Expanding Models Testing Media Contributions to Self-Sexualization

Although everyday exposure to mainstream media is theorized to be associated with women’s self-objectification, empirical tests of this association have yielded mixed results. There are several possible explanations for these mixed results. For example, it may be that mainstream media relate to women’s broader tendency to self-sexualize rather than to self-objectify, that newer forms of media (i.e., social media) should be incorporated into these models, or that these associations may operate indirectly. Accordingly, the current study examines how traditional media and Facebook investment are associated with women’s tendency to self-sexualize and tests whether these associations are mediated by the view of women as sexual objects, internalization of the thin ideal, and perceived pressure to conform to this ideal. Structural equation modeling demonstrated that magazine and Facebook investment predicted only internalization of the thin ideal, whereas reality television and sitcom exposure predicted all three mediators. Media use was indirectly related to self-sexualization through internalization of the thin ideal and the view of women as sexual objects. These results support the value of incorporating broader definitions of self-sexualization, diverse forms of media, and potential mediators into future models.


Original Research
Objectification theorists (Fredrickson & Roberts, 1997;McKinley & Hyde, 1996) propose that women in Western society must navigate a culture that views them not as active agents but rather as sexual objects, a collection of body parts whose value is determined by their physical appearance and sexual attractiveness to others. Over time, women come to internalize this third-person perspective and engage in selfobjectification, valuing the external, observable characteristics of their body more than the internal, unobservable characteristics (Noll & Fredrickson, 1998). As part of this process, women are believed to engage in body surveillance or body monitoring, whereby they continually assess the extent to which their external appearance aligns with culturally valued ideals. The maintenance of this so-called objectified body consciousness (McKinley & Hyde, 1996) has been theoretically and empirically linked to reduced awareness of internal bodily states, reduced flow (i.e., mental focus or concentration), disordered eating, and sexual dysfunction (Fredrickson & Roberts, 1997;Moradi & Huang, 2008).
One of the many socialization agents believed to contribute to women's development of an objectified body consciousness are the mainstream media, the content of which frequently equates women's value as people to their sexual attractiveness. The power of these portrayals in triggering self-objectification may come both from their prevalence and their accessibility. Indeed, sexually objectifying portrayals have been observed at high rates in primetime and reality television, in music videos, and in advertising (Aubrey & Frisby, 2011;Flynn, Park, Morin, & Stana, 2015;Messineo, 2008;Smith, Choueiti, Prescott, & Pieper, 2012;Turner, 2011). Moreover, adolescents and emerging adults, respectively, spend on average 8 and 12 hr consuming media each day, creating circumstances in which regular exposure to sexually objectifying images of women is likely chronic and virtually unavoidable (Rideout, Foehr, & Roberts, 2010;Ward, Seabrook, Giaccardi, & Zuo, 2015).
However, few studies have actually examined the direct link between everyday media use and self-objectification among youth (see Ward, 2016, for a review), and the pattern of results that has emerged is inconsistent. Within this small body of literature, self-objectification is typically operationalized as either the tendency to surveil one's own body (McKinley & Hyde, 1996) or as the tendency to value observable, appearance-related aspects of the body more than unobservable, competence-related aspects of the body (Noll & Fredrickson, 1998). A few studies have successfully linked composite measures of media consumption to valuing body appearance over body competence (Aubrey, 2006b;Nowatzki & Morry, 2009). Other studies have found associations between sexually objectifying television content and valuing body appearance over body competence (Aubrey, 2006a;Vandenbosch, Muise, Eggermont, & Impett, 2015). Finally, several studies have linked consumption of sexually objectifying magazines (particularly beauty/fashion magazines or other appearance-oriented magazines) both to valuing body appearance over body competence and to body surveillance (Aubrey, 2007;Fardouly, Diedrichs, Vartanian, & Halliwell, 2015;Morry & Staska, 2001;Vandenbosch & Eggermont, 2012, 2015. However, there are also studies that have demonstrated null results. For example, some studies have failed to find a link between sexually objectifying television and selfobjectification Vandenbosch & Eggermont, 2012;Zurbriggen, Ramsey, & Jaworski, 2011) or have linked media exposure to one component of sexual objectification but not others (i.e., valuing body appearance over body competence but not body surveillance; Aubrey, 2006aAubrey, , 2007. Other studies have found inconsistent results when examining fitness magazines specifically or sexually objectifying magazines more generally (Morry & Staska, 2001).

One Potential Culprit: Self-Objectification Versus Self-Sexualization
There are several possible explanations for the inconsistent results observed in this literature. One potential explanation is the fact that these studies emphasize self-objectification alone as the outcome of interest. Although these studies build upon prior literature by measuring body surveillance and self-objectification (operationalized as valuing body appearance over body competence), it may be that selfobjectification alone does not best capture how women respond when exposed to sexually objectified women in the media. Instead, it may be the case that chronic exposure to sexually objectified portrayals in mainstream media relates to women's tendency not only to self-objectify, but also to their tendency to self-sexualize, which includes elements that extend above and beyond self-objectification. The American Psychological Association [APA] Task Force on the Sexualization of Girls (APA, 2007) defines sexualization as occurring when a person's value comes only from his or her sexual appeal or behavior, to the exclusion of other characteristics; OR a person is held to a standard that equates physical attractiveness (narrowly defined) with being sexy; OR a person is sexually objectified-that is, made into a thing for others' sexual use; OR sexuality is inappropriately imposed upon a person. (p. 1) Under this definition, self-objectification is just one component of self-sexualization, which may more accurately represent how women respond to chronic exposure to sexual objectification in the media. Women may not only learn to self-objectify from the media but may also internalize messages about how their sex appeal contributes to their sense of self-worth. For this reason, it is possible that studies examining self-sexualization may find more consistent links than those examining self-objectification alone.
Two measures commonly employed in the examination of this broader construct of self-sexualization are the Enjoyment of Sexualization Scale (ESS; Liss, Erchull, & Ramsey, 2011) and the Sexual Appeal subscale of the Self-Worth Scale (SASW; Gordon & Ward, 2000). Following the definition of sexualization outlined by the APA Task Force on the Sexualization of Girls (APA, 2007), these measures work well to capture both women's potential enjoyment in being valued for their sexiness and women's valuing of themselves primarily for their sexual appeal to others. The ESS was designed to assess the extent to which women seek to and enjoy emphasizing their sexiness, two messages that are commonly purveyed through mainstream media. In developing the ESS, the authors found significant correlations between enjoyment of sexualization, body surveillance, self-objectification, and engagement in self-sexualizing behaviors, though these significant correlations do not exceed .5. Together, these data suggest that these items reflect similar, but not completely overlapping, constructs.
Similarly, the SASW has been used successfully as part of a latent variable capturing self-sexualization in three studies (Manago, Ward, Lemm, Reed, & Seabrook, 2015;Ward, Seabrook, Grower, Giaccardi, & Lippman, 2018;Ward, Seabrook, Manago, & Reed, 2016). Items on this scale reflect the extent to which an individual's self-worth is grounded in her or his romantic or sexual appeal to others. In each study in which it was used, correlations between SASW, ESS, and Body Surveillance were significant and ranged from .41 to .50. Again, the fact that correlations between these variables did not exceed .5 suggests that this scale may be capturing some unique component of self-sexualization.
Accordingly, we argue that the existing literature provides support for using these scales together to assess selfsexualization, providing a means to better understand how women may internalize different media messages and how these media messages may influence women's beliefs about themselves and their bodies. These multiple measures serve as a way of capturing the multi-dimensional construct of self-sexualization. If models assessing self-sexualization rather than self-objectification were to find more consistent results, this outcome would lend credence to the notion that including aspects of sexualization is an important next step in clarifying and understanding the link between everyday mainstream media consumption and women's feelings about their bodies.

A Second Potential Culprit: Traditional Media Versus Social Media
The second potential explanation for the inconsistent results found in the existing literature is the conceptualization and measurement of the media involved. Specifically, it may be important to move beyond measures of activity on social networking sites (SNS) and instead consider the extent to which a person is invested or involved with social media. Social media are increasingly vital components of individuals' media diets. At the time of data collection, Facebook was the most popular SNS among teens and young adults; estimates indicate that among 18-to 29-year-old women, 87% used Facebook in 2014 (Duggan, Ellison, Lampe, Lenhart, & Madden, 2015;Lenhart, 2015). More recent data show that 43% of teens use Facebook, with 15% reporting that they use it the most compared with several other SNS (Common Sense Media, 2018). Not only is social media use prevalent, but these media differ fundamentally from traditional media in their interactivity. Users not only consume but also construct their social media environments, suggesting that these media may have a particularly potent impact on women's tendency to self-sexualize. Moreover, the "characters" on social media consist of an individual's peers, perhaps serving to reinforce the notion that the ideals presented therein are particularly relevant for the self to obtain (Perloff, 2014). For these reasons, users of SNS may be particularly affected by the content.
Early research in this domain examined whether the amount of time spent using social media affected women's feelings about their body and their tendencies to self-objectify. Here, findings demonstrated that women who spent more time on SNS such as Facebook and Myspace reported greater endorsement of the thin ideal, a greater tendency to engage in social comparison, more body dissatisfaction, and higher levels of body surveillance when compared with women who spent less time on SNS Holland & Tiggemann, 2016;Tiggemann & Slater, 2014, 2015. Of particular note, some of this research implicated SNS as being more influential than traditional media in their potential associations with self-objectification. For example, Vandenbosch and Eggermont (2012) found that in a model with both traditional media and SNS, only the use of SNS was directly related to self-objectification. Girls in this study who reported viewing higher levels of sexually objectifying content on their feeds also reported higher endorsement of Western beauty ideals, a greater valuation of appearance in their self-worth, and higher levels body surveillance.
Other research has examined how different SNS or different features of these SNS might be associated with women's self-objectification tendencies (e.g., differential influences of photo-based SNS vs. text-based SNS; posting photos vs. status updates; Manago et al., 2015;Mingoia, Hutchinson, Wilson, & Gleaves, 2017;Saunders & Eaton, 2018). For example, in the case of photo-based activity, research has illustrated that women who report engaging in more photoheavy activity online also report higher levels of appearance self-worth, greater endorsement of the thin ideal, higher levels of body shame, a greater tendency to surveil their bodies, and are more likely to incorporate self-sexualizing elements in their own photos (based on profiles created in the context of an experiment; R. Cohen, Newton-John, & Slater, 2017;de Vries & Peter, 2013;Salomon & Brown, 2018). Women who reported viewing a lot of photos on SNS also reported greater weight dissatisfaction, stronger endorsement of the thin ideal, and higher levels self-objectification (Meier & Gray, 2014). Other work has shown that using SNS to monitor attractive peers, or having appearance-oriented conversations with peers, is linked to adolescents' self-reported tendency to self-objectify, to surveil their bodies, and to selfsexualize (Trekels, Ward, & Eggermont, 2018;.
This small body of work lends supports to the idea that SNS may provide particularly relevant models of sexually objectified women, even more so than other types of media. Given this pattern of results and the unique way in which women can engage with social media, we feel that it is important to incorporate not only social media usage but also social media investment into models linking media use, selfobjectification, and self-sexualization. By including diverse forms of media, researchers may better capture individuals' real-world media diets and thus construct a richer understanding of how exposure to sexually objectifying media relates to individuals' tendencies to self-sexualize.

Third Potential Culprit: Direct Versus Mediated Connections
As a final possible explanation for the inconsistent results in the current literature, we argue that it is possible that associations between women's media use and their tendency to selfsexualize may operate indirectly; because most studies have focused on direct effects models, they may not be fully capturing the process at hand. There is both theoretical and empirical precedent in the literature to suggest that the link between media use and other body image outcomes (e.g., body dissatisfaction) may involve several potential mediators, including body-related cognitions and affect. One mediator proposed is women's internalization of the thin ideal (Rousseau, Beyens, Eggermont, & Vandenbosch, 2017;. Thin-ideal internalization has been defined as the extent to which women are aware of the importance placed on appearance in Western cultures and internalize these beliefs by endorsing and desiring to emulate these social standards (Groesz, Levine, & Murnen, 2002;Thompson & Stice, 2001).
One common model examining the potential mediating role of thin-ideal internalization is the three-step process of self-objectification proposed by Vandenbosch and Eggermont (2012). These authors propose that women consider the thin ideal presented in the media as an appropriate standard for their own appearance and come to value the observable, appearance-related aspects of their body more than the unobservable, competence-related aspects of their body. This internalizing of the thin ideal and valuing body appearance over body competence relates to women's tendency to surveil their bodies in an ongoing effort to ensure their bodies are conforming to these media-driven standards of beauty (Vandenbosch & Eggermont, 2012, 2015. This model has been supported with empirical data and highlights the potentially vital of role of internalization as a cognitive process underlying the association between media and self-sexualization, as internalization of the thin ideal was correlated with almost all types of media assessed. In addition, body surveillance was seldom directly related to media use and instead operated through internalization of appearance ideals. Together, this work provides empirical evidence that the media serve as purveyors of the thin ideal, and that the extent to which women internalize this ideal may serve as a powerful mechanism undergirding their tendency to self-objectify and potentially to self-sexualize. A second potential mediator is the pressure women perceive to conform to media ideals. Given the way in which Western culture values the thin ideal, it has been theorized that women are likely to perceive pressure from various sources, including the media, their friends, and their family, to conform to this ideal (Myers & Crowther, 2007;Stice, 2002). Research has supported an association between these perceived sociocultural pressures and body dissatisfaction both experimentally and in correlational research (Groesz et al., 2002;Myers & Crowther, 2007;Stice & Whitenton, 2002). In a similar vein, researchers have proposed that perceived pressure to conform to media ideals might contribute to women's tendency to self-objectify. It stands to reason that chronic exposure to sexually objectifying media creates a sense of pressure for female consumers not only to obtain the bodies presented in media but also to construct notions of the self that are similarly reductive (i.e., objectified or sexualized). Although this association has only been tested empirically in a few studies (e.g., Tylka & Hill, 2004), the work that does exist provides preliminary support to the idea that pressure contributes to women's tendency to self-objectify. For example, Knauss, Paxton, and Alsaker (2008) found that greater perceived pressure from the media (both directly and through greater thin-ideal internalization) was linked to adolescents' tendency to monitor their bodies. In line with this thinking, we would anticipate that perceived sociocultural pressure might act as an important cognitive mechanism mediating the association between media use and women's tendencies to self-objectify and self-sexualize.
A third potential mediator may be women's view of other women as sexual objects. Objectification theorists posit that women observe other women being objectified in society at large and in the media, and that they then turn this tendency to objectify upon themselves. Though explicitly stated in objectification theory, this process of turning the objectifying gaze onto the self has yet to be empirically tested. This dynamic is supported not only by objectification theory but also by cultivation theorists, who would posit that chronic media exposure cultivates a shared sense of social reality and social norms (Gerbner & Gross, 1976). Applying cultivation theory to sexualized media, it is possible that exposure to sexually objectified women in the media would cultivate a sense that women are sexual objects. This cultivated belief, particularly over time, might normalize the expectation that women objectify themselves. Prior work supports the potential role of this mediating variable, as different types of media exposure have demonstrated correlations with women's tendency to endorse objectifying statements about women (Gordon, 2008;Hust & Lei, 2008), and with men's tendency to objectify women (Ward, Vandenbosch, & Eggermont, 2015). Even preferences for certain types of media (e.g., hip hop music; explicit sexual content online) seem to cultivate a particular view of women as sex objects (ter Bogt, Engels, Bogers, & Kloosterman, 2010). Empirically testing how women's thoughts about other women influence their own tendency to self-objectify and self-sexualize would shed light on the potential cognitive mechanisms undergirding the inconsistent media-body links in prior research.

The Current Study
Accordingly, the current study examined associations between media exposure and young women's selfsexualization, operationalized as the extent to which they report monitoring their appearance, report enjoyment of sexualization by others, and report grounding their self-worth in their sexual appeal. By including these broader measures of self-sexualization as opposed to self-objectification alone, we hoped to better capture the complex ways in which media use is associated with young adult women's beliefs, attitudes, and behaviors toward their own bodies Ward et al., 2016). In addition, the current study examined whether the link between media use and self-sexualization is statistically mediated by internalization of the thin ideal, perceived pressure to achieve these ideals, and the view of other women as sexual objects. In doing so, we built upon prior theory and research supporting internalization of the thin ideal as a mechanism linking media exposure and women's tendency to self-objectify and hoped to shed light on other potential cognitive mechanisms potentially undergirding this link. We chose to include a diverse set of media measures, including three genres of television, magazine consumption, and Facebook investment. By assessing facets of young women's use of traditional media, print media, and social media, we hoped to better capture the breadth of young women's modern media environment and to investigate whether different types of media demonstrate particular potency in their connections to young women's body image and objectification. We hypothesized that: Hypothesis 1: Heavier exposure to each of the different types of media would be associated with higher selfreported internalization of the thin ideal, greater subjective feelings of pressure to conform to cultural beauty ideals, and stronger endorsement of the view that other women are sex objects. Hypothesis 2: Higher belief that women are sexual objects, higher internalization of cultural beauty ideals, and higher perceived pressure to conform to these ideals would be associated with women's self-reported selfsexualization (as assessed by their self-reported tendency to self-objectify, enjoyment of sexualization, and stronger grounding of their self-worth in their sex appeal).
In addition, this article examined whether associations between media use and self-sexualization are most accurately represented as direct relations or as indirect relations. Although the literature would suggest a mediated association may be most appropriate, these alternative models were compared, as they are founded on different theoretical assumptions about how media use relates to women's tendency to self-sexualize. Support for a mediated model would highlight the role of several potential cognitive mechanisms in the self-objectification process.
Religiosity was assessed using three items, each scored on a 1 (not at all/never) to 5 (very religious/all the time) scale. Responses to the following three items were averaged: "How religious are you?" "How often do you go to religious services (like church or temple)?" and "How often do you pray?" The sample was moderately religious, with a mean score falling slightly above the midpoint (M = 2.69, SD = 1.94). Sexual experience was assessed with a single question, "How would you describe your current level of experience with dating and sexual relationships?" Response options ranged from 0 "just starting out (virgin)" to 10 "have had several sexual relationships." Average sexual experience level for the current sample was 4.32, roughly corresponding to the response option "1-2 sexual partners (no longer a virgin)."

Measures
Media exposure. To examine magazine consumption, participants were asked to indicate how many issues per year (between 0 and 12) they read of the following 12 monthly women's magazines: Allure, Cosmopolitan, Elle, Glamour, In Style, Marie Claire, Self, Seventeen, Shape, Vanity Fair, Vogue, and Women's Health. This selection of magazines was chosen partly from past research (e.g., Ward et al., 2018) and partly because it includes lifestyle magazines, fashion magazines, and health and fitness magazines, which may better capture the breadth of women's magazine exposure. A mean score was computed across the 12 magazines (Cronbach's alpha = .73).
To assess consumption of situation comedies, we provided participants with a list of 24 situation comedy (sitcom) programs. These programs represented all sitcoms currently airing on primetime or syndication on major networks (ABC, CBS, NBC, FOX, TNT, TBS, and CW) in our market. Participants indicated how often they watched each sitcom using the following response options, scored 0 to 3: never, a little/a few episodes, sometimes/some episodes, and a lot/ most or all episodes. A mean was calculated across the 24 programs.
To assess exposure to reality and drama exposure, we provided participants with a list of 36 popular lifestyle reality programs and 16 dramas (e.g., Real Housewives, The Bachelor, How to Get Away with Murder, Shameless) currently airing on network TV or basic cable networks (e.g., MTV, VH1, Bravo; NBC, Fox). These programs were chosen based on online rankings of popular programs (e.g., tv. com). As with situation comedies, participants reported how frequently they had ever viewed the 52 different programs using the following response options scored 0 to 3: a little/a few episodes, sometimes/some episodes, and a lot/most or all episodes. A mean across all 36 reality television programs and 16 dramas was calculated.
To assess Facebook investment, participants answered six items regarding their use of Facebook and its importance to their social life. This scale was adapted from the Facebook Intensity Scale (Ellison, Steinfield, & Lampe, 2007), with two items modified. Participants responded to the items on a 3-point scale, from disagree to agree. A sample item is "I feel like my social life would suffer if Facebook were shut down." A mean score was computed, with higher scores reflecting greater Facebook investment. Cronbach's alpha for the current sample was .80, slightly higher than the alpha reported in a prior study (Cronbach's alpha = .76; Manago et al., 2015).
Self-sexualization. Following the work of Ward et al. (2016), self-sexualization was measured via three scales to capture both traditional components of self-objectification, such as body surveillance, as well as components centered more on women's prioritizing and valuing of their sexual appeal. The first scale included in the sexualization latent variable was the Surveillance subscale of the Objectified Body Consciousness Scale, which examined the extent to which individuals engage in regular body monitoring. Respondents indicated their level of agreement with each of eight items (e.g., "During the day, I think about how I look many times") using a 6-point scale anchored by strongly disagree at 1 and strongly agree at 6. A mean score was computed such that higher scores indicate greater body surveillance. This scale was originally developed and validated among three samples of undergraduate women and their middle-aged mothers. Cronbach's alpha for the surveillance subscale among the three samples was .89, .79, and .76, respectively. Similarly, Cronbach's alpha for the current sample was .77.
The second scale included in the sexualization latent variable was the ESS (Liss et al., 2011), which measures the extent to which individuals attempt to and enjoy emphasizing their own sexiness. Participants indicated their agreement with eight items using a 6-point scale (1 = strongly disagree; 6 = strongly agree). Sample items include "I love to feel sexy" and "I like showing off my body." Mean scores were computed such that higher scores indicate greater enjoyment of sexualization. The ESS was originally developed and validated among two predominantly White samples of undergraduate women. The scale demonstrated strong construct and discriminant validity. Cronbach's alpha for the development and validation samples was .85 and .86, respectively. Cronbach's alpha for the current sample was .88.
The third scale included in the sexualization latent variable was the SASW (Gordon & Ward, 2000), which assesses the extent to which participants base their self-worth on their sexual appeal. Participants were first given the prompt "How would you feel about yourself if . . ." and were asked to indicate whether they would feel better or worse about themselves in 23 situations, 13 of which reflected their sexual appeal/ attractiveness. Sample items include the following: "You were asked to be a model for a calendar featuring college students" and "A car accident left you with a permanent, two-inch scar on your face." Responses were indicated using a 7-point scale anchored by -3 ("Ugh! I would feel really horrible about myself") and +3 ("Wow! I would feel really great about myself"). Higher scores, based on absolute values, reflect placing a stronger emphasis on sexual appeal in perceptions of one's self-worth. Cronbach's alpha for the 13 items assessing sexual appeal self-worth was .81. This level of internal consistency is similar to or higher than in previous studies using this scale (Cronbach's alpha = .70, .79, and .81, in Manago et al., 2015;Ward et al., 2018;Ward et al., 2016, respectively).

Potential mediators.
To assess their views of other women as sex objects, participants completed an extended version of the Women are Sexual Objects Scale (Ward, 2002), which included 13 items scored on a 6-point scale from 1 (strongly disagree) to 6 (strongly agree). Sample items include "Women who wear tight or revealing clothing are asking to be hit on by men" and "Girls should do whatever they need to (e.g., use make up, buy attractive clothes, work out) to look good enough to attract a date/partner." Items were averaged (Cronbach's alpha = .80) so that higher scores indicated a stronger endorsement of the belief that women are sex objects.
To assess women's internalization of the thin ideal, participants responded to the nine items of the Internalization-General subscale of the Sociocultural Attitudes Toward Appearance Questionnaire (SATAQ-3; Thompson, van den Berg, Roehrig, Guarda, & Heinberg, 2004). Participants rated their agreement on a 6-point scale (1 = definitely disagree; 6 = definitely agree) to nine statements including "I compare my body to the bodies of people who are on TV" and "I wish I looked like the models in music videos." Scores were averaged such that higher scores indicated a greater internalization of the thin ideal. Cronbach's alpha for the current sample was .93.
To assess women's feelings of pressure to look like media ideals, participants completed the seven items from the Pressure subscale of the SATAQ-3 ( Thompson et al., 2004). Participants rated their agreement on a 6-point scale (1 = definitely disagree; 6 = definitely agree) to eight statements including "I've felt pressure from TV or magazines to diet" and "I've felt pressure from TV and magazines to be thin." Responses to each item were averaged (Cronbach's alpha = .92) such that higher scores indicated a greater perception of pressure from the media to conform to media ideals of beauty. Both subscales from the SATAQ-3 were developed among a sample of 175 undergraduate women (age range = 17-25 years) and validated among a sample of 195 undergraduate women (age range = 18-22 years). Cronbach's alpha for the first sample was .96 for Internalization-General and .92 for Pressure. For the second sample, Cronbach's alpha was .92 for Internalization-General and .94 for Pressure.

Procedure
Participants were recruited from an introductory psychology subject pool and received research participation credit in return for completing the survey. Students completed the survey in small groups no larger than 10 participants on tablets in an oncampus research lab. Participants were informed that the study was interested in examining their media use and beliefs about social relationships. Other measures were assessed, including measures related to participants' sexual agency, gender beliefs, and mental health, but are not conceptually relevant to the research questions under examination here. It took approximately 45 min to complete the full packet of measures. Table 1 presents the means, standard deviations, and zeroorder inter-correlations for the key variables in the study.

Descriptive Statistics
Taking into consideration the skewness of the media variables, we conducted Spearman's correlation test, which is recommended for data with tailed distributions and outliers (de Winter, Goslin, & Potter, 2016). The three self-sexualization measures (body surveillance, sex appeal self-worth, and enjoyment of sexualization) were moderately to strongly correlated, with correlations ranging from .42 to .53. These moderate-to-strong correlations suggest both conceptual overlap and conceptual distinction of these constructs in line with the aforementioned prior research. Concerning the three mediators and three self-sexualization measures, the strongest correlations were between internalization of the thin ideal and pressure. A discriminant validity test, which assesses the extent to which a latent variable is distinct from other latent variables in the model, showed that there might be some conceptual overlap between these constructs (i.e., the square root of average variance extracted [AVE] in internalization was equal to its correlation with pressure). This overlap is likely due to the fact that both are subscales of the SATAQ-3. In the next step, we checked for vertical collinearity (i.e., collinearity between the mediating variables). We retrieved tolerance and variance inflation factor (VIF) values that take into consideration only the effects of these variables on themselves (Kock & Lynn, 2012). These values did not exceed 0.2 for tolerance and 4 for VIF, indicating no multicollinearity issues between these variables and their indicator items (Kline, 1998).

Testing the Hypothesized Model
The integrative model was tested with structural equation modeling (AMOS) using the maximum likelihood method. The model controlled for age and sexual experience by employing these variables as predictors for all of the hypothesized endogenous variables. First, multiple imputation was performed as the bootstrapping method does not allow the sample to include missing data. 1 Overall, 148 respondents (22.2%) had missing data; 84 of those respondents had only one missing value that had to be imputed. All variables had less than 5% missing data. According to skewness and kurtosis information, all variables, except for magazine exposure, were normally distributed on a univariate level (i.e., skewness values did not exceed 2 and kurtosis values did not exceed 7; J. Cohen, Cohen, West, & Aiken, 2003). However, Mardia's coefficient for the data was 36.57 with a critical ratio value of 16.67, indicating that the data were significantly non-normal on a multivariate level. To resolve this issue, we decided to perform bootstrapping in AMOS, which is an adequate way to deal with non-normal data in structural equation modeling (Nevitt & Hancock, 2001).
First, a measurement model was tested assessing the fit of the self-sexualization latent variable. The media variables and mediators (i.e., internalization, perceived pressure, and objectification of women) were measured with observed variables (i.e., total scale score). Self-sexualization was measured as a second-order latent variable comprising the enjoyment of sexualization, body surveillance, and appearance self-worth. Indicators for the three first-order measures were first assigned to parcels using the item-to-construct balance technique (Little et al., 2002). Specifically, a principal components analysis was performed for each scale (separately). SPSS was asked to extract one factor for each scale, and items were assigned to three parcels based on their factor loadings. The three highest factor loadings were distributed across the three parcels followed by the next three highest factor loadings until all items were assigned to one of the three parcels. The measurement model showed a good fit of the data, χ 2 (24) = 84.244, p < .001, root mean square error of approximation (RMSEA) = .062, comparative fit index (CFI) = .982. Second, the hypothesized structural model was tested and showed an adequate fit of the data (Figure 1). The model yielded a chi-square value of 363.633 with 110 degrees of freedom, p < .001, RMSEA = .059, CFI = .946, and χ 2 /df = 3.306. The results indicated that 7.7% of the variance of internalization of appearance ideals, 7.1% of the variance of pressure to conform to media ideals, and 4.3% of the view of women as sex objects were explained by media exposure. Moreover, 58.5% of the variance of self-sexualization could be explained by media exposure, internalization of appearance ideals, pressure, and objectification of women.
Magazine exposure was only marginally significantly related to the objectification of women, but was significantly related to pressure, and internalization of appearance ideals. Both sitcom exposure and reality TV exposure were significantly related to the view of women as sex objects, pressure to conform to media ideals, and thin-ideal internalization. Interestingly, drama program exposure was unrelated to all three proposed mediators. Finally, Facebook investment was significantly associated with both perceived pressure to conform to media ideals and with thin-ideal internalization, but unrelated to the view of women as sex objects.
User-defined estimands were created to examine the indirect effects. The relative contribution of each mediator (i.e., internalization, perceived pressure, and objectification of women) to the indirect relation between media exposure and self-sexualization is presented in Table 2. Both magazine exposure and Facebook investment had a significant indirect effect on self-sexualization through internalization of appearance ideals, whereas reality television use had a significant indirect effect on self-sexualization through both internalization of appearance ideals and through cultivating the view of women as sex objects. These results indicate that the strongest mediation was through internalization.
Alternative model. To better understand the role of mediators in the relation between media exposure and self-sexualization, we also tested a model that did not include the three mediators (see Figure 2). This model also showed an adequate fit, χ 2 (80) = 215.03, p < .001, RMSEA = .051, CFI = .964, χ 2 /df = 2.689; subsequently, we performed a nested model comparison test. An unconstrained model (i.e., a model in which the relation between the media variables and selfsexualization was mediated by internalization, perceived pressure, and objectification of women) was compared with a constrained model (i.e., a model in which the paths between the media variables, the mediators, and self-sexualization were set to zero and thus not estimated). This chi-square difference test showed that the model involving the mediators had a significantly better fit (Δχ 2 = 326.795, p < .001, df = 18).

Discussion
The current study investigated associations between diverse forms of media and women's tendency to self-sexualize. We extended prior research by including an array of media measures (three genres of television, magazines, and social media), while also including broader measures of self-sexualization. The current study also examined the role of three potential mediating mechanisms: internalization of the thin ideal, perceived sociocultural pressure to conform to the thin ideal, and the view of women as sex objects. Results demonstrate that women who reported more engagement with magazines and investment in Facebook also reported higher internalization of the thin ideal and greater perceived pressure to conform to the thin ideal. Women who report greater exposure to reality television also reported higher internalization of the thin ideal, higher perceived pressure to conform to the thin ideal, and higher endorsement of the view of women as sex objects compared with women who reported lower levels of reality television exposure. Greater sitcom exposure was also related to greater internalization of the thin ideal and pressure to conform to the thin ideal, but was related to lower endorsement of the view of women as sex objects. Dramas were not associated with any of the mediators examined. Of these varied media sources, magazines and Facebook investment were indirectly related to self-sexualization through internalization of the thin ideal, whereas reality television was indirectly related to self-sexualization through internalization of the thin ideal and the view of women as sex objects. In turn, the view of women as sex objects significantly related to self-sexualization, as did internalization of appearance ideals; perceived pressure did not.
Our findings contribute to the existing body of research in several meaningful ways. First, prior studies examining links between media use and women's tendency to self-sexualize have generally examined traditional or social media separately and have not considered how women may be uniquely invested with social media. Second, past research has often examined self-objectification without consideration of other components of women's bodily experiences in the world. By focusing on women's tendency to self-sexualize more broadly rather than their tendency to self-objectify alone, we advance the understanding of how sexualized depictions of women in the media are associated with women's self-reported view of their own bodies and sense of self. Third, by including more diverse measures of women's everyday media use and a broader assessment of self-sexualization, which includes both behavioral and cognitive components, we may more accurately capture how media use may contribute to women's experiences of their bodies. Finally, the inclusion of several potential mediators in our model contributes to a more nuanced understanding of the cognitive processes that may undergird women's tendencies to self-sexualize. Although examining direct links between media exposure and selfobjectification or self-sexualization has merit, examining potential mediators provides insight into the mechanisms through which exposure to sexualized bodies in the media may influence women's view of their own bodies. In our results, heavier media consumption most consistently predicted thin-ideal internalization and perceived sociocultural pressure, as all media measures except for exposure to television dramas significantly predicted these mediators. In contrast, only sitcoms and reality television cultivated a more general view of women as sex objects. It may be that the valuing of thin bodies and pressure to conform to those ideals are nearly omnipresent in media, but that viewers are more susceptible to the specific view that women are sex objects when the women being presented as sex objects are perceived to be more "real," as in reality television, but are less susceptible to the view of women as sex objects when this objectification occurs in the context of comedy.
It is surprising that dramas did not emerge as a significant predictor of any of the proposed mediators. One potential explanation for these null findings is that some dramas may actually contain less sexually objectifying content than other TV genres, or that consumption of certain programs (e.g., Law & Order: SVU) may actually cultivate women's sense of empowerment (Hust, Marett, Lei, Ren, & Ran, 2015;Kim et al., 2007). For example, women in Hust et al.'s (2015) study who frequently watched Law & Order: SVU reported increased intentions to request consent in their sexual encounters and endorsed rape myths less strongly than those who reported less frequent viewing of Law & Order: SVU. This work also highlights another potential explanation of these null results, which is that different programs in the same genre (Law & Order: SVU versus NCIS) may have opposing effects on participants' beliefs. Some programs may serve to empower women and to diminish their view of themselves as sexual objects, whereas other programs may serve to promote and reinforce these messages.
Concerning indirect effects, both magazines and Facebook investment acted indirectly on women's tendency to selfsexualize through internalization of the thin ideal, whereas reality television operated through internalization of the thin ideal and the view of women as sex objects to influence women's tendency to self-sexualize. The associations between different types of media use and internalization of the thin ideal are consistent with Vandenbosch and Eggermont's (2012) three-step process of self-objectification. There were no indirect associations between sitcoms or dramas and self-sexualization through any of the mediators, which is also similar to Vandenbosch and Eggermont's (2012) model, which found no link between primetime television and internalization of the thin ideal. Although these results are not as consistent as expected, they do support the need for future work to clarify the relation of these mediators in women's self-sexualization tendencies, as well as other potential mediators.

Limitations, Future Directions, and Conclusions
Although the current study had several strengths, it also had several limitations. The sample in the current study was relatively homogeneous, comprising primarily White young adult women from educated backgrounds. In addition, the correlational nature of this work means we are unable to make causal claims about the associations observed in our model; it may be that having a strong tendency to selfsexualize leads women to select particular media to consume that validate these tendencies. Finally, although we included assessments of both traditional and social media, the media assessed in this study may not accurately represent women's full media diets. In the time since data collection was completed, several other social media platforms have risen in popularity. In addition, these new social media platforms may have an even greater influence on women's tendency to self-objectify or to self-sexualize, as they are image-rather than text-based (e.g., Instagram). Though content analyses do support that even more text-based SNS contain objectified content (e.g., through profile pictures), image-based platforms are environments potentially rife with sexualized self-representations.
Future studies should address these limitations by testing this mediated model in a more diverse sample of women (with regard to age and racial composition), by using longitudinal data to verify the directionality of the proposed model, and by including even more diverse media measures (e.g., exposure to music genres, movies, or music videos; other forms of social media). In addition, future work could also expand on the current model by examining other mediators and moderators. For example, future work could examine the role of different gender ideologies or endorsement of sexual scripts as potential mediators in this media-body association. Although common in multivariate analyses (e.g., Hair, Black, Babin, & Anderson, 2009;Kock & Lynn, 2012), the Internalization and Pressure scales-both derived from the SATAQ 3-were highly correlated but not collinear. High correlations between subscales could suggest that participants may not actually differentiate between constructs that researchers theoretically consider to be distinct (Hyde, 1998) and should, therefore, be handled cautiously in future studies. In the realm of moderators, perceived realism, identification with characters in the media, and motives for viewing, has each been implicated in prior work examining media effects (e.g., Gordon, 2008;Ward & Carlson, 2013). Women who perceive television content as being highly realistic, who identify strongly with characters in the media, or who use media to learn about how the world operates may be particularly prone to self-sexualize themselves, as they view media as a more appropriate model on which to base their own behaviors.
The results of the current study have several implications for the lives of women. These results suggest that chronic exposure to various types of media may influence women's beliefs both about themselves (i.e., internalization of the thin ideal and perceived pressure to conform to this ideal) and about other women (i.e., the view of other women as sex objects). Taken together, these findings highlight how exposure to sexualized portrayals of women in the media is linked with the ways in which women act to feel worthy. The reduction in women's value to aspects of the self that are external, to the exclusion of other aspects of the self, by definition precludes women the possibility of navigating the social world as agents. The current study builds our understanding of the cognitive processes underlying media's associations with women's self-sexualization tendencies by identifying several potential mechanisms. Accordingly, these findings may prove useful for developing media literacy programs or other interventions targeting these mechanisms, intervening both prior to and in the process of women's problematic cognitions before they manifest in the form of reduction via self-sexualization.

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. Note 1. In multiple imputation, each missing value is replaced by a simulated value and m plausible alternative versions of the complete dataset are generated. In our study, nine new datasets were generated and we tested our hypothesized model with each dataset. These analyses did not significantly differ and the model that is reported in the article was analyzed with the first dataset that was obtained from the MI method.