More Evidence on the Effects of a Presenter's Attractiveness Some Cognitive, Affective, and Behavioral Consequences

Kathleen Debevec, University of Massachusetts
Jerome B. Kernan, University of Cincinnati
ABSTRACT - This study examines the effect of both male and female models' physical attractiveness on male and female respondents' reactions to a slide show, designed to gain verbal and behavioral support for a local to unity issue. Cross gender reactions to attractiveness are highlighted, since they have received little attention in past attractiveness research. Presenters' persuasive effectiveness is gauged by comparing reactions to a control version of the slide show (no model pictured) with treatment versions featuring models of varying attractiveness (all shows presenting identical information). Thus, central versus peripheral routes to persuasion (Petty and Cacioppo 1983) are evaluated within the context of subjects' and models' gender and the latter's physical attractiveness.
[ to cite ]:
Kathleen Debevec and Jerome B. Kernan (1984) ,"More Evidence on the Effects of a Presenter's Attractiveness Some Cognitive, Affective, and Behavioral Consequences", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 127-132.

Advances in Consumer Research Volume 11, 1984      Pages 127-132

MORE EVIDENCE ON THE EFFECTS OF A PRESENTER'S ATTRACTIVENESS

SOME COGNITIVE, AFFECTIVE, AND BEHAVIORAL CONSEQUENCES

Kathleen Debevec, University of Massachusetts

Jerome B. Kernan, University of Cincinnati

ABSTRACT -

This study examines the effect of both male and female models' physical attractiveness on male and female respondents' reactions to a slide show, designed to gain verbal and behavioral support for a local to unity issue. Cross gender reactions to attractiveness are highlighted, since they have received little attention in past attractiveness research. Presenters' persuasive effectiveness is gauged by comparing reactions to a control version of the slide show (no model pictured) with treatment versions featuring models of varying attractiveness (all shows presenting identical information). Thus, central versus peripheral routes to persuasion (Petty and Cacioppo 1983) are evaluated within the context of subjects' and models' gender and the latter's physical attractiveness.

BACKGROUND

A communicator's physical attractiveness has been investigated as a vehicle mediating both persuasion (Chaiken 1979; Horai, Naccari, and Fatoullah 1974; Maddux and Rogers 1980; Mills and Aronson 1965; Mills and Harvey 1972; Norman 1976; Snyder and Rothbart 1971) and behavioral influence (Dion and Stein 1978). Attractive communicators have proved to be more persuasive "sources" than unattractive or unpictured communicators, a finding not necessarily related to their perceived expertness (Horai, Naccari, and Fatoullah 1974; Snyder and Rothbart 1971) or trustworthiness (Snyder and Rothbart 1971). Physically attractive sources apparently need not rely on supporting arguments to persuade their audience while expert sources seemingly must (Norman 1976).

The prevalence of attractive models in advertising testifies to the general belief concerning their efficacy as a vehicle of promotion. Researchers have investigated individuals' perceptions of models themselves, of models relative to the ad in which they appear, and relative to the impressions formed of the product represented (Baker and Churchill 1977; Kanungo and Pang 1973; Smith and Engel 1968; Steadman 1969). Baker and Churchill (1977) found attractive models to be especially effective in altering individuals' affective impressions of products (i.e., the products' aesthetic qualities but not their cognitive and conative dimensions). Their results also indicate that the type of product presented as well as the gender of the perceiver interacted with the physical attractiveness variable. For example, males reacted more positively to a physically attractive female model when representing a romantically-oriented product (perfume/cologne) than when representing a functional product (coffee) and vice versa for the unattractive female model. In general, attractive individuals are perceived in a highly positive light and reflect very favorably on the products they represent.

In addition to realizing that attractive models and communicators can enhance the persuasive power of a message, it is important to understand the judgmental process they prompt by perceivers. An individual's physical attractiveness serves as a cue for perceivers. An individual's physical attractiveness serves as a cue for perceivers to make inferences about his/her personal characteristics, abilities, and motivations. A physical attractiveness stereotype exists such that attractive persons tend to be evaluated in a favorable light while unattractive persons tend to be viewed rather negatively (Dion, Berscheid, and Walster 1972). There also appears to be a high degree of agreement among perceivers (regardless of their gender) in evaluating attractiveness (Berscheid et al. 1971; Cavior and Dokecki 1971; Kopera, Maier, and Johnson 1971; Murstein 1972; Walster et al. 1966). Although a considerable amount of evidence has been compiled with regard to judgments made of others based on their physical attractiveness, individuals are not apt to admit that they make such judgments (Hudson and Henze 1969; Miller and Rivenbark 1970; Tesser and Brodie 1971).

The present research is an attempt to understand the inferences perceivers make about models based on the model's attractiveness and to ascertain the impact of these inferences on perceivers' evaluations. The study investigates the effect of models' physical attractiveness relative to the gender of the perceiver in a real life promotional context. Attitudes toward the advocated issue, verbal compliance, and impressions of the model are solicited. While the former measures have been sought previously in marketing studies, the latter impressions have received attention only in psychological research.

In addition, within- and across-gender attractiveness effects have received little attention in advertising research. Excepting the Baker and Churchill (1977) study, female models have been the primary stimuli for subjects' perceptions and, instead of manipulations of their physical attractiveness, models' sexiness (or lack thereof) has taken precedence. The present study examines the effectiveness of male and female models' physical attractiveness as perceived by both male and female subjects--effects which have been largely overlooked in both the consumer and psychological literature.

To explicate these many relationships, consider the following scenario. The Cincinnati Zoo wished to encourage support for an upcoming tax levy whose proceeds were designated for its operating expenses. Zoo management, in concert with the authors, agreed that a slide show would be prepared and presented to various community groups. The purpose of the presentations was to enlist volunteer help aimed at securing voter passage of the levy.

The verbal content of the presentations was decided without much argument. [A complete script of the presentation is available from the authors.] There was some concern, however, over the issue of "presenters" -- whether models should be depicted in the presentations (Petty and Cacioppo 1983) and, if so, whether they should be physically attractive or ordinary looking. [Unattractive models were thought to have no efficacy and therefore were not considered further.] Additionally, there was the question of whether the models should be male or female. The intended audience for the slide-show presentation contained both males and females.

These issues prompted the experimental design depicted in Figure 1. As the figure shows, there were eight treatment groups (1-8) and two control groups (9-10).

FIGURE 1

EXPERIMENTAL DESIGN FOR CINCINNATI ZOO SLIDE-SHOW PRESENTATION

Hypotheses

The following working hypotheses (based on the extant literature) guided the study (numbers in parentheses refer to cells in Figure 1):

H1: Attractive male and female models will generate more favorable impressions, attitudes, and compliance among both male and female subjects than their average-looking counterparts (1>3; 2>4; 5>7; 6>8).

H2: Female subjects will form more positive impressions and attitudes when viewing a presentation featuring an attractive male than when viewing one featuring an attractive female; vice-versa for male subjects (2>6; 5>1).

H3: Female subjects will form more positive impressions and attitudes when viewing a presentation featuring an average-looking male than when viewing one featuring an average-looking female; vice versa for male subjects (4>8; 7>3).

H4: Among subjects viewing a presentation with an attractive female model, males will form more positive impressions and attitudes than females; vice-versa for subjects viewing a presentation with an attractive male model (5>6: 2>1).

H5: Presentations featuring attractive male or female models will generate more favorable impressions, attitudes, and compliance than those featuring no model (1>9; 5>9; 2>10; 6>10).

H6: Presentations featuring average-looking male or female models will generate less favorable impressions, attitudes, and compliance than those featuring no model (9>3: 9>7; 10>4; 10>8).

METHODOLOGY

Subjects

A convenience sample of adults--members of community groups in the Cincinnati area--served as subjects for the research The majority of respondents were married (80%) with at least one child (60%), owned a home (77%), and had family incomes over $20,000/year (75%). They ranged in age from the 20's to the 60's and their educational level was well above average (80% attended college).

Procedure

As noted, the research was conducted with the cooperation of the Cincinnati Zoo under the auspices of recruiting volunteers to campaign for a Zoo levy to appear on the ballot in the upcoming election. An experimenter, posing as a Zoo volunteer, attended the groups' regularly scheduled meetings to show them a slide presentation (supposedly prepared by the Zoo) which would give them information about the forthcoming levy, its value and importance for the community and to solicit their help. A taped narrative accompanied the slide presentation. Questionnaires were distributed prior to the slide show. Subjects responded to questions prior to the experimental manipulation and following it. The purpose of the pre-experimental measures was to determine whether the groups were approximately equivalent regarding their prior attitudes in selected areas since groups of subjects rather than individuals were randomly assigned to treatments.

Stimulus Materials and Pretesting

The stimulus photographs used in the attractiveness manipulations were pretested to obtain an operational definition of attractiveness. Two expert judges screened the initial stimulus pool of 60 models eliminating those who were inappropriate because of their age (models between the ages of 25 and 40 were optimal). Photographs of 20 men and 20 women remained in the final pretest. They were rated on an 11-point scale (1=unattractive, 11=attractive) by 10 men and 10 women who were demographically similar to the experimental participants. Four photographs were selected (one of each gender to represent both the highly attractive and average conditions) based on their mean ratings and minimal variation in raters' evaluations. The attractive woman chosen had a mean score of 9.7 (0.8 standard deviation) while her average counterpart was rated at 5.2 (1.7 standard deviations). The attractive and average men had mean scores of 8.5 (1.4 standard deviations) and 5.3 (1.8 standard deviations), respectively.

Manipulations

The study involved 179 adult participants, 111 males and 68 females. Since subjects were approached in groups, it was not possible to balance the number of subjects in each treatment condition. A minimum of twelve individuals was exposed to any one treatment.

The hypotheses were tested in a 2x2 randomized block design in which the models' level of attractiveness (highly attractive/average) and gender were manipulated while the gender of the perceiver served as the blocking variable. The physical attractiveness manipulation consisted of exposing some of the subjects to an attractive model and the rest to an average-looking model, both depicted as zoo volunteers acting out the types of activities in which respondents could engage should they make the decision to help the zoo. In the gender manipulation, some of the experimental subjects were exposed to a model of the same gender while the rest viewed a model of the opposite gender. In the control condition, subjects (male and female) saw a presentation without a model.

Subjects were given a questionnaire prior to the slide show so that attitudinal differences between groups could be assessed (since groups rather than individuals were randomly assigned to treatments). Three questions were included specifically to test for these between-group differences and subjects responded to these same three measures both before and after the slide show. In particular. subjects were asked how they felt about tax support for the Cincinnati Zoo, if they felt the Zoo had sufficient funds with which to work, and how they felt about the passage of a Zoo levy in the upcoming election. Subjects responded in each case on a 7-point scale.

Following the slide show, subjects were asked for their perceptions of the presentation and of the model as well as for varying degrees of compliance. First, subjects judged the slide presentation on eight pairs of adjective characteristics appearing as 7-point bipolar scales. These measures were factor-analyzed later and were designed to tap two of the three components of attitudes: cognitive and affective. In addition, subjects were asked for their "overall reaction to the slide presentation" on 7-point scales ranging from "favorable" to "unfavorable."

Reactions to the model were assessed on 46 bipolar adjective pairs (which were selected from experiments reported in the literature). They were designed to examine perceptions subjects held of the models with regard to their personality characteristics and effectiveness and credibility as communicators--to assess the stereotypic perceptions and expectations individuals are thought to form of others based on physical appearance and to provide potential explanations for the other measures. Subjects were also asked for their "overall reaction to the volunteer pictured in the slide presentation." Their response was recorded once again on a 7-point bipolar adjective scale ranging from "favorable" to "unfavorable."

Three measures of verbal compliance were sought. First, subjects were asked if they would be interested in attending one of the volunteer meetings the Zoo had arranged for organizational purposes (as mentioned in the slide show). Subjects recorded their response on a 7-point bipolar scale ranging from "yes-definitely" to "no-definitely not." Second, the meeting dates and times were included on the last page of the questionnaire and subjects were instructed in the slide presentation to tear off this page and take it home with them if they might be interested in attending a meeting. Whether subjects tore the page off was a secondary measure of compliance. Last, subjects were given the opportunity to leave their phone number in a space provided on the questionnaire in the event that a meeting would be canceled and rescheduled. This was the third and highest level of comPliance measured.

RESULTS

Preliminary Considerations

Given the nature of the research design, in which groups rather than individuals were assigned randomly to treatments, it was necessary to conclude that the groups participating were not significantly different from one another. The pre-post measures were included in the design to determine whether the treatment groups were equivalent prior to the treatment manipulations with regard to individuals' attitudes toward tax support for the Zoo, its adequacy of funds, and toward support of the levy. A one-way analysis-of-variance was done to test for prior attitudinal differences between groups. No two groups were found to be significantly different at the 0.05 level by the Scheffe multiple-comparison test for any of the three variables.

Although not reported in this study (for the sake of brevity), an additional attempt was made to control for individual sources of variation. In particular, a covariance analysis was conducted to determine the extent to which demographic variables affected responses. None of the covariates, however, were significant and thus subjects' responses were attributed to the treatment manipulations. The demographic characteristics of treatment groups are available from the authors.

Prior to the experimental analysis, it also was necessary to determine the underlying dimensions in each of the two sets of bipolar adjectives in which respondents expressed their feelings toward the slide presentation and the model (when featured). In order to reduce the data to a manageable level, a principal components analysis with varimax rotation was used. The number of dimensions was ascertained by a consideration of eigenvalues, screes, and factor interpretability.

The adjectives describing the narrative were viewed in terms of two underlying factors, an affective and a cognitive component of attitudes. The affective component accounted for 49.9 percent of the variance (eigenvalue of 3.9) and was composed of the adjective pairs impressive/ unimpressive, interesting/boring, and forceful/forceless. The second factor, the cognitive component, accounted for 13.2 percent of the variance (eigenvalue of 1.05) and consisted of the adjective pairs believable/unbelievable, informative/uninformative, and logical/illogical. Together, the two factors explained 63.1 percent of the variance.

Respondents' impressions of the models can be understood through three factors--"trust," "credibility," and "sociability." These factors accounted for 45.7 percent (eigenvalue of 21.04), 4.7 percent (eigenvalue of 2.1), and 4 percent (eigenvalue of 1.8) of the variance, respectively. "Trust" consists of the adjective pairs likeable/dislikeable, trustworthy/untrustworthy, honest/dishonest, friendly unfriendly, pleasant/unpleasant, agreeable/disagreeable, sincere/insincere, kind/unkind. "Credibility" is composed of expert/inexpert, experienced/inexperienced, authoritative/unauthoritative, assertive/submissive, qualified/unqualified, and informed/uninformed. Lastly, "sociability" encompasses happy/sad, sociable/unsociable, socially adept/ socially awkward, appealing/unappealing, poised/awkward, agreeable/argumentative, and industrious/lazy. In total, the cumuLative variance explained by these three factors was 54.5 percent. Five additional factors were extracted (with eigenvalues greater than 1.0) but were not used in further analysis because their total contribution to explained variance was a mere 13 percent and none of them was interpretable.

Manipulation Check

Although the pretest results confirmed a satisfactory physical-attractiveness manipulation, a manipulation check was included to ensure that the attractiveness of the models was perceived similarly in the actual experiment. Subjects were asked to rate the relative attractiveness of the model they viewed on a 7-point scale (l=attractive, 7=unattractive) while they were rating the model on the 45 other characteristics. Contrary to expectations, the only significant difference between attractive and average models was among female perceivers viewing the male models (Table 1). While the attractive and average-looking female models were not rated significantly different from one another, results would later indicate differential reactions to these models.

TABLE 1

MEAN ATTRACTIVENESS RATINGS FOR MODELS BY SUBJECTS' GENDER

Hypothesis 1 (Attractive vs. Average-Looking Models)

The first hypothesis examined the effectiveness of the attractiveness manipulation and proposed that attractive male and female models would generate higher mean ratings across the criterion measures than their average-looking counterparts among all subjects. As Table 2 indicates, males had significantly more favorable reactions to the attractive rather than average female, while female subjects tended to react more favorably to the attractive male rather than his average counterpart across selective measures. In particular, males rated the attractive female model significantly greater overall than the average female (although the difference was marginal p<.10) and the attractive female generated more support for the levy (p<.01) and meeting interest (p<.05) than the average female model among male perceivers. On the other hand, female perceivers reacted significantly more favorably overall to the model (p<.05) and to the presentation (p<.05) when viewing the attractive male volunteer as opposed to the average male. Relatively no differences among criterion measures were found when perceivers viewed models of the same gender.

TABLE 2

ATTRACTIVE VS. AVERAGE-LOOKING MODELS (HYPOTHESIS 1)

Hypothesis 2 (Cross-Gender Perceptions of Attractive Models)

The second hypothesis proposed that among perceivers viewing attractive models, females would react more positively to the presentation when the model was male than when of the same gender and vice versa for males. It is testing for gender preferences in models while holding attractiveness constant. Results partially supported the hypothesis (Table 3). It was supported among male perceivers but not among female perceivers. All subjects tended to favor viewing the attractive female model over the attractive male model, although the preference was stronger for male subjects than for female subjects. Female subjects were significantly more likely to be interested in attending a meeting (p<.05) and to tear off the schedule of meeting dates (p<.05) when viewing the attractive female than when viewing the attractive male. Similarly, males had the same reaction to these measures and in addition, they were more likely to favor passage of the levy (pe.05), exhibited greater trust for the model (p<.10) and rated the model more favorably overall (p<.01) when it was an attractive female rather than an attractive male.

TABLE 3

CROSS-GENDER PERCEPTIONS OF ATTRACTIVE MODELS (H2) AND AVERAGE-LOOKING MODELS (H3)

Hypothesis 3 (Cross-Gender Perceptions of Average-Looking Models)

The third hypothesis proposed that when subjects viewed average-looking models, they would react less positively to models of the same gender than models of the opposite gender. The results, once again partially support the hypothesis as shown in Table 3. Unexpectedly, female subjects were more interested in attending a meeting (p<.05) and were more likely to leave their phone number (p<.05) when viewing the presentation with the average female model than with the average male model. They also reacted more favorably to the cognitive aspects of the slide presentation (px .05) and perceived the model as more trustworthy (p<.05), and sociable (p<.10), as well as reacting more positively to the model overall when that model was an average female rather than an average male.

In support of the hypothesis, males reacted more favorably to the average female than to the average male model. Their affective and overall reaction to the slide presentation (p<.10) was greater when viewing the average female and she was perceived as more trustworthy, credible, and sociable (p<.01) than the average-looking male. Among male perceivers, there were no significant differences in the compliance measures. however.

Hypothesis 4 (Male vs. Female Perception of Attractive Models)

Hypothesis four asserted that when an attractive female model was pictured, she would be responded to more favorably by men than by women and vice versa when an attractive male model was pictured. Unexpectedly, male and female subjects did not significantly differ from one another in their reactions to the attractive female model (Table 4). Women participants did have more favorable overall reactions to the attractive male model and perceived him as more trustworthy than did male participants, results supportive of the hypothesis (p<.01 and p<.05, respectively).

TABLE 4

MALE VS. FEMALE PERCEPTIONS OF ATTRACTIVE MODELS (HYPOTHESIS 4)

Hypotheses 5 and 6 (Models vs. No Models)

In terms of the effectiveness of attractive and average models in a presentation relative to an identical presentation without a model, hypotheses Five and Six have suggested that attractive models of either gender would generate more favorable reactions while average models of either gender would prompt less favorable reactions than the presentation alone (the control group). Hypothesis five received stronger support among male perceivers than among female perceivers while hypothesis six was supported only by female perceivers (Table 5). Male perceivers were more likely to favor passage of the levy when exposed to the presentation with the attractive female model than the version without a model (p<.01). They also judged the slide presentation more favorable overall and relative to its affective components, although the differences between the two versions was marginal (p<.10). No significant differences were found between the control presentation and the versions portraying average models of each gender as rated by males. In one instance, hypothesis five was refuted. Female participants' overall reaction to the slide presentation was greater when no model was present than when viewing the version with the attractive male model (p<.01).

TABLE 5

MODELS VS. NO MODELS (HYPOTHESES 5 AND 6)

Among subjects viewing the average male and female models, females were significantly more likely to assign higher affective, cognitive, and overall ratings to the presentation without a model than the version featuring an average male model (p<.01, .05, .01, respectively). In the same instance, they also were marginally more likely to be interested in attending an organizational meeting (p<.10). Females' overall reaction to the slide presentation also was greater overall when exposed to the control presentation than the presentation featuring the average female model.

DISCUSSION

Although subjects did not rate the attractive and average-looking models significantly different in all cases with regard to physical attractiveness, they did respond differentially to the attractive and average-looking models, both in their perceptions of the slide presentation and the mode: and in verbal compliance measures. In addition, male and female perceivers displayed different sensitivities to the attractiveness dimension which also surfaced in cross-gender preferences. For example, females' overall reaction to the slide presentation and to the model were more positive when they viewed an attractive male model than an average male model; however, their attitudes toward the advocated issue and compliance were not differentially affected. Males, on the other hand, were significantly more interested in attending a meeting and favored passage of the levy when they viewed an attractive female model than when exposed to the average female model; thus, their verbal compliance and attitudes were affected. Neither males nor females were apt to judge models of their own gender differentially. Consequently. it appears that the dominance of attractive femAle models in advertising is well founded when attempting to reach a male market, but in the present study, little support for the efficacy of an opposite (male models for female audiences) strategy was found.

In terms of general model effectiveness, attractive models were perceived in a more favorable light than average-looking models and attractive female models generated greater verbal compliance than did attractive males among all subjects--but especially among males. When average models alone were compared, the female model was more effective than the male model. She was trusted more than her male counterpart by all subjects, thought to be more credible by males and generated more meeting interest among females.

Similar results were found when evaluating the effectiveness of using models in the slide presentation versus not using them. Women respondents reacted no more positively to the slide presentation when attractive models were featured than when no model was featured, but they did react less positively to the presentation when an average male was featured than when no model was featured. Men did not respond differentially to the control than to the average models but they did respond significantly more favorably to the attractive female model. She was more persuasive than the message alone in prompting males to favor passage of the levy.

In conclusion, advertisers attempting to persuade a male audience may be most effective when using an attractive female model than an attractive male model, average models, or advertisements void of a model. Among females, however, results are not supportive of a similar cross-gender strategy, utilizing male models. In addition, results suggest that the use of average male models be avoided by advertisers, especially when attempting to persuade women.

In terms of central versus peripheral routes to persuasion, it is apparent that the peripheral route is effective, but that its efficacy is greater among male audiences. Characteristics beyond the message itself appears to be more important to male perceivers than to females. Physical attractiveness is only one example of this peripheral route but deserves further investigation.

The present study measures global impressions of physical attractiveness and investigates perceivers' impressions in a limited "product" and situational context. While it does not provide conclusive results with regard to the use of models in advertising or to cross-gender preferences for models based on their attractiveness, it is an initial step toward understanding these preferences and their apparent efficacy.

REFERENCES

Baker, Michael J. and Gilbert A. Churchill (1977),"The Impact of Physically Attractive Models on Advertising Evaluations," Journal of Marketing Research, 14, 538-555.

Berscheid, Ellen, Karen Dion, Elaine Walster, and G. William Walster (1971),"Physical Attractiveness and Dating Choice: A Test of the Matching Hypothesis," Journal of Experimental Social Psychology, 7, 173-189.

Cavior, Norman and Paul R. Dokecki (1971), "Physical Attractiveness Self-Concept: A Test of Mead's Hypothesis,' in Proceedings of the 79th Annual Convention of the American Psychological Association, Vol. 6, 319-320.

Chaiken, Shelly (1979), "Communicator Physical Attractiveness and Persuasion," Journal of Personality and Social Psychology, 37, 1387-1397.

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

Dion, Karen K. and Steven Stein (1978> "Physical Attractiveness and Interpersonal Influence," Journal of Experimental Social Psychology, 14, 97-108.

Horai, Joann, Nicholas Naccari, and Elliot Fatoullah (1974) "The Effects of Expertise and Physical Attractiveness Upon Opinion Agreement and Liking," Sociometry, 37, 601-606.

Hudson, John U. and Lura F. Henze (1969), "Campus Values in Mate Selection: A Replication," Journal of Marriage and the Family, 31, 772-775.

Kanungo, Rabindra N. and Sam Pang (1973), "Effects of Human Models on Perceived Product Quality," Journal of Applied Psychology, 57, 172-178.

Kopera, Anthony A., Richard A. Maier, and James E. Johnson (1971), "Perception of Physical Attractiveness: The Influence of Group Interaction and Group Coaction on Ratings of the Attractiveness of Photographs of Women," in Proceedings of the 79th Annual Convention of the American Psychological Association, Vol. 6, 317-318.

Maddux, James E. and Ronald W. Rogers (1980), "Effects of Source Expertness, Physical Attractiveness and Supporting Arguments on Persuasion: A Case of Brains Over Beauty," Journal of Personality and Social Psychology, 39, 235-244.

Miller, Howard L. and W.H. Rivenbark (1970), "Sexual Differences in Physical Attractiveness as a Determinant of Heterosexual Likings," Psychological Reports, 27, 701-702.

Mills, J.S. and Elliot Aronson (1965), "Opinion Change as a Function of the Communicator's Attractiveness and Desire to Influence," Journal of Personality and Social Psychology, 1, 137-177.

Mills, J.S. and John Harvey (1972), "Opinion Change as a Function of When Information About the Communicator is Received and Whether He is Attractive or Expert," Journal of Personality and Social Psychology, 21, 52-55.

Murstein, Bernard I. (1972), "Physical Attractiveness and Marital Choice," Journal of Personality and Social Psychology, 22, 8-12.

Norman, Ross (1976), "When What is Said is Important: A Comparison of Expert and Attractive Sources," Journal of Experimental Social Psychology, 12, 294-300.

Petty, Richard E. and John T. Cacioppo (1983), in Advertising and Consumer Psychology, L. Percy and A. Woodside, eds., Lexington, MA: Lexington Books, in press.

Smith, George H., and Rayme Engel 1968), "Influence of a Female Model on Perceived Characteristics of an Automobile," in Proceedings of the 76th Annual Convention of the American Psychological Association, 681-682.

Snyder, Mark and Myron Rothbart (1971),"Communicator Attractiveness and Opinion Change," Canadian Journal of Behavioral Science, 3, 377-387.

Steadman, Major (1969), "How Sexy Illustrations Affect Brand Recall," Journal of Advertising Research, 9, 15-19.

Tesser, Abraham and Michael Brodie (1971), "A Note on the Evaluation of a Computer Date," Psychonomic Science, 23, 300.

Walster, Elaine, Vera Aronson, Darcy Abrahams, and L. Rottman (1966), "The Importance of Physical Attractiveness in Dating Behavior," Journal of Personality and Social Psychology, 4, 508-516.

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