Physical Attractiveness and Personality in Advertising: More Than Just a Pretty Face?

Anne M. Brumbaugh, Duke University
ABSTRACT - Generally accepted in advertising is the adage that "what is beautiful is good," and the use of attractive spokespeople and models is de rigueur. Various hypotheses have been put forth to explain how attractiveness affects various measures of advertising effectiveness. Within the literature, however, it appears that for every significant result supporting attractiveness as affecting attitude towards a brand or product, another study fails to show the effect. This paper suggests that people's perception of an advertisement is affected not only by the spokesmodel's physical appearance, but also by personality inferences made by the viewer about the model.
[ to cite ]:
Anne M. Brumbaugh (1993) ,"Physical Attractiveness and Personality in Advertising: More Than Just a Pretty Face?", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 159-164.

Advances in Consumer Research Volume 20, 1993      Pages 159-164


Anne M. Brumbaugh, Duke University

[The author wishes to thank Debra Stephens, Ron Hill, Morris Holbrook, and Cindy Hanson for the use of their data.]


Generally accepted in advertising is the adage that "what is beautiful is good," and the use of attractive spokespeople and models is de rigueur. Various hypotheses have been put forth to explain how attractiveness affects various measures of advertising effectiveness. Within the literature, however, it appears that for every significant result supporting attractiveness as affecting attitude towards a brand or product, another study fails to show the effect. This paper suggests that people's perception of an advertisement is affected not only by the spokesmodel's physical appearance, but also by personality inferences made by the viewer about the model.


Advertising featuring attractive spokespersons and models fills television screens and print media, presumably because attractive people sell more products (Dion, Berscheid, and Walster 1972). Rarely is an unattractive person featured prominently in an ad, save the occasional "character" actor. However, empirical evidence is mixed in its support for the contention that physical attractiveness in a person pictured in an ad serves to increase ad effectiveness. Within the literature, it seems that for every significant result supporting physical attractiveness as affecting attitude towards a brand or product, another study fails to show the effect.

As people view an advertisement, and the person pictured in it, they form inferences about that person and his or her personality. Social psychology literature suggests that such inferences occur spontaneously and frequently as we observe others. This paper suggests that these personality inferences mediate the effect of physical attractiveness on ad effectiveness, that such personality inferences also influence directly how effective the ad will be, and that the formation and application of these inferences are affected by the gender of the observer.

The first section of this paper reviews some work on physical attractiveness in advertising. Sections two and three discuss the formation of personality inferences based on physical appearance and how advertising might be affected by such inferences. Section four proposes five hypotheses about the effects of physical attractiveness and personality inferences on attitude towards a product. Methodology and results of a study follow in sections six and seven, concluding with a discussion of the findings in the final section.


Advertisers have long accepted the idea that "beauty sells" and have utilized attractive celebrity endorsers, spokespeople, and models in their advertisements. Empirical studies bear out this phenomenon, showing that physical attractiveness of a person shown in an ad increases advertiser believability (Kamins 1990), willingness to purchase (Petroshius and Crocker 1989; Kahle and Homer 1985), direct mail response rate (Caballero and Pride 1984), attitude towards the product (Kahle and Homer 1985), and actual purchase (Caballero and Solomon 1983). Furthermore, this effect is found when both male and female models are used (Petroshius and Crocker 1989), for print advertising (Kamins 1990), for point of purchase displays (Caballero and Solomon 1984), for actual communicators in one-on-one interactions (Chaiken 1979), and for celebrity endorsements (Kamins 1990).

However, for each piece of evidence in support of the effect of physical attractiveness on attitude towards the ad or other measures of advertising effectiveness, there seem to exist several that fail to uphold the hypothesis. For example, Petroshius and Crocker (1989) show physical attractiveness to influence ratings of ad characteristics (interesting, appealing, impressive, attractive, eye-catching), but not measures of product information (believable, informative, clear) or product quality. Baker and Churchill (1977) show similar results.

While Kamins (1990) shows that the differential levels of physical attractiveness offered by celebrities Tom Selleck and Telly Savalas affect advertiser and spokesperson believability and spokesperson credibility, the difference in physical attractiveness does not affect arguably more important indicators of advertising effectiveness: attitude towards the brand, credibility of the advertisement, and purchase intention.

Caballero and Solomon (1984) find that physical attractiveness of a model pictured in a point of purchase display affects actual purchase of facial tissues, but not beer. Interestingly, the effect was not in the expected direction: the less attractive model yielded higher facial tissue sales than did the more attractive model. Their explanation for such seemingly incongruent results centered on the fact that the unattractive model may have attracted more attention to the point of purchase display than did the attractive model.

While some studies fail to show main effects of physical attractiveness on advertising effectiveness measures, some do show significant physical attractiveness by gender of subject interaction effects (Caballero and Solomon 1984; Kahle and Homer 1985). Other research specifically investigates gender differences in the impact of physical attractiveness on ad effectiveness (Debevec and Kernan 1984). In addition to the negative correlation between model attractiveness and facial tissue sales as found by Caballero and Soloman (1984), Kahle and Homer (1985) found that attractive sources were more effective with female subjects (but not male) in recognition scores for ads promoting disposable razor blades, and unattractive sources were related to lower recall scores for male subjects (but not female) for toothpaste ads.

Debevec and Kernan (1984) specifically try to assess the impact of the gender of target by gender of speaker interaction as well as model attractiveness on the effectiveness of a slide presentation soliciting support for a levy raising funds for the Cincinnati Zoo. Their results illustrate a number of gender differences in a variety of affective and behavioral measures. For example, females reacted more positively to the slide presentation picturing an attractive male model than to one showing an average male model, but were not more disposed to attending meetings or passing the levy. Conversely, male subjects' were affected on these dimensions when an attractive female model was pictured. Other results show that attractiveness affected different measures to different degrees, depending on the sex of the respondent and sex of the model featured (Debevec and Kernan 1984).


The mixed results shown in studies of physical attractiveness of a model on attitude towards the ad, source persuasiveness, attitude towards the product, purchase intent, coupled with the evidence of gender differences when the effect is shown to exist, suggest that something may be mediating the effect of physical attractiveness of the model on evaluative measures of the ad. Such a mediator might neutralize the persuasive effect of an otherwise attractive model, may enhance the effect of an average looking model, and might explain the effectiveness of less attractive models in advertising (Kamins 1990).

Physical attractiveness, or lack thereof, is a very salient, highly visible cue that observers use to form impressions of another person (Schneider, Hastorf, and Ellsworth 1979; Chaiken 1986). Based on this outward appearance, we make all sorts of inferences about the agents we observe. Among these inferences are judgments about the agent's personality (Chaiken 1979; Funder and Colvin 1988; Winter, Uleman, and Cunnif 1985; Winter and Uleman 1984; Debevec and Kernan 1984) and status (Kalick 1988). People spontaneously and unintentionally make inferences about others' personalities even after only a brief period of exposure. Even when other data about the agent are available, for example occupation, role and trait information, observers continue to rely heavily on physical appearance (Deaux and Lewis 1984). Furthermore, these inferences are surprisingly accurate. Funder and Colvin (1988) found that strangers' assessments of each other, made after only five minutes of exposure to each other, agreed with judgments made by close acquaintances as well as with self-assessments. Readily observable personality traits like extraverted, sociable, talkative, good-natured, funny, poised, status, interesting, sexually warm and responsive, and kind (Joseph 1982; Albright, Kenny, and Malloy 1988; Funder and Colvin 1988; Maddux and Rogers 1980) seem to be the most highly correlated with physical attractiveness.

Kalick (1988) had subjects match pictures of males and females of varying degrees of physical attractiveness with descriptions of different levels of ascribed (conferred through privileged background) and achieved (earned through hard work or applied talent) status. Irrespective of the status manipulation used (individual or family status, earned or inherited, rich achieved or ascribed), high status descriptions were associated with pictures of physically attractive people, and low status descriptions with less attractive photos. In assessing differences between achieved and ascribed status, he found the relationship between physical attractiveness and ascribed status to be stronger than that between attractiveness and achieved status. These assessments made solely on the basis of physical attractiveness are shown to exhibit gender differences as well. Albright et al (1988) showed that observers rated females as more conscientious when using only physical appearance as evidence of personality. Kalick (1988) found that female subjects tended to match more physically attractive photos with all status descriptions than did male subjects, and that sex of the stimulus person produced different results. When analyzing data of male stimulus persons, he found that achieved status was associated with higher physical attractiveness ratings, while for female stimulus persons achieved status was associated with lower attractiveness scores.


That people can't help themselves from forming inferences about others on the basis of appearance seems unequivocal. Do these inferences, once formed while viewing a model in an advertisement, affect the effectiveness of the advertisement?

Advertisers frequently try to convey certain personality characteristics through their choice of actors and the traits the actors display. Trustworthiness, credibility, and expertness are all traits that seem to influence positively the scores of evaluative measures of advertisements (Ohanian 1990). Perhaps advertisers also need to be mindful of the personality inferences people make spontaneously, in addition to those personality inferences that advertisers want to induce. Presuming the existence of personality as a mediator of physical attractiveness might help explain why some studies find no main effect for physical attractiveness or otherwise unexpected results, while others offer full support for its main effects and interactions.

For example, Caballero and Solomon's (1984) finding that a less attractive model was more effective in selling facial tissues than a more attractive model might not be just because the less attractive model was more noticed. Perhaps the result might be explained in terms of personality. People viewing the less attractive model may have perceived her to be less active, less outgoing, and consequently less likely to be healthy, while they perceived the more attractive model to be extraverted and healthy. The less attractive model's perceived greater experience with illness, then, made her a much better endorser for facial tissues than the attractive, healthy model.

Results of studies in which physical attractiveness was shown to impact affective measures, but not cognitive or conative measures (Petroshius and Crocker 1989; Baker and Churchill 1977; Caballero, Lumpkin, and Madden 1989), might also be explained by personality inferences as mediators of physical attractiveness. Physical attractiveness may have elicited perceptions of extraversion and kindness, for example, which caused observers to like the ad and the spokesperson. However, kindness may not have been an appropriate trait for an endorser of computers or luxury cars (Kamins 1990), thereby reducing any positive effect due to the sheer aesthetic vision offered by the source.

Baker and Churchill's (1977) seemingly inconsistent results might also be explained by personality as a mediating variable. Male subjects' high purchase intent for perfume may have resulted from their ascribing sociable and sexual warmth traits to the attractive female model pictured, traits they might look for in a romantic partner. Conversely, their high purchase intent for coffee after viewing a less attractive model might result from their ascribing to her traits like hard-working and intelligent, perceiving her to be a knowledgeable, expert coffee drinker.

Gender differences might also be explained by the different ways men and women perceive others and utilize physical cues in making inferences. Both men and women tend to confer on men stereotypically male traits like ambitious and aggressive, while they tend to attribute to women traits like caring and frivolous (Schneider et al 1979; Deaux and Lewis 1984). This natural tendency to resort to a conventional stereotype in the absence of additional information might cause viewers to ascribe personality traits to a model pictured in an ad which, while consistent with the stereotype, are not congruent with the product. Furthermore, even if men and women perceive the same personality traits, they might subsequently interpret them differently in the context of the advertisement. For example, while both men and women rate a female model high on sociability for an ad for exercise equipment, purchase intention for the men is higher than for the women because they associate sociability with health and health with exercise, while women's intent is lower because competence, not sociability, is the trait that would make them believe that exercise equipment is a smart investment for themselves.

This paper provides an initial attempt at teasing apart the effects of personality inferences from the effects due to physical attractiveness. One advantage of this paper is the structure of the data. Most studies on appearance in advertising manipulate physical attractiveness by pretesting a number of photos or ads featuring models of different levels of attractiveness and picking as stimuli the two photos rated most and least physically attractive (or three photos including an average rating). While manipulation checks verify that physical attractiveness ratings within this type of study differ significantly in the expected direction, such a manipulation doesn't allow for individual differences in preference. In this study, each subject makes his or her own ratings of twenty models, which are then used as measures of attractiveness in this repeated measures design.


Based on the idea that personality mediates the effect of source physical attractiveness on evaluative measures of the ad, a number of hypotheses will be examined in the following sections. First, H1 establishes that physical attractiveness of the source enhances the attitude towards the ad or, in this experiment, the "product," women's clothing.

H1: The physical attractiveness of a model pictured in an ad positively influences the attitude towards the product.

Next, it is necessary to show that attractiveness of the source is a cue that observers use to make inferences about the model's personality, particularly in the absence of other information.

H2: The physical attractiveness of a model pictured in an ad influences the formation of perceptions about the model's personality.

Once it has been established that physical attractiveness indeed influences the attitude towards the model's clothing and the personality inferences made about her, it may be hypothesized that gender of the observer plays a role in the formation of the personality perceptions as well.

H3: Gender of the subject influences the formation of perceptions about the model's personality.

It is further hypothesized that, not only does physical attractiveness affect observers' attitude towards the model's clothing, but that the personality inferences also have their impact.

H4: Perceptions about the model's personality affect the attitude towards the product.

Finally, it is hypothesized that there will be gender differences in how personality inferences affect attitude towards the product.

H5: Gender of the subject moderates the influence of personality inferences on the attitude towards the product.


The data for this experiment were obtained from the first of two sessions to assess the impact of women's body satisfaction/ dissatisfaction on their evaluation of advertisements picturing female models differing in physical attractiveness. In the first session, subjects were shown slides of 20 models and were asked to assess models' clothing, physical attractiveness, and personality. During the second session, two questionnaires measuring subjects' body dissatisfaction were administered. Only data from the first session are used in this paper.


Pictures of twenty models were obtained from a number of American and European fashion magazines and were selected to provide a variety of physical characteristics (hair, face, complexion, physique). Pretests with students showed that students varied in their assessment of physical attractiveness and differed in ascribing personality traits to the models based on the pictures. Tests of means on current data revealed that models differ significantly in attractiveness, with means ranging from 2.356 (unattractive) to 4.596 (attractive).


Evaluations of models' clothing, physical attractiveness, and personalities were obtained from 90 male and 88 female undergraduate subjects. Subjects were shown 20 pictures of 20 different models three times each. During the first exposure, subjects were shown each picture for approximately 20 seconds and were then asked to rate each model's clothing. The second exposure also lasted approximately 20 seconds, after which subjects were asked to rate the models' physical attractiveness. The third and final exposure to each slide of the models lasted approximately 30 seconds, after which subjects were asked to make judgments about the models' personalities.


Models' clothing was rated using six five-position semantic differential scales: unfavorable/favorable, neat/sloppy, like/dislike, not stylish/stylish, bad/good, and tasteful/tasteless. A single measure of clothing attractiveness was obtained by averaging the five semantic differential scales. This score served as the dependent variable in the analyses.

Two sets of measurements were obtained to assess physical attractiveness of the model. In the first measurement, subjects were asked to rate the model's hair, face, complexion, and physique on a five-position scale anchored by the phrases "not at all attractive" to "very attractive." A single composite attractiveness rating was obtained by averaging these four ratings. In the second measure, subjects completed the statement "This model is more physically attractive than ___ percentage of all females I've ever seen." This measure served as a manipulation check for the mean attractiveness rating, showing it to be a reliable reflection of the model's attractiveness.

Subjects recorded their personality inferences on twenty-two five-point semantic differential scales anchored by the following pairs: active/passive, not intelligent/intelligent, hard-working/lazy, snobbish/down-to-earth, withdrawn/outgoing, happy/unhappy, feminine/unfeminine, unpopular/popular, unsuccessful/successful, healthy/unhealthy, trustworthy/not trustworthy, self-conscious/self-confident, independent/dependent, nice/mean, not sophisticated/sophisticated, weak-willed/disciplined, free spirited/traditional, fearful/brave, neat/sloppy, not likeable/likeable, boring/interesting, exciting/dull. This list of personality traits was based on inventories found in previous studies (Brenner and Hinsdale 1978, Harris, Harris and Bochner 1982, Staffieri 1972), and was modified based on students' responses to pretests of the models. The adjectives were factor analyzed with a principal components analysis with a varimax rotation to yield four factors. Factor one, labeled "sociable," includes active, outgoing, happy, popular, healthy, confident, interesting, exciting, and free spirited. Factor two, labeled "capable," includes intelligent, hardworking, independent, disciplined, and brave. Factor three, labeled "poised," includes feminine, successful, neat, and sophisticated. Finally, factor four, labeled "friendly," includes down-to-earth, trustworthy, nice, and likeable. Cronbach coefficient alphas equal .88, .76, .75, and .76, respectively.




The impact of model attractiveness (MdlAttr), subject gender (Gender), and personality inferences (Sociable, Capable, Poised, Friendly) on subjects' attitude towards models' clothing (Acloth) is assessed by forming a series of Fadd calculations (Lutz 1977 p. 203). The Fadd statistic measures the additional contribution of one or more variables to a full regression model. To test main effects of gender, attractiveness, and personality on attitude towards clothing, a full model shown in equation (1) is used, with variables being omitted in turn as appropriate to test the various hypotheses.

"cloth =Gender Subj MdlAttr Sociable

Capable Poised Friendly (1)

To demonstrate that the formation of personality inferences is affected by subject gender and model attractiveness, a MANOVA analysis is performed using the four personality indices as dependent variables, and gender, model attractiveness, and the interaction between the two as independent variables. For further analyses of the gender main effect and interactions with personality indices, series of Fadd statistics are calculated by omitting variables from the full model shown in equation (2).

"cloth =Gender Subj Mdlattr MdlAttr*Gender

Sociable Capable Poised Friendly

Sociable*Gender Capable*Gender

Poised*Gender Friendly*Gender (2)

Interactions in a regression of equation (2) that were significant at the p=0.05 level are explored further by analyzing parameter estimates yielded by a regression of equation (3).

"cloth =MdlAttr*Gender Sociable*Gender

Capable*Gender Poised*Gender

Friendly*Gender (3)

To test hypothesis 1, physical attractiveness' influence on attitude towards clothing, a regression of equation (1) yields a significant contribution of model attractiveness on the total variance explained (F=440.93 (1, 3548), p<=0.0001). An Fadd statistic formed by omitting model attractiveness from the full model confirms the effect of model attractiveness (Fadd=77.31 (1, 3548), p<=.01), providing further support for hypothesis 1.

Hypotheses 2 and 3 were tested by performing a MANOVA analysis on the four personality indices, with model attractiveness, gender, and their interaction serving as independent variables. The table provides data from the MANOVA analysis which shows that model attractiveness and the interaction between gender and model attractiveness both contribute to the formation of three of the four personality indices. Gender alone does not impact the personality assessments. The third factor, poised, is not significantly influenced by the interaction between gender and model attractiveness.

To test hypothesis 4, the effect of the four personality factors on attitude towards clothing, a series of Fadd statistics were calculated as above. First, eliminating all four personality indices from the full model in equation (1) showed significant effects of personality on attitude towards clothing (Fadd=154.25 (4, 3548), p<=0.01). Separate Fadd calculations for reduced models which omitted each personality index individually provided support for the personality traits of capable (Fadd=3.70 (1, 3548), p<=.0562) and poised (Fadd=35.62 (1, 3548), p<=.01). The effects of the personality traits of sociable (Fadd=0.26) friendly (Fadd=0.06) were not statistically significant.

Support for hypothesis 5 comes from tests of the interactions between gender and the personality indices. Equation (2) shows the full model for the test of subject gender as a moderator of the influence of personality on the attitude towards the models' clothing. Significant interactions included capable (F=9.74 (1, 3547), p<=.0018) and friendly (F=4.32 (1, 3547), p<=.0378). To investigate the nature of the two interactions, a regression was run on the reduced model shown in equation (3) to obtain parameter estimates of the interactions involving gender. The capable by gender interaction was significant only for female subjects (estimate=.14, t=4.10, p<=.0001), while the friendly by gender interaction was significant only for male subjects (estimate=.06, t=2.34, p<=.0192).


Support for all five hypotheses was obtained. From these results, it appears that people use both physical appearance and their spontaneous impression of the source's personality when making judgments relating to the source. Physical appearance was shown to influence attitude towards clothing directly (H1). Both physical appearance and its interaction with gender of subject influenced the formation of personality inferences about the model (H2 and H3). These inferences, in addition to the direct effect of physical appearance, affected the subjects' overall impression of the models' clothing (H4). Finally, gender differences were found in both the formation of personality inferences, as well as in their use in forming attitude towards clothing (H5). Specifically, for women, a model's apparent competence provided a positive influence on their subsequent evaluation of her clothing, while for men, their perceived friendliness of the model caused them to make more favorable evaluations of her clothing.

These results suggest that personality inferences made by viewers of an ad mediate the influence of physical attractiveness of the model on evaluative measures of the ad's effectiveness, an idea that is consistent with other theories of when and how attractive models should be utilized. Results of studies that demonstrate different effects of physical attractiveness based on level of involvement of the viewer with the product (Kahle and Homer 1985) and source expertise (Maddux and Rogers 1980) might also be explained by influence of personality traits conveyed by the source. For example, Kahle and Homer's (1985) result that an attractive source leads to higher recognition scores only for females for razor blades (the high involvement product) may be because females view razor blades as more of a beauty product, rather than a daily necessity as they might be for men, and the personality traits conveyed by the attractive source reinforced that image of the razor blades. Maddux and Rogers' (1980) manipulation of source expertise and attractiveness showed there was no significant main effect for appearance on agreement with the source message, and that expert sources produced greater agreement than nonexpert sources. A further analysis of several of their secondary analyses might explain the results. They found that expert sources were rated higher on the attribute sincere, and that experts were more effective in eliciting agreement when they were less attractive. In terms of personality, the traits of sociable, status, outgoing, sexually warm and responsive, etc., which were found to be associated with attractive sources, may have in fact detracted from the credibility of the source, while the sincere trait contributed to source credibility. Rather than attributing their findings solely to the attractiveness and expert manipulations, intervening personality attributions may have played a mediating role.

Of course, limitations exist with this study. Female models were selected from magazines as stimuli, and as one might expect, none was very unattractive. Utilizing models of both genders that span the full range of physical appearance might yield richer results. The order in which data were collected might have also influenced the results. Here, subjects were first asked about their attitude towards the product, which may have affected their subsequent perceptions of the models' attractiveness, and in turn the models' personalities. Changing the order in which questions were posed might yield different results. Finally, by forcing subjects to make inferences about the models' personalities might actually cause them to make assessments that they might not have occurred spontaneously. Using open-ended measures might have avoided this pitfall.

The contribution of this paper to consumer research is to point out that physical attractiveness of a source pictured in an ad, by itself, does not guarantee that ad's success. Rather, the impressions that viewers make about the source and their subsequent application of those inferences are also relevant. Further research is needed to determine how these inferences are made in an advertising context, whether certain personality traits are more effective for viewers of one sex or another, which ascribed traits are desirable and which are undesirable, and whether these impressions can be effectively controlled to increase ad effectiveness. Kamin's "match-up" hypothesis (1990) and the incongruent results of other studies suggest that product class might also interact with physical attractiveness and personality inferences, making the three-way interaction the variable of interest in future research.


Albright, Linda, David A. Kenny, and Thomas E. Malloy (1988), "Consensus in Personality Judgements at Zero Acquaintance," Journal of Personality and Social Psychology, 55 (3), 387-395.

Baker, Michael J. and Gilbert. A. Churchill, Jr. (1977), "The Impact of Physically Attractive Models on Advertising Evaluations," Journal of Marketing Research, 14 (November), 538-55.

Brenner, David and Gary Hinsdale (1978), "Body Build Stereotypes and Self-Identification in Three Age Groups of Females," Adolescence, 13, 551-61.

Caballero, Marjorie J., James R. Lumpkin, and Charles S. Madden (1989), "Using Physical Attractiveness as an Advertising Tool: An Empirical Test of the Attraction Phenomenon," Journal of Advertising Research, August/September, 16-22.

Caballero, Marjorie J. and William. M. Pride (1984), "Selected Effects of Salesperson Sex and Attractiveness in Direct Mail Advertisements," Journal of Marketing, 48 (January), 94-100.

Caballero, Marjorie J. and Paul J. Solomon (1984), "Effects of Model Attractiveness on Sales Response," Journal of Advertising, 13 (1), 17-23.

Chaiken, Shelly (1979), "Communicator Physical Attractiveness and Persuasion," Journal of Personality and Social Psychology, 37 (August), 1387-97.

Chaiken, Shelly (1986), "Physical Appearance and Social Influence," in Physical Attractiveness, Stigma, and Social Behavior: The Ontario Symposium, Vol. 3, ed. C. Peter Herman, Mark P. Zanna, and E. Tony Higgins, Hillsdale, NJ: Laurence Erlbaum, 132-77.

Deaux, Kay and Laurie L. Lewis (1984), "Structure of Gender Stereotypes: Interrelationships Among Components and Gender Label," Journal of Personality and Social Psychology, 46 (5), 991-1004.

Debevec, Kathleen and Jerome B. Kernan (1984), "More Evidence on the Effects of a Presenter's Physical Attractiveness: Some Cognitive, Affective, and Behavioral Consequences," in Advances in Consumer Research, Vol. 11, ed. Thomas C. Kinnear, Provo, UT: Association for Consumer Research, 127-132.

Dion, Karen K., Ellen Berscheid, and Elaine Walster (1972), "What is Beautiful is Good," Journal of Personality and Social Psychology, 24, 285-290.

Funder, David C. and C. Randall Colvin (1988), "Friends and Strangers: Acquaintanceship, Agreement, and the Accuracy of Personality Judgment," Journal of Personality and Social Psychology, 55 (1), 149-158.

Harris, Mary B., Richard J. Harris, and Stephen Bochner (1982), "Fat, Four-Eyed, and Female: Stereotypes of Obesity, Glasses and Gender," Journal of Applied Social Psychology, 12, 503-16.

Joseph, W. Benoy (1982), "The Credibility of Physically Attractive Communicators: A Review," Journal of Advertising, 11 (3), 15-24.

Kahle, Lynn R. and Pamela M. Homer (1985), "Physical Attractiveness of the Celebrity Endorser: A Social Adaptation Perspective," Journal of Consumer Research, 11 (March), 954-61.

Kalick, S. Michael (1988), "Physical Attractiveness as a Status Cue," Journal of Experimental Social Psychology, 24, 469-489.

Kamins, Michael A. (1990), "An Investigation into the "Match-Up" Hypothesis in Celebrity Advertising: When Beauty May Be Only Skin Deep," Journal of Advertising, 19 (1), 4-13.

Lutz, Richard J. (1977), "An Experimental Investigation of Causal Relations Among cognition, Affect, and Behavioral Intention," Journal of Consumer Research, 3 (March), 197-208.

Maddux, James E. and Ronald. W. Rogers (1980), "Effect of Source Expertness, Physical Attractiveness, and Supporting Arguments on Persuasion: A Case of Brains Over Beauty," Journal of Personality and Social Psychology, 39 (August), 235-44.

Ohanian, Roobina (1990), "Construction and Validation of a Scale to Measure Celebrity Endorsers' Perceived Expertise, Trustworthiness, and Attractiveness," Journal of Advertising, 19 (3), 39-52.

Petroshius, Susan M. and Kenneth E. Crocker (1989), "An Empirical Analysis of Spokesperson Characteristics on Advertisement and Product Evaluations," Journal of the Academy of Marketing Science, 17 (Summer), 217-225.

Schneider, David A., Albert H. Hastorf, and Phoebe C. Ellsworth (1979), Person Perception, Reading, MA: Addison-Wesley.

Staffieri, J. Robert (1972), "Body Build and Behavioral Expectancies in Young Females," Developmental Psychology, 6, 125-27.

Winter, Laraine and James S. Uleman (1984), "When Are Social Judgments Made? Evidence for the Spontaneousness of Trait Inference," Journal of Personality and Social Psychology, 49 (2), 237-252.

Winter, Laraine, James S. Uleman, and Cathryn Cunnif (1985), "How Automatic Are Social Judgments?" Journal of Personality and Social Psychology, 49 (4), 904-917.