White Women Wanted? An Analysis of Gender Diversity in Social Justice Magazines

The role of media in collective action repertoires has been extensively studied, but media as an agent of socialization in social movement identity is less understood. It could be that social movement media is normalizing a particular activist identity to the exclusion of other demographics. For instance, Harper has identified white-centrism in anti-speciesist media produced by the Nonhuman Animal rights movement and supposes that this lack of diversity stunts movement potential. Using the lesser-studied Nonhuman Animal rights movement as a starting point, this study investigates two prominent Nonhuman Animal rights magazines. We compare those findings with an analysis of comparable leftist movements also known to exhibit diversity strains. A content analysis of Nonhuman Animal rights, women’s rights, and gay rights magazine covers spanning from 2000 to 2012 was undertaken to determine the manifestation of gender, race, body type, and sexualization. We find that the Nonhuman Animal rights media in our sample overwhelmingly portrays white women with a tendency toward thinness, but with low levels of sexualization as comparable to that of the other movements. All three movement samples unevenly depicted gender, overrepresented whites, and underrepresented non-thin body types.


Introduction
For social movements, media presence matters [1,2]. Exactly who is presented and how they are presented also matters. Diversity is an important strategy for effective and influential social movements, but not all movements are successful in this regard. This article will explore the intersections of media mobilization and diversity strains as may be relevant to movement resonance.
Notably, the American Nonhuman Animal rights movement is dominated by middle-class white women [3][4][5]. The reasons for this are complex. The low level of African American participation, for instance, has been speculated by some to be a consequence of social disadvantage [6], cultural barriers [5], and ongoing trauma from historical experiences with violence and discrimination that creates discomfort with insensitive trans-species comparisons [7]. It may also result from historical constructions of "whiteness" as an indicator of compassion and "Blackness" with an inability to empathize with others or engage humaneness [8]. Early reformers working in an era of abolition positioned concern for other animals as a trait that was inherent to whites but impossible for people of color, a characteristic that would be used to legitimize systemic racism. Limited African American participation is sometimes thought a consequence of off-putting tactics which are rounded in racial insensitivities, as well as the For example, a demographic measure of One Green Planet contributors indicates that 90% of the 137 featured authors present as white. 5 Likewise, of 67 Animal Rights Zone podcast guests, 93% appear to be white-identified. 6 Indeed, the Nonhuman Animal rights movement might be said to abide by post-racial epistemologies that, as Harper [3,4] suggests, are insufficient in recognizing the very real consequences of socially created racial ascriptions: "My critique is that there are those (white and non-white) who believe 'race is a feeble matter' in animal rights activism. Such people are producing 1 A reviewer has suggested that this disconnect could also stem from a long history of white supremacist brutality against Black communities via the use of dogs. 2 Bryant Terry is a food justice advocate for communities of color; his activism is discussed in the cookbooks referenced. Food and cooking are integral to African American cultures of political resistance [14]. They have also been essential to feminist and environmentalist efforts [15]. 3 Although veganism and activism for other animals are often distinct, the two concepts are often used interchangeably. Most Nonhuman Animal rights efforts idealize veganism as foundational to anti-speciesist efforts, and the public tends to conflate the two. The Nonhuman Animal rights literature included in this study happens to be vegan-based (non-vegan magazine sources were not deemed suitable for analysis due to low numbers of human subjects). For these reasons, veganism will sometimes be used to reference Nonhuman Animal rights. 4 The Nonhuman Animal rights movement is considered to be a "post-citizenship" movement [19]. Participants of post-citizenship movements generally advocate on behalf of others or for non-material values and morals. They also tend to be well-integrated and relatively privileged in their society [20,21]. This may partially explain why disadvantaged demographics are underrepresented in the Nonhuman Animal rights movement. 5 This analysis was conducted in 2013 and was based on contributor avatars listed on One Green Planet's website: http://www.onegreenplanet.org/channel/about-us. One Green Planet is an influential online webzine and community featuring many of the most prominent authors and activists in the movement. 6 This analysis was conducted in 2013 according to the guests listed on the ARZone podcast page: http://arzone.ning.com/page/podcasts. Guests who made repeat appearances in more than one episode were counted once for each appearance. The regular hosts were excluded. ARZone is an academically-focused online community that also features the most prominent authors and activists in Nonhuman Animal rights mobilization. and practicing their own 'post-racial' epistemologies and praxis of AR/VEG 'cruelty free ethics'" [3] (p. 17). Adding to this, Adams [22], Gaarder [24], and Wrenn [29] have suggested that the movement also exploits gender inequality in an effort to advance Nonhuman Animal rights.
The apparent lack of race and gender consciousness in the Nonhuman Animal rights movement is a problem that should be a concern to those agitating for species justice in particular, but also for those working for social justice in general. Inequality reproduced in social movement media may have serious implications for coalition-building in social justice efforts. One magazine included in our sample has been cited as an example of such problematic representations: Though VegNews surely has its largely upscale market and audience in mind, the magazine does little to effectively counter the prevailing notion of veganism as the exclusive practice of upper-class, new-agey "bourgies", and it does little to promote solidarity or affinity based anything beyond buying cool "green" stuff [30] (p. 136). 7 Though class is not specifically explored in this study, it is a variable that very often intersects with gender, race, and body size. Implicit class representations in social movement media may also work to prevent strategic alliances among left-leaning movements. For these reasons, inequality in movement-produced material is incompatible with social justice goals and is also detrimental to necessary coalition-building.
This article seeks to address the manifestation of race and gender in American social movement media, specifically looking at the Nonhuman Animal rights movement and comparing it to the similar movements: the women's rights movement and the gay rights movement, as both the women's rights movement and gay rights' movement have shared histories of homogeneity and exclusion [31][32][33]. The Nonhuman Animal rights movement was chosen as it has been relatively understudied in the social sciences and recent developments in intersectionality perspectives in its ranks make this research especially timely. The comparison with other leftist movements will be helpful in ascertaining the significance of the findings. Beyond the rough similarities in profile, the comparison movements were chosen based on availability of regular magazine publications that feature human subjects (we had difficulty locating a suitable sample from the environmental movement, for instance).
The findings could illuminate the possibility that social movement media mobilization is hindering resonance and response from some marginalized populations by excluding them from media representations. We expect that the findings will inform Harper's observation of a superabundance of thin, white women in Nonhuman Animal advocacy media spaces. The Nonhuman Animal rights movement may not just reflect, but also reinforce and cultivate a largely female-identified and white constituency with a problematic emphasis on thinness and sexualization. This study does not investigate the actual influences this media may be having on audiences (nor is it framed to account for trans or agender visibility), but it does act as a starting point for future research. Comparing anti-speciesist media to feminist and queer media will allow us to differentiate the Nonhuman Animal rights movement as either unique or comparable in its representations. That is, whether or not the Nonhuman Animal rights movement is an outlier or simply part of a larger social movement trend could be evidenced in the study's outcome.
While investigating how feminist and queer social movement media impacts the demographic makeup of their constituencies is beyond the scope of this study, we will offer some brief suggestions. Specifically, we speculate that social movement media mobilization is, to some degree, responsible for reproducing an identity inclusive of a specific demographic. This could seriously hinder any social movement interested in attracting a diverse participant base. It may also be a concern to a social movement that desires to dismantle social inequalities and purports to value intersectionality. 7 Vegan consumer items of convenience and comfort are, arguably, helpful in the transition to veganism and sustaining a vegan lifestyle. Torres' critique, however, points to the problematic ways in which the consumer focus of veganism undermines its political capacity.

Literature Review
Social movements rely heavily on the media to increase visibility [34], extend claims-making, and to recruit participants [35][36][37]. Through media representations, movements can foster a group identity and all-important solidarity [38]. Unfortunately, the usefulness of media representation is often undermined by considerable bias that often works to the movement's disadvantage. The media's coverage of a movement is largely dependent upon political context and a myriad of other variables such as type of protest, size of protest, and fickle media attention cycles. Not all movements are likely to receive coverage equally, if they receive any at all [2,[39][40][41].
To some extent, movements are able to circumvent media bias by promoting their own media. Online entrepreneurship, for example, offers movements more freedom in recruitment and some control over how their organizations and claims are viewed [42,43]. As with those representations outside the movement's control, the media that is movement-controlled is also vital for volunteer recruitment and other mobilization efforts.
As an agent of socialization, the media is important in influencing social behaviors and attitudes. This is because media works to "[ . . . ] maintain boundaries in a culture" [44] (p. 225). Modeling theory, for instance, posits that individual action can be guided through observing the normalized and institutionalized behavior of others [45]. It is thought that, through exposure to the media, individuals and groups absorb norms and values, appropriate behavior, and identity. Media, however, is generally constructed to reinforce hegemonic powers. Indeed, the media tends to reflect dominant culture at the expense of oppressed groups [46]. As one example, the National Association for Multi-Ethnicity in Communications [47] reports that ethnic minorities and women are still significantly underrepresented as writers, managers, executives, and other media producers.
In fact, the underrepresentation of minorities in the media has been heavily documented. Early research finds that Black appearances constituted less than 9% of prime time television programming. Cross-racial interactions constituted less than 2% of that time and generally displayed less intimacy and less shared decision-making [48]. More recent studies find that representation for African Americans has improved, but whites are still overrepresented while other minorities continue to be underrepresented and negatively depicted [49][50][51]. In addition to this failure to diversify representations, the media often fails to adequately cover issues of racism in content as well, sometimes to the effect of impeding productive discourse [52]. These societal trends provide the context for social movement activity. It is not a stretch to suppose that movements may be replicating social inequality already saturating media spaces here described.
Media can perpetuate prejudice against oppressed groups by exaggerating cultural differences and depicting negative portrayals [45,46,53]. This can jeopardize the self-concepts of minority children [54] and their mental health in general [55]. On the other hand, as the media increases in sophistication and becomes more sensitive to a public distaste for degrading ethnic images, it has the potential to overcompensate and create an unrealistic image of success which dilutes real social issues [56]. Indeed, vegan feminist Aph Ko [57] has taken issue with this problematic countermeasure in the Nonhuman Animal rights movement whereby tokenized images of Black and Brown persons are used superficially to illustrate narratives that essentially remain white-centric. "All too often", she explains, "diversity discourse is used as an uncritical way to satisfy our culture's desire to see minorities in predominately white-dominated fields without changing the actual white framework". This process can lead to the erasure of difference and the replication of racism.
This review is not to suggest that audiences have no agency in their relationship with media. Certainly, viewers are frequently active participants in what they consume and interact with. For one, the ability to identify with the subject matter and to experience a feeling of similarity to those demographics represented can attract an audience. This attraction or detraction is one space for the audience to engage agency. So, in addition to theories of modeling and observational learning, it is helpful to recognize that audiences are not passive consumers. Instead, audience members are active Societies 2016, 6, 12 5 of 18 participants in their media experience. Social cognitive theory, for example, sees social behavior as a product of cognitive and personal factors working reciprocally with environmental factors [58].
Thus, if some degree of agency in media consumption is recognized, an increase in positive and realistic representations of a movement's diverse demographic would be needed to improve resonance. If an audience has the power to tune out or tune off, media producers would presumably be vested in accommodating consumer demand. Researchers explain: "Media have their greatest effect when they are used in a manner that reinforces and channels attitudes and opinions that are consistent with the psychological makeup of the person and the social structure of the groups with which he or she identifies" [46] (p. 44). For social movements, this could mean the self-produced media that reflects a limited demographic could be boxing in a limited demographic. These boundaries could not reasonably be expected to expand beyond that catered identity without some media cooperation. For example, one Hollywood diversity study finds that programming that features more diverse actors and writers is more popular, but the programs are "[ . . . ] nonetheless woefully underrepresented [ . . . ]" [59]. Social movements, then, would probably not be advantaged in presuming their audiences to be passive recipients; a social movement's audience, like any other, can pick and choose what it will tune into.
In spite of this audience agency, the media nonetheless both reflects and reinforces social attitudes, behaviors, and identities. A social movement concerned with social justice has good reason to be concerned with the social consequences of the media it produces. Media constructs a social reality and acts as an interpretive package for the public by giving meaning to social issues [1]. This not only pertains to racial categories and stereotypes, but also to gender roles. For example, women's homemaker role is emphasized in commercials for domestic products (like vacuums and air fresheners) that overwhelmingly depict women as the consumers. Commercials for household goods (like refrigerators and tools), on the other hand, consistently underrepresent women [60]. In the world that much of the mass media presents, women's subservient place is clear.
Another such trend is the increasing sexualization of women in the media [53,61]. This pattern is thought problematic because media could foster anti-feminist attitudes and increase aggressive behavior toward women [62][63][64][65][66][67]. Furthermore, the media's glorification of thin bodies and its sexualization of women is linked to increased body dissatisfaction and decreased self-esteem [68][69][70][71]. Weight discrimination is also linked to hiring and promotional discrimination in the workplace and lower overall earnings, especially for women [72,73]. Hence, a thin and sexualized identity is perpetuated for women largely in the service of their societal devaluation. Social movement media that engages in the sexualization of women and the normalization of thinness might actually be contributing to inequality. If countering inequality is integral to that movement's goals, as is true of the movements explored in this study, then accommodating these harmful norms in movement media could be particularly contradictory and problematic.
Certainly, some producers acknowledge the benefits to challenging stereotypes and privilege in media. Many advertisers attempt to increase their resonance by actively targeting minority populations, which in turn, has increased media multiculturalism. Social movement media operates similarly to that of advertising because it is generally intended to gain support and financial contributions from its audience. Therefore, it would presumably be advantageous to social movements to target marginalized demographics as well. Such a tactic would be in the interest of increased multiculturalism as a social justice matter, but, as a practical matter, it would also present greater potential for a movement's frame resonance.

Methods
This study examines the demographical frequencies of some readily identifiable diversity indicators-race, gender, body type, and sexualization-in a sample of magazines representing some of the more prominent leftist movements. To accomplish this, the Nonhuman Animal rights movement, the feminist movement, and the gay rights movement are compared. The movements were chosen based on their similarity in justice-based, egalitarian claims-making, and their apparent interest in or tendency toward inclusivity. Other movements, such as the Black liberation movement and the environmental movement, might also have been useful to the study, but were not included due to restrictions in available or suitable sample material, but also in the interest of keeping the sample size manageable. The unit of analysis for this study is the human subject as depicted on covers of two major social movement publications from each of the three movements. The statistical analysis was conducted using SAS software. The questions proposed are considered in two parts: in regard to the sexualization of the magazine cover subjects, and in regards to the representation of gender and racial groups among the three movements. A more sophisticated content analysis exploring the content of these magazines could prove insightful. However, for the exploratory purposes of this study, we have chosen to focus specifically on covers under the assumption that the covers are designed to grab attention and quickly convey a specific message and image to the public.

Profile of Samples
People become vegan for all manner of reasons [5]. Subsequently, this study focuses on Nonhuman Animal rights rather than veganism exclusively. To represent this movement, we have analyzed PETA's Animal Times (circulation of approximately 350,000) and VegNews (circulation of 76,000). Animal Times was selected as a representation of Nonhuman Animal rights outreach and movement culture, as PETA is one of the largest and most well-known organizations; VegNews is included as a representation of "normalized" vegan living. Indeed, because veganism is so intertwined with Nonhuman Animal rights, the behaviors of VegNews (the only mainstream vegan magazine at the time of this writing with a circulation of 76,000) should have implications for Nonhuman Animal liberation mobilization. Of note, the first two years of VegNews (2000-2001) and a number of issues in the years following were specifically related to Nonhuman Animal liberation and not the vegan lifestyle alone, thus speaking to its underlying anti-speciesist framework. Other magazines were not considered because they failed to regularly include human subjects on their covers. 8 Compassion Over Killing's Compassionate Action, VIVA! UK, and the Humane Society of the United States' All Animals, for example, feature mostly Nonhuman Animal subjects on their covers. Vegetarian Times, on the other hand, primarily features food.
For the feminist movement, we sampled the covers of Ms. Magazine and BUST Magazine. Considered a "landmark institution" in women's rights, Ms. has been in publication since the 1970s and represents the "second wave" feminist movement. It currently has a circulation of over 100,000. BUST is a newer publication and is reflective of a more contemporary generation of feminists, often those who identify as "third wave". We specifically chose representatives of both waves to gain a more comprehensive look at a highly fractured movement. Ms. has a long history in the field, but BUST is a somewhat more modern magazine, active only since 1993 and catering to a younger demographic.
Representative of the gay rights movement, we chose Equality (published by the Human Rights Council) and The Advocate. Equality was selected because it is published by the largest gay rights organization in the United States, equivalent in size and structure to the Nonhuman Animal rights movement's PETA. The HRC bills it as the largest LGBT magazine in circulation. The Advocate was selected because it is has a long history of publication (since the 1960s) and is advertised as world's leading gay and lesbian advocacy publication. The Advocate has a circulation of almost 200,000. 8 It is worth considering that nonhuman subjects (who are less likely to present gender norms) could have an equalizing effect and may actually welcome underrepresented groups. Furthermore, nonhuman imagery is heavily utilized by the Nonhuman Animal rights movement to stimulate empathy and mobilization. This relationship between nonhuman cover subjects and propensity for participation would be worth exploring in future research. Data was located through three sources: (1) Some issues are available free online (as are some issues of Equality); (2) Some covers were available in the back issues page of the magazines' websites; and (3) Those that could not be obtained online were obtained through our institutions' library services. We were unable to locate several issues, which were coded as missing. 10 We excluded six issues of Animal Times, 24 issues of VegNews, 21 issues of Ms., 11 issues of Equality, and 20 issues of The Advocate because they failed to portray a human subject or a discernible human subject. 11 Animal Times, for example, often featured Nonhuman Animals, while VegNews often utilizes images of food. If, however, a nonhuman subject dominated the cover, but there was a thumbnail image featuring a human face; these human subjects were included. A few issues of The Advocate featured a dozen or so couples on the cover, all of which were coded individually. Regardless of scale, any human subjects depicted are presumed to be influential in shaping the magazine's demographic identity.
These were not the only issues to arise when coding the covers. In one case, 12 a thumbnail image was reused on a subsequent cover and this was omitted. Another cover featured several equally prominent subjects with only one visible face; only the subject whose face was visible was coded. 13 Several covers featured a large image of a human subject whose face was not to the camera. Because they were the only subjects on the cover and we felt we could reasonably ascertain their race from skin and hair, we included those. One issue featured infants. In this case, only their race could be determined, but several other images featuring adult humans were on that cover, so infants were omitted while the other subjects were not. 14 Background subjects and cartoons were excluded, however very realistic drawings were included. In three issues, only a woman's torso is pictured. 15 Because it was impossible to accurately code her race or degree of sexualization (according to the scale we have chosen), these images were omitted.
The final subject count came to 48 for Animal Times, 104 for VegNews Magazine, 69 for Ms., 64 for BUST, 109 for Equality, and 564 for The Advocate. Only one author coded the data, but a reliability check was performed on 10% of the sample with the second author resulting in a 71% agreement. The bulk of the disagreement resulted from racially ambiguous subjects and subjects that were on the border of sexualized and not sexualized. One androgynous-presenting subject resulted in disagreement as well. 9 Three times in 2000, 10 times in 2001, 9 times in 2002, and 5 times in 2004. 10 Bust Magazine (11); Equality (23); and The Advocate (6). 11 For example, the Winter 2007 issue of Equality depicted a blurred face and an outreaching hand. Gender was undeterminable, and because the library scan was in black and white, the race of the subject was also unknown. 12 VegNews 2001, January (4). 13 VegNews 2001, February (5). 14 VegNews 2001, May/June (8,9). 15

Race and Gender
While race and gender are social constructions, and are therefore relatively subjective, an exploration into other content analyses of a similar nature did not uncover a coding scheme for either variable [51,61,74,75]. Therefore, we have depended on our personal socialization to discern race and gender. Race was coded as white, African American, Asian, Latinx, 16 other, or unknown/undetermined. Gender was coded as female-presenting or male-presenting. Transgender individuals and persons in drag were coded according to the gender they were presenting. 17 As there is no inherent order to gender or racial categorizations, a logistic model is not necessary. Instead, a two-way contingency table analysis was conducted for each variable against movement.

Body Type
We used Johnson's [76] body type scale that divides bodies into five types: round bodies with excess body fat (endomorph), bodies with some muscle definition and also excess fat (endo-mesomorph), muscle-toned bodies with a lack of excess body fat (mesomorph), fit and thin bodies (ecto-mesomorph), and thin bodies with stringy muscles (ectomorph). Several subjects were coded as the default (ecto-mesomorph) if only their face was pictured (with no indication of facial fat) or their clothing covered up their body and it was not possible to detect if they were toned or not. The inability to gauge some non-thin body types on covers that feature only faces will limit the study. We analyzed under the presumption that the face is likely to indicate the presence of a round body type, but this approach could not be said to be fully valid or reliable.

Sexualization
To determine degree of sexualization, we draw on Hatton and Trautner's [6] coding scale that rates images as not at all sexualized, sexualized, or highly sexualized. 18 Their system entails a 23 point additive scale with 11 separate variables: clothing/nudity; touch; pose; mouth; breast/chest, genitals, and buttocks; text; head vs. body shot; display of sex act; and sexual role play. The total of points received indicates the level of sexualization in the image.
Depending on the amount of nudity, a cover subject could score between 0 and 5 points. Here, 0 indicates unrevealing clothing; 1 indicates slightly revealing clothing (such as shirts with low necklines or exposed arms); 2 indicates somewhat revealing clothing (demonstrated by exposed midriffs); 3 indicates highly revealing and/or skin-tight clothing; 4 indicates swimsuits and lingerie; and 5 indicates no clothing at all. Images that showed subjects from the chest up with an indication that the person was not wearing clothes were coded as 5.
The depiction of touch was also used as a marker of sexualization, with subjects scoring between 0 and 3 points. Those subjects neither touching nor being touched scored 0, those engaging in casual touching scored 1, those engaging in provocative touching scored 2 (lifting one's shirt, for example), and those engaging in explicit sexual touching scored 3. The Hatton and Trautner study intended the touch scale to refer to human subjects touching themselves or other humans, but we have extended this variable to include the touching of other animals as well.
Based on the subject's pose, the individual could score between 0 and 2 points, with 0 indicating a nonsexual pose; 1 indicating a slightly suggestive pose (including raised arms, leaning, or sitting); and 2 indicating an overtly suggestive pose. The subject's mouth is also examined on a scale of 0-2, with 16 This is the gender-neutral term meant to include trans and gender fluid persons. 17 Admittedly, the framework of the analysis and its focus only on cover subjects makes addressing gender variance difficult; it also reinforces a limiting gender binary. 18 Please see the original study for further explanation on how variables we've included were coded and why.
Societies 2016, 6, 12 9 of 18 0 indicating a mouth that does not suggest sexual activity; 1 indicating a slightly suggestive mouth (including parted, non-smiling lips); and 2 indicating an overtly suggestive mouth. The breasts/chests, genitals, and buttocks of subjects were also examined on a scale of 0-2 with 0 assigned when body parts were either not visible or not the focal point; 1 assigned when one or more body part was emphasized; and 2 assigned when that body part was the central focus.
Text was analyzed on a scale of 0-2 when text is directly related the subject. If unrelated to sexuality, the image was scored zero. If it was in some way sexually suggestive, for example, "On Fire: Joaquin Phoenix today's hottest veg celebrity" (VegNews Magazine, Issue 40), it scored 1. Images would be scored 2 if an explicitly sexual reference was made, but no images met this criterion. Body shots, which Hatton and Trautner have argued to be more sexualized, scored 1 point, whereas headshots scored none. Sex acts and sexual role playing were included in our analysis as part of the Hatton Trautner methodology, but none of our images applied. According to Hatton and Trautner's coding scheme, the scores from these 11 categories should be tallied to determine overall sexualization. Images that scored 0-4 points were labeled nonsexualized, those that scored 5-9 were labeled sexualized, and those that scored 10 or more points were labeled as hypersexualized.
Two categorical models were built to measure the difference in sexualization of subjects. The independent variables (movement, race, gender) are all nominal variables, while the dependent variable (sexualization) is an ordinal variable. Therefore, a logistic model is most appropriate, as it models an ordinal variable as a function of one or several predictors. In this case, it was used to predict the probability of magazine cover subjects-that of varying races, genders, and representing three different movements-being not at all sexualized, sexualized, or very sexualized. The first of the two models established a baseline for comparison; sexualization as a function of gender and race. The second model built on those results, with movement added to determine its unique contribution, if any, to the variability in sexualization of the magazine models.

Gender
The contingency table analysis for association between gender and movement were significant (p < 0.001). That is, there are significant changes in the representation in gender associated with the different movements. Figure 1 illustrates where those differences arise: the gay rights movement predominantly used male-presenting models in the sample of covers studied (68.9%), while the Nonhuman Animal rights and feminist movements used predominantly female-presenting models (60.7% and 81.2%, respectively).
Societies 2016, 6, 12 9 of 18 mouth (including parted, non-smiling lips); and 2 indicating an overtly suggestive mouth. The breasts/chests, genitals, and buttocks of subjects were also examined on a scale of 0-2 with 0 assigned when body parts were either not visible or not the focal point; 1 assigned when one or more body part was emphasized; and 2 assigned when that body part was the central focus.
Text was analyzed on a scale of 0-2 when text is directly related the subject. If unrelated to sexuality, the image was scored zero. If it was in some way sexually suggestive, for example, "On Fire: Joaquin Phoenix today's hottest veg celebrity" (VegNews Magazine, Issue 40), it scored 1. Images would be scored 2 if an explicitly sexual reference was made, but no images met this criterion. Body shots, which Hatton and Trautner have argued to be more sexualized, scored 1 point, whereas headshots scored none. Sex acts and sexual role playing were included in our analysis as part of the Hatton Trautner methodology, but none of our images applied. According to Hatton and Trautner's coding scheme, the scores from these 11 categories should be tallied to determine overall sexualization. Images that scored 0-4 points were labeled nonsexualized, those that scored 5-9 were labeled sexualized, and those that scored 10 or more points were labeled as hypersexualized.
Two categorical models were built to measure the difference in sexualization of subjects. The independent variables (movement, race, gender) are all nominal variables, while the dependent variable (sexualization) is an ordinal variable. Therefore, a logistic model is most appropriate, as it models an ordinal variable as a function of one or several predictors. In this case, it was used to predict the probability of magazine cover subjects-that of varying races, genders, and representing three different movements-being not at all sexualized, sexualized, or very sexualized. The first of the two models established a baseline for comparison; sexualization as a function of gender and race. The second model built on those results, with movement added to determine its unique contribution, if any, to the variability in sexualization of the magazine models.

Gender
The contingency table analysis for association between gender and movement were significant (p < 0.001). That is, there are significant changes in the representation in gender associated with the different movements. Figure 1 illustrates where those differences arise: the gay rights movement predominantly used male-presenting models in the sample of covers studied (68.9%), while the Nonhuman Animal rights and feminist movements used predominantly female-presenting models (60.7% and 81.2%, respectively).

Race
After combining human subjects of unknown and undetermined race with "other", a chi-squared goodness of fit test including all 956 subjects from the three movements was conducted to determine whether the racial makeup of the images is the same as that of the US population [77]. The test was significant (p < 0.0001), indicating that the models were not representative of the population at large. Figure 2 demonstrates that white models are over represented in the sample, relative to the population, while the other races are underrepresented to varying degrees. The test for significant association between race and movement was significant (p = 0.0139), indicating that there are non-random changes in the representation of different races across the different movements. The distributions presented in Figure 2 illustrate where those differences arise: while whites and African Americans are represented most often in each of the movements in fairly consistent proportions, the ordering of Asians and Latinx of color in the three movements varies, as do their respective proportions across movements. Nonetheless, the Nonhuman Animal rights movement's media does lay claim to the highest percentage of white subjects at 87.3%. This frequency is 4.4 percentage points higher than the gay rights movement media, and 12.9 percentage points higher than the feminist media.

Race
After combining human subjects of unknown and undetermined race with "other", a chi-squared goodness of fit test including all 956 subjects from the three movements was conducted to determine whether the racial makeup of the images is the same as that of the US population [77]. The test was significant (p < 0.0001), indicating that the models were not representative of the population at large. Figure 2 demonstrates that white models are over represented in the sample, relative to the population, while the other races are underrepresented to varying degrees. The test for significant association between race and movement was significant (p = 0.0139), indicating that there are non-random changes in the representation of different races across the different movements. The distributions presented in Figure 2 illustrate where those differences arise: while whites and African Americans are represented most often in each of the movements in fairly consistent proportions, the ordering of Asians and Latinx of color in the three movements varies, as do their respective proportions across movements. Nonetheless, the Nonhuman Animal rights movement's media does lay claim to the highest percentage of white subjects at 87.3%. This frequency is 4.4 percentage points higher than the gay rights movement media, and 12.9 percentage points higher than the feminist media. In two cases, African Americans were underrepresented. African Americans comprise approximately 13% of the American population [77], but only 8.8% of the gay rights movement media sample and only 7.3% of the Nonhuman Animal rights sample. The feminist movement was relatively representative at 12.8%. About 5% of the American population is Asian, meaning that the 9% frequency for Asians in the feminist sample's is somewhat over-representative, the Nonhuman Animal rights 4.7% frequency is somewhat representative, while the gay movement's 2.8% frequency is somewhat lacking. Although Latinx persons comprise 17% of the population, they were seriously underrepresented in all three cases. The gay and feminist samples included about 3% Latinx subjects of color, while the Nonhuman Animal rights movement featured Latinx of color in less than 1% of the sample covers.

Sexualization
All three movements featured comparable levels of sexualization: 15.8% of the feminist sample featured sexualized subjects (an additional 0.75% were hypersexualized), 9.06% of the gay rights sample was sexualized (an additional 1.63% were hypersexualized), and 11.3% Nonhuman Animal rights subjects were sexualized (an additional 1.3% were hypersexualized). The initial model of sexualization vs. gender and race indicates that there is a significant difference in sexualization In two cases, African Americans were underrepresented. African Americans comprise approximately 13% of the American population [77], but only 8.8% of the gay rights movement media sample and only 7.3% of the Nonhuman Animal rights sample. The feminist movement was relatively representative at 12.8%. About 5% of the American population is Asian, meaning that the 9% frequency for Asians in the feminist sample's is somewhat over-representative, the Nonhuman Animal rights 4.7% frequency is somewhat representative, while the gay movement's 2.8% frequency is somewhat lacking. Although Latinx persons comprise 17% of the population, they were seriously underrepresented in all three cases. The gay and feminist samples included about 3% Latinx subjects of color, while the Nonhuman Animal rights movement featured Latinx of color in less than 1% of the sample covers.

Sexualization
All three movements featured comparable levels of sexualization: 15.8% of the feminist sample featured sexualized subjects (an additional 0.75% were hypersexualized), 9.06% of the gay rights sample was sexualized (an additional 1.63% were hypersexualized), and 11.3% Nonhuman Animal rights subjects were sexualized (an additional 1.3% were hypersexualized). The initial model of sexualization vs. gender and race indicates that there is a significant difference in sexualization between women and men (p = 0.0139). The estimated odds ratio of sexualization for comparing male-presenting models to female-presenting models is 0.609, indicating that female-presenting models have lower odds of being sexualized than do male-presenting models (Figure 3). Race was not a significant predictor of sexualization (p = 0.6830), so it was removed from the subsequent model. To test for differences in sexualization due to movement, over and above that due to gender, another model was developed. Movement was not a significant predictor of sexualization (p = 0.6242), though gender remained significant (p = 0.0496), with an estimated odds ratio for females to males of 0.652.
Societies 2016, 6, 12 11 of 18 between women and men (p = 0.0139). The estimated odds ratio of sexualization for comparing male-presenting models to female-presenting models is 0.609, indicating that female-presenting models have lower odds of being sexualized than do male-presenting models (Figure 3). Race was not a significant predictor of sexualization (p = 0.6830), so it was removed from the subsequent model. To test for differences in sexualization due to movement, over and above that due to gender, another model was developed. Movement was not a significant predictor of sexualization (p = 0.6242), though gender remained significant (p = 0.0496), with an estimated odds ratio for females to males of 0.652. There is some evidence of increase in sexualization over time (Figure 4). The Spearman correlation between sexualization and publication year is significant (p = 0.0004), but relatively weak (r = 0.113); the large sample size is likely a driver of its significance. Modeling the relationship with a proportional odds model to determine probabilistic differences in the levels of sexualization over time likewise found a positive trend (p = 0.0032). The estimated odds ratio, regardless of movement, indicates that each additional year increases the likelihood of sexualization of a magazine cover model by a factor of 1.09. A plot of the relationship between the levels of sexualization and time supports the weak Spearman correlation and low odds ratio.   There is some evidence of increase in sexualization over time (Figure 4). The Spearman correlation between sexualization and publication year is significant (p = 0.0004), but relatively weak (r = 0.113); the large sample size is likely a driver of its significance. Modeling the relationship with a proportional odds model to determine probabilistic differences in the levels of sexualization over time likewise found a positive trend (p = 0.0032). The estimated odds ratio, regardless of movement, indicates that each additional year increases the likelihood of sexualization of a magazine cover model by a factor of 1.09. A plot of the relationship between the levels of sexualization and time supports the weak Spearman correlation and low odds ratio. between women and men (p = 0.0139). The estimated odds ratio of sexualization for comparing male-presenting models to female-presenting models is 0.609, indicating that female-presenting models have lower odds of being sexualized than do male-presenting models (Figure 3). Race was not a significant predictor of sexualization (p = 0.6830), so it was removed from the subsequent model. To test for differences in sexualization due to movement, over and above that due to gender, another model was developed. Movement was not a significant predictor of sexualization (p = 0.6242), though gender remained significant (p = 0.0496), with an estimated odds ratio for females to males of 0.652. There is some evidence of increase in sexualization over time (Figure 4). The Spearman correlation between sexualization and publication year is significant (p = 0.0004), but relatively weak (r = 0.113); the large sample size is likely a driver of its significance. Modeling the relationship with a proportional odds model to determine probabilistic differences in the levels of sexualization over time likewise found a positive trend (p = 0.0032). The estimated odds ratio, regardless of movement, indicates that each additional year increases the likelihood of sexualization of a magazine cover model by a factor of 1.09. A plot of the relationship between the levels of sexualization and time supports the weak Spearman correlation and low odds ratio.

Body Type
There was a significant difference in body type representation across the three movements (p < 0.0001), though all three movements had a majority of fit and thin models on their covers ( Figure 5). The gay rights magazines had a greater diversity of body types in their models, including more muscle-toned models and models of size than either of the other two movements; 76.5% of these models were of the fit and thin body type, and the remaining quarter were fairly evenly distributed among the remaining types. The feminist and Nonhuman Animal rights movements (92.5% and 92.7%, respectively) each had over 90% of the cover models fall within the fit and thin body type representation, with the remaining body types ranging from muscle-toned to round. The Nonhuman Animal rights sample was the least likely to include subjects of size.

Body Type
There was a significant difference in body type representation across the three movements (p < 0.0001), though all three movements had a majority of fit and thin models on their covers ( Figure 5). The gay rights magazines had a greater diversity of body types in their models, including more muscle-toned models and models of size than either of the other two movements; 76.5% of these models were of the fit and thin body type, and the remaining quarter were fairly evenly distributed among the remaining types. The feminist and Nonhuman Animal rights movements (92.5% and 92.7%, respectively) each had over 90% of the cover models fall within the fit and thin body type representation, with the remaining body types ranging from muscle-toned to round. The Nonhuman Animal rights sample was the least likely to include subjects of size.

Conclusions
In representation of gender, the Nonhuman Animal rights movement more closely resembles that of feminist media, with more women than men represented. However, it also resembles queer media in that it presents gender ratios that do not closely match actual activist ratios. Queer media in our sample disproportionately features male-presenting subjects, which would inaccurately downplay the presence and involvement of lesbians. The relatively few number of women presented in gay rights magazines may speak to a history of lesbian exclusion in the movement [78] and lesbian underrepresentation in gay media [31,79]. Though the vegan media sample demonstrates higher than average numbers of women, these numbers do not come close to the actual presence of women in Nonhuman Animal rights spaces. This likely relates to patriarchal norms in the movement which position men as highly visible leaders and women as caregivers, organizers, or other support providers expected to remain behind the scenes [24]. Relatively few men were pictured on the feminist magazines, which may be indicative of the feminist goal of creating a woman-normative space that resists patriarchal domination. To be sure, the positive and dignified representation of women, especially as leaders and role models, is important in a patriarchal society; however, failure to include more male-presenting subjects may be hindering the feminist movement's ability to attract male-identified allies. Alternatively, the feminist sample featured considerable sexualized subject matter, particularly BUST. This may reflect the third-wave focus on sex positivity, as well as the patriarchal co-optation of feminist spaces. These pressures sometimes increase the emphasis on sexual liberation to the effect of sexualizing feminism [80].

Conclusions
In representation of gender, the Nonhuman Animal rights movement more closely resembles that of feminist media, with more women than men represented. However, it also resembles queer media in that it presents gender ratios that do not closely match actual activist ratios. Queer media in our sample disproportionately features male-presenting subjects, which would inaccurately downplay the presence and involvement of lesbians. The relatively few number of women presented in gay rights magazines may speak to a history of lesbian exclusion in the movement [78] and lesbian underrepresentation in gay media [31,79]. Though the vegan media sample demonstrates higher than average numbers of women, these numbers do not come close to the actual presence of women in Nonhuman Animal rights spaces. This likely relates to patriarchal norms in the movement which position men as highly visible leaders and women as caregivers, organizers, or other support providers expected to remain behind the scenes [24]. Relatively few men were pictured on the feminist magazines, which may be indicative of the feminist goal of creating a woman-normative space that resists patriarchal domination. To be sure, the positive and dignified representation of women, especially as leaders and role models, is important in a patriarchal society; however, failure to include more male-presenting subjects may be hindering the feminist movement's ability to attract male-identified allies. Alternatively, the feminist sample featured considerable sexualized subject matter, particularly BUST. This may reflect the third-wave focus on sex positivity, as well as the patriarchal co-optation of feminist spaces. These pressures sometimes increase the emphasis on sexual liberation to the effect of sexualizing feminism [80].
In representations of race, all movements fail to adequately represent minorities to some extent, though the feminist sample was most favorable to African Americans and Asians. All movements seriously underrepresented Latinx of color. The results of the gay rights sample are consistent with a long history of the gay media excluding people of color [31]. The Nonhuman Animal rights sample was the most white-centric, with almost nine out of ten subjects appearing to be white-identified.
While some of Hatton and Trautner's sexualization variables did not surface in the data (the performance of a sex act and sexual role play), indicating potentially lower levels of sexualization in social justice magazines as compared to some of the mainstream media they analyzed, the substantial number of images that did score as sexualized is relevant. Despite its focus on sexual orientation and sexual liberation, queer media in our sample ranked lowest in sexualized subjects. Because the gay media is known to underrepresent female-presenting subjects and these subjects are more likely to be sexualized, this is understandable. The gay rights sample did, however, demonstrate the highest percentage of hyper-sexualization among sexualized subjects. This sexualization may be concerning given research which indicates that gay men are also susceptible to self-objectification, body shame, and restricted eating as a result of existing in a subculture that prioritizes physical appearance [81]. Probably reflecting third-wave "sex-positive" trends in feminism, our feminist sample ranked the highest in sexualization. Nonhuman Animal rights fell in between the two, though all three movements were relatively comparable in sexualization (levels ranged between 11% and 17%). Sexualization appears to be growing more prevalent across time, though only minimally and perhaps due to the large sample size. This trend does correspond with other findings regarding the increase of sexualization in the media [61,82,83]. It is also important to mention here that PETA, the publisher of Animal Times, is notorious for sexually objectifying the subjects in its media [24,27,28], which may have inflated the prevalence of sexualization in the results.
Regarding body type, the predominance of thinness in our samples may reflect the publishers' interest in portraying a conventionally desirable body image. This may be intentionally cultivated by the gay rights movement following stereotypes of gays and lesbians as weak or ill. For the Nonhuman Animal rights media, the predominance of thinness may also be a response to perceived healthiness or unhealthiness of vegetarian and vegan diets that are often associated with the movement. Furthermore, Johnson [76] finds that negative personality ascriptions are associated with body types that "deviate" from thinness. Understandably, our sampled magazines are likely hoping to portray their respective movements as positively as possible. Though, in doing so, they create an unrealistic body type ideal. They also foster an identity that can potentially alienate other body types, specifically impacting people of color. Women of color often exhibit beauty aesthetics based in their own cultural legacies, or even nurture embodied difference as a form of resistance [84]. As a result, Black and Brown bodies can be pathologized or experience stigma in a thin-privileging media space.
To summarize, the Nonhuman Animal rights media in our sample, when compared to other leftist social justice movements, appears to be somewhat more gender-balanced, somewhat more white-centric, moderately sexualized, and less diverse in body type representations. Nonhuman Animal rights, women's rights, and gay rights media could be reflecting a perceived demographic or they may be facilitating an idealized demographic. In utilizing media spaces, social change actors must overcome societal biases that tend to favor hegemonic status quo and invisiblizes social movement claims-making and activities. Social movements may be exacerbating this disadvantage in fostering their own bias. The media is essential in movement recruitment and claims-making: it constructs a social reality, and shapes public opinion. If a social movement's own media representations are creating a limiting identity, they, alongside mainstream media, may be just as responsible for biased representation.
One major limitation in this study is the visibility of the magazines utilized. We selected magazines based on their accessibility to the public, and we chose covers because of their role in marketing the magazine's theme and purpose expeditiously to passing customers. Some non-profit affiliated magazines, such as Animal Times and Equality, are issued only to subscribing members with only a few select issues available for online viewing and thus may be less influential to the public. For the most part, readers would have to subscribe to see the magazines, so these magazines may be more reflective of their existing demographic and would wield less power in advertising. However, PETA and HRC's publications are far from top secret as they are advertised and shared, meaning that the identity maintained through these covers likely help foster a particular constituency. VegNews, Ms., BUST, and The Advocate on the other hand, are publically available publications. Yet, they, too, are limited in their reach because they cost a nominal fee and they are distributed in very specific sale locations. VegNews, for example, is a specialty magazine with a focus on an alternative diet and lifestyle, and tends to be offered in natural foods grocers. Natural foods grocers, in turn, tend to cater to middle class whites who are more likely to be privileged with the income and access for healthful shopping.
Nonetheless, as dominant representations of Nonhuman Animal rights culture, the white, feminized identity that is nurtured by Animal Times and VegNews could undoubtedly reach beyond a simple reflection of their existing demographic and may actually begin to police that identity through the exclusion of others. It would also be interesting to ascertain what impact this movement-nurtured identity is having for the majority demographic in regards to how they relate to the marginalized. As Harper bemoans, a false consciousness about diversity (sometimes referred to as "color-blindness") can shut down critical discussions about persistent discrimination within activist spaces (and society in general).
This study is also limited in the small, non-random sample size which cannot be wholly representative. Movements are large, diverse, and factionalized, meaning that the prominent magazines we chose will present a relatively biased view. Likewise, magazines represent only one form of movement produced media. Additional analyses that include other publications and other forms of media (commercials, websites, music, etc.) could broaden the scope of this study's implications for social movement diversity.
The Nonhuman Animal rights media thus examined favors thin white women, 19 more so than that of the feminist movement and the gay rights movement. However, other movements are also recreating similar problems that reflect stereotypes about their demographics. This study cannot speak specifically to the impact these demographic representations are having on potential recruits or existing activists, but it does offer an important starting point for future research into the relationships between social movement media and social movement success. Further research into viewer interpretations, minority experiences in the ranks, and movement motivations and goals regarding diversity would be beneficial to this topic. Sexualized (5-9) 2 Highly sexualized (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)