percentage of his deodorant purchases during the last year were Arrid, Right Guard and Sure, respectively. The next question in this section asked the respondent to name any other brands of deodorant he purchased during the last year, and to indicate the percentage of time each was purchased. All brand loyalty questions elicited unaided percentage responses with the stipulation that the percentage must sum to one-hundred percent. The focus of this section was to provide a surrogate measure of brand loyalty toward the advertised brand in the copy manipulation and toward the competing brand (Arrid) in comparative copy treatments.

In the second section of the questionnaire, the respondent was exposed to a fictitious piece of deodorant copy for either Right Guard, Sure, or a fictitious new brand (Secure). The copy explicitly stated the market position of the advertised brand, with Right Guard having the number one market position, Sure occupying the number three market position, and Secure being identified as a new brand. In the comparative copy manipulation, Arrid was always the competing brand and identified as occupying the number two market position. The comparative copy stated that the advertised brand was better than Arrid for exactly two product characteristics: ability to stop wetness and odor. Other product characteristics mentioned in all copy manipulations without a comparison were: goes on dry, keeps you dry, long-lasting, and never stings or stains. One-half of the claims in the copy manipulations were substantiated with the statement, "An independent testing agency's results have proven that..." while the other half of the copy manipulations did not include any claim substantiation statement. Three separate but parallel themes were included in the design to avoid copy testing and to increase the external validity of the study. Thus, the respondent could receive any one of thirty-six different copy manipulations included in the factorial design (2x3x2x3). There were two levels of advertisement copy type (comparative or non-comparative), three levels of competitive position (new brand, number one brand or number three brand), two levels of claim substantiation (substantiated or unsubstantiated), and three levels of copy theme. Respondents were randomly assigned to treatments and except for the inclusion of the experimental manipulations, all treatments were parallel.

The third section of the questionnaire asked the respondent to indicate how likely he would be to purchase the advertised brand. Purchase intention ratings were measured on a horizontal nine-point scale with the highest extreme labeled "very likely" and the lowest extreme labeled "very unlikely."

The last section of the questionnaire obtained information concerning the relative importance of product characteristics for which brand comparisons were made and other characteristics which were mentioned or not mentioned in the advertisements. Nine characteristics were presented to the respondent: price, scent, stops odor, stops wetness, applies dry, does not stain, does not' sting, long-lasting and over-all effectiveness, respectively. Respondents were asked to indicate on a horizontal nine point scale (highest extreme labeled "very important") how important each characteristic was to them when purchasing a deodorant.

The data from the first three sections of the questionnaire were submitted to analysis of covariance (Bart and Goodnight, 1972). Degree of brand loyalty toward the advertised brand and degree of brand loyalty toward the competing brand in the comparative advertisement were the covariates in this analysis. Thus, the influence of brand loyalty upon purchase intention ratings could be controlled. The product characteristic importance ratings were submitted to analysis of variance (Dixon, 1970).

VALIDATION STUDY

One limit to the external validity of this study is the use of copy manipulations. Copy manipulations were used to limit the amount of stimuli the respondent was exposed to and to allow as clean an attribution as possible that the results were a reaction to manipulations rather than to other features of a full advertisement. However, it is difficult to assume that consumers would react to full comparative advertisements in the same way as they would respond to comparative copy manipulations. For this reason, a validation study was conducted with full advertisements currently in print media to investigate differences in consumer reactions to comparative and non-comparative advertisements and copy.

The design of the questionnaire was similar to that used for the major study. The questionnaire obtained exactly the same information concerning the respondent's degree of brand loyalty toward both the advertised brand and the competing brand, as well as purchase intentions toward the advertised brand. The experimental manipulation consisted of one of two forms of an advertisement appearing in a local Sunday newspaper. The comparative advertisement for Dial deodorant was constructed so that the comparisons appeared in the upper left corner and could easily be eliminated for the non-comparative treatment. The validation study was administered to 77 university juniors and seniors with respondents randomly assigned to either the comparative or non-comparative treatment. The data were submitted to analysis of co-variance (Bart and Goodnight, 1972).

RESULTS

A summary of the analysis of covariance for purchase intention ratings toward the advertised brand appears in Table 1. Although all interactions were included in the analysis, only the results of the significant interactions and main effects are included in the summary table. The terms included in the model explain thirty-eight percent of the variation of purchase intention ratings. The probability of explaining this amount of the variation by chance is less than one percent.

TABLE 1

ANALYSIS OF COVARIANCE FOR PURCHASE INTENTIONS

There is no significant main effect of advertisement copy type. [The correlation coefficient for degree of brand loyalty toward the competing brand and advertisement copy type is .35. The effect of this may be downward biased F-values for advertisement copy type, thus these results should be interpreted with caution. For the comparative treatment, the partial correlation of the degree of brand loyalty toward the competing brand with purchase intentions is .08. Since there was no competing brand in the non-comparative copy, all respondents in this treatment received a zero for degree of brand loyalty toward the competing brand.] Thus, the purchase intention ratings elicited by the comparative advertisement are not significantly different from those of the non-comparative advertisement. There is also no significant main effect of competitive position, [Respondents receiving an advertisement for the fictitious brand were coded a zero for degree of brand loyalty toward the advertised brand. The three levels of the competitive position treatment were divided into two dummy variables co obtain a multiple correlation between competitive position and degree of brand loyalty toward the advertised brand. The multiple correlation coefficient is .33. This may result in downward biased F-values for the advertiser's competitive position.] claim substantiation or theme.

Two interaction effects are significant: the interaction of advertisement copy type and theme and the interaction of competitive position and theme.

Table 2 presents the mean purchase intention ratings for the interaction of advertisement copy type and theme adjusted for the covariates. The lowest mean purchase intentions occur for the interaction of the non-comparative advertisement and theme two, while the highest purchase intention mean occurs for the interaction of the comparative advertisement and theme two. Thus, although there are no relative differences in the impact of comparative and non-comparative copy, the particular theme used with a comparative advertisement may make it more effective than a non-comparative advertisement. Both themes two and three produce higher purchase intentions with a comparative advertisement than with a non-comparative advertisement.

TABLE 2

ADJUSTED MEAN PURCHASE INTENTIONS FOR THE INTERACTION OF ADVERTISEMENT COPY TYPE AND THEME

The adjusted mean purchase intention ratings for the interaction of the advertiser's competitive position and theme are reported in Table 3. The highest purchase intention ratings occur for the interaction of the number three brand and theme three, and the lowest purchase intention ratings occur for the interaction of the number one brand and theme three. Thus, different themes employed by the same advertiser may differently affect the influence of his advertisement. These interaction effects allude to the value of copy testing.

TABLE 3

ADJUSTED MEAN PURCHASE INTENTIONS FOR THE INTERACTION OF ADVERTISEMENT COPY TYPE AND THEME

The summary of the analysis of variance for product characteristic importance ratings is presented in Table 4. The results indicate a significant difference in mean importance ratings across respondents. The summary of mean importance ratings is presented in Table 5.

TABLE 4

ANALYSIS OF VARIANCE FOR PRODUCT CHARACTERISTIC IMPORTANCE RATINGS

TABLE 5

MEAN IMPORTANCE RATINGS FOR PRODUCT CHARACTERISTICS

The two product characteristics for which copy comparisons were made between the advertised brand and the competing brand in the comparative advertisements are stops wetness and odor. The highest mean for the importance ratings occurs for the characteristic stops odor. Stops wetness has the fourth highest mean rating and is higher than the grand mean. The second and third highest mean occur for over-all effectiveness and long-lasting, respectively. Overall effectiveness was not mentioned in any copy treatments, but the characteristic long-lasting appeared in all copy treatments and was not used for comparisons. Except for the product characteristic does not stain, all other means are lower than the grand mean. Thus, in the aspects of the copy which were not comparative, both relatively important and relatively unimportant characteristics were used, while the characteristics used for comparison appear to be relatively important to the respondents. This feature of the copy should serve to strengthen the comparisons.

The purpose of the validation study was to provide a measure of external validity concerning the generalizations that could be made from copy manipulations to full advertisements in print media. The results of the analysis of covariance for the validation study are summarized in Table 6. The terms included in the model explain twenty-one percent of the variation of purchase intention ratings. The probability that these results occurred by chance is less than one percent. The full advertisements used in the validation study produce no significant main effect of advertisement copy type. Thus, purchase intentions for the comparative advertisement do not differ significantly from those for the non-comparative advertisement. The results of the validation study are similar to the results presented in Table 1 for the variables in common.

TABLE 6

VALIDATION STUDY: ANALYSIS OF COVARIANCE FOR PURCHASE INTENTIONS

CONCLUSIONS AND IMPLICATIONS

Comparison of the results from the full advertisements and the copy manipulations tends to indicate that consumers react to full advertisements in essentially the same manner as they react to copy manipulations. Thus, under the conditions of this study, conclusions can be drawn from the copy manipulations which may be generalized to represent consumer reactions to full advertisements.

When isolated from other variables, a comparative advertisement is no more effective an influence upon purchase intentions than is a non-comparative advertisement. The results of the product characteristic ratings indicate that the characteristics used for comparison in the comparative copy were important to the respondents. Thus, any lack of significant differences in comparative and non-comparative purchase intentions probably cannot be attributed to the lack of importance of product characteristics used for comparison. It appears that, given the conditions of this experiment, an advertiser is neither going to increase nor decrease purchase intentions relative to a non-comparative advertisement by using a comparative advertisement.

However, in conjunction with specific themes a comparative advertisement may have a relatively stronger influence upon purchase intentions than a non-comparative advertisement. Thus, the choice of an advertising strategy should be mediated by the specific themes under consideration. An advertiser cannot simply assume that a non-competitive advertisement will have no relative effect upon purchase intentions toward his product without considering the impact of his theme upon the effect of his advertising strategy.

The theme of the advertisement appears to be an important influence upon the effects of other variables on purchase intentions. Not only does the theme temper the relative impact of a comparative advertisement, but the theme also interacts with the advertiser's competitive position to differentially influence purchase intentions. Thus, certain themes may be more effective for advertisers in a particular competitive position than are other themes. An advertiser should consider his market position as unique and use the theme which most favorably affects the impact of his competitive position upon consumer response.

Test results appearing in advertisements are receiving scrutiny by public policy makers and major television networks due to the possibility that deceptive claims may mislead consumer expectations and purchasing. However, the results of this study tend to indicate that substantiated claims are relatively no more effective in influencing purchase intention ratings than are unsubstantiated claims. This does not mean that public policy makers and other groups should not be concerned about the impact of deceptive claims or test results. The statements for claim substantiation only referred to an independent testing agency and not to a specifically named agency. Possibly, the inclusion of a named agency would make the claim substantiation stronger to the extent that it would be more effective than unsubstantiated claims. In addition, repeated exposure to an advertisement may produce differences in the effects of different strategies incorporated into this study. Finally, since consumers have been so frequently exposed to claim substantiation, inoculation may have occurred, and the assumptions surrounding the effectiveness of claim substantiation may not be valid.

No significant main effects occur for any of the uncontrolled independent variables. Overall, the lack of significant main effects and the presence of significant interactions tends to imply that the advertising industry is operating in a multivariate world. The effects of the variables in this study do not constitute a "simple'' situation, and "theories" or assumptions based on main effects only are not rich and do not adequately describe the situation. Basically, what this research implies is that general statements may be made concerning comparative advertisements only as indicated by the recurring underlying conditions and contexts.

Advertisers have been using comparative advertising and public policy makers, as well as other groups, have been endorsing it without empirical evidence regarding its effects upon the consumer. The endorsements are based upon the assumption that comparative advertising supplies the consumer with more information and more useful information than non-comparative advertising. The accuracy of these assumptions may have been alluded to by increased purchase intentions for comparative advertising. This was not the case. More information is needed concerning the impact of comparative advertising with regard to the assumptions and controversy surrounding the phenomenon. Future research should investigate the effects of comparative advertising upon other dependent variables for other products over-time.

REFERENCES

A. Barr and J. Goodnight, A User's Guide to the Statistical Analysis System (Raleigh: North Carolina University, 1972), 114.

Tom Bradshaw, "Comparative Ads: What's Their Status Now?" Television/Radio Age (April 29, 1974).

Maurene Christopher, "NBC Spells Out New Formal Guidelines for Comparative Spots," Advertising Age (January 28, 1974), 1.

W. Dixon, Biomedical Computer Programs (2nd ed.) (Berkeley: University of California Press, 1970), 586-600.

Don Grant, "Be Hard on Comparison Ads: Roberts to Four A's," Advertising Age (November 19, 1973).

Jack Roberts, "Comparative Advertising...I'm O.K... You're Not O.K.," Speech delivered to American Association of Advertising Agencies (New York: November 15, 1973).

George Szybillo and Richard Heslin, "The Television Code," (Washington, D.C.: National Association of Broadcasters, April, 1973).

S. Tannenbaum, "Policy Statement and Guidelines for Comparative Advertising," American Association of Advertising Agencies (April, 1974).

Stanley Ulanoff, "No Names Please: The Case for Comparative Advertising," Paper presented at the Annual Meeting of the American Academy of Advertising (University of Oklahoma: March. 1972).

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