Sex Roles: A Journal of Research, May 1997 v36 n9-10 p551(22) Advertising’s effects on men’s gender role attitudes. Jennifer Garst; Galen V. Bodenhausen. Author’s Abstract: COPYRIGHT 1997 Plenum Publishing Corporation We posited that media images of men influence the gender role attitudes that men express soon after exposure to the images. A total of 212 men (87% European American, 7% Asian or Asian American, 3% African American, and 3% other) viewed magazine advertisements containing images of men that varied in terms of how traditionally masculine vs.
androgynous they were and whether the models were the same age or much older than the viewers. Men who had initially been less traditional espoused more traditional attitudes than any other group after exposure to traditionally masculine models, although they continued to endorse relatively nontraditional views after exposure to androgynous models. These findings suggest that nontraditional men’s gender role attitudes may be rather unstable and susceptible to momentary influences such as those found in advertising. Full Text: COPYRIGHT 1997 Plenum Publishing Corporation In the average American household, the television is turned “on” for almost seven hours each day, and the typical adult or child watches two to three hours of television per day.
It is estimated that the average child sees 360,000 advertisements by the age of eighteen (Harris, 1989). Due to this extensive exposure to mass media depictions, the media’s influence on gender role attitudes has become an area of considerable interest and concern in the past quarter century. Analyses of gender portrayals have found predominantly stereotypic portrayals of dominant males and nurturant females within the contexts of advertisements (print and television), magazine fiction, newspapers, child-oriented print media, textbooks, literature, film, and popular music (Busby, 1975; Durkin, 1985a; Leppard, Ogletree, ; Wallen, 1993; Lovdal, 1989; Pearson, Turner, ; Todd-Mancillas, 1991; Rudman ; Verdi, 1993; Signorielli ; Lears, 1992). Most of the research to date on the effects of gender-role images in the media has focused primarily on the female gender role. A review of research on men in the media suggests that, except for film literature, the topic of masculinity has not been addressed adequately (Fejes, 1989). Indeed, as J. Katz (1995) recently noted, “there is a glaring absence of a thorough body of research into the power of cultural images of masculinity” (p.
133). Katz suggests that studying the impact of advertising represents a useful place to begin addressing this lacuna. Of the few existing studies, a longitudinal content analysis of nine magazines in 1959, 1969, and 1979, found that advertisements featuring men are slowly moving toward decreased gender role stereotyping.
However, in traditionally male magazines such as Esquire and Field ; Stream, the proportion of advertisements depicting men in “manly” activities did not decrease as much as it did in traditionally female and general interest magazines (Skelly ; Luridstrom, 1981; see also England ; Gardner, 1983). In fact, Jacobson and Mazur (1995) posit that current advertising promotes a “masculine ideal” that encourages men to “exude an aura of physical strength, power, dominance, and detachment” and to “repress, and loath, their ‘feminine’ traits (such as vulnerability and compassion)” (p. 80). Thus, although there may be some trends toward less stereotypic images of masculinity in advertising in some print genres, the traditional, agentic man is still a ubiquitous positive cultural representation. GENDER-ROLE ATTITUDES AS TEMPORARY CONSTRUCTIONS Although media researchers have typically been interested in relatively enduring attitudinal consequences of media exposure, recent research suggests it may be more fruitful to focus on the short-term impact of viewing media depictions. Wilson and Hodges (1992; see also Schwarz ; Bless, 1992a) proposed that individuals’ attitudes can vary, depending on the current context and the information that is salient in it.
For instance, research has shown that incidental exposure to a well-liked African American celebrity, such as Oprah Winfrey or Michael Jordan, can influence subsequently reported racial attitudes (Bodenhausen, Schwarz, Bless, & Wanke, 1995). Likewise, Schwarz and Bless (1992b) have found that activating thoughts about politicians who had been involved in a scandal influenced respondents’ subsequent evaluations of politicians’ trustworthiness in general, as well as the trustworthiness of specific politicians. It is thus quite plausible that mediated exemplars can have at least a transitory impact on social beliefs and attitudes. For many relatively complex attitudinal topics, including gender role attitudes, people are likely to possess a mixture of relevant knowledge and beliefs, not all of which are mutually compatible (Wilson & Hodges, 1992; Smith, 1992). The attitude that one holds at any given time will depend in part on which subset of information is most salient at the moment. Although these context-induced attitudinal shifts may not always be of great magnitude, they indicate that attitudes are relatively dynamic entities.
Rather than simply residing in some fixed form in memory, they appear to be constructed “on the fly,” at least in part, on the basis of currently salient considerations. Shifts in attitudes would be expected to be particularly noticeable when greater ambiguity or ambivalence characterizes the issue in question. In contemporary society, messages about appropriate standards for male behavior are mixed at best. Some socializing influences emphasize the importance of traditionally masculine, agentic qualities, while others emphasize the desirability of developing the communal side of the male psyche. It is thus quite plausible that for many men, attitudes about appropriate male behavior are based on somewhat conflicting ideas and prescriptions. For this reason, these attitudes may be susceptible to momentary influences that emphasize either more traditional or less traditional images of masculinity.
Schwarz and Bless (1992a) offer an “inclusion/exclusion model” to study the context dependency of attitudinal judgments. Applied in the current context, the model implies that when viewers see media images of men, their subsequent judgments about the male role may be assimilated toward that particular media image if the viewers incorporate the image into their current understanding, or representation, of the male role. However, if the viewers exclude a certain male media image from their gender-role representations, the viewers’ judgments will either not be affected by the media image at all, or they will use the media image as a standard of comparison against which other attitudinally relevant information will be contrasted. Thus, for example, the behavior of a male exemplar who is seen as too feminine might be excluded from the mental representation a viewer forms of the male role.
Moreover, judgments of role-appropriate behavior may actually become more traditionally masculine, if this atypical exemplar is used as a standard of comparison. Whether individuals include or exclude a particular exemplar from their representations of the male role will depend on a number of factors, including the width or breadth of the individuals’ representations. As the width of viewers’ representations of the male role increases, it becomes more likely that they will assimilate available male media models into their representations of the male role. Lastly, it is important to note that media depictions will only influence viewers’ subsequent judgments if the depictions are sufficiently different from the viewers’ previously held representations. If the media representations are largely consistent with the viewers’ prior understanding of the male role, then the images should produce no impetus for viewers to change their representations (or judgments based on them). FACTORS MODERATING THE IMPACT OF MEDIA EXEMPLARS Although it appears that media images can influence viewers’ attitudes (e.g.
, Gels, Brown, Jennings (Walstedt), ; Porter, 1984), it is also clear that not all viewers are uniformly affected. Durkin (1985b), in an extensive review of television and gender-role acquisition, has criticized previous research for failing to take into account how much variability there is in how viewers respond to messages or images in the mass media. According to Durkin, most research on the impact of gender roles only looks at the overall impact of the medium but does not investigate the individual difference variables that may moderate this impact. Lending support to Durkin’s view is a study of racial attitudes showing that viewers’ preexisting attitudes moderated their reactions to a television program. In an investigation of reactions to the character of Archie Bunker in “All in the Family,” it was found that high-prejudice as compared with low-prejudice white viewers tended to identify more with the character of Archie Bunker, saw him as winning rather than as ridiculed, and saw nothing wrong with his use of racial and ethnic slurs (Vidmar ; Rokeach, 1974). Prior Male Gender Role Attitude We predict that the impact of male media models will be moderated by the male gender role attitudes of the viewers. For instance, it has been found that boys and girls with feminine preferences tend to identify more with a male counter-stereotypical portrayal than do masculine children (Eisenstock, 1984).
Although less traditional men hold less traditional attitudes, they are also part of a culture that responds positively to men displaying masculine gender role behaviors and negatively to men displaying feminine gender role behaviors. In fact, Kanner, (1990, May 21) suggests that men are not willing to give up their traditionally masculine style to become “domesticated wimps” (p. 20).
To support her argument, she cited Barbara Durham, associate director of consumer resources for a large advertising firm, as saying “roles are changing, but we found it’s important not to rob men of their masculinity” (p. 20). One possible implication of this state of affairs is that compared to men with traditional attitudes toward the male gender role, less traditional men may have rather wide representations of the male role, incorporating both traditional and nontraditional elements. As such, less traditional men’s gender role attitudes may be relatively less stable and more susceptible to momentary influences such as those found in advertising, because of the greater complexity and breadth of their attitudes.
Traditional men, on the other hand, may have relatively more narrow and sharply defined images of the male role, and they may therefore exclude from their mental representations any exemplars who do not fit this narrow conception. If so, it should be the case that the attitudes of more traditional men are less affected by momentary influences, because they are only likely to incorporate into their models of masculinity those exemplars who are already fairly consistent with their more narrow world view. Moreover, research suggests that traditional or conservative thought is associated with more authoritarian personality characteristics (Dillehay, 1978; Duncan, Peterson, & Winter, 1997) and that, in turn, authoritarian individuals respond defensively to counter-attitudinal information. Research by D. Katz (1960) indicates that ego defensiveness is a major obstacle to attitude change, and indeed, Wagman (1955) found that counter-attitudinal information had a boomerang effect of making authoritarian people even more prejudiced. Likewise, the gender schema literature suggests that highly gender-typed individuals may respond negatively to deviations from the normative standard (Ruble and Stangor, 1986).
Signorella and Liben (1984) found that highly stereotyped children recalled more traditional than nontraditional pictures and more often reconstructed nontraditional pictures as traditional rather than the reverse. Thus, for a variety of reasons it is likely that more traditional men, as compared to less traditional men, have narrower and less complex male gender role attitudes and will, therefore, be less susceptible to discrepant male media images. Alternatively, it may be the case that less traditional men have a greater resistance to momentary influences on their gender-related attitudes. Perhaps the very fact that they espouse relatively nontraditional views demonstrates that they have developed defenses against the ubiquitous but more limiting traditional prescriptions for male behavior. The current study was designed to investigate the question of whether the traditionalism of viewers’ pre-exposure attitudes is related to the likelihood that their attitudinal reports will be influenced by exposure to an advertising image.
Attractiveness of Media Models The influence of a media model may also depend on whether the exemplar captures the viewer’s attention. According to Bandura’s (1986) cognitive social learning theory, the influence of an observed model is related to the “sway of attention” (p. 54). This notion suggests that observers will tend to seek out and imitate models who are interesting and otherwise rewarding, whereas they ignore or reject models who have unattractive qualities. If androgynous individuals, who display both masculine and feminine behaviors, are perceived as more attractive than individuals high in only one category (Spence, Deaux, ; Helmreich, 1985), it might be expected on the basis of social learning theory that viewers will be most likely to imitate these media models.
However, Schwarz and Bless (1992a) argue that assimilation is dependent merely on an exemplar being categorized on a dimension relevant to the judgment being made. Thus, it could be that liking for a media model is largely irrelevant to whether men include the models in their mental representations of the male gender role. The present research investigates the issue of whether liking for a media model provides a basis for assimilating that model into one’s representations. Similarity of Media Models Similarities between the viewers and the media models are also expected to enhance the impact of media images. The attitude change and attraction literature suggests that observers tend to be attracted to models who are most similar to themselves (e.
g., Berscheid, 1985; McGuire, 1985). Similarity is thought to increase persuasion through its impact on source credibility and attractiveness. Hass’s (1981) review of the impact of source factors in persuasive messages suggests that speaker-audience similarities result in greater persuasion because they increase the attractiveness of the message source.
Likewise, Tan (1985) suggests that when many diverse characteristics are shared, receivers perceive the source to be “one of us” or “our kind of people” and thus infer that they share common needs and goals with the message source. In a similar vein, McGuire (1985) proposes that people infer that sources who differ demographically also differ ideologically from themselves. Thus, it is expected that models who are demographically similar to university students (e.
g., in early twenties versus late sixties) will be perceived as more similar and attractive and hence promote greater acceptance of the modeled behavior. In terms of the inclusion/exclusion model, dissimilar exemplars may simply not be included in viewers’ mental representation of the male gender role. Taken together, the available theory and evidence suggests that various images of men used in advertising (highly androgynous, mildly androgynous, or traditionally masculine images) can influence the gender role attitudes that men express soon after exposure to these images. However, this influence may be moderated by the gender role attitudes of the participants (more versus less traditional) and the demographic similarity of the media models to the viewers (college age versus retirement age). Several specific hypotheses were investigated: 1.
It is expected that the impact of the male advertising images will depend on the initial gender role attitudes of the male viewers. Two competing possibilities were considered. On the one hand, less traditional men’s gender role attitudes may be more affected by the momentary influences of all types of gender role portrayals than more traditional men.
This pattern is expected to occur if less traditional men have wider representations of the male gender role than more traditional men. Alternatively, less traditional men may be more resistant to the impact of traditionally masculine images than are more traditional men because less traditional men may have developed defenses against the more limiting traditional prescriptions for male behavior. 2. The present research also seeks to determine whether the tenets of social learning theory apply to the impact of male media models on viewer’s male gender role attitudes. Based on social learning theory, it is expected that male viewers’ gender role attitudes will be more affected by media models whom the viewers perceive favorably versus unfavorably. Alternatively, viewers’ liking for the media models may be largely irrelevant to the models’ influence on the viewers’ male gender role attitudes, if factors other than liking per se determine whether a model is included or excluded from gender representations. 3.
Based on research suggesting that age may serve as an important similarity or attractiveness cue, it is expected that models who are the same age as the viewers will promote greater change of the viewers’ male gender role attitudes than will older models. METHOD Participants and Design A total of 267 male undergraduates from introductory psychology classes participated in the main experiment in exchange for course credit. Responses of 55 participants were excluded from the data analysis because they could not be matched with a previously completed screening survey. Thus, a total of 212 students participated, ranging in age from 17 to 46 years, with a mean age of 19.
4 years, and whose ethnic composition was 87% European American, 7% Asian or Asian American, 3% African American, and 3% other. More traditional and less traditional participants were randomly assigned to one of six experimental conditions, comprising a 3 (type of media portrayal: highly androgynous, mildly androgynous, and traditionally masculine) x 2 (age of media model: young versus old) x 2 (initial gender role attitudes: more traditional versus less traditional) between-subjects factorial design. Materials Stimulus Magazine Advertisements. Stimulus materials were professional quality, color, full page mock magazine advertisements that pictured a “typical” male product user along with his written biographical profile. The quality and format of the stimulus materials were modeled from Dewar’sR scotch advertisements that were circulating in magazines at the time the study was run.
Advertisements featuring two different products were developed for each of six conditions. For each condition, one advertisement pictured a cup of coffee being filled with “Delvecchio’s gourmet coffee” (Scenario 1), while the other featured a computer represented to be an “EPS Technologies personal computer” (Scenario 2). Both advertisements listed the typical male product user’s name (e.g., John Hadon), home (e.g.
, Chicago, Illinois), hobbies (e.g., enjoys going to concerts), recent achievement (e.g., received a service award from the social club he belongs to), name of the product he uses (e.g.
, Delvecchio’s gourmet coffee), outlook on relationships, philosophy regarding marriage, age, and profession. For each product advertisement, the content of the model’s “outlook on relationships” as well as “philosophy regarding marriage” varied across the highly androgynous, mildly androgynous, and traditionally masculine media portrayals. Four masculine and/or feminine interests/role behaviors were used to form both the outlook on relationships and philosophy regarding marriage categories. Each behavior or interest had been previously identified by Orlofsky, Ramsden, and Cohen (1982) in the Revised Sex-Role Behavior Scale (SRBS-2) as more typical of one gender but considered desirable by both genders.
The age similarity of the male media models, compared to the participants, was also varied. Half the participants viewed advertisements that pictured a young, male college student who was represented to be either 18 or 20 years of age. The other half examined advertisements that pictured an older, male retiree represented as either 65 or 66 years old. Photos of both the younger and older targets for each scenario were matched on perceived level of masculinity in previous pilot testing.
Attributes other than outlook on relationships, philosophy regarding marriage, age, and profession were held constant within each product advertisement. Based on pilot testing, the common text of both scenarios across the six experimental conditions was matched based on perceptions of the models’ similarity to the viewer, attractiveness, masculinity, femininity, and power. Gender Attitude Inventory (GAI). A total of thirty-four items were selected from the 109 items of the Gender Attitudes Inventory (Ashmore, Del Boca, & Bilder, 1995) to measure participants’ attitudes toward relationships between men and women and the roles played by men in contemporary American society.(3) Responses were assessed using a 7-point scale with “1” signifying “agree strongly” and “7” signifying “disagree strongly.
” Items were worded so that agreement sometimes reflected traditional attitudes and sometimes reflected less traditional attitudes. The items were recoded so that higher values always indicated more liberal (i.e., less traditional) attitudes. Items from six of the fourteen subscales identified by Ashmore et al.
were used: (1) sexual initiative (e.g., the initiative in dating should come from the man); (2) female superiority (e.g.
, women are morally superior to men); (3) traditional stereotypes (e.g., men are more competitive than women); (4) chivalry (e.g., chivalrous gestures toward women on the part of men should be encouraged); (5) family roles (e.
g., the husband should have primary responsibility for the support of the family); and (6) differential work roles (e.g.
, all occupations should be equally accessible to both men and women). Model Perception Scales (MPS). Participants’ perceptions of the men portrayed in the magazine advertisements were measured by nine 5-point questions. The measures focused on participants’ liking for each model (e.
g., how much do you think you would like the person in the advertisement?), perceived similarity to each model (e.g., how similar do you perceive the person in the advertisement to be to you?), and the social attractiveness of each model (e.
g., how well liked do you think this person is by others?). In addition, a series of ten 5-point scale items anchored by not at all x to very x, were used to assess participants’ perceptions of each model on a number of unique dimensions (e.g.
, masculine, feminine, and similar to me). Procedure Screening. A total of 738 male introductory Psychology students were given the thirty-four item subset of the GAI during the first week of the semester. The participants were divided by using a median split of their average GAI scores (Mdn = 4.5 on 7-point scale); students in the upper half were classified as “less traditional,” while those in the lower half were classified as “more traditional.” The GAI scores showed acceptable reliability (Cronbach’s Alpha = 0.
85). Main Experiment. Participants reported to the laboratory for a study of magazine advertising. No mention was made of the screening questionnaire they had completed more than a month beforehand. Each session included groups of up to 15 males with male and/or female experimenters. After being greeted by the first experimenter, participants were told that due to the short length of the study, they would also be asked to fill out a questionnaire needed as part of a graduate student’s dissertation. The experimenter then explained that the first study concerned magazine advertising and that they would be asked to look over and provide their perceptions of and reactions to two advertisements.
Participants were then told that the purpose of the study was to determine which features of magazine advertisements most effectively promoted purchase intentions. Lastly, participants were asked to read and sign a consent form and to begin going through the experimental booklets. Each participant received an experimental booklet and a folder containing a set of two color stimulus magazine advertisements (Scenario 1 and 2) from one of the six experimental conditions. Thus, each folder contained two advertisements with either highly androgynous, mildly androgynous, or traditionally masculine biographical profiles of models whose ages were either young or old.
The order of the advertisements within each folder and the random assignment of the six experimental folders was accomplished by use of a pre-established randomized schedule. In an attempt to draw attention to the media models and amplify the models’ impact on viewers’ attitudes, participants filled out the Model Perception Scale items after viewing each advertisement. In addition, in order to hide the true intent of the experiment, participants were also asked to answer questions assessing their purchase intention of each product, their reactions to each product, and their reactions to each advertisement. Lastly, participants completed standard demographic items. When all participants had finished the questionnaire, the experimenter thanked them for participating in the advertising study, introduced them to a different experimenter who was in charge of the ostensibly unrelated second study, and then left the experiment room.
The second experimenter(4) began by thanking the participants for helping the graduate student with her study and explained that the second questionnaire concerned basic attitudes and values currently held by college students. The experimenter administered new consent forms after which he/she handed out a questionnaire that used a format and type face that differed from that used in the “first” study. The new questionnaire contained the same subset of the GAI items previously completed by the participants during the first week of the semester as a screening questionnaire. The GAI questions were interspersed with 21 gender unrelated attitude measurement items. After the participants completed the questionnaire, they were given an educational debriefing, thanked, and excused. RESULTS Manipulation Checks To determine whether the participants thought that there was a continuum of increasing masculinity and correspondent decreasing femininity across the highly androgynous, mildly androgynous, and traditionally masculine media portrayals, and to ascertain whether participants perceived the younger models to be more similar to themselves than the older models, repeated-measures analyses of variance (ANOVAs) were used. Specifically, participants’ perceptions of the masculinity, femininity, and the similarity of the two media portrayals, as measured by single-item 5-point scales, were examined as a function of the type of media portrayal, age of the models, and the original gender attitudes held by the participants.
As expected, participants differentiated between the types of media portrayals as a function of the perceived femininity of the models, F(2, 199) = 10.28, p less than .001 (see Table I).
Highly androgynous media portrayals were viewed as more feminine than the mildly androgynous portrayals, F(1, 140) = 6.23, p less than .05, and the mildly androgynous portrayals, in turn, were viewed as significantly more feminine than the traditionally masculine portrayals, F(1, 139) = 3.96, p less than .
05. Overall, participants did not differentiate among the three types of media portrayals in terms of the perceived masculinity of the models, a separate dimension from the models’ perceived femininity, F(2, 200) = 1.40, p greater than .15. However, there was a two-way interaction between the age of the models and participants’ original gender role attitudes, F(1, 200) = 8.03, p less than .
01 (see Table II, top section). Whereas less traditional participants did not differentiate between the masculinity of younger and older models, F(1, 108) = 2.24, p greater than .10, more traditional participants perceived the younger models to be significantly more masculine than they perceived the older models to be, F(1, 100) = 5.50, p less than .05. As expected, participants perceived the younger models to be more similar to themselves than the older models, F(1, 200) = 5.
67, p less than .05 (see Table II, middle section). However, this difference was qualified by an unexpected interaction between the age of the models and the original gender role attitudes of the participants, F(1, 200) = 6.
94, p less than .01. Although more traditional participants perceived the younger models to be significantly more similar to themselves than the older models, F(1, 100) = 15.
50, p less than .001, less traditional participants did not differentiate between the younger and older models’ similarity to themselves, F less than 1, ns. It was also found that participants differentiated between the highly androgynous, mildly androgynous, and traditionally masculine portrayals in terms of the perceived similarity of the media models to themselves, F(2, 200) = 3.03, p less than .05 (see Table I). Participants rated the traditionally masculine portrayals as significantly less similar to themselves than the highly androgynous portrayals and as marginally less similar than the mildly androgynous portrayals, TABULAR DATA FOR TABLE I OMITTED F(1, 138) = 5.51, p less than .
05 and F(1, 140) = 3.54, p less than .07, respectively. There was no difference between the perceived similarity of the highly androgynous and mildly androgynous portrayals, F less than 1, ns.(5) Table II. Mean Perceptions of Media Models’ Masculinity,Similarity, and Overall Perceptions by Age of Media Model andOriginal Gender Role AttitudeDependent Measure:Age of Media ModelInitial Gender RoleAttitudeYounger Older p less thanMasculinityLess traditional 3.
31 (0.67)3.52 (0.
75)nsMore traditional 3.44 (0.69)3.06 (0.94).
05SimilarityLess traditional 2.47 (0.84)2.49 (1.11)nsMore traditional 2.82 (0.81)2.17 (0.84) .001Overall perceptions (MPS Composite)Less traditional 2.80 (0.68)2.95 (0.80)nsMore traditional 3.02 (0.59)2.79 (0.69).07a Sample sizes range from 49 to 56 across condition; standarddeviations appear in parentheses following each mean. Significanceof the simple main effect of model age is indicated in the thirdcolumn.To ensure that the advertisements were perceived as being of equal quality across experimental conditions, composite measures of participants’ intentions to purchase each product (3 items), reactions to each product (3 items), and their reactions to each advertisement (3 items) were examined. The reliabilities for all composites for each scenario were acceptable, with Cronbach’s Alphas ranging from 0.74. to 0.88. Repeated-measures analyses of variance (ANOVA) revealed that participants’ perceptions of the advertisements did not vary systematically across the manipulated conditions (all ps = ns). The validity of using a single model perception composite measure for the 9-item model perception scale (MPS) was verified by a principle-axis factor analysis (retaining eigenvalues greater than 1, oblique rotation) of reactions to each of the two scenarios separately. The obtained component for each scenario focused on participants’ liking for the model, perceived similarity to the model, and the social attractiveness of the model. Reliabilities of the composites were high, with Cronbach’s Alpha = 0.91 for the advertisement for Delvecchio’s coffee (Scenario 1) and Cronbach’s Alpha = 0.90 for the computer advertisement (Scenario 2). A higher average score value on the MPS composite indicates that the participants perceived the models more favorably compared to those who had lower scores on the average MPS composite. Overall Perceptions of Media Models A question of the present study is to determine whether favorable perceptions of media models will increase the likelihood of participants including the modeled behaviors and interests in their representations of the male role (Hypotheses 2 & 3). In order to assess whether all participants, regardless of their original male gender role attitudes, perceived the androgynous versus masculine and younger versus older media models more favorably (i.e., they liked the models more, found them more similar to themselves, and found them more attractive), participants’ average MPS composites for both of the advertisements were examined using repeated-measures analysis of variance (ANOVA). Participants’ overall perceptions of the models, as measured by the MPS composites, varied as a function of the type of media portrayal viewed, F(2, 196) = 12.58, p less than .001. Participants reported more favorable perceptions of the highly androgynous (M = 3.13, SD = 0.63) and mildly androgynous media portrayals (M = 2.98, SD = 0.70) than the traditionally masculine portrayals (M = 2.58, SD = 0.66), F(1, 136) = 24.88, p less than .001, and F(1, 137) = 11.89, p less than .001, respectively. Perceptions of the highly androgynous and mildly androgynous portrayals did not differ significantly, F(1, 137) = 1.71, p greater than .15. The effect of the type of media portrayals viewed was not qualified by an interaction with the male gender role attitude originally held by the participants, F less than 1, ns. Contrary to the thought that male participants would perceive the younger models more favorably than the older models, participants’ overall perceptions of the media models did not vary as a function of the models’ ages, F less than 1, ns. However, the male gender role attitudes of the participants did interact with the age of the models in a pattern close to the one found when participants’ ratings of the models’ similarity were examined, F(1, 196) = 4.23, p less than .05 (see Table II, bottom section). Whereas overall perceptions of the younger versus older models did not differ among less traditional participants, F(1, 104) = 1.03, p greater than .15, more traditional participants perceived the younger models somewhat more favorably than the older models, F(1, 100) = 3.45, p less than .07. Effects of Male Media Models on Subsequent Male Gender Role Attitudes A central question of the present investigation was to determine whether male media images can have an immediate effect on men’s gender role attitudes and whether such effects are moderated by the traditionalism of participants’ male gender role attitudes and the demographic similarity of the media models to the viewers. In order to determine whether participants’ male gender role attitudes were affected by viewing the media models, participants’ attitudes immediately after viewing the advertisements were analyzed using analysis of co-variance (ANCOVA). The average Gender Attitude Inventory scores were examined as a function of type of original gender role attitude held (more traditional versus less traditional), type of media portrayal (highly androgynous, mildly androgynous, or traditionally masculine), and age of the media models (young versus old), with the participants’ original male gender role attitudes covaried out. Not surprisingly, original male gender role attitude was a significant covariate (p less than .001) of the gender role attitudes they expressed after they had viewed the two advertisements. As expected, the impact of advertising images of men was moderated by the male gender role attitudes of the viewers. The average posttest GAI scores of more versus less traditional participants were differentially influenced by the type of media portrayals they viewed, F(2, 199) = 4.31, p less than .05 (see Table III).(6) While the difference in less traditional participants’ responses to the three different portrayals was significant, F(2, 107) = 4.80, p less than .01, there was only a marginal difference in more traditional participants’ responses to the media portrayals, F(2, 99) = 2.37, p less than .10. Thus, less traditional participants were indeed more susceptible to media influences than their more traditional counterparts. Compared to those respondents who viewed the highly and mildly androgynous portrayals, less traditional participants who viewed the traditionally masculine portrayals endorsed more traditional male gender role attitudes, F(1, 72) = 7.97, p less than .01, and F(1, 70) = 6.62, p less than .05, respectively. The average adjusted GAI scores for less traditional participants who saw the highly androgynous and mildly androgynous media portrayals did not differ from one another, TABULAR DATA FOR TABLE III OMITTED F less than 1, ns. Contrary to the possibility that less traditional men may have developed defenses against traditional prescriptions for male behaviors, less traditional men’s gender role attitudes became more traditional, as compared to more traditional men’s attitudes, after viewing traditionally masculine male images, F(1, 68) = 5.29, p less than .05. Although the influence of the advertising content on the post-exposure gender role attitudes of the more traditional respondents was only marginally significant overall, it may be useful to examine, somewhat tentatively, the pattern of responses seen among these individuals. The more traditional participants who saw the mildly androgynous portrayals displayed significantly greater traditionalism than those who saw the highly androgynous portrayals, F(1, 66) = 4.12, p less than .05, and revealed marginally greater traditionalism relative to those who viewed the traditionally masculine portrayals, F(1, 68) = 3.58, p less than .07. There were no differences between more traditional participants who saw the media portrayals that were gender-typed as highly androgynous and traditionally masculine, F less than 1, ns. Contrary to the tenets of social learning theory, it appears that viewers’ liking for the highly and mildly androgynous media models may be unrelated to these models’ influence. Although there was a nonsignificant trend for type of media portrayal, F(2, 199) = 2.78, p less than .07, this was qualified by participants’ original gender role attitudes, F(2, 199) = 4.31, p less than .05 (see Table III). Likewise, contrary to Hypothesis 3, it appears that even though the age of the media models tended to serve as a similarity and attractiveness cue for more traditional participants, more versus less traditional participants were not more likely to emulate the behaviors and interests of younger versus older models, F less than 1, ns. DISCUSSION It appears that men’s gender role attitudes can be influenced by the images of men they regularly see in the mass media. Rather than being fixed in memory, it seems that gender role attitudes are dynamic entities that are susceptible to momentary influences that emphasize either more or less traditional images of masculinity. However, it also appears evident that the specific impact that media images have is dependent on how traditional the male viewers are and the type of gender role behaviors and interests to which they are exposed. Contrary to the thought that less traditional men may have developed defenses against the more limiting traditional prescriptions for male behavior, the current investigation suggests that even less traditional men are not immune, at least in the short-term, to the influence of images that reinforce the cultural norm of strong, highly masculine men. Granted that the influence of traditionally masculine images may only be short-term, this pattern still makes one wonder about the malleability of less traditional men’s gender role attitudes, particularly given that men are generally portrayed in a stereotypic fashion in the mass media (e.g., Busby, 1975; Pearson et al., 1991). Wilson and Hodges (1992) have posited that when people hold complex attitudes, as less traditional men are likely to do, the most salient part of the attitude will have the most influence when a choice is made. Although less traditional men hold less traditional attitudes, they are also part of a culture that places a high value on masculine behaviors and interests and a somewhat questionable value on men displaying feminine behaviors and interests. Since less traditional men’s representations of men are rather wide, including both traditional and nontraditional elements, less traditional men are susceptible to images that make more traditional elements of their attitudes salient. The salience of less traditional images, on the other hand, fails to induce change. This lack of influence may occur because the media representations are not sufficiently different from less traditional men’s preexisting attitudes or even that less traditional men are less sensitive to nontraditional masculine images that are largely unsupported, if not discouraged, by strong cultural norms. More traditional men, on the other hand, appear less susceptible to the influence of media images of men. Unlike less traditional men, more traditional men did not fully incorporate any of the male portrayals into their representations. In fact, if anything, it appears that more traditional men react against media models who do not fit into their rather narrow pre-established male gender role representations (e.g., Dillehay, 1978; D. Katz, 1960; Ruble & Stangor, 1986; Wagman, 1955). Since more traditional participants may not be able to classify the intermediate, mildly androgynous models as prototypes of either traditional, “manly” masculine men or “with it” highly androgynous “modern” men, they may feel uncertain or uncomfortable about assimilating the behaviors and interests expressed by the male models into their own lives. This uncertainty may cause them to react in a more traditional manner since more traditional attitudes are probably more safe and comfortable to the more traditional participants than are less traditional attitudes. In the same vein, more traditional participants may feel more comfortable expressing less traditional attitudes when they view media models that can be classified and understood to represent a specific type of masculine role (i.e., masculine or highly androgynous). Even though social learning theory suggests that observers tend to seek out and emulate models who are interesting and rewarding and reject or ignore models who are not, it appears that this straightforward hypothesis does not provide an adequate explanation for the influence of media depictions. Though the men in this study were more favorably disposed toward the highly and mildly androgynous media models, there was not a consistent tendency to emulate the behaviors and interests displayed by them. In fact, the traditionally masculine images appeared to reinforce masculine cultural norms for less traditional participants while more traditional men appeared to reject or respond defensively to images that did not fit their definitions of what are appropriate role behaviors and interests for men. In sum, the argument that people emulate media portrayals that they find attractive may be too simplistic within the domain of men’s gender role attitudes. Furthermore, it appears that reported affinity to a media model is largely irrelevant to the manner in which men categorize media images and, thus, has little influence on the likelihood of men including the models into their representations. This study also reinforces the idea proposed by others (e.g., Hass, 1981; Simons, Berkowitz, & Moyer, 1970) that the demographic similarity-persuasion link is not straightforward. First, the college age participants did not uniformly perceive the younger models to be more likable, similar, and attractive than the older models; rather, only the participants whose gender role attitudes were more traditional showed a somewhat more positive response to the younger model. Furthermore, in another instance in which liking for a media model is largely irrelevant to the models’ influence, even though more traditional participants responded more favorably to the younger versus old media models, their male gender role attitudes were not significantly influenced by the age of the models. The results of the present research suggest that more traditional participants have a greater sensitivity to issues of age and, more specifically, a bias toward youth, or ageism, than less traditional participants. In fact, it was only the more traditional participants who perceived the younger models to be more masculine, similar to themselves, and somewhat more attractive than the older models. Research has found that traditionalism is tied to authoritarian personality characteristics and that, in turn, authoritarianism is tied with prejudiced attitudes toward many social groups (Dillehay, 1978). Thus, it is tempting to speculate that more traditional individuals may also tend to be more ageist than less traditional individuals. The short-term influence of media depictions is especially interesting considering the prevalence of largely stereotypic media images in our daily lives (e.g., Busby, 1975; Durkin, 1985a; Lovdal, 1989; Pearson et al., 1991; Signorielli & Lears, 1992). Although the present research only examined the short-term influence of media images, it is interesting to speculate whether repeated exposure to traditional images will, over the long term, translate to stable attitude change for less traditional men. This is an even more pertinent question given that media images are continuing to invade more and more facets of our lives. Televisions now play in airports and in schools, advertisements precede video and cinema movie presentations, and audio advertisements now even play while listeners are “on hold” on the telephone. In addition, even short-term attitude change can have important implications for human behavior. For example, for a period of time after exposure to images reinforcing traditional masculine behavior, even less traditional men may revert to more dominating, sexist reactions toward women. This possibility represents an especially important avenue for future research. The overall implication of this study is that male gender role attitudes are complex and that it is difficult to change these attitudes in all men using the same methods. In line with Durkin’s (1985b) argument that individual difference variables moderate the impact of mediated models on gender role attitudes, the present research suggests that the influence of male media images is contingent on viewers’ prior gender role attitudes. While it appears that less traditional men’s attitudes are susceptible to male media images, the attitudes of more traditional men, on the other hand, are less flexible, causing them to set up strict boundaries or definitions of what are appropriate male interests and behaviors. Anything that conflicts or contradicts this definition may be rejected. Thus, the present research casts doubts on the feasibility of creating dramatic changes in more traditional men’s gender role attitudes through mass media channels because these men appear to reject images that do not fit their frameworks for appropriate male roles. In fact, it is interesting to speculate whether more traditional men’s gender role attitudes may be more stable than less traditional men just because they do tend to reject more images of masculine behaviors and interests. Another factor that may be important to the influence of images of masculine behaviors and interests in advertising is the product that is associated with the masculine images. In the present research the products being advertised, gourmet coffee and a personal computer, are upscale products that are not generally considered either highly masculine or highly feminine products. It is interesting to consider whether men’s perceptions of the media models would differ and even whether the effect found in the current research would have been different if the products associated with the male media models emphasized more traditionally masculine interests and behaviors (e.g., a truck) or more traditionally feminine interests and behaviors (e.g., a household appliance). Once again, it appears that any attempt to influence gender role attitudes through advertising is a very complex process. Not only is it important to examine the impact of the specific behaviors and interests that a male media model displays, but it is also very important to examine what influence the entire image of a media representation has on its viewers. Future research also needs to address the long-term influence of male media images and the effect of individual difference variables on this process. Additionally it would be interesting to further explore whether less traditional men do have a wider representations for the male gender role than do more traditional men. Examination of the impact of mass media portrayals in more realistic settings (i.e., with more incidental exposure) is also desirable, in order to provide a more authentic view of the actual transmission of male gender role attitudes via mass media sources. Much more remains to be discovered about the impact of media images on human behavior patterns. This paper is based on a thesis submitted by the first author in partial fulfillment of the requirements for the degree of master of arts. We are grateful to Norbert Kerr and Linda Jackson for ideas and suggestions concerning this research; to Ann Baumgardner, Rick DeShon, and the anonymous reviewers for their helpful comments on earlier drafts; and to Lisa Campbell, Brian Curneal, Nickie DeLaRosa, Moonson Elliott, Corrie Higgins, Leslie Rabaut, Brian Richardson, Nicole Rieger, Christopher Rocco, Adam Stevenson, Glenn Stutzky, and Patrick Taylor for their assistance in data collection. 3 A few of Ashmore et al.’s (1995) items were modified slightly in order to enhance the focus on the roles that men versus women play in contemporary American society. Items from eight of Ashmore et al.’s subscales were not used (e.g., homosexuality, female sex initiative, male violence, political leadership, social change, day care, abortion, and individual change). 4 preliminary analyses found that there was not a main effect for the gender of the experimenter present when the GAI measures were collected nor was there any interactions involving the gender of the experimenter and the theoretically central variable of type of media portrayal. Therefore, experimenter gender effects will not be discussed further. 5 Unexpectedly, participants appeared to respond differently to the models in the advertisements that featured Delvecchio’s coffee versus EPS computers on some measures. Since these findings are not related to the type of media portrayal viewed, the central focus in this study, and because these differences cannot be easily explained because the advertisements vary on a number of features (e.g., photographs, products, name of the model, etc.), these occasional differences in reactions to the two advertisements will not be discussed further. 6 A regression analysis that maintained participants’ original gender role attitudes as a continuous variable supported the same conclusions as when gender role attitudes were dichotomized. REFERENCES Ashmore, R. D., Del Boca, F. K., & Bilder, S. M. (1995). Construction and validation of the Gender Attitude Inventory, a structured inventory to assess multiple dimensions of gender attitudes. Sex Roles, 32, 753-785. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Berscheid, E. (1985). Interpersonal attraction. In G. Lindzey & E. Aronson, (Eds.), The handbook of social psychology (3rd ed., Vol. 2), New York: Random House. Bodenhausen, G. V., Schwarz, N., Bless, H., & Wanke, M. (1995). Effects of atypical exemplars on racial beliefs: Enlightened racism or generalized appraisals? Journal of Experimental Social Psychology, 31, 48-63. Busby, L. J. (1975). Sex-role research on the mass media. Journal of Communication, 25, 107-131. Dillehay, R. C. (1978). Authoritarianism. In H. London & J. E. Exner, Jr., (Eds.), Dimensions of personality. New York: John Wiley & Sons. Duncan, L. E., Peterson, B. E., & Winter, D. G. (1997). Authoritarianism and gender roles: Toward a psychological analysis of hegemonic relationships. Personality and Social Psychology Bulletin, 23, 41-49. Durkin, K. (1985a). Television and sex-role acquisition: I. Content. British Journal of Social Psychology, 24, 101-113. Durkin, K. (1985b). Television and sex-role acquisition: II. Effects. British Journal of Social Psychology, 24, 191-210. Eisenstock, B. (1984). Sex-role differences in children’s identification with counter-stereotypical televised portrayals. Sex Roles, 10, 417-430. England, P., & Gardner, T. (1983). Sex differentiation in magazine advertisements: A content analysis using log-linear modeling. Current Issues and Research in Advertising: Original Research and Theoretical Contributions, 253-268. Fejes, F. (1989). Images of men in media research. Critical studies in mass communication, 6, 215-221. Geis, F. L., Brown, V., Jennings (Walstedt), J. J., & Porter, N. (1984). TV commercials as achievement scripts for women. Sex Roles, 10, 513-525. Harris, R. J. (1989). A cognitive psychology of mass communication. Hillsdale, NJ: Erlbaum. Hass, R. G. (1981). Effects of source characteristics on cognitive responses and persuasion. In R. E. Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion. Hillsdale, NJ: Erlbaum. Jacobson, M. F., & Mazur, L. A. (1995). Marketing madness: A survival guide for a consumer society. Boulder, CO: Westview. Kanner, B. (1990, May 21). Big boys don’t cry: Returning the macho message. New York, 23, pp. 20-21. Katz, D. (1960). The functional approach to the study of attitude. Public Opinion Quarterly, 24, 163-204. Katz, J. (1995). Advertising and the construction of violent white masculinity. In G. Dines & J. Humez (Eds.), Gender, race, and class in media. Thousand Oaks, CA: Sage. Leppard, W., Ogletree, S. M., & Wallen, E. (1993). Gender stereotyping in medical advertising: Much ado about something? Sex Roles, 29, 829-839. Lovdal, L. T. (1989). Sex role messages in television commercials: An update. Sex Roles, 21, 715-724. McGuire, W. J. (1985). Attitudes and attitude change. In G. Lindzey & E. Aronson, (Eds.), The handbook of social psychology (3rd ed., Vol. 2). New York: Random House. Orlofsky, J. L., Ramsden, M. W., & Cohen, R. S. (1982). Development of the revised sex-role behavior scale. Journal of Personality Assessment, 46, 632-638. Pearson, J. C., Turner, L. H., & Todd-Mancillas, W. (1991). Gender and communication. Dubuque, IA: Wm. C. Brown. Ruble, D. N., & Stangor, C. (1986). Stalking the elusive schema: Insights from developmental and social-psychological analyses of gender schemas. Social Cognition, 4, 227-261. Rudman, W. J., & Verdi, P. (1993). Exploitation: Comparing sexual and violent imagery of females and males in advertising. Women and Health, 20, 1-14. Schwarz, N., & Bless, H. (1992a). Constructing reality and its alternatives: An inclusion/exclusion model of assimilation and contrast effects in social judgment. In L. L. Martin & A. Tesser (Eds.), The construction of social judgments. Hillsdale, NJ: Erlbaum. Schwarz, N., & Bless, H. (1992b). Scandals and the public’s trust in politicians: Assimilation and contrast effects. Personality and Social Psychology Bulletin, 18, 574-579. Signorella, M. L., & Liben, L. S. (1984). Recall and reconstruction of gender-related pictures: Effects of attitude, task difficulty, and age. Child Development, 55, 393-405. Signorielli, N., & Lears, M. (1992). Children, television, and conceptions about chores: Attitudes and behaviors. Sex Roles, 27, 157-170. Simons, H. W., Berkowitz, N. N., & Moyer R. J. (1970). Similarity, credibility, and attitude change: A review and a theory. Psychological Bulletin, 73, 1-16. Skelly, G. U., & Lundstrom, W. J. (1981). Male sex roles in magazine advertising, 1959-1979. Journal of Communication, 31, 52-57. Smith, E. R. (1992). The role of exemplars in social judgment. In L. L. Martin & A. Tesser (Eds.), The construction of social judgments. Hillsdale, NJ: Erlbaum. Spence, J. T., Deaux, K., & Helmreich, R. L. (1985). Sex roles in contemporary American society. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed., Vol. 2), New York: Random House. Tan, A. S. (1985). Mass communication theories and research. New York: John Wiley & Sons. Vidmar, N., & Rokeach, M. (1974). Archie Bunker’s bigotry: A study in selective perception and exposure. Journal of Communication, 24, 36-47. Wagman, M. (1955). Attitude change and authoritarian personality. Journal of Psychology, 40, 3-24. Wilson, T. D., & Hodges, S. D. (1992). Attitudes as temporary constructions. In L. L. Martin & A. Tesser (Eds.), The construction of social judgments. Hillsdale, NJ: Erlbaum.