Taking It Off All Over Again: The Portrayal of Women in Advertising Over The Past Forty Years

Magazine advertisements from six publications in 2004 were analyzed to replicate an original, seminal study that assessed the use of sexual imagery in advertising in 1964 and 1984. Research indicates that media, advertising in particular, can create and sustain unrealistic images and stereotypes, especially of women. It is important to examine how the portrayal of women in advertising has changed over the past decades. Preliminary data analysis shows that across all magazine genres, in 2004, males appeared demurely dressed 83.5 percent of the time, while women are only shown as demurely dressed a third of the time. This enormous discrepancy in the level of dress between male and female models indicates that women are portrayed much more often as sexual objects than are men. This study will hopefully advance feminist scholarship through the thorough documentation of the trend of inequality in the portrayal of sexuality and gender roles of females in the past and today. Sex in advertising 3 Taking It Off All Over Again: The Portrayal of Women in Advertising Over The Past Forty Years The use of sexual imagery in advertising is nothing new. Past studies suggest that the level of sex in advertising has grown more prevalent, more explicit, and more diverse over time (Soley & Reid, 1988; Soley & Kurzbard, 1986; Reichert, Lambiase, Morgan, Carstarphen and Zavoina,1999). Cultivation research shows us that frequently repeated themes and behaviors seen in the media can influence our perceptions of social reality (Gerbner, 1990). One mechanism to explain how this happens is that the constant exposure to these themes and behaviors make them more accessible from memory (Shrum, 2003) – and when asked to make judgments in real life, those who are more exposed to certain mediated constructs are more likely to use them (Rhodes, Edison, & Bradford, 2004). Is the level of sex in advertising increasing? How are women portrayed in advertising? This research seeks to address these important questions in order to quantitatively document the mediated role of women in advertising. Two analyses assessing the level of sexuality in magazine advertising found that the level of sexual dress in magazine advertisements increased from 1964 to 1984 (Soley and Reid, 1988) and that there was also a greater reliance on both visual and verbal sexual imagery in 1984 than in 1964 (Soley and Kurzbard, 1986). In order to assess change in the portrayal of women over time, the same coding procedures and content categories were used to update the seminal Soley and Reid study. Although Soley and Reid looked only at how level of dress changed over time, Soley and Kurzbard also examined headlines for sexual references and physical contact between models. By adding headline sexual reference and level of physical contact as additional content categories, this study will allow us to determine how advertisers’ use of sexual imagery has changed over time and how women in Sex in advertising 4 advertising are represented today. In addition, other variables investigating the role of models (especially female models) in magazine advertisements will also be examined and discussed. Sex in Advertising Since the 1850s, sexual imagery has been widely used to sell products, services, and ideas. During this time period, the development of cheaper, better printing technologies made using illustrations in newspapers and magazines more affordable, and therefore more accessible to advertisers and to the public (Folkerts & Lacy, 2004). For example, tobacco companies were using suggestive or even nude images of females during the later 1800s (Reichert, 2003). However, during the last 50 years, social changes have become rapidly apparent in acceptable dress in real life – and in the media. The level of dress in advertising content today ranges from mere sexual suggestion to full nudity and depiction of intercourse. However, sex in advertising can also include “sexual imagery, innuendo, and double entendre” (Courtney and Whipple, 1983, p. 103). This study will assess other sexual content characteristics in addition to level of dress. Content Analyses In addition to the original, seminal works by Soley and Reid (1988) and Soley and Kurzbard (1986), which examined sex in advertising from 1964 to 1984, more recent studies have examined sex in advertising over the past twenty years (Reichert et al., 1999; Reichert & Carpenter, 2004). A content analysis by Reichert and colleagues (1999) found that the incidence of sexual dress in magazine advertisements has risen over the past few decades. In addition to the level of dress, contact between models became more sexual over time. A recent update of the Reichert et al. study found that models in ads in 2003 were no more likely to be explicitly dressed or in sexual contact than they were in 1993 (Reichert & Carpenter, 2004). Not surprisingly, sexual ad content differs between magazine genres. For example, sexual ads are most apt to appear in Sex in advertising 5 women’s and men’s magazines compared to general-interest magazines such as Newsweek and Time (Soley & Reid, 1988; Soley & Kurzbard, 1986; Busby & Leichty, 1993, Reichert & Carpenter, 2004). Cultivation Effects and Construct Accessibility Why do we care how prevalent sexual imagery is in the media? The cultivation hypothesis posits that televised social reality influences viewers’ social perceptions and beliefs (Hawkins & Pingree, 1990). Although original cultivation hypothesis research examined only effects on perceptions of reality due to television use, other forms of media are also associated with distortions of reality (Rhodes et al., 2004). Cultivation “implies long-term, cumulative consequences of exposure to an essentially repetitive and stable system of messages” (Morgan & Signorielli, 1990, p. 18). For example, a recent study showed a correlation between reported viewing of romantic television programs and idealistic expectations for marriage (Segrin & Nabi, 2002). The frequent portrayal of women as sexual objects in all forms of the mass media can influence men’s perceptions of women and women’s perceptions of themselves. Cultivation effects can be explained through models of construct accessibility (Shrum, 1996, 2002). Using more of a specific media genre can subtly influence perceptions and interpretations of the world and the people encountered in it by making mediated constructs/behaviors more accessible from memory than experiential realities. Accessibility is based on how often activation of a construct occurs, how recently it has occurred, and how vivid the information is. Accessibility is also dependent on availability – if a construct has not been encountered previously, or was not stored in memory, it cannot be available for activation (Sanbonmatsu & Fazio, 1991). Highly accessible information is more likely to be used when forming judgments than is less accessible information (Shrum, 2002). Since exposure, or repetition, is an important part that determines how accessible a construct is, it makes sense that Sex in advertising 6 when people are repeatedly exposed to an idealized, sexualized image of the female body through the media, a real-world situation would be likely to activate the vicariously-learned image. “The media are instrumental in developing the constructs that are available in memory, and in determining which of those constructs are chronically accessible” (Sanbonmatsu & Fazio, 1991, p. 57). As Harris (1999) suggests, the mental reality experienced through the media becomes the basis for attitudes and behaviors. Research Questions and Hypotheses First, we compared ads from 2004 to the original study by Soley & Reid (1988). These authors examined the amount of clothing models were wearing in six high-circulation mainstream magazines over two decades (Cosmopolitan, Redbook, Esquire, Playboy, Newsweek, and Time). They found that although women were dressed more sexually in 1984 than in 1964, the incidence of sexualized images of men also increased. However, more recent studies suggest that the increase in sexuality in advertising is tapering off (Reichert & Carpenter, 2004). Comparing 2004 ads to the original Soley and Reid study will show how the use of sexual appeals in advertising has changed over the past four decades and where it stands today. RQ1: How has the level of dress for male and female models in magazine advertisements changed over the past forty years? RQ2: How has the level of dress for male and female models in magazine advertisements changed by magazine type over the past forty years? Soley and Kurzbard (1986) also examined headlines in magazine advertisements for verbal sexual references. They found that from 1964 to 1984, verbal sexual reference had increased. We will assess amount of change in this content category over the last four decades, therefore: Sex in advertising 7 H1: The frequency of headlines utilizing verbal sexual references will have increased over the past forty years. In addition, we examined several other variables neglected in the previous study. First, we addressed current gender roles portrayed by models in advertising. Research focusing on the stereotypical portrayals of women in advertising is particularly important. In addition to being shown as young and good-looking (Goldman, 1992), women are more likely than men to be portrayed as domestic providers or sex objects in advertising (Courtney & Lockeretz, 1971). Many advertisements concentrate on the female “ideal,” an identity to which all women should aspire. Shields (1997) explains, “These messages, used to sell everything from cosmetics to cars to beverages, provide a prescription for how women should look and be looked at, how they should feel, and how they are expected to act. In short, these messages prescribe particular gender identities... and how men should relate to women” (p. 71). Geis, Brown, Jennings and Porter (1984) also suggested how gender representations in the media could influence attitudes and behaviors. First, on the macro level, they are broadcast publicly, and so appear to represent “everyone.” Second, on the micro level, the characters portrayed are shown as accepting their role relationship, thus appearing to validate the complementary roles within the scenario (p. 513). In other words, some females may view ads as life scripts. Geis et al. concluded that many women’s personal self-concept is determined by the sex typing shown in advertising. When analyzing the six publications in 2004 for sexual content, we addressed the following research questions: RQ3: How do gender roles enacted by models in magazine advertisements differ for males and females? RQ4: How do gender roles enacted by models in magazine advertisements differ between magazine types for males and females? Sex in advertising 8 Finally, physical contact between models is also becoming more sexual and explicit (Soley & Kurzbard, 1986; Reichert et al., 1999; Reichert & Carpenter, 2004). Although the sexuality of dress in magazine advertisements may no longer be increasing as rapidly as in previous decades, advertisements are becoming more sexual in other ways – primarily through physical contact between models. Therefore, we addressed this issue by comparing model contact in 2004 to model contact in 1964 and 1984 (Soley & Kurzbard): H2: There will be increasing explicit sexual contact between male and female models over the past four decades. Method Sample The present study analyzed sex in magazine advertising in 2004 by replicating the procedures and coding scheme employed in Soley & Reid’s (1988) content analysis. As in the original study, all full-page or larger ads in the same six consumer magazines were analyzed for model dress level. Product category, sex role, contact level, headline sexuality, and homosexual reference were also analyzed. The magazines were initially chosen because they were the highest-circulation publications that represented three distinct editorial profiles (women’s, Cosmopolitan and Redbook; men’s, Esquire and Playboy; general interest, Newsweek and Time; Soley & Reid, 1988). Also as in the original study, ads in three issues of each magazine (January, August, and November) were coded. Because two of the magazines selected are weekly publications (Newsweek and Time), the first issue of the month was chosen for analysis. Only ads featuring discernable adult models were included in the analysis of sexual content. All ads featuring body parts, such as feet or hands, were coded as not containing a model. All duplicate ads were included in the analysis. Sex in advertising 9 The 2004 sample, which consisted of 18 magazine issues, included 960 full-page or larger ads. Of the 960 total ads, 568 contained discernable adult models. While 280 ads contained female models, 154 ads contained male models, and 134 ads contained both female and male models. The previous content analysis of 1964 magazines reported 586 full-page or larger advertisements, with 367 including adult models. The analysis of 1984 magazines reported 1,112 ads, 645 including adult models (Soley & Reid, 1988). Content Variables As in the original study, each discernible adult model was classified into one of four ordinal categories of dress (Soley & Reid, 1988). Dress differs primarily by the amount of clothing worn by models in the advertisements, and has four levels: (1) demure, (2) suggestive, (3) partially-clad, and (4) nude. Demure dress was defined as “everyday dress.” This category included all typical clothing, including walking shorts, but excluded short-shorts, underwear, or bathing suits. Suggestive dress included clothing that partially exposed the upper body, such as unbuttoned blouses showcasing cleavage, or tight “muscle” shirts. Models were considered partially clad if they were shown in underwear or bathing suits. If nudity was implied (models were holding only a towel or if the photograph was composed so that genitals were concealed), models were coded as nude. If the ad contained multiple models of the same sex, the model featuring the least amount of clothing was used to code that advertisement, as was done in the original study (Soley & Reid). When at least one female and one male model were present in an ad, the level of physical contact between them was classified into one of four ordinal categories: (1) no contact, (2) simple contact (e.g., holding hands), (3) intimate contact (e.g., kissing and embracing), and (4) very intimate contact (e.g., the depiction or suggestion of sexual intercourse or behavior). The behavior within each category becomes more sexually explicit from Level 1 to Level 4. If the Sex in advertising 10 advertisement only featured models of one sex, the ad was coded as Level 0: no contact, one model. In addition to these variables, this study also looked at gender stereotypes. The five-level sexism scale developed by Pingree, Hawkins, Butler and Paisley (1976) was used to investigate stereotypical scenarios displayed in each full-page magazine advertisement featuring a discernable adult model. The operational definitions for each level in the Pingree scale were applied to the primary character in the ad. Level I is described for females as putting her down, the dumb blond, the sex object or being a two dimensional image (p. 194). Male Level I ads show a sexy male model for a product which doesn’t necessarily help with explanation (p. 198). Level II is described for females as showing traditional female activities, characterized as “keeping females in their place.” Women are shown as wives, mothers, secretaries, teachers and nurses; for men, Level II “is the world of work and traditional ‘manly’ activities... in the business or sports contexts and, if they are shown working around the home, it is mowing the lawn or taking the garbage out” (p. 199). Level III for women gives her two places (p. 194). Women can work out of the home, but always maintain their work at home first. Level III men are shown as competent both at work and at home with work always the first priority. For females, Level IV images do not remind us that housework and mothering are top priority (p. 195). Level IV for men is exactly the same for women; here women and men must be seen as equals. The last level of portrayal, Level V, is nonstereotypical. “Individual women and men are viewed as superior to each other in some respects, inferior in other respects” (p. 195). In addition, we also examined product category for each advertisement, and presence or absence of verbal sexual reference in headlines, and presence or absence of visual homosexual innuendo. Coding Procedure. Two coders were trained in the content categories. The training session involved agreement between coders on all variables and variable levels and practice Sex in advertising 11 coding until high intercoder reliability was achieved. The practice ads were not part of the analyzed sample. Working independently, the coders analyzed all ads in the sample frame. Using Scott’s Pi analysis, average intercoder reliability across the variables product category, model gender, sex role, dress, contact level, headline sexuality, and homosexual reference was .98. Scott’s Pi is a test of reliability that statistically corrects for chance agreement when computing reliability (Riffe, Lacy, & Fico, 1998). Scott’s Pi for product category was .95 and .92 for sex role. All other variables achieved perfect intercoder reliability.