Sunteți pe pagina 1din 6

An experiment revealed the...

Effectiveness of Celebrity Endorsers

Charles Atkin and Martin Block


This a widely used type of source in contemporary advertising, the celebrity endorser. The study focuses on alcohol advertising and young audiences, since this has been a controversial issue in terms of the social effects of advertising. The sponsoring company is the underlying source of any advertising message, but the individual models depicted in the advertisment serve as the more visible comthoroughly studied source quality is credibility. Research conducted by social psychologists over the past 30 years demonstrates that a source perceived as highly credible is more persuasive than a low credibility sender (Hovland and Weiss, 1951; McGuire, 1969; Hass, 1981). The sources that companies use to present their advertising message typically attempt to project a credible image in terms of competence, trustworthiness, or dynamism. Celebrity endorsers are considered to be highly dynamic, with attractive and engaging personal qualities. Audiences may also trust the advice provided by some famous persons, and in certain cases celebrities may even be perceived as competent to discuss the product. According to Friedman, Termini, and Washington (1977), a celebrity is a person widely known to the public for accomplishments in domains unrelated to the product class. They cite a 1975 study showing that celebrities are featured in 15 percent of prime-time TV commercials. A later survey reported that this proportion was up to 20 percent {Advertising Age, 1978). The most widely used celebrity endorsers are sports figures, actors, or other types of entertainers. There are several reasons why a famous endorser may be influential: they attract attention to the advertisement in the cluttered stream of messages, they are perceived as more entertaining, and seen as trustworthy because of apparent lack of self-interest. This final element is due to the widespread attribution that major stars do not really work for the endorsement fee, but are motivated by a genuine affection for the product (Kamenera/., 1975). Despite the frequent use of famous endorsers, there is little published evidence regarding effectiveness. In one experiment, an advertisement for a flctitious brand of sangria wine featured an endorsement attributed to either a celebrity (actor Al Pacino), a professional expert, the company president, a typical consumer, or no source (Friedman, Termini, and Washington, 1977). College students read the ad and gave ratings on 0 to 10 scales of believability, probable taste, and intent to purchase. Across these three measures, the celebrity condition produced the highest scores. While the "no source" control group had a purchase intention rating of 2.7, the subjects exposed to the actor scored 3.9. Believability was rated 2.8 by control subjects versus 4.1 by those seeing the celebrity endorsement. For taste, the baseline of 4.0 compares to the celebrity group score of 5.6. None of the other three endorsers were as influential as the celebrated person. Miller's use of former athletes has been cited as a major reason for the success of Lite Beer campaigns on television over the past few years (Edwards, 1978). 57

Journal of Advertising Research Kamcn, Azhari, and Kragh (1975) reported that an oil company's advertising presenting popular singer Johnny Cash led to greater awareness of the ads and a positive change in company image. The type of endorser may interact with the type of product being advertised. Friedman and Friedman (1979) propose that the celebrity would be more effective for products with high psychologieal or social risk, involving such elements as good taste, self-image, and opinions of others. The expert endorser was seen as most appropriate for products associated with high financial, performance, or physical risk, while the typical consumer was considered to be the best source for low-risk products. In an experiment with housewives, actress Mary Tyler Moore was the celebrity endorser. The results on believabihty, ratings of the ad, attitude toward the product, and purchase intent showed the expected pattern of interactions between the three types of sources and three products representing differing risks. Regardless of the product class, the celebrity produced greater recall of both the advertisement and brand name. Since alcohol involves substantial psychologieal and social risk but minimal financial and performance risk, celebrity endorsers should be effective in promoting this type of product. Alcohol is an important type of produet to study because of the socially significant implications of advertising, particularly for young people in the audience. Adolescents and young adults are in a formative stage of developing orientations toward alcohol, and may be especially susceptible to advertising messages promoting beer, wine, and liquor. Since youth tend to respect famous entertainers and athletes, the celebrity endorser may be highly effective in influencing their attitudes. Social critics and public officials have raised questions regarding the impact of alcohol advertising, as refleeted by the following statement by Senator Hathaway in Senate hearings on the subject: "Many millions of .American youth are bombarded every day with many thousands of messages about drinking from many hundreds of glamorous, friendly, healthy, adventuresome, sexy, even famous people telling them of the joysandbenefitsof drinking. How much of it is actually designed to encourage non-drinkers to start drinking?" (Hathaway, 1976). In an attempt to substitute empirical evidence for hythe impact of celebrity sources in the context of alcohol advertising. Based on the literature, it is predicted that a celebrity source will have a greater impact than a noncelebrity on responses to the advertisement and to the advertised product. It is expected that the celebrity will be seen as more credible and that the message will be rated more favorably along evaluative dimensions. This positive response is expected to produce a more favorable attitude toward the product and a greater intention to use it. Furthermore, it is anticipated that ado\t^^^^t, will demonstrate the strongest ^^^^^^^^ ^^ ^^^ ^^^^^^^^^ endorsers, rel ^^j^^ ^^ adults _ _ _ _ ^ _ ^ _ _ _ _ _ _ Experimental Method i^n^i-^^-i^^^.^ The experimental manipulation varied the celebrity status of the source endors ing an alcohol brand. Multiple operationalizations of the stimulus treatment were employed to avoid idiosyncratic results specific to a particular brand, creative execution, medium, or endorser. Three versions of nearly identieal pairs of ads featured either a celebrity or a noncelebrity character. In each case, an existing ad depicting a famous person vvas compared with a modified version of the same ad using a noncelebrity per son to endorse the product. The original advertisements had not been extensively disseminated, and none had appeared in the year preceding the study. Less than one-tenth of the subjects reported seeing any of the messages before the study. The first of the test stimuli was a whiskey magazine advertisement featur ing actor Telly Savalas. In the nonceleb rity version of this message, the picture o\ an unfamiliar middle-aged man in identical dress and po.se replaced the original picture. In a second variation ot the stimulus message, the celebrity con dition used a newspaper beer ad showing former basketball player Happy Hair ston; an unfamiliar tall black man was pictured in the noncelebrity version ot the ad. The third variation used a magazine ad for a whiskey brand portraying model C^heryl Tiegs before she had bcThe celebrity treatment involved label

Charles K. Atkin is professor in the Department of Communication at Michigan State University, where he has taught since 1971. His PhD is from University of Wisconsin in Mass Communication. In addition to alchohol advertising research, he has conducted studies of children's advertising and political advertising.

Martin Block is Chairman of the Department of Advertising at Michigan State University. He received his PhD in Mass Media from Michigan State. Block is an associate of the ELRA Group.

Volume 23, No. 1, February/March, 1983 ing her by name, with the identifier "world famous model." The noncelebrity version contained no identification and altered her hair color to disguise her identity. Subjects were exposed to the test ad during a testing session where they were asked to evaluate a series of seven unrelated advertisements. After a brief exposure to the message, they completed a three-page questionnaire dealing with the advertisement. The instrument contained the following items to measure the dependent variables: (a) ad rating: The 18 semantic differential scaling items listed in Table 1 were presented on a page of the questionnaire, with each pair of polar-opposite adjectives anchoring the ends of a seven-step scale; (b) believability: Respondents were asked, "On a rating scale from 0 to 10, how believable is this ad?" followed by a continuum with numbers from 0 to 10 anchored by the labels "not believable" and "very believable"; (c) character rating: The "person in the ad" was rated along three 0 to 10 scales anchored by "not trustworthy" and "very trustworthy," "not attractive" and "attractive," and "not competent" and "very competent"; (d) product image: The product was rated on six pairs of semantic differential adjectives along a sevenstep scale, as shown in Table 1; and (e) behavioral intention: A single item asked, "On a rating scale from 0 to 10, how likely would you be to get this product?" with the eleven-step scale ranging from "not likely" and "very likely." This question did not refer to purchase or consumption because these behaviors are constrained for the younger subjects in the sample. The sample was composed of 196 subjects who were randomly assigned into either the celebrity or noncelebrity treatments. Half were underage adolescents between 13 and 17 years old, while the other half were predominantly young adults ranging in age from 18 to 77. Judgmental sampling was used to obtain participants with diverse characteristics and backgrounds representative of the broader population. Junior and senior high schools provided the major source of adolescents, supplemented by shop-

Mean Ratings for Celehrity and Noneelebrity Conditions Response variable Advertisement rating: (1 to 7 scale) Useless-Useful Dislike-Like Bad-Good Tasteless-Tasteful Dishonest-Honest Unpleasant-Pleasant False-True Unenjoyable-Enjoyable Irresponsible-Responsible Boring-Interesting Traditional-Modern Unsexy-Sexy Weak-Strong Indecent-Decent Ineffective-Effective Unimportant-Important Drunk-Sober Poor-Rich Believability of ad: (O = not, to 10 = very) Character rating: (O = not, to 10 = very) Trustworthy Attractive Competent Product image rating: (Ito 7 scale) Dislike-Like Bad-Good Unpleasant-Pleasant Unenjoyable-Enjoyable Inferior-Superior Unsexy-Sexy Behavioral intention: (O = not, to 10 = very) *Significant difference, p<.05 ping mall intercepts. For those in the young adult age segment, about half were chosen from college or trade school classes. For categories of people not available in school setfings, participants were solicited through newspaper classified advertising, direct mailing to randomly selected households, shopping mall intercepts, and social clubs. Demographically, the sample is fairly typical in terms of sex (49 percent male and 51 percent female), race (84 percent White, 12 percent Black, and 4 percent Chicano or Oriental), community size 59 4.17 4.17 4.35 4.41 4.03 3.73 2.82 3.71 3.87 3.86 3.91 3.86 3.36 2.36 1.82 1.30 2.09* 2.05* .84* 1.43 1.15 3.56 126 107 4.46 3.56 4.61 3.51 4.39 3.86 4.14 4.76 3.86 4.81 .11 4.59 .26 4.90 4.96 3.82 3.42 4.00 3.82 4.11 3.90 4.28 3.65 4.33 4.03 3.36 4.38 3.54 4.16 5.08 3.89 2.76 4.79 4.79 3.97 .56 .95 1.06 1.45 -1.45 1.39 - .60 .25 - .92 2.92* 1.47 .97 2.57* 2.08* 2.54* 2.08* .45 .69 - .41 Celebrity character (N=98) Noncelebrity character (N = 98) ratio

5.79 6.08 6.08

5.02 5.77 5.45

2.28* .54* 1.98*

Journal of Advertising Research (18 percent large urban city, 30 percent medium urban city, 31 percent suburban/college town, and 21 percent small town or rural area), occupational status of self or parent (35 percent upper middie cla.ss, 39 percent lower middle cla.ss, and 26 percent working class or lower ela.ss), age (17 percent 7-8th grade, 19 percent 9-lOth grade, 18 percent ll-12th years old, 16 percent 15-16th grade or 21-22 years old, and 13 percent 23-77 years old), and church attendance (30 percent regular attenders, 23 percent DCcasional, 25 percent seldom, and 22 percent never). Geographical di.spersion was attained by supplementing the MidSouth, and East (52 percent Michigan, 17 percent California, 17 percent New York. and 14 percent Georgia). ^ - Findings ^ ^ ' ^ Advertisements featuring celebrity figures produce consistently more favorable impact than the noncelebrity ads, as shown in Table 1. For 15 of the 18 advertisement rating scales, the mean score is higher for subjects exposed to the famous source; the difference is significant on strong, interesting, effective, and important. Summing all 18 semantic differential scales, the total celebrity score of 76.7 is significantly higher than the noncelebrity total of 72.3. The average rating score across all scales is 4.26 versus 4.02. The difference is small and nonsignificant on the scale measuring believability of the message. Celebrity characters are perceived as significantly more trustworthy and competent, and slightly more attractive. Summing the three eredibihty rating scales, the overall difference is significant. On the average, the celebrated character is rated 5.98 while the noncelebrity is rated 5.41 on the 0-to-lO continuum. Regarding product image, all six ratings are higher in the celebrity condition. pleasant and enjoyable measures. The ^^^^^ .^^^^^ _^ ^^ You'nger^and Older Subjects Celebrity 4.23 4.28 4 06 3 3-

Advertisement ratings: (18-item average) Believability:

Younger Older Younger Ql^jg^

' Product image: (6-item average) Behavioral intention:

Younger Older Younger Older

4.1.3 4.16 2.82 2.83 i.

cantly greater than the total of 22.2 for the noncelebrity condition. The average difference across the six ratings is 4.14 versus 3.76 on the l-to-7 scale. Finally, the likelihood of actually getting the alcohol product is slightly greater for those exposed to the famous endorser, but the half-point difference on this scale is not significant. Table 2 presents the findings for younger versus older subjects. The youth subsample is more impressed by the celebrity characters. On 17 of 18 ad rating scales, the famed model generates the higher score. The total rating is significantly different. The younger subjects also hold a significantly more positive evaluation of the product endorsed by the celebrity, and the likelihood rating is almost one point higher than in the noncelebrity condition. For these three sets of items, the treatment differences for the adult subsample are not significant. Both age have a significant difference on ter ratings, and neither group significantly on the believability groups characdiffers score.

than the almost-identical versions with a noncelebrity. particularly for the adjectives strong, interesting, effective, and important. The image of the product tends to be more favorable when a famous endorser is shown: readers are especially likely to rate the alcohol brand as enjoyable and pleasant. However, the behavioral inten^ tion is only slightly greater than tor an unfamiliar character, Most of the differences on advertise ment ratings and product images occur among the youth subsample. .Adults are only slightly intluenced by the famous sources. While the older subjects rated the celebrities favorably, this apparently did not transfer to evaluations of the advertising message or the endorsed prod^ act. For teenagers, the impact ot the celebrity on the source credibility ratings extends to the message and product This study provides mixed evidence of the influence of celebrity sources in the alcohol advertising context. The limited impact on adults may reflect strongly established orientations toward alcohol ads and products, such that it makes little difference whether a famous endorser is associated with the message, .Adults may di.scount the relevance of celebrities to alcohol, or may not be impressed by the fame of the individuals m these ads. In addition, the experimental situation artificially heightens iralizing the attention-getting advantage

aiM^ii^^K^K^^^^^^^^Hi^KMKiBH Conclusion ^ ^ ^ ^ ^ ^ ^ ^ _ ^ _ _ _ _ _ _ _ _ _ The use of famous persons to endorse alcohol products is highly effective with teenagers, while the impact on older per.sons is limited. For all age group.s, the competent and trustworthy. Ads featur-

Volume 23, No. 1, February/March, 1983 of the celebrity source. The clear impact on adolescents suggests that young people may be more readily persuaded by a famous name linked to a product. Teenagers may be at a more impressionable stage of development, where the endorsement of a celebrity is regarded as a meaningful factor in evaluating ads and products. The finding that famous sources enhance the impact of alcohol ads indicates that the social implications of this practice should be given closer examination. To the extent that celebrity endorsements for beer, wine, and liquor encourage underage youth to have a more favorable disposition toward alcohol, there is a basis for concern by responsible advertisers. Key Trend." Advertising Age (February 27, 1978): 32. Edwards, Larry. "John Murphy of Miller is Adam ofthe Year." Advertising /Ige (January 9, 1978): 1. Friedman, Hershey, and Linda Friedman. "Endorser Effectiveness by Product Type." Journal of Advertising Research, 19, 5 (October 1979): 63-71. Friedman, Hershey, Salvatore Termini and Robert Washington. "The Effectiveness of Advertisements Utilizing Four Types of Endorsers." Journal of Advertising, 6 (1977): 22-24. Hathaway, William. Opening Statement. Media Images of Alcohol: The Effects of Advertising and Other Media on Alcohol Abuse. Washington: Government Printing Office, 1976. Hass, R.G. Effects of Source Characteristics on Cognitive Responses and Persuasion. In R. Petty, T. Ostrom, and T. Brock (Eds.), Cognitive Responses to Persuasion. Hillsdale, N.J.: Erlbaum, 1981. Hovland, Carl, and Walter Weiss. "The Influence of Source Credibility on Communication Effectiveness." Public Opinion Quarterly, 15 (1951): 635-650. Kamen, Joseph, Abdul Azhari and Judith Kragh. "What a Spokesman Does for a Sponsor." Journal of Advertising Research, 16, 2 (April 1975): 17-24. McGuire, William. "The Nature of Attitudes and Attitude Change." In G. Lindzey and E. Aronson (Eds)., Handbook of Social Psychology, Reading, Mass.: Rand McNally, 1968.

"Celebrities in TV Ads Spotlighted s

The computer and statistical support need

ycu

Whether your problem involves simple survey tabulation or a sophisticated market simulation, problem, caii us. Crosstabs Editing, cleaning, data entry Conjoint and tradeoff analysis Perceptuai mapping Location models Market simulation Custom programming Sample design and selection Call or write for information

ITO Marketing Inc. Research,


214/385/9170 6750 HILLCREST PLAZA DR. SUITE 214 DALLAS, TEXAS 75230

S-ar putea să vă placă și