On the Effectiveness of Repeated Positive Expressions As an Advertising Strategy

Paul M. Herr, Indiana University
Russell H. Fazio, Indiana University
[ to cite ]:
Paul M. Herr and Russell H. Fazio (1991) ,"On the Effectiveness of Repeated Positive Expressions As an Advertising Strategy", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 30-32.

Advances in Consumer Research Volume 18, 1991      Pages 30-32


Paul M. Herr, Indiana University

Russell H. Fazio, Indiana University

[The present work was supported by a grant from the Ogilvy Center for Research and Development and by Research Scientist Development Award MH00452 from the National Institute of Mental Health. The authors thank Alexander Biel, Director of the Ogilvy Center, for all his assistance.]

The present research focuses upon the effectiveness of an ad that employed what appears to be a particularly interesting strategy. The ad was for a Mirage candy bar, which was both novel, and unavailable, to our subject population. Consequently, the present research examines attitude formation.

The ad itself begins with a close-up visual of Mirage candy bar and identification as such by a narrator. What ensues is a collection of endorsements by different individuals for Mirage. First, a voice is heard to say, "Incredible." This is followed by a second voice saying, "It's better than Mom's apple pie." This is followed by a third voice saying, "Exquisitely satisfying." The narrator then concludes, "It still seems to defy description; in a word, perfection."

The strategy that underlies this ad seems to be one of inducing the audience to associate the product with very positive evaluations-- in other words to build an association in memory between the brand and a positive evaluation. It seeks to do so by repeatedly pairing the brand with a positive evaluation.

We found this ad interesting because a number of different perspectives suggest that the strategy might be effective. First, from a classical conditioning perspective, a positive attitude toward the brand may be created by these repeated pairings with positive evaluative terms. Indeed, the ad is reminiscent of classic experiments on the conditioning of attitudes (e.g., Staats and Staats, 1958; Zanna, Kiesler, and Pilkonis, 1970). Second, from an attributional perspective (Kelley, 1967), it is interesting to note that the repeated pairings stem from three clearly different individuals rather than a single individual. This implies some consensus for ascribing the positive evaluation to the brand. Such assumed consensus should increase the likelihood of attributing causality for endorsements to the entity in question-- namely, the brand. Finally, from a cognitive, verbal learning perspective, the repeated pairings increase the likelihood that a viewer who is constructing a judgment of the advertised brand at some later time will recall and consider the positive descriptor as a basis for judgment. Obviously, the more such pairings, the greater the likelihood that a positive associate will be recalled and considered.

For all of these reasons, we wondered whether this ad was indeed more effective in developing positive attitudes toward Mirage than other potential ads would be. To examine this possibility, we made comparisons of this target ad (hereafter referred to as the Repeated Pairings condition) to two other commercials. One was simply an edited version of the present commercial. Editing removed the positive expressions from the three voices, leaving only the introduction and the narrator's concluding statement (hereafter referred to as the Single Pairing condition). Because this edited version is considerably shorter than the original, we also made comparisons to another Mirage commercial that did not employ the strategy of depicting positive evaluative descriptors, but was- comparable in length. The ad consisted of a "striptease" of a Mirage candy bar. Little by little, accompanied by 'The Stripper" and shots of excited on-lookers, the wrapping of a Mirage candy bar was provocatively removed until the candy bar inside was revealed. (This ad will be referred to as the control condition). The presence of this ad serves merely to permit a comparison of the original ad with a very different ad of comparable length. Thus, we are able to control for the amount of time that people are exposed to advertising for the product.

The major dependent variable is the subject's attitudinal rating of the Mirage Candy Bar following exposure to one of the 3 ads just described. We also considered a second possible consequence of the Repeated Pairings ad-- one that is particularly appropriate as an advertising goal when introducing a novel brand. How well, if at all, have viewers learned after a single exposure to an ad, that the brand belongs to a particular product category? In this case, have viewers learned that Mirage is a member of the candy bar category? The importance of this category to brand association has been stressed in much previous research (e.g., Baker, et al., 1986; Howard and Sheth, 1969), and is generally accepted. It is sufficient to note here that an objective of advertising is often to increase top-of-mind awareness of the advertised brand, so it will be part of a consumer's consideration set at the time of brand choice.

In the present case, why might the repeated pairings condition be especially effective in developing the candy bar-Mirage association in memory? From a purely functional perspective, any new object that has been described so positively is worth taking note of. That is, if the object really is as excellent as described, it is in the best interest of the viewer to categorize it appropriately, if for no other reason than the viewer can test the veracity of the advertising claims.


134 subjects took part in an experiment ostensibly sponsored by an ad agency. Up to 4 subjects participated in any given session. Each subject was seated in an individual booth equipped with a color monitor. Subjects in all conditions were exposed to a videotape containing 21 commercials, which were divided into 3 segments of 7 commercials each. Subjects were led to believe that each segment contained commercials prepared by a different agency (identified only as Agency A, B, and C). The subjects were asked to consider the commercials of each agency in turn. More specifically, after each agency's segment, subjects were asked to rate how unusual or distinctive the ads seemed and to describe the agency's "personality". The purpose of this "cover story" was to ensure that subjects attended to all of the ads while focusing their attention away from the true purpose of the experiment.

Among the 21 commercials presented was one Mirage commercial. Subjects saw a videotape that included one of the 3 Mirage commercials described earlier: the Repeated Pairings, the Single Pairing, or the Control version.

Following viewing of the tapes, subjects engaged in a task designed to measure whether the brand was now viewed as a member of the product category. Category membership was measured via direct inquiries about the membership of a given brand in a specified product category. The monitors in each subject's booth were now used to present stimuli for this task. First, a category label was presented (for example "KETCHUP"), remaining on the screen for 750 milliseconds. The label was replaced by a brand name which may or may not have been a member of that product category (for example "HEINZ" or 'TIDE"). The subjects' task was to press, as quickly as possible, one of two buttons on a response box. If the brand was not a member of the preceding category, subjects were to press a button labelled "NO". If the brand was indeed a member of that category, then pressing the button marked "YES" was the correct response. The brand names remained on the screen until all subjects had responded (up to 6 seconds). Trials were separated by three seconds. Each of the product categories represented by the commercials just shown was presented four times; once followed by the brand featured in the commercial, once followed by another member of that category (filler brands), and twice followed by brands which were not members of that category. Included among these trials was the critical "CANDY BAR: MIRAGE" trial. In addition, fifteen other categories were included, followed by an actual category member in half the trials and by a nonmember in the other half. The order of presentation of category-brand trials was randomized by the same computer program that controlled the recording of the response latencies.

The instructions stressed both speed and accuracy in response. Given 1) the rushed nature of the response task, and 2) the limited learning that could reasonably be expected to take place following one exposure to the Mirage ad (among 20 other advertised products), how well subjects learned the category membership (if at all) might be detected by differential error rates and/or differential latencies given correct responses. So, if subjects have not learned the category membership well, they may either make an error, (respond that Mirage is not a candy bar) while attempting to respond quickly or they may take considerable time to respond correctly that it is a candy bar.

In the final segment of the experiment, subjects completed an attitude questionnaire. Subjects were asked to express their attitudes toward a list of products, including Mirage candy bars, on a 7-point scale with the lower endpoint labelled "very much dislike" and the higher endpoint labelled "very much like". Unlike the category membership measure (gathered first) Mirage was clearly noted as a candy bar. That is, the attitude object to be evaluated was "Mirage candy bar".


Category Membership

Analysis of category-brand responses indicated that the repeated pairings strengthened the category-brand association, relative to both the single pairing ad and the control. The proportion of subjects in each condition who correctly indicated that Mirage was a candy bar was .47, .07 and .30 for Repeated Pairings (n=47), Single Pairing (n=43), and Control (n=44) conditions, respectively. A Chi-square analysis revealed significant differences across condition, X2(2) =17.56, p<.001. Planned comparisons revealed the Repeated Pairings condition to be significantly different from both the Single Pairing condition, X2(1)=17.73, p<.001, and the Control condition, X2(1)= 2.86, p<.10.

Given that so few subjects were capable of correctly identifying Mirage as a candy bar (particularly in the Single Pairing condition), analysis of the response latencies is of questionable value. In any event, for those subjects who did correctly identify Mirage as a candy bar, response latencies were transformed prior to being subjected to a one-way Anova with ad a between subjects factor. - As response latency distributions are relatively skewed, it is necessary to first subject response latencies to a reciprocal transformation. This took the form of l/(latency + 1). Response latencies to both the Mirage candy bar and the filler products (brands that were members of the advertised product categories, but not featured in the commercials) were subjected to this transformation. The mean transformed response latency to the filler products provide us with a baseline measure of response latency. The difference between these two transformed latencies (Mirage latency minus mean transformed filler latency) constitutes our strength of association measure.

Overall, the resulting F was marginally significant, F(2,35)=2.86, p=.07. Planned comparisons of the mean latencies revealed that subjects in the Single Pairing condition were significantly slower (kl= -171 milliseconds) than those in the Control condition (M= -55 milliseconds), t(35)= 2.39,; < .05. Single Pairing subjects were also slower than those in the Repeated Pairing condition (M= -74 milliseconds), (35)= 2.08, p< .05. The Repeated Pairing condition latencies were not reliably faster than the Control latencies. Keep in mind, however, that because of the very high error rates, this analysis is based on very few subjects. Consequently, the analysis focusing on the percent of subjects who correctly responded that Mirage is a candy bar is more valid.


Attitudes were significantly more favorable toward the candy bar in the Repeated Pairing condition (M= 4.02) than the Control condition (M.= 3.21) and the Single Pairing Condition (g= 3.24). The one-way Anova with Ad as a between subjects factor revealed a significant effect for ad, F(2, 123)= 4-39, I?= .01. Planned comparisons revealed the expected effect for the Repeated Pairings ad creating significantly more favorable attitudes than did either the Single Pairing, t(121)= 2.47, 2< .05, or the Control condition, t (121) = 2.56,; < .05.


The results of an experiment in which the content of ads was systematically manipulated provide support for the notion that attitude favorability and the likelihood of correct category identification may be increased through repeated pairings of a brand and evaluations of the brand in a television ad. Although we are in no position to identify which of the processes described in the introduction was responsible for the effects, it is clear that repeated pairings can be a highly effective technique for creating favorable brand attitudes.

The data also indicate that this repeated pairings strategy is an effective technique for creating appropriate category-brand associations in memory. The increase in correct identification is intriguing, as it suggests an increased likelihood that the brand might "come to mind" when the product category is considered. Hence, not only can attitudes be made more favorable but the likelihood of a brand surfacing in an individual's consideration set can also be influenced through repeated pairings.


Baker, William J., Wesley Hutchinson, Danny Moore, and Prakash Nedungadi (1986), "Brand Familiarity and Advertising: Effects on the Evoked Set and Brand Preference," in Advances in Consumer Research, Vol. 13, ed. Richard J. Lutz, Provo, UT: Association for Consumer Research, 637-642.

Howard, John A. and Jagdish N. Sheth (1969), The Theory of Buyer Behavior, New York: John Wiley.

Kelley, Harold H. (1967), "Attribution Theory in Social Psychology," In D. Levine (Ed.), Nebraska Symposium on Motivation. Lincoln, Neb. University of Nebraska Press.

Staats, Arthur A. and Carolyn K. Staats (1958), "Attitudes Established by Classical Conditioning," Journal of Abnormal and Social Psychology, 57, 37-40.

Zanna, Mark P., Charles A. Kiesler, and P. A. Pilkonis (1970) "Positive and Negative Attitudinal Affect Established by Classical Conditioning," Journal of Personality and Social Psychology, 14, 321-328.