Adolescent evaluation of gender role and sexual imagery in television advertisements.

Procedures

Respondents who answered publicized appeals were administered human subjects consent materials and randomized into groups. They viewed the television programs and ads in groups of two to six persons. Respondents did not communicate to one another while viewing. Each respondent reacted to six ads: four beer and two non-beer. Participants watched an excerpt approximately 20 minutes long (including the 90 seconds of ads), with the ads embedded approximately 5 minutes apart, in an attempt to create a more realistic viewing experience (exact placement depended on the flow of the game or story narrative to some extent). After each ad appeared, the videotape paused for a few minutes, and participants were encouraged to write down the "thoughts and feelings [they] had while watching the commercial" (cf., Cacioppo, Harkins, & Petty, 1981 ; Petty & Cacioppo, 1986). They wrote as much or as little as they could recall of their thoughts and feelings while watching each ad stimulus. We included in the cognitive response section probe questions asking for their comments on ad presentation, content, and the people portrayed. Respondents noted their demographics and media use, using 5-point Likert-type responses to questions about how much attention they paid to beer ads, and how often they see beer ads on television. Finally, participants were debriefed, paid, and excused.

Coding of Cognitive Responses

Coding of the cognitive responses involved categorizing any comment about the gender, sexuality, sexual orientation, and attractiveness of models. The single thought was the unit of analysis, with a single thought consisting of a conceptual declaration or question, whether or not it was a complete sentence. Coders were found to demonstrate an agreement of r = .93 (p < .001) on the number of thoughts coded.

Responses were coded as sexist, neutral, or nonsexist in valence with respect to what respondents were relating about themselves, as well as how respondents evaluated an ad. "Sexist" comments emphasized exclusion, limitation, or offensive language toward one gender over another. "Neutral" comments were simple reports describing the self or activities that were about gender roles, but without blatant references to the people as sexual objects or in limited roles. "Nonsexist" comments were statements that revealed an egalitarian view of the two genders. Examples of these follow.

Comments revealing something about the participants themselves as sexist, neutral, or nonsexist:

Comments about the advertisements or the advertisers' use of sex, sexism, or gender role content as sexist, neutral, or nonsexist:

Comments coded as criticizing the ads for sexism had to meet one of the following criteria: reaction to the objectification (e.g., use of people in the ads as "things" or "objects"); comments about presentation of parts of bodies that dehumanize or remove subjectivity and identity of a model; reference to possession or ownership of people in the ad (e.g., "They just show women as men's property"); reactions to exploitation linking the product with males or females, such as suggesting that a female "comes" with the purchase or consumption of the product (e.g., "They're just using women to sell beer"); comments about invisibility (e.g., "They didn't show women playing billiards, but we like it, too"); claims of unfairness of gender roles (e.g., "Women drink beer, too"); and criticism about sexuality, such as the use of unfair sexual themes ("e.g., They showed half-naked women to get your attention").

Attractiveness of models was also coded, again, with "sexist," "neutral," and "nonsexist" as the responses. These were responses to the ad actors' or models' looks, comments that designated a general or specific liking for a male or female in the ad, but not in reference to sexual anatomy or sexuality--e.g., "The guys were good looking," and "Advertisers think beautiful people will sell beer."

Counterarguments, which were direct challenges to the ads' persuasive premises, were also coded. Examples include, "They'd all be real fat if they really drank beer all the time," and "I don't think that tastes good." Two coders achieved a Cohen's Kappa of .77 for categories, and .87 for valences, using 15% random responses. These are considered "substantial" for intercoder reliability (Landis & Koch, 1977).

Content Analysis of Advertisements

In this study, we attempted to derive fairly parsimonious categories that could be used to address our hypotheses and research questions regarding common gender role and sexuality themes. Ads were coded as "sexual but not sexist" (yes/no) if they displayed revealing parts of the bodies of both males and females (e.g., in bikinis, wearing other revealing clothing like low-cut tops and skimpy shorts, underwear); people were shown as modeling or exhibiting their bodies with little difference in male and female modeling; the ad's language used sexual terms, but not exclusive to one gender; or the behavior of the models was sexually provocative, or involved an implicit or explicit sexual act.

"Sexist without sexual content" (yes/no) included ads that did not include implicit or explicit sexual content where people appeared in limited roles (e.g., the women were dancing in the background and the men were engaged in sports and drinking); people are shown only in traditional roles (e.g., females cleaning, cooking, or caring for children and males going off to work or participating in recreational and sports activities); females were in submissive roles and males were in dominant, active roles. "Sexual and sexist" (yes/no) combined the above "yes" categories in one advertisement--e.g., an ad featuring nudity for the females but not the males, with the males as protagonists in active roles and the females in decorative minor roles, even objectified as props.

"Targeted aged" used the categories "younger," "older," and "indeterminant," with the following guidelines for a "younger" code: the behavior of people in the ad disrupted adult routines or lifestyles; there were no clear work, profession, homemaking, or other responsible adult images in the ad; older people in the ad were used only as background--i.e. they were out of the action of the ad; they looked uninvolved, disinterested, even bored by what was going on in the ad's central action, while the younger people in the ad were the principal actors/protagonists; freedom, fun, and youthful activities were present (e.g., amusement parks, play, high action sports like rafting, sky diving, snow boarding, and other symbols of youth like school or camp); youth products are advertised--e.g., toys, candy; and youth styles of clothing. "Older" included people whose age was generally determined as over 25, and they did not fall into the above description of "younger"; models were presented in "older" roles, such as working at mature, adult careers (e.g., lawyers, doctors, parents); more adult-types of activities were portrayed--e.g., "slower" sports like softball, bowling, hiking; and more evidence of passive recreation, with a slower pace of activities, generally. An "indeterminant age" category was used for an animated, nonhuman ad.

The gender targeting focus was a 5-point scale that coders ranked, with 1 being male dominated activities or theme, 2 being males in the forefront of activities or importance, 3 being balanced activities and the importance of males and females, 4 being females in the forefront of activities or importance, and 5 being female-dominated activities or theme. This measure was later collapsed to male, female, and balanced targeting.

Two coders achieved a Cohen's Kappa of .84 for the above categories, using a 12% subsample. This is considered an excellent reliability coefficient for categorical coding (Landis & Koch, 1977).

Data Analysis

Descriptive statistics were used for the research questions, including frequencies, cross-tabulations, and correlations. Chi-square analysis, t-tests, and analysis of variance were used to test the hypotheses.

Results

We first examined the research questions to determine the nature of the ad categories. RQ1 asked the nature of the reliable categorical distinctions about gender role and sexual portrayals (Table 1). Most of the ads contained traditional gender/ sexist, but not sexual, imagery. Seven advertisements (10%) contained sexual but not sexist/traditional gender role role imagery, and 65 (90%) did not; 8 ads (11%) presented both sexual and traditional gender/sexist imagery and 64 (89%) did not. Although a few ads targeted younger people, a majority of the ads targeted older people. While males were targeted more than females, about one third demonstrated balanced gender targeting.

RQ2 addressed the nature of beer vs. non-beer ads. A substantial majority of beer ads were found to contain traditional gender role/sexist content. Beer ads with sports content revealed even higher levels. Non-beer ads were split evenly relative to sexual and sexist content. Looking at perceived target audiences for the beer ads, we found beer ads were either male-directed or balanced, with almost no female targeting. As for beer ads with sports imagery, the majority were found to be male-targeted, with the remaining balanced between males and females. Females also were seldom explicitly the target of non-beer ads, as a majority of non-beer ads were found balanced in their gender targeting. As for age, beer ads tended to be targeted toward older audience members. Beer ads with sports content tended to be targeted to youth, with a little more than half geared toward younger people, and fewer geared toward older people. Non-beer ads were almost exclusively found to target mature audience members.

Eliminating the one advertisement for which an age target was indeterminant, cross tabulations were run against the other coded ad characteristics. Although the chi-square test may be unreliable for testing tables where 20% of the cells have less than 5 instances, patterns of relationship and their significance are reported here. A relationship between focus of ad targeting (young vs. old) and whether ads were sexist without sexual content showed that ads targeting older audience members contained more sexist content ([chi square] = 5.07, p < .05). Also, a relationship between ad targeting and ad type suggested ads targeted toward younger people are more likely lifestyle ads ([chi square] = 10.83, p < 001).

Print Get the Mag Weekly Updates