Abstract
Across two studies, we investigated the hypothesis that exposure to objectifying television in which women are shown as sexual objects increases the likelihood of harassing conduct. In both studies (Ns = 141; 120), male participants were exposed to one of the three TV clips in which women were portrayed (a) as sexual objects (objectifying TV), (b) in professional roles, or (c) excluded (a nature documentary). Study 1 showed that men exposed to objectifying TV reported greater proclivity to engage in sexual coercion and manifested more gender-harassing behavior than participants in the other conditions. Study 2 further demonstrated that exposure to objectifying TV increased participants’ conformity to masculine gender role norms, which, in turn, mediated the relation between experimental condition and gender harassment. Together, the two studies suggest that media content plays a central role in activating harassment-related social norms, which in turn encourage or inhibit harassing conduct.
Sexual harassment in its various forms constitutes a widespread phenomenon that affects people at work as well as in informal social contexts (see Maass, Cadinu, & Galdi, 2013, for a review). Although the phenomenon has been observed across different historical periods and in different geographical areas, incidence rates of sexual harassment vary greatly across countries. Some authors have argued that differences in the likelihood to sexually harass, as well as the tolerance of sexual harassment by victims, are due to fundamental differences in norms and values both in the society at large (Luthar & Luthar, 2007) and within specific contexts and organizations (see Willness, Steel, & Lee, 2007, for a review). Together, these findings suggest that cultural differences in behavioral standards and social norms are key components in shaping the varying conceptions of sexual harassment across countries and social contexts (Barak, 1997; Luthar & Luthar, 2007).
The dissemination of restrictive versus permissive social norms concerning sexual harassment depends, to a large degree, on available channels of communication. What is accepted and what is not in a given society not only is communicated by common socializing agents (such as family, school, and workplace), but also is conveyed by the media that play an important role in defining prevailing social norms. Television, in particular, which is widely accessible and intentionally appealing and engaging, makes massive use of stereotypical messages that the majority of people can easily understand. TV’s shorthand depictions of reality are transmitted both verbally and nonverbally, as well as directly and indirectly, and represent often subtle but influential tools by which knowledge, attitudes, and values are conveyed (Cialdini & Trost, 1998).
Although in the last two decades important research has been conducted to determine social norms and cultural risk factors associated with sexual harassment both in educational and in work settings (Fitzgerald, Drasgow, & Waldo, 1999; Ilies, Hauserman, Schwochau, & Stibal, 2003; Meraviglia, Becker, Rosenbluth, Sanchez, & Robertson, 2003; O’Hare & O’Donohue, 1998), to date no known experimental research has investigated the relationships among messages conveyed by television, social norms, and sexual harassment. In Italy, the overwhelming diffusion of television programs characterized by the presence of stereotypical depictions of women as sexual and decorative objects led us to investigate whether the mere exposure of men to objectifying TV programs (i.e., in which women are presented and treated as bodies) increases the likelihood of sexual harassment. The main hypotheses of the present work are (a) that exposure to objectifying TV programs encourages male viewers’ proclivity to engage in sexual coercion and gender-harassing behavior (Study 1) and (b) that such an increase in sexual harassment after the exposure to objectifying TV programs derives from a shift in the endorsement of traditional masculine norms that may condone harassing conduct (Study 2).
Over the last 20 years, television has conveyed an overwhelming diffusion of messages that stereotype and objectify women by portraying them as passive, dependent on men, compliant, and sexual objects—namely mere bodies (Bartky, 1990; Fredrickson & Roberts, 1997). For example, a consistent finding from content analyses of TV programming is the strong difference in the degree to which women’s and men’s bodies are sexualized. Across many TV genres, women, in contrast to their male counterparts, are typically presented as decorative elements whose value is based solely on their physical appearance. Women talk less and appear less frequently in the role of experts; conversely, they are more likely to be scantily clad, suggestively dressed, and posed in sexually exploitative and submissive postures (Andsager & Roe, 1999; Arnett, 2002; Ward, 1995). Similar gender differences are also typical of televised commercials. For example, Lin (1998) analyzed 505 commercials and found that women, in comparison to men, were more likely to be shown scantily dressed, with an emphasis on their bodies, and were more often depicted as sexual objects.
Objectification of women in television not only is visual but also is expressed by explicit as well as subtle verbal acts. Here, objectifying content is typically conveyed through messages in which men make comments on women’s physical appearance or in a humorous context through jokes, gags, and double meanings (Fouts & Burggraf, 2000; Ward, 1995). At times, such events represent actual harassing conduct. For example, Montemurro (2003) analyzed a total of 56 episodes of five different situation comedies for a period of 5 months. Findings demonstrated that there was an average of nearly 3.5 incidents of gender harassment per episode, with behaviors ranging from uninvited sexual teasing to suggestive looks or gestures, and an average of 0.5 incidents of sexual harassment, with behaviors spanning from uninvited sexually inappropriate touching to uninvited pressures for dates and specific requests for sexual favors. In a similar vein, Grauerholz and King (1997) conducted a content analysis of 48 hours of prime time TV programs and found that 84% of the shows contained at least one incident of sexual harassment. Specifically, the most frequent behaviors were using sexist terms to describe women (e.g., blondie, babe, chick, broad), sexual comments focusing on women’s bodies or body parts, and body language generally involving men or boys leering at women or girls. Interestingly, the authors noted that none of these acts was labeled as sexual harassment. Instead, they were presented in humorous ways, and victims were generally portrayed as unharmed and as being able to deal effectively with the harassment.
These findings suggest that even prime time TV programs convey the message that sexual harassment is not serious and that victims should be able to handle the situations themselves. In summary, the objectification of women on TV programs ranges from more explicit and direct forms, such as the presence of provocatively dressed female models or sexual comments, to more subtle and indirect forms, such as the presence of women as merely decorative elements or the use of degrading jokes and double entendres about women.
Exposure to objectifying-media, including TV contents, may have important social consequences. For example, exposure to sexual and sexist media content is associated with greater acceptance of stereotypical attitudes. In this regard, there is evidence that exposure to magazine ads featuring women as sexual objects produces stronger acceptance of gender-role stereotyping and of rape myths among male undergraduates (Lanis & Covell, 1995; MacKay & Covell, 1997). Consistently, survey research indicates that more regular exposure to TV genres such as soap operas and music videos—which typically deal with sexual contents and sexual feelings or impulses (Gruber & Grube, 2000; Media Report to Women, 2001)—is associated with more stereotypical sexual attitudes, with dysfunctional beliefs about relationships, and with greater acceptance of sexual advances, especially among women and adolescent girls (Haferkamp, 1999; Strouse, Goodwin, & Roscoe, 1994). Likewise, Ward (2002) has found that, regardless of exposure levels, young women exposed to prime time TV images depicting men as sex-driven and women as sexual objects showed stronger endorsement of such stereotypes than did women exposed to control clips. Other experimental results support these findings, showing that women exposed to sexual and sexist media content manifest stronger endorsement of stereotypical attitudes about sex than do women exposed to nonsexual content (Ward, 2002; Ward, Hansbrough, & Walker, 2005).
In a similar vein, research drawing from priming and schema theories (Fiske & Taylor, 1991; Jo & Berkowitz, 1994; Srull & Wyer, 1989) has demonstrated that even viewers’ interpretation of ambiguous stimuli and impressions about real-world interactions may be affected by the nature of the TV content just watched. For example, Hansen and Hansen (1988) tested the hypothesis that stereotypical gender-role depictions in music videos prime corresponding gender-role stereotypic schemata, which, in turn, distort appraisals of a subsequently observed interaction between a man and a woman. In their study, participants were exposed to music videos that did (or did not) incorporate gender-role stereotypic depictions of women and men. Then, participants evaluated the taped interactions of a male and a female job applicant. Results showed that participants who had watched the neutral music videos later interpreted the man’s sexual advances toward the female applicant as unwanted sexual attention, whereas participants who had viewed the stereotypical music videos perceived the man’s sexual advances as appropriate and thought less favorably of the applicant when she rejected the man.
Similar results were obtained by Rudman and Borgida (1995). Male participants performed a lexical decision task after watching TV commercials portraying women as sexual objects or control TV commercials. The authors hypothesized that exposure to TV commercials portraying women as sexual objects would boost the accessibility of the category of women as sexual objects. As a result, ambiguous stimuli would be more likely to be interpreted consistently with the primed construct. Findings demonstrated that participants in the sexist condition interpreted the ambiguous stimuli included in the lexical decision task (i.e., double entendres such as cherry, easy, strip) consistently with the construct primed by the sexist commercials (i.e., women as sexual objects), whereas participants in the control condition interpreted the same words as neutral stimuli.
What these studies show is that exposure to objectifying messages distorts viewers’ impressions and affects their interpretations of messages. Yet, their findings do not directly speak to the hypothesis that objectifying TV may affect viewers’ interpersonal behavior. Indeed, most of the research conducted to explore the effects of media images on interpersonal behavior has focused on the role of media violence on aggression (see Freedman, 2002, for a review). To our knowledge, only two studies have examined the impact of exposure to objectifying female models on interpersonal behaviors (Hitlan, Pryor, Hesson-McInnis, & Olson, 2009; Rudman & Borgida, 1995). For example, Hitlan, Pryor, Hesson-McInnis, and Olson (2009, study 1) demonstrated that male participants who were primed with a video showing women portrayed in sex-stereotypic terms were more likely to ask sexist questions during a subsequent mock job interview than men who were not primed. Similar results were obtained by Rudman and Borgida (1995) in the study we briefly described previously. The authors, indeed, measured the effects of temporary accessibility of the category of women as sexual objects not only on cognition, but also on behavior. Men exposed to TV commercials portraying women as sexual objects not only were more likely to think of women as sexual objects, but also exhibited more sexualized behaviors in a subsequent interaction with a female confederate than participants in the control condition. Specifically, compared to participants in the control condition, those who were exposed to the objectifying TV commercials sat closer to the female confederate and asked more sexist and inappropriate questions during the interaction. Interestingly, both female confederates and independent observers judged participants’ behavior in the objectifying condition as more sexualized than in the control condition.
What these findings suggest is that harassing behavior might depend not only on people’s personal predisposition to view women as sexual objects, but also on mere TV exposure to objectified women that enhances the accessibility of women as sexual objects, as well as other strictly related constructs. A theoretical perspective supporting this idea is spreading-activation theory (Collins & Loftus, 1975). According to this theory, the presentation and processing of a stimulus with a particular meaning activates semantically related schemas, or constructs, and calls them to mind. Such activation makes the construct-related information temporarily highly accessible, thus increasing the likelihood that subsequently encountered persons, stimuli, or events will be appraised in the context of the activated constructs. Hence, one might predict that TV depictions of objectified women will promote the accessibility not only of the construct of women as sexual objects but also of related constructs, such as gender-role stereotyped beliefs and norms. If this were the case, then much of the potential impact of TV exposure on behavior would be due to its effects on belief and norm systems that define which behaviors are appropriate in a given situation. But what specific norms may be influenced by exposure to objectifying TV programs and how do these norms affect harassing behavior?
In the present work, we focused on men and the role played by traditional masculinity ideology (Levant et al., 1992; Pleck, 1995), namely the constellation of social norms, behavioral standards, and expectations associated with the traditional male role. As a case in point, research has shown that the belief system associated with traditional masculinity ideology is common throughout most of the world (Pleck, 1995) and represents one of the factors most clearly associated with sexual harassment in its various forms as well as with rape-supportive attitudes and violence against women (Davis & Liddell, 2002; De Judicibus & McCabe, 2001; Pavlou & Knowles, 2001; Wade & Brittan-Powell, 2001; White & Kurpius, 2002). Furthermore, the endorsement of traditional masculine norms has been found to be correlated with the Likelihood of Sexual Harassment (LSH) scale (Pryor, 1987), a broadly used measure of sexual harassment proclivities in men (see O’Leary-Kelly, Bowes-Sperry, Arens Bates, & Lean, 2009; Pina, Gannon, & Saunders, 2009, for reviews).
Extending these findings to the present research, we therefore hypothesized that, regardless of viewers’ personal predispositions and personality traits, the mere exposure to objectifying TV images would enhance male viewers’ accessibility to norms associated with masculinity and men’s roles. As a result, these viewers would be more likely to conform to traditional masculine norms, which, in turn, would lead to an increase in harassing conduct.
Under the concept of sexual harassment, most psychological research distinguishes three related categories of harassing behaviors (Fitzgerald, Gelfand, & Drasgow, 1995; Fitzgerald, Swan, & Magley, 1997; Leskinen, Cortina, & Kabat, 2011): gender harassment, unwanted sexual attention, and sexual coercion. In this classification, sexual coercion refers to instances in which an individual attempts to use his or her power over a subordinate to engage in sexual activity. This form of harassment includes behaviors such as sexual blackmail, bribes, or threats aimed at receiving sexual cooperation. The category of unwanted sexual attention refers to unwelcome, unreciprocated, and offensive acts of sexual interest and encompasses verbal and nonverbal behaviors such as objectifying gaze (scrutinizing another person’s body parts), unwanted touching, or pressure for dates. Finally, gender harassment refers to behaviors aimed at creating an intimidating, offensive, or hostile environment and includes sexist acts aimed at conveying hostility (in particular toward individuals who violate gender ideals) rather than sexual interest (Berdahl, 2007a). Common examples of this form of harassment are the use of derogatory terms of address, unprofessional sexist comments, sexist jokes, and displaying sexual materials.
Although research on sexual harassment has mainly focused on unwanted sexual attention/coercion that occurs within unbalanced power relationships (Bartling & Eisenman, 1993; Pryor, 1987), there is strong evidence showing that gender harassment between peers constitutes the most common form of harassment (Barak, 2005; Langhout et al., 2005; Shepela & Levesque, 1998). For example, recent survey data (American Association of University Women Educational Foundation, 2011) show that about half of 12- to 17-year-old American students report to have experienced some form of sexual harassment at school during the 2010–2011 school year. Importantly, according to these survey data, gender harassment (e.g., unwelcome sexual comments, jokes, and gestures) appears to be the most common type of sexual harassment, and the majority of harassed students (54%) have identified a schoolmate as their harasser. Similarly, gender harassment is common in the workplace, and coworkers have been found to be the most frequent perpetrators within organizations (Keyton, 1996; Leskinen & Cortina, in press; Leskinen et al., 2011). Therefore, in the present work we investigated the effects of the mere exposure to TV programs on both proclivity to sexual coercion and actual gender-harassing behavior by presumably equal-status men.
Study 1
In Study 1, we investigated the effects of exposure to one of the three sets of video clips—which depicted objectified women (the Objectifying TV condition), nonobjectified women professionals (the Female Professionals condition), or no women in nature video clips (Control condition), all taken from real Italian TV programs—on male viewers’ proclivity to sexually coerce and actually engage in gender-harassing behavior. Our specific predictions were that (a) participants would show higher proclivity to sexual coercion (measured by the LSH scale; Pryor, 1987) and would perform more gender-harassing behavior (measured by the number of jokes supporting offensive and degrading depictions of women sent to a female interaction partner) after watching the Objectifying TV clip as compared to the Control clip and (b) even though men may experience female professionals as threatening (Maass et al., 2013), mere exposure to a video showing successful Female Professionals would not be sufficient to lead to a significant increase in either proclivity to sexual coercion or gender-harassing behavior as compared to the Control condition.
We included the Female Professionals condition to show that not just any kind of portrayal of women would induce harassment, rather only a specific, objectifying depiction. Previous research has shown that even professional, successful, and independent women may be perceived as threatening by men and that such masculine identity threat may result in greater levels of gender harassment (Berdahl, 2007b; Gruber & Bjorn, 1982; see also, Maass et al., 2013 for an overview). In fact, men tend to harass women more if they do not conform to gender stereotypes (Dall’Ara & Maass, 2000; Maass, Cadinu, Guarnieri, & Grasselli, 2003) or if women outperform them on male-gendered tasks (Hitlan et al., 2009, Study 2). However, these studies are conceptually different from ours in that their threat derived directly from an interaction partner. In contrast, our research focuses on the mere exposure to TV programs and, specifically, on media portrayals that are likely to set norms, rather than to pose a personal identity threat. We therefore decided to include the Female Professionals video clip as a comparison condition to test the hypothesis that, as compared to a neutral Control condition, male participants’ proclivity to sexual coercion and engagement in gender harassment would increase specifically in reaction to their mere exposure to the objectification, but not to other portrayals, of women in TV.
Importantly, one could speculate that sexual coercion and gender harassment are very different forms of sexual harassment: Whereas sexual coercion may be related in part to the perpetrator’s goal of gaining sexual satisfaction, this is not the case for gender harassment, which includes sexist behaviors aimed at degrading and conveying hostility rather than sexual interest. Nevertheless, there is evidence showing that these two forms of sexual harassment are strongly related to each other, that they tend to co-occur in work settings, and that each of them is a predictor of the other (Maass et al., 2003; Rudman & Borgida, 1995; Siebler, Sabelus, & Bohner, 2008). Thus, if it is true that sexual objectification of women’s bodies promotes sexual harassment, then such an effect should be evidenced on both proclivity to sexual coercion and gender-harassing behavior.
Method
Participants and Procedure
We tested the above hypotheses with a diverse sample of 141 men living in Northern Italy, including students and employed men, whose age ranged from 18 to 48 years (M = 25.26, standard deviation [SD] = 4.61). In order to avoid a solely student sample, participants were recruited by one of the two female experimenters at different university libraries and study rooms, among neighbors, and in gyms (for a similar procedure, see Galdi, Arcuri, & Gawronski, 2008; Galdi, Gawronski, Arcuri, & Friese, 2012). The experiment was run using a laptop computer in quiet rooms at the university (student sample; n = 76) or at the participants’ homes (nonstudent sample; n = 65). Throughout the data collection, the experimenter kept at a distance so that the participants could respond without feeling observed. In the case of the nonstudent sample, to avoid any risk of exposing the experimenters to unpleasant and hence ethically problematic experiences, only male participants known by the experimenters were contacted (none of whom were aware of the true aims of the study). All participants volunteered in the study without monetary compensation. Each experimental session took about 30 minutes. The procedure of the experiment and the main dependent variables were administered in the same order in which they are presented in the following.
Cover Story
Participants were told that the aim of our study was to investigate their short-term memory for TV images. The cover story was designed to provide a rationale for watching the video clips. This cover story also allowed us to present the Computer Harassment Paradigm (Siebler et al., 2008) and the LSH scale (Pryor, 1987) as unrelated intermediate tasks to let some time pass between the video and the memory test. Therefore, participants were told that they would watch a brief TV video; then, they would perform two distractor tasks (which were in reality our main dependent variables); and, finally, they would receive a number of memory questions about the TV images that they had watched at the beginning of the experiment.
Experimental Manipulation
An equal number of participants was randomly assigned to one of the three experimental conditions (47 participants in each condition), and with similar proportions of students and nonstudents, by inviting them to watch a brief video composed of TV segments approximately 3 minutes long. For the Objectifying TV condition, we created a video that included televised segments selected from common public and private Italian prime time TV shows and sharing a common type of women—namely silent, passive, scantily clad, or suggestively dressed women whose presence in the show is solely due to their physical appearance. The Female Professionals condition video included a selection of TV segments from common public and private Italian programs portraying young, successful women in domains such as science, culture, and business. The Control condition video included segments from a TV documentary on the beauties of nature (images of animals, flowers, and plants) and hence included no images of women (or men). Because we were interested in objectifying visual images rather than texts, the dialogue was excluded from all three experimental videos and replaced with the same background music across conditions.
Gender-Harassing Conduct
Gender-harassing behavior was measured employing a modified version of the Computer Harassment Paradigm (Maass et al., 2003; Mitchell, Hirschman, Angelone, & Lilly, 2004; Siebler et al., 2008), which simulates a less severe, but very common form of gender harassment—namely telling sexist/sexual jokes in which women are depicted in an offensive or degrading way. Thus, after the manipulation, as an alleged distractor task, participants were instructed to participate in a computer chat in which they could exchange jokes with a (fictitious) female partner. Ten jokes (pretested as equally funny) were used for the task. Of them, five were sexist/sexual jokes supporting offensive or sexually degrading depictions of women (e.g., “Why are there parking spots for only women? To not damage men’s cars”; “A woman without bosom is like jeans without pockets: you never know where to put your hands”), and five were neutral, that is, without sexist or sexual content (e.g., “A child asks: ‘Daddy, would you give me carnival confetti?’ Response: ‘No, last year you threw all of them away’”). On each trial, participants were presented with a pair of jokes (one sexual/sexist and the other neutral) and asked to choose one to send to their female chat partner. The dependent behavioral variable was the total number (ranging from 0 to 5) of sexual/sexist jokes sent during the chat interaction.
Harassment Proclivity
Participants’ proclivity to sexually coerce was measured through the LSH scale (Pryor, 1987), which assesses men’s intention to engage in what the law calls quid pro quo sexual harassment, a broadly accepted form of sexual harassment proposed by the U.S. Equal Employment Opportunity Commission (1980) that is roughly equivalent to sexual coercion. For the present study, we employed a short version of the adaptation of the LSH scale used in other research with Italian men (Dall’Ara & Maass, 2000; Maass et al., 2003). In its current form, the scale contained three hypothetical scenarios, in which men, described as holding powerful positions (a film director, a TV news director, and a successful university student) had the opportunity to exert power over a female subordinate (an actress, an anchor woman, a needy student, respectively).
Following each scenario, four increasingly blatant courses of action were listed. For example, for the film director scenario, the options were (a) “I think I would give the role to the sexiest actress,” (b) “I wouldn’t mind if the actress I like were to dedicate her attention to me; in this case I may consider giving her the role,” (c) “I think I would ask the actress I like most to discuss the role with me after dinner,” and (d) “She would become the leading actress if she had sex with me.” For each hypothetical scenario, participants had to place themselves in the role of the male protagonist and to indicate on a 5-point scale, ranging from 1 (not at all likely) to 5 (very likely), the likelihood that they would perform each of the four harassing courses of action if they were assured that no negative consequences would result from their choices. Of prime interest to the present research were the last two blatantly harassing alternatives. Therefore, participants’ responses to these two courses of action (c and d above) for each of the three hypothetical scenarios of the LSH scale were averaged into a single harassment proclivity score (Cronbach’s α across all three scenarios = .84.)
Perceptions of the Chat Partner
The first part of the questionnaire included 10 items assessing the Perceived Competence and Flirtatiousness of the (fictitious) chat partner on scales ranging from 1 (not at all) to 6 (a lot). To assess Perceived Competence, participants were asked to rate how intelligent, insightful, brilliant, professional, and intellectual the chat partner seemed to be. To assess Perceived Flirtatiousness, participants projected how sexy, attractive, mischievous, frivolous, and teasing the chat partner was. For each participant, responses for Perceived Competence (α = .84) and Perceived Flirtatiousness (α = .80) were averaged into two single scores such that higher scores projected greater competence/flirtatiousness. In addition, a single item, ranging from 1 (not at all) to 6 (a lot), assessed the partner’s Perceived Likeability.
Evaluation of the Video and Concluding Information
In the last part of the questionnaire, 4 items, ranging from 1 (not at all) to 5 (a lot), assessed how interesting, offensive, educational, and entertaining the video segment was perceived to be. Respondents were then asked to estimate the duration of the video segment (in minutes) and to briefly describe its content (which all participants did adequately). At the end of the experimental session, participants were asked to report their age and gender. Finally, they were thanked and thoroughly debriefed regarding the actual aims of the study.
Results
Perceptions of the Video and Chat Partner
The Objectifying TV video was perceived as less educational, F(2, 138) = 37.84, p < .001, η p 2 = .35, more offensive, F(2, 138) = 59.70, p < .001, η p 2 = .46, and less interesting, F(2, 138) = 11.78, p < .001, η p 2 = .15, than the other videos. In addition, the Female Professionals video was judged as less entertaining than videos in the other two conditions, F(2, 138) = 7.36, p = .001, η p 2 = .10 (see Table 1).
Descriptive Statistics and Mean Comparisons for All Measures, Study 1.
Note. n = 47 men per cell. Mean values across rows that do not share the same subscript are significantly different at the p ≤ .05 level (Bonferroni-adjusted).
We conducted a series of three one-way analyses of variance (ANOVAs) on participants’ scores of Perceived Competence, Perceived Flirtatiousness, and Perceived Likeability of the (fictitious) interaction partner, again using condition as the independent variable for each. No effects emerged either for Perceived Flirtatiousness or for Perceived Likeability (ps > .07). However, for Perceived Competence a significant main effect of condition was found, F(2, 138) = 9.77, p < .001, η p 2 = .12. As shown in Table 1, post hoc tests with a Bonferroni correction showed that the interaction partner was perceived as more competent in the Female Professionals condition than in the Control condition (p = .05) and in the Objectifying TV condition (p < .001). Importantly, the interaction partner was perceived to be less competent in the Objectifying TV condition than in the Control condition (p = .05).
Harassment Proclivity
A one-way ANOVA with three conditions and exploring harassment proclivity was conducted. Results showed a significant effect of condition, F(2, 138) = 5.37, p = .006, η p 2 = .07. As shown in Table 1, post hoc tests with a Bonferroni correction showed that participants reported higher harassment proclivity after watching the Objectifying TV than after watching the Female Professionals (p = .004) or the Control segments (p = .05). As predicted, no difference was found between the Female Professionals and the Control conditions.
Gender-Harassing Conduct
Using the total number of sexist/sexual jokes sent during the chat interaction as a measure of gender-harassing conduct, we conducted a one-way ANOVA to compare the three experimental conditions. As predicted, condition significantly affected the number of sexual/sexist jokes sent, F(2, 138) = 4.84, p = .009, η p 2 = .07. Post hoc tests with a Bonferroni correction showed that participants sent a higher number of sexual/sexist jokes in the Objectifying TV condition than in the Control (p = .01) or in the Female Professionals condition (p = .05; see Table 1). As predicted, no difference was found between the Female Professionals and the Control conditions. Furthermore, and as expected across the three experimental conditions, participants’ harassment proclivity and the gender-harassing conduct scores were significantly correlated, r(141) = .33, p < .001.
Discussion
The predicted effect of exposure to objectified female models on men’s proclivity to sexual coerce women and actual gender-harassing behavior was demonstrated in Study 1. As expected, men reported a higher intention to engage in sexual coercion (assessed by the LSH scale) and were more likely to send sexist/sexual jokes to a female chat-line partner after watching Objectifying TV as compared to a Control video. Furthermore, these outcomes did not extend to a nonobjectifying video that featured professional women. Importantly, in line with our hypothesis that harassment would increase specifically in reaction to the mere exposure to objectifying, but not to other portrayals of women in TV, no difference was found between the Female Professionals and the Control conditions—either on participants’ harassment proclivity or on gender-harassing conduct scores.
In addition, the present results showed that viewers’ impressions of real individuals may be affected by the type of TV just watched. It was found that exposure to the Objectifying TV video significantly affected not only participants’ harassing conduct but also participants’ impression of the (fictitious) interaction partner. Although there was no logical connection between the women portrayed in the objectifying TV clip and the unknown female chat partner, participants perceived the interaction partner as less competent after watching the Objectifying TV clip as compared to the other two conditions. In line with previous research drawing from priming and schema theories (Hansen & Hansen, 1988; Rudman & Borgida, 1995), this finding suggests that exposure to TV depictions of objectified women promotes the accessibility of the construct of women as sexual objects, according to which women are not only submissive and sexually acquiescent but also incompetent (Ashmore, 1981; Ashmore & Del Boca, 1981). As a result, participants’ subsequent impressions of the (fictitious) interaction partner were distorted in the direction of the activated construct.
Study 2
Results from Study 1 showed that male participants report greater intentions to engage in sexual coercion and are more likely to engage in gender-harassing behavior after viewing TV depictions of objectified women as compared to nature images or nonobjectified professional women. But how does this change in behavioral intentions and in actual behavior come about? In line with the spreading-activation theory (Collins & Loftus, 1975), one possibility is that, for a male audience, Objectifying TV promotes the accessibility of women as sexual objects and also related constructs, such as specific social norms about how men are expected to behave. As a consequence, exposure to Objectifying TV programs would increase the conformity of men to traditional masculine norms, which, in turn, would affect gender-harassing conduct (Levant & Richmond, 2007). Thus, in Study 2 we tested the hypothesis that men would conform to masculine gender-role norms more in the Objectifying TV condition than in the Control condition. Importantly, in line with the findings of Study 1, we expected that the mere exposure to the Female Professionals video clip would not represent a sufficient threat to participants’ masculine identity to lead to a significant increase in their adherence to traditional masculine norms as compared to the Control condition.
Participants’ conformity to traditional masculine norms was assessed through the Male Role Norms Inventory–Revised (MRNI-R; Levant et al., 2007) because some of the norms measured by the MRNI-R have been found to be common throughout most of the world and to be associated with attitudes conducive to sexual harassment (Levant & Richmond, 2007). Specifically, we used five dimensions of the MRNI-R: Non-relational Attitudes toward Sex, Aggression, Dominance, Extreme Self-reliance, and Restrictive Emotionality. Of primary importance for our study was Non-relational Attitudes toward Sex, which was expected to be the dimension most directly linked to the objectification of women. The dimension of Aggression was also expected to be involved in harassing behavior (Pryor, 1987), and Dominance was used in line with the idea that sex and power are closely linked (Bargh, Raymond, Pryor, & Strack, 1995). Thus the hypothesized increase in conformity to the three traditional masculine norms of Non-relational Attitudes toward Sex, Aggression, and Dominance was expected to mediate the relation between experimental condition and gender-harassing behavior (again assessed via the Computer Harassment Paradigm). Because the dimensions of Extreme Self-reliance and Restrictive Emotionality were not specifically related with sexual harassment, they served as indices of discriminant validity and, as such, they were expected to be unaffected by video condition.
Method
Participants
A new sample of residents of Northern Italy was asked to participate in a study investigating the short-term memory of TV scenes. As in Study 1, participants completed the relevant tasks individually on a laptop computer in a quiet room either at the facilities of the university (in the case of students) or at their homes (in the case of the nonstudent subsample). The sample consisted of 120 men (average age 24.5, SD = 5.44, range = 18–49 years) including 67 (56%) university students and 53 (44%) employed. Participants volunteered to participate in the study without receiving monetary compensation.
Procedure and Materials
As in Study 1, participants were recruited by one of the two female experimenters and were told that the aim of the study was to investigate their short-term memory of TV images. An equal number of participants was randomly assigned to one of the three experimental conditions (40 participants in each condition), by inviting them to watch the Objectifying TV, the Female Professionals, or the Control video. They were then asked to complete two alleged distractor tasks (in reality, our main dependent variables) and a questionnaire.
The first “distractor” task was a short version of the MRNI-R (Levant et al., 2007), one of the most common measures used to assess individuals’ endorsement of seven norms of traditional masculinity ideology. Given the lack of previous empirical work using the MRNI-R with Italian men, the items of the inventory used in the present study were translated into Italian. In the present abridged form, the MRNI-R contained 31 items measuring five dimensions: Non-relational Attitudes toward Sex (6 items; e.g., “A man should always be ready for sex”; α = .82), Aggression (6 items; e.g., “I think a man should try to be physically tough, even if he’s not”; α = .77), Dominance (6 items; e.g., “A man should always be a boss”; α = .81), Extreme Self-reliance (5 items; e.g., “A man should know how to repair his car in case of damage”; α = .60), and Restrictive Emotionality (8 items; e.g., “A man should not react if someone cries”; α = .82). Participants rated each item on a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). Single averaged scores were calculated for each of the five dimensions of interest, with higher scores indicating greater adherence to each masculinity norm. For each participant, we also calculated a score of overall conformity (i.e., Overall Masculine Ideology score) to the three masculine norms expected to be the most directly linked to harassment, that is Non-relational Attitudes toward Sex, Dominance, and Aggression.
Given that the administration of the MRNI-R took considerable time (approximately 10 minutes), which might weaken the effects of the experimental manipulation, all participants subsequently watched again a brief (30 seconds) segment of the same TV video that they had watched at the beginning of the experiment. This assured that the temporal distance between TV exposure and the primary dependent variable (chat) was identical to that of Study 1. This procedure was justified to the participants by claiming that it served to reactivate their memory (see cover story). Afterward, as in Study 1, they were invited to participate in the computer chat. Finally, participants completed the same questionnaire as in Study 1, including 11 items assessing the Perceived Competence (5 items), Perceived Flirtatiousness (5 items), and Perceived Likeability (1 item) of the (fictitious) interaction partner; 5 items assessing the Evaluation of the videos; and a request for a brief description of the content of the video. At the end of the experimental session, participants were asked to report their age, gender, and profession. Participants were then thanked and thoroughly debriefed regarding the true aims of the study.
Results
Perceptions of the Video and Chat Partner
As shown in Table 2, the Objectifying video was perceived as less educational, F(2, 117) = 26.09, p < .001, η p 2 = .31, more offensive, F(2, 117) = 31.35, p < .001, η p 2 = .35, and less interesting, F(2, 117) = 4.84, p < .010, η p 2 = .08, than the other two videos. The Female Professionals video was judged less entertaining than the other two conditions, F(2, 117) = 8.11, p = .001, η p 2 = .12. As for the chat partner, separate one-way ANOVAs were conducted on participants’ Perceived Competence, Perceived Flirtatiousness, and Perceived Likeability of the interaction partner, using condition (Objectifying TV, Female Professionals, Control) as the independent variable. As can be seen in Table 2, no significant effects emerged (all ps > .12).
Descriptive Statistics and Mean Comparisons for All Measures, Study 2.
Note. MRNI-R = Male Role Norms Inventory–Revised. n = 40 men per cell. Mean values within rows that do not share the same subscript are significantly different at the p ≤ .05 level (Bonferroni-adjusted).
aOverall Masculine Ideology score averaged the MRNI-R subscales of Non-relational Attitudes toward Sex, Aggression, and Dominance.
Gender-Harassing Conduct
The total number of sexist/sexual jokes sent by each participant during the chat interaction was subjected to a one-way ANOVA, using condition as the independent variable. As predicted, a significant effect of condition was found, F(2, 117) = 3.26, p = .04, η p 2 = .05. As shown in Table 2, participants harassed more in the Objectifying TV condition than in the Control (p = .04), with harassment in the Female Professionals condition occupying an intermediate position that did not differ reliably from the other two conditions. Thus, confirming results from Study 1, in the present study participants were more likely to manifest gender-harassing behavior after watching TV depictions of objectified women as compared to nature images, and no difference was found between the Female Professionals and the Control condition. However, different from Study 1, the difference between Objectifying TV and Female Professionals was not significant (p = .20).
Masculinity Norms
Zero-order correlations among the five dimensions of the MRNI-R we assessed are presented in Table 3. Analyses were first conducted on the score of Overall Masculine Ideology (a composite of three of the subscales), followed by specific analyses on each of the five subscales of the MRNI-R. The Overall Masculine Ideology score was analyzed with a one-way ANOVA, using condition as the independent variable. As shown in Table 2, post hoc tests with a Bonferroni correction showed that participants’ score of Overall Masculine Ideology, F(2, 117) = 5.72, p = .004, η p 2 = .09, was higher after watching the Objectifying TV than after watching the Control neutral video (p = .003). Participants’ conformity to Overall Masculine Ideology in the Female Professionals condition fell in the middle, such that no difference was found with respect to either the Objectifying TV or the Control condition.
Correlations Among MRNI-R Dimensions, Study 2.
Note. MRNI-R = Male Role Norms Inventory–Revised.
*p < .001.
We next conducted separate one-way ANOVAs for each subscale of the MRNI-R using condition as independent variable. A significant effect of condition was found on Non-relational Attitudes toward Sex, F(2, 117) = 7.03, p = .001, η p 2 = .11, indicating that participants’ conformity to masculine norms regarding Non-relational Attitudes toward Sex (see Table 2) was higher after watching the Objectifying TV clip than after watching the Control clip (p < .001). No difference emerged between the Objectifying TV and the Female Professionals conditions as well as between the Female Professionals and the Control conditions.
As shown in Table 2, participants manifested higher conformity to masculine norms regarding Dominance, F(2, 117) = 4.32, p = .01, η p 2 = .07, after watching the Objectifying TV clip as compared to the Control condition (p = .008), with the Female Professionals condition occupying an intermediate position that did not differ reliably from the other two conditions. A similar pattern of results emerged for participants’ conformity to masculine norms regarding Aggression, F(2, 117) = 3.48, p = .03, η p 2 = .06. Participants showed higher conformity to masculine norms regarding Aggression after watching the Objectifying TV clip as compared to the Control condition (p = .03), with no difference between the Objectifying TV and the Female Professionals or between Female Professionals and the Control conditions. As predicted, no significant main effects of condition were found for both Extreme Self-reliance and Restrictive Emotionality (ps > .07).
Overall, participants conformed more to masculine gender-role norms after exposure to TV messages that objectify women than after watching a neutral Control video. Moreover, in all analyses with the three MRNI-R constructs we expected to be related to sexual harassment (Non-relational Attitudes toward Sex, Dominance, and Aggression), adherence to traditional masculinity norms in the Female Professionals condition consistently fell between the Objectifying TV and the Control conditions. Thus, confirming our hypothesis, the mere exposure to the Female Professionals video did not represent a sufficient threat to participants’ masculine identity to predict an increase in their adherence to traditional masculine norms as compared to the Control condition.
Mediation Analyses
Given the parallel patterns of relationships across our three video conditions (Objectifying TV, Female Professionals, and Control) with both gender-harassing conduct and the three expected masculine norms, we tested whether participants’ conformity to Non-relational Attitudes toward Sex, Aggression, and Dominance had a mediating role in the prediction of harassment as a function of condition. Following Preacher and Hayes’ (2008) recommendations, we included these three mediators in the same model. Given that our predictor (video condition) was a categorical variable with three levels, we created two dummy-coded variables, with the Objectifying TV condition as the reference group (Hayes & Preacher, 2013). Specifically, Contrast 1 tested the effect of the Objectifying TV (coded 0) versus Control condition (coded 1), with the Female Professionals condition also coded 0. Contrast 2 tested for the residual difference between the Objectifying TV (coded 0) and the Female Professionals condition (coded 1), with the Control condition also coded 0.
Consistent with the univariate analyses reported above, the effect of Contrast 1 on gender-harassing conduct was significant, β = −.25, t(78) = −2.45, p = .01, whereas the effect of Contrast 2 fell short of significance, β = −.21, t(78) = −1.84, p = .07. Similarly, the effect of Contrast 1 on all three mediators was significant: for Non-relational Attitudes toward Sex, β = −.39, t(78) = −3.74, p < .001; for Dominance, β = −.33, t(78) = −3.07, p = .003; and for Aggression, β = −.29, t(78) = −2.63, p = .009 (see Figure 1). However, the effect of Contrast 2 was significant only on Non-relational Attitudes toward Sex, β = −.23, t(78) = −2.12, p = .04; this contrast was not significant for Dominance, β = −.11, t(78) = −1.06, p = .29, and for Aggression, β = −.12, t(78) = −1.10, p = .27. Moreover, when the three mediators and the two contrasts were entered simultaneously in the model predicting gender-harassing conduct, only the effect of Non-relational Attitudes toward Sex was significant, β = .35, t(77) = 3.11, p = .003, indicating that participants’ conformity to masculine norms regarding Non-relational Attitudes toward Sex, but not Dominance or Aggression, affected the number of sexist/sexual jokes sent to the chat partner. Importantly, neither the effect of Contrast 1 (Objectifying TV vs. Control condition) nor the effect of Contrast 2 (Objectifying TV vs. Female Professionals condition) was significant (ps > .20).
Results of mediation analysis testing the indirect effects of video condition (0 = Objectifying TV, 1 = Control) on gender-harassing conduct via Non-relational Attitudes toward Sex, Dominance, and Aggression. *p < .05. **p < .01. ***p < .001.
To test for significance of the indirect effects of Contrast 1 and Contrast 2 on gender-harassing conduct through each of the three mediators (Non-relational Attitudes toward Sex, Dominance, and Aggression), we calculated bias-corrected 95% confidence intervals (CIs) using a bootstrapping technique (Preacher & Hayes, 2008). Because the null hypothesis of no mediation states that the indirect effect is zero, the null hypothesis is rejected when the CI does not include zero. For Contrast 1, the CIs (with 5,000 resamples) for the estimate of the indirect effect on gender-harassing conduct through both Dominance and Aggression included zero: Dominance’s 95% CI [−0.19, 0.19]; Aggression’s 95% CI [−0.12, 0.20]. Conversely, the CI (with 5,000 resamples) for the estimate of the indirect effect on gender-harassing conduct through Non-relational Attitudes toward Sex did not include zero (95% CI [−0.51, −0.01]), thus confirming that Non-relational Attitudes toward Sex mediated the relation between Objectifying TV versus Control condition and participants’ gender-harassing conduct.
For Contrast 2, the CIs (with 5,000 resamples) for the estimate of the indirect effect on gender-harassing conduct through all the three mediators included zero (Non-relational Attitudes toward Sex’s 95% CI [−0.37, 0.009]; Dominance’s 95% CI [−0.10, 0.10]; Aggression’s 95% CI [−0.06, 0.1]), thus showing that none of the three masculine norms mediated the relation between Objectifying TV versus Female Professionals condition and participants’ gender-harassing conduct. In sum, these results showed that Non-relational Attitudes toward Sex mediated the effect of Objectifying TV (vs. Control) on harassing behavior, whereas Dominance and Aggression were not significant mediators.
Discussion
Replicating Study 1, male participants were more likely to harass an unknown female chat partner after viewing Objectifying TV than a neutral nature video. More importantly, we also tested the role that masculinity norms play in this context. As hypothesized, our results showed that exposure to Objectifying TV programs (vs. neutral Control programs) increased participants’ conformity to traditional masculinity norms regarding Non-relational Attitudes toward Sex, Dominance, and Aggression, whereas no effect was found for those masculine norms that are not logically linked to sexual harassment, namely Restrictive Emotionality and Extreme Self-reliance. Thus, only the specific masculine norms related to sexual behavior, dominance, and aggression were affected by the Objectifying TV video. Importantly, only the norm of Non-relational Attitudes toward Sex was found to be responsible for the greater gender-harassing behavior in the Objectifying TV condition, whereas Dominance and Aggression norms were not significant mediators.
Results from the present study also supported the hypothesis that harassment would increase in reaction to the mere exposure to objectifying content but not to other portrayals of women in TV. Indeed, although no significant difference emerged between the Objectifying TV and the Female Professionals conditions, nonetheless, the mere exposure to a clip showing competent women in professional roles also did not lead to increased conformity to the masculine norms of Non-relational Attitudes toward Sex, Dominance, and Aggression as compared to a nature video. Importantly, men in the Female Professionals condition did not engage in gender-harassing behavior any more than in the Control condition.
At first sight, the last finding seems to contradict prior research on gender harassment suggesting that, to protect or restore a threatened gender identity, men tend to harass independent women who do not conform to gender stereotypes (Dall’Ara & Maass, 2000; Maass et al., 2003; see Maass & Cadinu, 2006, for an overview) or those who are better than them on a “masculine” task (Hitlan et al., 2009, Study 2). However, this apparent contradiction can be explained by considering the different experimental manipulations employed in the two lines of research. Whereas in Dall’Ara and Maass (2000), in Maass, Cadinu, Guarnieri, and Grasselli (2003), and in Hitlan et al. (2009) participants interacted with the woman who explicitly expressed a feminist view or who outperformed them, in the present studies participants simply watched a TV video of women who performed successfully in typically male professions without directly threatening men with their superior performance. Then, our participants interacted with another woman who had no connection to the female professionals. Thus, even if men may experience female professionals as threatening when they interact directly with them (Maass et al., 2013 for an overview), the current research suggests that the exposure to media portrayals of female professionals may not be sufficient to pose a personal identity threat to men or to increase the likelihood of gender harassment. In sum, the results of Study 2 showed that visually portraying women as objects may encourage men to endorse traditional masculine norms, which, in turn, affect the way they treat available women, even though these generalized women have no direct connection to the objectified women portrayed on TV.
General Discussion
The aim of the present research was to explore whether exposure to TV images portraying women as sexual objects may increase the likelihood of harassing conduct. Both studies showed that participants were more likely to harass a readily available woman chat partner after watching a TV video in which women were objectified. Male participants were more likely to engage in gender harassment, here operationalized as sending sexist/sexual jokes to an unknown female interaction partner, in the Objectifying TV as compared to the neutral video Control condition. Moreover, Study 1 provides evidence that Objectifying TV also may encourage more blatant forms of harassment, given that men in our first study reported a greater intention to engage in sexual coercion after observing women being treated as objects in a TV video.
Importantly, our studies also speak to the processes through which Objectifying TV programs may encourage harassment. In Study 2, we found that exposure to Objectifying TV programs promotes the accessibility of specific norms regarding traditional masculinity. Of the dimensions of the MRNI-R hypothesized to be specifically related with harassing behavior (i.e., Non-relational Attitudes toward Sex, Aggression, and Dominance) only the Non-relational Attitudes toward Sex dimension, one of the basic dimensions of the male role (Levant et al., 1992), was found to play a mediating role. This may be due to the fact that Objectifying TV programs generally show passive and compliant women, explicitly portrayed as objects of sexual gratification, thus likely to evoke non-relational sexual attitudes that do not require consensual attraction or intimacy. Specifically in this condition, conformity to Non-relational Attitudes toward Sex led to gender-harassing behavior. This finding is in line with previous theories postulating that a lack of reciprocal attraction is one of the defining features of harassment (see DeSouza, 2011, for an overview of current legal definitions), thus making Non-relational Attitudes toward Sex a likely mechanism driving sexual harassment.
Finally, although not originally predicted, a carryover effect emerged in Study 1 such that, after watching the objectifying portrayal of women, men perceived their female interaction partner as less competent. In line with schema theories (Hansen & Hansen, 1988; Rudman & Borgida, 1995), the present finding suggests that the model of woman activated by the Objectifying TV video and characterized by low intellectual ability was generalized to women as a whole, including the unknown chat partner. However, given that this finding was not replicated in Study 2 and that we have no process evidence to support our interpretation, a direct test of this process should be the goal of future investigation.
Limitations and Future Directions
Given the wide use and great impact of television, the present study tested the effects of objectifying TV content on sexual harassment. One obvious limitation of the present study is that TV was the specific focus of the study, and therefore its results may not be generalizable to other media. Notwithstanding this limitation, there is reason to speculate that objectifying portrayals of women are likely to encourage harassing behavior regardless of the specific modus (e.g., advertisements) and medium (e.g., magazines, video games) through which they are transmitted. As a case in point, research has demonstrated that long-term exposure to violent video games is correlated with greater tolerance of sexual harassment and greater rape myth acceptance (Dill, Brown, & Collins, 2008). More recent research has also shown that even short-term exposure to video game violence has tangible negative effects, in this case on moral reasoning (Gabbiadini, Andrighetto, & Volpato, 2012). Given that female characters in video games are often objectified or mistreated (e.g., the videogame Grand Theft Auto allows players to earn points by raping women), the latter results strongly suggest that playing with objectifying (vs. nonobjectifying) video games might have detrimental effects also on the cognitive representation of women as a category and on subsequent (harassing) behavior towards women in general.
Another limitation of the present study is also concerned with the external validity of the present findings. Specifically, gender harassment was demonstrated during online chats between male participants and a fictitious female target. Although these effects may not generalize to real interaction contexts, it should be noticed that, especially in the Western world, virtual exchanges have become an important part of “real” interactions between people, commonly experienced both during work and leisure time. Still concerning external validity, the main findings of the present study were obtained after a 3-minute video constructed by assembling segments of various Italian TV programs, thus making up a rather artificial condensation of real TV programs. Nevertheless, because a large part of Italian TV programs frequently proposes such scenarios, the present results are likely to be an underestimation of daily exposure to objectified women via Italian TV programs.
In addition, a specific feature of the present study is that participants were all volunteers due to the fact that it is uncommon to pay participants in Italy. It remains to be seen whether similar results would be obtained when paying or rewarding participants as is often done in other countries (e.g., the United States). We suspect that the present effects might even increase in the latter case due to experimental demands. A final potential limit of our research is that the MRNI-R had not previously been validated in Italy so its psychometric properties and predictive validity are largely unknown. However, we consider it encouraging that the structure emerging from factor analysis was quite similar to that of the original scale employed in the United States.
Practice Implications
At the societal level, sexual harassment is influenced by social norms, but the dissemination of such norms depends to a large degree on broadly available channels of communication. By linking media culture, social norms, and sexual harassment, our results show that TV exposure to objectified women—even if for only 3 minutes—enhances the endorsement of traditional masculine norms, which in turn promotes sexual harassment behavior. Thus, the present study suggests that restrictive versus permissive norms regarding sexual harassment may be communicated through television programs, underlining both the dangers and the potential of media in shaping gender norms. These findings have important implications for both policy and practice. First of all, they attest to the responsibility of the media industry in creating a cultural environment that can be either derogatory or respectful of individuals’ gender identity. Advocates of gender-fair media production worldwide have long argued for the need of professional ethics codes and for constant monitoring of media content (see Global Media Monitoring Project, 2013). The current set of studies provides a strong argument for this position and suggests to closely monitor objectifying media content, given its tangible effects on harassment.
Moreover, the connection between sexual harassment and media identified here is important for critical media education programs intended to increase the awareness of media consumers. School-based media literacy programs generally teach critical viewing skills and have been shown to be (partially) effective for female viewers exposed to objectifying media content (e.g., Choma, Foster, & Radford, 2007). Considering that young girls are the most vulnerable targets of sexualized media role models (Brookes & Kelly, 2009), our findings may be particularly important for those media literacy interventions that address relatively young media consumers (<14 years of age). Importantly, our studies further suggest that such media literacy programs should specifically address male participants, whose norms and behaviors may be directly affected by objectifying media content. Together, practitioners from a wide range of professions (including teachers, counselors, game developers, journalists, and advertisement specialists) should be aware that media represent a powerful vehicle for sexualized social norms by (a) providing legitimacy to potential male harassers and (b) increasing the vulnerability of women and young girls as possible targets of sexual harassment.
Conclusion
Our research started from the observation that, over the past decades, Italian TV has become increasingly sexist (Zanardo, 2010), while at the same time sexual harassment of women has become a widespread phenomenon. Although not confined to Italy, an overwhelming diffusion of objectified female models has specifically distinguished Italian television programs in the last 10 years (Zanardo, Malfi Chindemi, & Cantu, 2009). During the same time period, there has been an increase in harassment and violence against women. A recent analysis of the Italian National Institute of Statistics (Istituto Nazionale di Statistica, 2010) revealed that 52% of Italian women have been victims of at least one episode of gender harassment or unwanted sexual attention across their life, with adolescents and young women being the prime targets. The same report has estimated that 20% of Italian women aged 14 to 65 have been personally harassed at work or, more generally, have personally experienced some form of sexual harassment at least one time between 2006 and 2009. The Italian National Institute of Statistics has also calculated that both subtle and blatant harassing behaviors have increased steadily in Italy in the last years. For example, the prevalence of verbally harassing behaviors has shown a tendency to increase from 25% in 2002 to 28% in 2009. At the same time, a report by Telefono Rosa (help phone line for women; Corriere della Sera, 2012) revealed that, from 2010 to 2011, physical attack charges filed by women have increased by 34%, and rape charges have increased greatly, from 430 to 578 episodes. Such statistics document that in Italy sexual harassment and violence against women constitute a very serious and widespread phenomenon.
The main question investigated here is whether the two phenomena, objectifying TV and harassment, are causally linked. Going beyond the simple question of co-occurrence, we therefore designed two experimental studies that used video materials of actual TV shows, thus preserving a certain level of mundane realism and external validity but also allowing to investigate the causal relationship between TV exposure and harassing conduct. Together, our findings suggest that degrading TV portrayals of women play a causal role in both gender harassment and sexual-coercion intentions and that this relationship with gender harassment is at least in part attributable to a shift in masculinity norms. In other words, objectifying TV productions seem to create a normative context that is conducive to gender harassment, attesting to the responsibility of the mass media in promoting a climate that is hostile toward women. Because the endorsement of traditional male ideology is a crucial link between objectifying media and gender harassment, future research should investigate those social and cultural factors that play a fundamental role in shaping those norms and values. These conceptions go beyond harassment by men but involve society at large, including responsibilities of women and men to promote a different set of values. On a positive note then, the same TV programs can also promote a change in social and cultural norms, including those related to masculinity, which suggests that they can be used as means to reduce sexual harassment behavior and more generally to promote respect for women and men.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.