The Effects of Recall on Belief Change: the Corrective Advertising Case

Neil K. Allison, University of South Carolina
Richard W. Mizerski, Federal Trade Commission
ABSTRACT - This study reports the association of product or brand recall with various beliefs consumers hold about a product subject to the corrective remedy. Respondents recalling exposure to a test product's corrective ad exhibited belief change for only the target belief of the corrective message. Those subjects who were unable to recall exposure had overall beliefs similar to a control group not exposed to corrective advertising.
[ to cite ]:
Neil K. Allison and Richard W. Mizerski (1981) ,"The Effects of Recall on Belief Change: the Corrective Advertising Case", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 419-422.

Advances in Consumer Research Volume 8, 1981      Pages 419-422

THE EFFECTS OF RECALL ON BELIEF CHANGE: THE CORRECTIVE ADVERTISING CASE

Neil K. Allison, University of South Carolina

Richard W. Mizerski, Federal Trade Commission

[The authors would like to acknowledge the financial assistance of the University of Cincinnati, College of Business, for support of this research.]

ABSTRACT -

This study reports the association of product or brand recall with various beliefs consumers hold about a product subject to the corrective remedy. Respondents recalling exposure to a test product's corrective ad exhibited belief change for only the target belief of the corrective message. Those subjects who were unable to recall exposure had overall beliefs similar to a control group not exposed to corrective advertising.

INTRODUCTION

The 1970's have seen the Federal Trade Commission (FTC) expand its role in consumer protection. An area of FTC jurisdiction concerns advertising that the Commission has defined to be false or misleading. This type of advertising is considered to be an unfair method of competition for which the Commission is able to enforce several remedies, the most severe and controversial being an order for corrective advertising. Such a ruling not only requires the advertiser to stop the false or deceptive communication but, in addition, to engage in a corrective advertising campaign that seeks to "... eliminate the residual effects of inaccurate or misleading information" (Wilkes and Wilcox 1974, p. 58) presented to the consumer in the previous advertising.

The most important action involving corrective advertising, in terms of both the size of the settlement and the legal precedent, was levied against Warner-Lambert's Listerine Antiseptic Mouthwash. The case against Warner-Lambert began in 1971 with an FTC complaint that Listerine ads falsely promoted the product as an "aid in preventing and ameliorating colds," because medical evidence failed to support this claim. Unlike previous corrective advertising cases, Warner-Lambert refused to sign a consent decree. In response, the FTC ordered Warner-Lambert to judicial review. This action set the legal testing ground for the FTC's power to issue and enforce the corrective advertising remedy. The order was upheld with slight modifications by the Court of Appeals for the District of Columbia in 1977, and in 1978 the U.S. Supreme Court refused to hear the case, thus establishing the legality of corrective advertising.

The modified corrective advertising order required Warner-Lambert to include the corrective message in all Listerine advertising until it had spent $10 million, an amount equal to the average annual advertising expenditures for the previous ten-year period. The content of the message was limited to "Listerine does not cure or prevent colds, or lessen their severity." The purpose of this and all other corrective messages was to change only consumers' falsely formed beliefs about the product and not to favorably or unfavorably affect other beliefs about the product, the manufacturer, or the competition. This additional impact would be considered punitive and beyond the legal bounds of the FTC.

Since corrective advertising was first proposed in 1969 as a remedy for false or deceptive advertising, the FTC, and the courts in the Listerine case, have been continually developing guidelines for its use and implementation. The following are a sample of significant decisions that have had an effect on the corrective advertising rule as it is presently enforced (Maddox and Zanot 1979).

- Corrective advertising can be justified in situations where there are residual effects of an advertising campaign that is no longer running (Firestone Tire and Rubber Company Case).

- Corrective advertising must at least reach the same target with similar impact as the original deceptive advertising (Sugar Association Case).

- Corrective advertising must appear in the same media and in the same manner as the original deceptive advertising (Lens Craft Research and Development Company Case).

- Corrective advertising need not include the phrase "contrary to past advertising" (Listerine Case).

- Background sounds or music during the corrective disclosure are not allowed (Listerine Case).

The evolution of the application of corrective advertising still leaves many issues unanswered. In general, the FTC lacks objective criteria for its decision making in the corrective advertising area. For example, expenditures for a corrective campaign have been mandated in different cases to be either 25% of all media expenditures, one in four ads for corrective advertising, or the $10 million figure in the Listerine case. Another issue that is a concern of both industry and government involves the impact of corrective advertising on the consumer. There have been a number of empirical studies with conflicting findings that address this issue (Hunt 1973a, 1973b, Dyer and Kuehl 1978, Mazis and Adkinson 1974). A recent study (Mizerski, Allison and Calvert 1980) concluded that multiple exposure to an actual corrective radio commercial had a significant influence on lowering the subjects' target belief for which that remedy was developed. At this point, it appears that corrective advertising, as it is presently implemented, is capable of removing a false belief without harming the offending advertiser or its competition.

As with any persuasive communication, standards for establishing the effectiveness of corrective advertising must be defined. A common industry technique for advertising effectiveness is recall testing. This includes aided and unaided recall performed on a day-after-exposure basis or with a more substantial time delay. (For a discussion of recall techniques, see Darrell Blaine Lucas and Stuart Henderson Britt, Measuring Advertising Effectiveness, Chapter 4). There have been many questions raised as to the validity of recall testing as a measure of advertising effectiveness. There is evidence that, over time, advertising may induce attitude change even though the specifics of the verbal content of the message itself are not recalled (Percy 1978). If this phenomena is true, then a measure of general advertising recall may be a necessary and sufficient criteria for establishing the persuasive impact of a communication, rather than a measure of detailed information recall. This study investigates the value of a measure of general advertising recall for establishing the effectiveness of corrective advertising.

Based on previous studies of corrective advertising and research on advertising recall, the following hypotheses are proposed:

1. Experimental subjects who recall hearing a corrective advertisement will have a weaker belief that Listerine "fights colds and sore throats" (target belief) than either experimental subjects who do not recall exposure to a corrective advertisement or control subjects.

2. Experimental subjects who recall hearing a corrective advertisement will not have different levels of beliefs concerning non-target attributes of Listerine than either experimental subjects who do not recall exposure to a corrective advertisement or control subjects.

3. Experimental subjects who recall hearing a corrective advertisement will not have a different belief concerning the honesty of Listerine's manufacturer than either experimental subjects who do not recall exposure to a corrective advertisement or control subjects.

4. Experimental subjects who recall hearing a corrective advertisement will not have a different affect/ attitude toward Listerine than either experimental subjects who do not recall exposure to a corrective advertisement or control subjects.

METHODOLOGY

[The data reported in this study are part of a more detailed investigation into the effect of corrective advertising. For a more complete discussion of the methodology see Mizerski, Allison and Calvert 1980.]

Sampling Plan

In order to more realistically investigate the effects of corrective advertising on consumers' cognitive structure than in previous studies (Hunt 1973a, 1973b, Dyer and Kuehl 1978, Mazis and Adkinson 1974) the natural setting of an on-campus radio station was used as the medium of the corrective advertising message. The station reaches only dormitory residents of a large mid-Western university by direct transmission via carrier current and does not broadcast to the general public, thus allowing for a controlled field experimental environment. The programming of the station follows the format of a typical contemporary rock commercial station with music, disc jockeys, news, commercials, games and giveaways.

A systematic random sampling plan from a sample frame of all dormitory residents was used to make initial contact with students. Potential respondents were telephoned and asked to participate in a study to help improve the programming and format of the on-campus radio station. This guise for the corrective advertising experiment was supported by an article in the campus newspaper discussing the need for a radio station survey that was to be conducted by university personnel.

All potential subjects were first asked whether or not they were able to receive the station on the radio in their own dormitory room. Because of the nature of the carrier current transmission and how it is affected by a building's structural design, individuals next to each other in the same dorm differed in ability to receive the station. Those students who reported they were able to receive the station were considered potential experimental group subjects and those who could not were considered potential control group subjects.

Potential experimental subjects were informed that participation would require four one-half hour listening periods over the course of a week. In order to increase participation, they were to be paid $5 for their efforts. As a manipulation check, during each of the four experimental sessions, the disc jockey announced a "listener survey password" that had to be presented to the researchers at the end of the study in order for the subject to receive his/her pay. To make participation in the survey even more attractive, students were given a chance to win their choice of record albums in the course of the four listening sessions. Twice during each half-hour period a student's social security number was randomly drawn and read out over the air. The respondent had five minutes to telephone the radio station and claim his/her prize, thus providing an additional manipulation check. After attrition due to extensive missing data, response bias or incomplete participation, there were 49 experimental subjects available for data analysis.

Those students who could not receive the radio station and who agreed to participate were told they would need to fill out a questionnaire concerning the on-campus radio station. The four one-half hour test sessions were not mentioned to this group of control subjects. For their participation, they received $0.50 coupons to a local fast-food restaurant. There were 51 control group subjects after removing those with obvious response bias or extensive missing data. Upon conclusion of the entire study, debriefing sessions showed that no control subjects were exposed to the experimental treatments.

Experimental Treatments and Measurements

The study employed a Solomon Four-group experimental design incorporating a pre-test measurement for half of the experimental subjects and half of the control subjects. No data collected from the pre-test will be used in the analyses presented in this paper and, therefore, it will not be discussed.

The experimental subjects were required to listen to the on-campus station in their own dorm rooms for at least the four designated one-half hour periods that ran from Monday to Thursday of one week. They were requested to listen and behave during these listening periods as they normally do while listening to the radio. This was done to prevent an inordinate amount of attention on involvement in the radio broadcast, thus keeping the deception and the subjects' behavior as realistic as possible.

The corrective ad was a professionally prepared adaptation from a current Listerine TV commercial that was neither corrective nor deceptive in nature. Although the theme of this ad was not being used on radio at that time, it was eventually extended to this medium after the experiment was completed. The actual Listerine corrective advertising campaign was not aired until data collection had been completed. The corrective message was placed near the middle of three of the four half-hour test periods. This was done to help disguise the true nature of the study. Commercials for other products and services were also aired during the test sessions.

Upon completion of the experimental treatments, the subjects were requested to go to the radio station's office to fill out a questionnaire concerning their feelings and attitudes toward their listening experiences. All respondents were required to report to the station within 11 to 13 days of their last experimental exposure. This spacing of the exposure from the data gathering was done to reduce demand and to lower sensitization to the experimental treatment as well as provide a more realistic measure of advertising recall.

The post-test questionnaire consisted of demographic data, opinions and attitudes toward the radio station's programming, and attitudes and beliefs toward commercial products. There were twenty-four belief statements, four each for a brand of beer, a local audio store, the University bookstore, a brand of toothpaste, a brand of soap, and Listerine. All belief statements were presented with a seven-point horizontal scale ranging from how "unlikely" to "likely" the belief statement represented the respondent's feelings about the product or service. The belief statements for Listerine were:

1. Listerine is effective for killing germs.

2. Listerine leaves mouth feeling refreshed.

3. Listerine fights colds and sore throats (target belief of the corrective advertisement).

4. Listerine has long lasting effects.

Perceptions of honesty for Listerine and the other product/service manufacturers were measured on a seven-point horizontal belief statement ranging from "likely" to "unlikely" while the seven-point affect/attitude measure for each product/service had scale labels of "like" to "dislike." Purchase intention as based on a 0% to 100% probability the respondent would buy the particular brand of product/service the next time the need to purchase that type of product/service arose.

Upon completion of the questionnaire, each experimental subject was interviewed by one of the investigators. They were first asked for any general comments and suggestions about the radio station. Unaided recall for all commercials and advertisements was then elicited followed by aided recall of any ads not mentioned under unaided recall. Finally, the subjects were asked what they thought to be the true nature and intent of the study. All respondents thought the sole intention was a programming study for the campus radio station without any underlying commercial or other research purpose.

Control subjects received a questionnaire similar to the experimental group questionnaire in both content and structure. All questions pertaining to product/service and demographic information remained the same. Information elicited about the radio station was phrased in terms of their "favorite local radio station."

Upon completion of the data collection, a de-briefing letter was sent to all experimental and control subjects. They were informed of the true nature of the study and that the Listerine ads were test ads, although the corrective statement would be similar once the company began its corrective advertising campaign.

ANALYSES AND RESULTS

Based on the experimental subjects' response to the Listerine ad during the exit interview, respondents were classified as either no recall (N=36), unaided recall (N=11), or aided recall (N=2). For purpose of analysis, unaided and aided recall subjects were grouped together into one recall category (N=13). No subjects who reported recall of the Listerine ad under the aided or unaided condition were able to remember any specific corrective or noncorrective information from the advertisement thus providing a measure of only source or general recall. Due to the self-selective nature of the recall variable, a series of chi-square analyses were computed in order to check for any inherent bias between people who could recall the Listerine ad and those who could not recall the ad. No differences between recall groups were found for product usage (three levels: nonuser, regular Listerine user, and regular other brand user), sex, class standing, age, ethnic origin or experimental grouping (pre-test vs. no pre-test group of the Solomon Four group design). These data suggest that there was no previous bias in the two groups that would affect their processing of a corrective advertisement (Table I).

TABLE I

CHI SQUARE ANALYSES

TABLE II

ANALYSES OF VARIANCE FOR BELIEF VARIABLES

Weighted one-way analyses of variance to adjust for unequal cell sizes were computed to test all hypotheses (Winer 1971, Nie, et. al, 1975). These data are presented in Table II. The only variable for which significant differences existed between recall, no recall, and control subjects was the target belief of "fights colds and sore throats," thus supporting all four hypotheses. In order to investigate whether significant differences existed between groups, a Scheffe test was performed on the third belief variable. The Scheffe test, which is an appropriate statistic to use for unequal cell sizes, is a strict or conservative post-hoc test. In order to report statistical significance, differences have to be rather substantial (Nie, et. al. 1975, Kerlinger 1973). At the P <.05 alpha level, the Scheffe procedure found the recall group subjects to be different from both the no recall and the control group subjects, and the no recall group to be equal to the control group.

SUMMARY, CONCLUSIONS AND IMPLICATIONS

Multiple exposures to an actual corrective radio advertisement had significant influence on lowering recall subjects' target belief of "fights colds and sore throats" while those subjects who did not recall the corrective ad were not effected by the experimental manipulation. No differences between recall, no recall, and control subjects were found for any non-target beliefs, perception of honesty, or affect/attitude toward Listerine. It appears that corrective advertising can be effective for removing or negating false beliefs formed and developed by deceptive advertising without affecting other beliefs and attitudes about the product. This appears true when the consumer is able to report or recall the ad, but can occur without awareness or recall of the specific corrective information presented in the communication.

Limited external validity of these data exists due to the product and brand studies, the student sample, and the procedure and the incentives used to induce cooperation and to elicit cognitive responses. Nonetheless, the validity of these findings and significance are enhanced by the field experimental nature of this study as opposed to the laboratory settings of all other empirical research assessing the impact and effectiveness of corrective advertising. At this point, it appears that corrective advertising, as it is implemented in the Listerine case, is capable of removing a false belief without harming the offending advertiser or its competition if the consumer has had sufficient exposure and cognitive activity to develop recall of the message. Such data should be useful to policy makers in guiding future enforcement of corrective advertising in terms of amount and endurance of the corrective advertising campaign.

REFERENCES

Dyer. R. F. and Kuehl, P. G. (1978), "A Longitudinal Study of Corrective Advertising," Journal of Marketing Research, 15, 39-41.

Hunt, H. K. (1973a), "Source Effects, Message Effects, and General Effects in Counteradvertising," Proceedings, Association for Consumer Research, 370-381.

Hunt, H. K. (1973b), "Effects of Corrective Advertising," Journal of Advertising Research, 13, 15-24.

Kerlinger, F. N. (1973), Foundations of Behavioral Research, 2nd ed. New York: Holt, Rinehart and Winston, Inc.

Lutes, D. B. and Britt, S. H. (1963), Measuring Advertising Effectiveness. New York: McGraw-Hill Book Company.

Maddox, L. M. and Zanot, E. J. (1979), "Corrective Advertising: Review and Prognosis," Current Issues and Research in Advertising, Division of Research, Graduate School of Business Administration, University of Michigan, 53-63.

Mazis, M. B. and Adkinson, J. E. (1976), "An Experimental Evaluation of a Proposed Corrective Advertising Remedy," Journal of Marketing Research, 13, 178-183.

Mizerski, R. D., Allison, N. K., and Calvert, S. (1980), "A Controlled Field Study of Corrective Advertising Using Multiple Exposures and a Commercial Medium," Journal of Marketing Research, 17:3 (August).

Nie, N. H., Hull, C. H., Jenkins, J. G., Steinbrenner, K., and Bent, D. H. (1975), Statistical Package for the Social Sciences, 2nd ed., New York: McGraw-Hill Book Company.

Percy, L. (1978), "Some Questions on the Validity of Recall Testing as a Measure of Advertising Effectiveness," Current Issues and Research in Advertising, Division of Research, Graduate School of Business, University of Michigan, 121-130.

Wilkes, R. E. and Wilcox, J. B. (1975), "Recent FTC Actions: Implications for the Advertising Strategist," Journal of Marketing, 38, 55-61

Winer, B. J. (1971), Statistical Principles in Experimental Design, 2nd ed., New York: McGraw-Hill Book Company.

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