Introduction

“Gay men are as misogynistic as straight men, if not more so…” quipped actor Rose McGowan during a podcast interview in 2014 (Friess 2014, para. 3). The comment made headlines and sparked online debates about gay men and their attitudes towards women (e.g. Murphy 2014; Selby 2014; Strudwick 2014). Even though McGowan later apologized for generalizing gay men (Hare 2014), the incident challenged the myth gay men cannot be misogynistic, or more generally sexist. Misogyny is defined as a hatred of women, whereas sexism is discrimination based on sex (Merriam-Webster.com). This study focuses on forms of sexism that overlap with misogyny.

News articles (e.g. Daley 2016; Liveris 2016; Power 2016) and academic research (e.g. Murgo et al. 2017; Warriner et al. 2013; Zheng and Zheng 2015) have explored potential sexist attitudes of gay men. Comprised are specific examples such as negative stereotypes of women that are sometimes enacted in popular culture like on Rupaul’s Drag Race (Tabberer 2017), the firing of a Gay Times magazine editor for sexist and anti-Semitic tweets (Rudgard 2017), Stephen Fry’s public comments about women including his view on why women do not enjoy sex (Newkey-Burden 2016), Milo Yiannopoulos’ constant misogyny against women (Lopez 2017), and so forth. Television and film have also normalized male privilege through prominent white gay cisgender male characters, overlooking lesbian, bisexual, and transgender viewpoints (Shugart 2003).

HuffPost blogger Donovan (2017) argues that given the abundance of negative gay stereotypes, these examples should not be labeled as gay sexism, but rather sexism by gay men. In other words, sexism is sexism, regardless of the source’s sexuality or gender. Further research is needed to understand whether or not forms of sexism differ according to sexuality, along with potential predictor variables. However, complications exist when discussing the shortcomings of an already marginalized group, such as gay men (Donovan (2017). The challenge for researchers is to identify potentially harmful negative attitudes towards women by gay men without overgeneralizing, scapegoating, or forgetting the heterosexism gay men face (Simoni and Walters 2001).

Further confounding this issue is previous research that shows sexism and homophobic or anti-gay attitudes are often correlated. In the US, sexism positively predicted anti-lesbian attitudes (Wilkinson 2008). In the US (Aosved and Long 2006) and UK (Davies et al. 2012), researchers found connections between homophobic attitudes, sexism, and rape myth acceptance. Also in the UK, sexism has been correlated with anti-gay rights attitudes (Davies 2004; Masser and Abrams 1999). In Canada, hostile sexism in men and hostile and benevolent sexism in women predicted negative attitudes towards gay and lesbian adoption (Rye and Meaney 2010). In Italy, sexist, homophobic, and transphobic attitudes in teachers were positively correlated with straight conservative men who were religious (Scandurra et al. 2017). In Turkey, a correlation between anti-gay attitudes and hostile sexism was also found, especially in men (Sakalli 2002). Overall, men dismiss the severity of sexual harassment more than women, even in the cases of man-to-man sexual harassment (DeSouza and Solberg 2004).

Given these findings, it seems reasonable to conclude gay men and women would shun problematic attitudes such as sexism and homophobia. Nevertheless, based on anecdotal and research evidence, there is a need to better understand how already marginalized groups can still hold biases against other marginalized groups. Such an examination draws upon intersectionality, which emphasizes that marginalized groups could better support each other through identifying problematic attitudes/behaviors—often connected to societal power structures (Cole 2008; Montgomery and Stewart 2012). In other words, intersectionality classifies multiple factors that comprise one’s identity, in order to more accurately define privilege and inequality (Crenshaw 1991).

Accordingly, this study uses the ambivalent sexism inventory (ASI) and several predictor variables in a survey of self-identified gay men in the UK and the US to investigate if and how sexism manifests. Glick and Fiske (1996) created ASI in order to explicate the commonly acknowledged hostile sexism, but to also bring attention to benevolent sexism, which masks as positive in cherishing women with chivalry. Both are harmful to women because the former harbors negative views of women as inferior to men, while the latter reveres women but only in a traditional gendered role (Glick and Fiske 1996, 2001a, b). It is the contrast and coinciding correlation between hostile and benevolent sexism, which creates ambivalence (Glick et al. 1997).

In-line with previous research on ambivalent sexism (Hellmer et al. 2018), this study draws upon several predictor variables to better explain potential ambivalent sexism. Religiosity is incorporated because it is often used as a potential predictor variable of both hostile and benevolent sexism (Glick et al. 2016; Mikołajczak and Pietrzak 2014). Christianity is the predominant religion in the US (Newport 2011) and the UK (White 2012), which has particularly been associated with benevolent sexism in men and women (Glick et al. 2002; Mikołajczak and Pietrzak 2014), and hostile sexism in men (Maltby et al. 2010). Previous research on religion and sexism mostly overlooks an individual’s sexuality. Some religions have anti-LGBT + policies and beliefs (Johnson and Vanderbeck 2014), therefore it is imperative to investigate the intersectionality of such identities. This study adds to previous literature on sexism by including religion as a predictor variable.

Political ideology is also used based on previous studies which show it influences levels of ambivalent sexism in men and women (Christopher and Mull 2006). Conservativism, for instance, often promotes ideas of traditional gender roles and accordingly, benevolent sexism is found in right-leaning people (Glick et al. 1997). Also, used in this study are two predictor variables which are closely related to conservatism: Nationalism and anti-immigration attitudes. Measuring nationalism and anti-immigration attitudes is particularly pertinent, due to recent findings in the US and UK connecting conservative led movements such as voting for Brexit (Corbett 2016) or supporting Donald Trump (Major et al. 2018) to nationalism and anti-immigration sentiment. The inclusion of the above predictor variables can help to build upon previous research of conservatism and ambivalent sexism. For example, conservativism, in a similar fashion to religion, has been associated with anti-LGBT + policies (e.g. Hunt 2011; Wang et al. 2016). This study seeks to provide a more robust understanding of the phenomenon and its predictor variables.

Finally, unique to this study, is the inclusion of mediated news to further explore the relationship between news consumption and ambivalent sexism. Ambivalent sexism is common for political journalists (Blumell 2018), in many newsrooms (IWMF 2013), and in news content (Attenborough 2013). Consequently, more is needed to understand the possible connection to news consumption and ambivalent sexism.

The US and UK were chosen as they share similarities. Both countries accepted populist notions in 2016 by voting yes for Brexit and Donald Trump as president. They have similar levels of ASI (Glick and Fiske 2001a) and masculinity (hofstede.com). There is also male privileging in the US (Cole 2009) and in the UK (Atewologun and Sealy 2014). Consequently, comparing and contrasting ASI in both countries can provide insight into common predictor variables and provide better understanding of international patterns of sexism among gay men. It is important to consider the intersectionality of various variables in order not to overgeneralize attitudes of gay men; furthermore, to pinpoint more accurately what factors correlate with ASI in gay men. By doing so, this study also examines to what extent predictor factors identified in a general population apply to gay men.

Male Privilege and Sexism

While this study focuses on ambivalent sexist attitudes of gay men in the US and UK, identifying male privilege under the theoretical lenses of intersectionality and hegemonic masculinity can provide context to why sexism has been found in gay men, and what factors contribute to it. Additionally, previous research shows a connection between privilege and sexism in both men and women (Grubbs et al. 2014), thus inquiry that goes beyond gender and includes sexuality is warranted.

Case et al. (2012, p. 3) describe privilege as “unearned benefits for certain groups.” They further explain that societal norms advantage individuals belonging to certain groups, particularly those “perceived to be male, white, heterosexual, or middle class.” At the heart of this privilege is an individual’s obliviousness to inherited gains and the oppression of others not possessing such inheritances (Johnson 2006; Pratto and Stewart 2012). Social norms uphold privilege, often making it invisible and consequently unchallenged (Wildman and David 1994). For instance, within lesbian and gay scholarship, gay men (particularly those who are white) have received the most attention; whereas, lesbian perspectives are often overlooked (Ellis and Peel 2011; Houston and Kramarae 1991; Savin-Williams and Diamond 2000). A gendering of sexualities also exists, evidenced by the stark contrast between the stereotyping of gay men and that of lesbian women (Edwards 2004).

Intersectionality

Though developed throughout the twentieth century, Crenshaw (1991) is primarily credited as the first to use the word intersectionality to describe the multiple factors that contribute to individual lived experiences. Intersectionality, specifically, is used to identify privilege and inequality (Gopaldas 2013). It addresses the basic questions: Who is included within a social group? How are members different or similar because of the inequalities they face? (Cole 2009). Cole (2009) explains by recognizing social hierarchies, researchers can better identify how social groups intersect (such as race, gender, sexuality, SES, ability, etc.)—shedding light on previously ignored or assumed experiences. Failing to acknowledge how members within specific groups differ from one another has led to favoring some marginalized groups over others in the quest for human rights. For example, political intersectionality describes the precedence of focusing on one disadvantage of a group that is otherwise privileged, thus excluding those of multiple subordinate identities (Cole 2008; Crenshaw 1995). Broadly speaking, different social political movements have advantaged certain people over others such as white women in women’s suffrage, white working class men in unions, black men in civil rights, and white cisgender men in gay rights. The purpose of intersectionality is not, however, to pit marginalized groups against each other—but rather to address power structures and form allies within and between groups (Cole 2008; Montgomery and Stewart 2012).

Hegemonic masculinity

Hegemonic masculinity contends societal norms privilege men to dominate women (Connell 1987). By doing so, it aggrandizes an ideal masculinity, even if not all men enact it (Connell 1995). Hegemonic masculinity also privileges men in public and private social, political, and economic institutions, thus creating a patriarchal status quo (Connell and Connell 2000). Hegemonic masculinity can be used to describe how gay men and women (bearing in mind the importance of intersectionality) share similarities and differ, since it was in part inspired by the women’s and gay liberation movements (Connell and Messerschmidt 2005). The prejudice, and even violence, gay men have experienced from straight men evidences a hierarchy of masculinities, positioning straight, cisgender males at the top (Connell and Messerschmidt 2005). This hierarchy negatively impacts men who do not ascribe to traditional masculinity, including prejudice from straight and gay men (Taywaditep 2002). For example, Cohen et al.’s (2009) experiment found that heterosexual men responded more positively to gay men who displayed typical masculine traits than to gay men who did not, and both men and women responded more positively to lesbians who portrayed typical feminine traits.

Hegemonic masculinity manifests in gay men thorough a process called mascing, where gay men reinforce their own masculinity by policing their rhetoric and actions, as well as maintaining masculine norms by seeking out masculine partners (Rodriguez et al. 2016). Those not as masculine are othered and ridiculed, thus creating a hierarchy of masculinities within the gay community. Nevertheless, hegemonic masculinity relies “on the evidence of the overall privileging of men and masculinity over women and femininity” (Connell 2014, p. 8). Moreover, Keiller (2010) found male narcissism is most strongly connected to hostile sexism towards straight women over lesbian women and gay men.

Ambivalent Sexism Inventory

To measure the potential sexism in gay men, we use Glick and Fiske’s (1996) Ambivalent Sexism Inventory (ASI). The inventory defines two categories of sexism: hostile and benevolent. Just as the patriarchal status quo maintains traditional gender roles delineated in hegemonic masculinity and male privilege in general, both hostile and benevolent sexism are enforced by male power structures, but in different ways (Glick and Fiske 1996). Hostile sexism antagonistically berates

women as unfit for any kind of public positions of power; whereas, benevolent sexism lovingly, but assuredly, defines a women’s place in the home (Glick and Fiske 1996). The juxtaposition of placing women on a pedestal through revered ideals of domesticity (benevolent sexism) and/or in the gutter through centuries of institutionalized abuse and oppression (hostile sexism) explains the ambivalence of sexism (Glick et al. 1997, p. 1323).

Ambivalence is further exasperated by men having a higher status than women, but ultimately depending on women for at least reproduction and intimacy—making sexism unique in that the offender and the target can be closely connected (Lee et al. 2010a, b). Glick and Fiske (2011b, p. 532) later explained, “Benevolent sexism was the carrot aimed at enticing women to enact traditional roles and hostile sexism was the stick used to punish them when they resisted.” Backlash from hostile sexism is also extended to non-traditional men (Glick et al. 2015).

While hostile sexism is often recognized as negative, the impact of benevolent sexism is underestimated and consequently endorsed by men and women (Bosson et al. 2010). For instance, a 19 country study showed men had higher levels of hostile sexism than women, but women had similar or sometimes higher levels of benevolent sexism as men (Glick and Fiske 2001a). Interestingly, research shows hostile sexism can be significantly reduced through diversity education, but the same is not true for reducing benevolent sexism (Case 2007). Moreover, hostile sexism has been shown to correlate with anti-women’s rights attitudes, but benevolent sexism correlated with support for women’s rights (Masser and Abrams 1999).

Benevolent sexism is sub-divided into three sub-scales: protective paternalism, complementary gender differentiation, and heterosexual intimacy. Glick and Fiske (1996) developed the three sub-scales in their original study to address different aspects of benevolent sexism. Firstly, protective paternalism situates women as needing a male authority to watch over them, as expected of a father to his children (Glick and Fiske 1996). Complementary gender differentiation distinguishes women as having specific feminine traits, such as purity and morality, which balance masculine traits of power and dominance (Glick and Fiske 1996). Heterosexual intimacy creates a “dyadic dependency” (493) for men, who depend on a subordinate group for important sources of pleasure and family life (Glick and Fiske 1996). This study eliminates heterosexual intimacy since the sampled population is gay men.

Over the past three decades, ambivalent sexism has been observed in various scenarios including amongst gay men. For instance, Zheng and Zheng (2015) used ambivalent sexism in a sample of gay men to understand if gender expectations established by heterosexual men and women were applicable. Notably, they found that hostile sexism was significantly correlated to a preference for traditionally masculine faces, but not to benevolent sexism (Zheng and Zheng 2015). Also using ambivalent sexism, Zheng et al. (2017) found hostile sexism significantly predicted gay and bisexual men’s preference of a complementary sexual partner. This meant those with higher levels of hostile sexism and a preference of being a “top” during sex, sought a “bottom” partner or vice versa (Zheng et al. 2017).

To investigate ambivalent sexism levels, this study asks two overarching research questions:

RQ1: To what extent does benevolent sexism manifest in the US and UK among gay men?

RQ2: To what extent does hostile sexism manifest in the US and UK among gay men?

The purpose of asking two general research questions is to account for the various predictor variables that will be analyzed through the multiple statistical analyses of the benevolent sexism and hostile sexism scales. Since this is an exploratory study, which seeks to understand how variables that predicted ambivalent sexism in general populations manifest in gay men, research questions were only used and not hypotheses.

Predictor Variables of Ambivalent Sexism

Also important to ambivalent sexism research are various predictor variables such as religiosity, political ideology, personality, education, and even facial hair (Hellmer et al. 2018). Within the two broad research questions, this study looks at two commonly tested predictor variables: religiosity and political ideology to understand how they manifest in gay men. It also looks at three overlooked, but perhaps significant factors that may result in significance: nationalism, anti-immigration attitudes, and news consumption.

Sexism and Religion

Most major world religions rely on a patriarchal status quo which subjugates women and, accordingly, previous research shows correlation between ambivalent sexism and religiosity. For Turkish students, Islamic religiosity for both genders, benevolent sexism in women, and hostile sexism in men, all predicted endorsing women’s subordination to men (Glick et al. 2016). Furthermore, Islamic religiosity correlates with benevolent sexism for men and women, and hostile sexism for men (Tasdemir and Sakalli-Ugurlu 2010). The same was partially true for Israeli Jewish men and women in terms of benevolent sexism, but hostile sexism actually decreased for men with religiosity (Gaunt 2012). In Christianity, correlations have been found in benevolent sexism, but not hostile (Glick et al. 2002; Mikołajczak and Pietrzak 2014). Maltby et al. (2010) found that as Christian religiosity in men increased, so did their ambivalent sexism.

Sexism and Political Ideology

Just like many religions, conservative values also uphold a patriarchal status quo, which has been connected to ambivalent sexism (Christopher and Mull 2006; Feather and McKee 2012). Specifically, benevolent sexism found in men and women correlates with conservative ideology which genders women to traditional roles (Glick et al. 1997; Viki et al. 2003; Sibley et al. 2007). Nonetheless, hostile sexism and traditional gendered attitudes were predictor variables for voting Donald Trump in the 2016 US presidential election (Bock et al. 2017). Although not all research is neatly divided between political ideologies. For instance, Huang et al. (2016) found that benevolent sexism, regardless of political party membership, predicted anti-reproductive rights attitudes.

Sexism, Nationalism, and Anti-immigration Attitudes

Part of political ideology are nationalism and anti-immigration attitudes, which are understudied as predictor variables of sexism. Thobani (2000) did assert that anti-immigration attitudes and nationalism in Canada were linked to a history of racism and sexism, namely against aboriginal women and women of color. In an experiment, Sarrasin et al. (2015) found that benevolent sexism played a role in anti-immigration attitudes when participants were exposed to pictures of threatening male immigrants.

Sexism and News Consumption

News consumption as a predictor of sexism is understudied, especially when considering news organizations have a history of producing sexist content, and tolerating sexism in the workplace (Attenborough 2013). For example, news coverage has consistently dismissed the seriousness of sexism and sexual misconduct (O’Hara 2012; Pennington and Birthisel 2016; Romaniuk 2015) or engaged in rape culture tropes such as blaming the victim while excusing the accused (Jordan 2012). Not to mention the long list of high profile news personnel who have faced recent accusations of various forms of sexual harassment (for a list see e.g. Corey 2017).

Method

This study employed a single-wave survey on self-identified gay men in the UK and the US. Recruitment was performed by Qualtrics, a third-party web-based data collection. Qualtrics partners with over twenty web-based panel providers to supply “diverse, quality respondents” based on the researchers’ inclusion criteria (Ibarra et al. 2018, p. 3). Inclusion criteria for participants included that participants must self-identify as (1) male; (2) gay; and (3) be of legal voting age. Participants were asked questions that focused on media consumption, religion, political ideology, nationalism, anti-immigration attitudes, and ambivalent sexism. The survey (N = 291) was hosted online and participants for this study were recruited from the US (n = 161) and the UK (n = 130). Data was collected over a ten-day time period in October 2017.

In regards to the sample: (1) 94.5% of the sample were registered to vote; (2) 24.8% identified as conservative, 24.8% as moderate, and 50.3% as liberal; (3) 71.5% identified as White, 11.4% as Asian, 7.6% as African American/Black; 4.8% as Latino/Hispanic, 0.7% as Native American/Indigenous, 0.3% Middle Eastern, and 3.4% as other; (4) 20.8% identified as Mainline Protestant/Christian, 17.4% as Catholic, 14.6% as atheist, 9% as agnostic, 1.4% as Mormon, 1% as Evangelical, 1% as Jewish, 1% as Muslim, and 33% had no religious affiliation; (5) 9.7% served in the armed forces; (6) 100% identified as gay; (7) ranged in age from 21 to 79 years old; (8) household income (in USD) spanned 25.1% less than $30,000, 25.2% between $30,001 and $60,000, 24.1% between $60,001 and $100,000, and 25.6% above $100,001; and (9) education consisted of 11.7% with a high school diploma or less, 28.6% with some post-secondary education, 32.8% with a bachelor’s degree, and 26.9% with a post-graduate degree.

Measures

Ambivalent Sexism

An eighteen-item sexism scale was adapted from Glick and Fiske (1996). Participants were asked their agreement on a seven-point scale—strongly disagree to strongly agree. In the original study, four variables formed a sub-scale of benevolent sexism titled heterosexual intimacy. Since the participants in this study identified as gay, this subscale was eliminated. The remaining variables were run in a Promax factor analysis with a strong Kaiser–Meyer–Olkin result of .87, and a significant Bartlett’s test of sphericity (χ2 (153) = 3476.58, p < .001). Table 1 outlines how the items loaded. For hostile sexism, all but three variables loaded as expected. These variables were eliminated and the remaining combined for the hostile sexism scale (Cronbach’s α = .94, M = 4.62, SD = 1.47). For the two subscales of benevolent sexism, protective paternalism loaded as expected, with the exception of the reverse coded variable. All variables loaded as expected for the complementary gender differentiation scale. These subscales were then combined to form benevolent sexism (Cronbach’s α = .87, M = 4.68, SD = 1.30).

Table 1 Summary of exploratory factor analysis results for hostile and benevolent sexism (protective paternalism, complementary gender differentiation, and heterosexual intimacy)
Full size table

Religiosity

A twelve-item revised religious scale was adapted from Altemeyer and Hunsberger (2004) and participants were asked their agreement on a seven-point scale—strongly disagree to strongly agree (Cronbach’s α = .92, M = 2.82, SD = 1.38).

Level of Conservativeness

Participants were asked their agreement on a seven-point scale—strongly disagree to strongly agree on their level of conservativeness (M = 2.61, SD = 1.12). The skewness of the variable was .23 (SE = .14), and the kurtosis was − .86 (SE = .28).

Nationalism

A seventeen-item nationalism scale was adapted from Todosijević (2001) and participants were asked their agreement on a seven-point scale—strongly disagree to strongly agree (Cronbach’s α = .93, M = 2.74, SD = .80).

Anti-immigrant Attitudes

Five statements were adapted from Watson and Riffe (2012) and participants were asked their agreement on a seven-point scale—strongly disagree to strongly agree (Cronbach’s α = .95, M = 3.41, SD = 1.87).

Media Consumption

Participants were asked on a five-point scale how much time they spent consuming conservative television news, which consisted of Fox News for US participants and Sky News for UK participants (M = 2.20, SD = 1.44). They were also asked about their consumption of moderate television news, which was CNN for US participants and BBC for UK participants (M = 3.23, SD = 1.50). Participants also indicated their consumption of liberal television news, which was MSNBC for US participants, and Channel 4 news for UK participants (M = 2.70, SD = 1.47). Participants were also asked to indicate their consumption of conservative (M = 2.08, SD = 1.25), moderate (M = 2.63, SD = 1.28), and liberal (M = 2.55, SD = 1.43), online news. The television and online variables were then combined to form three separate variables labeled: conservative media, moderate media, and liberal media.

Results

To answer the research questions, which sought to understand how benevolent sexism (RQ1) and hostile sexism (RQ2) compared between the US and UK, several statistical tests were run.

Firstly, to understand overall levels of ambivalent sexism, independent t-tests were run. There was significance for benevolent sexism, t(289) = − 2.81, p ≤ .01, and hostile sexism, t(289) = 2.54, p ≤ .01. US participants (M = 4.49 and SD = 1.28), had significantly lower benevolent sexism than UK participants (M = 4.91 and SD = 1.29). UK participants (M = 4.38 and SD = 1.42), on the other hand, had significantly lower levels of hostile sexism than US participants (M = 4.81 and SD = 1.48). The results show that ambivalent sexism manifests differently between countries. Even though the mean for both scales was within 1 scale point, it was significant.

Multiple regressions were then performed to investigate religiosity, political ideology, nationalism, anti-immigration attitudes, and media consumption (conservative, moderate, and liberal) in ambivalent sexism. This includes controlling for race, education, and household income. Table 2 shows Pearson Correlations between the predictor variables and ASI.

Table 2 Intercorrelations for predictor variables and ASI
Full size table

Table 3 shows results for the predictor variables and benevolent sexism in which there was significance for the US, F(13, 157) = 2.64, p ≤ .01, and the UK, F(13, 129) = 4.59, p ≤ .001. The only significant predictor for both countries however, was religiosity. Some of the variables had a negative correlation for both countries, such as political ideology, but none were significant. Other variables were positive for one country, while being negative for another, but again without significance. The importance of this analysis shows how the traditional views of cherishing women and positioning women and men in traditional gendered norms (Glick et al. 2002; Mikołajczak and Pietrzak 2014) are similar for gay men.

Table 3 Ordinary least squares (OLS) regression model predicting levels of benevolent sexism in gay men in the UK and the US
Full size table

Multiple regressions were then run with the same predictors and hostile sexism (see Table 4). Once again there was significance for the US, F(13, 157) = 10.99, p ≤ .001, and the UK, F(13, 129) = 5.27, p ≤ .001. Nationalism and anti-immigration attitudes were significant for both countries. Religiosity was significant for the UK. In the US, consuming conservative news media was significantly positive, but consuming moderate media was significantly negative. Finally, Asian participants in the US also had significance. For hostile sexism, political ideology was not significant as it has been with other ASI studies (Christopher and Mull 2006; Feather and McKee 2012). Interestingly, nationalism and anti-immigration attitudes, often associated with conservativism, were significant. This shows the importance of connecting sexuality beyond political ideology alone to also include other potential, yet often overlooked, predictor variables.

Table 4 Ordinary least squares (OLS) regression model predicting levels of hostile sexism in gay men in the UK and the US
Full size table

Discussion

This study measured ambivalent sexism of gay men in the US and UK. Its purpose was to (1) understand how hostile and benevolent sexism manifest in gay men, and (2) test predictor variables, which can influence levels of ambivalent sexism. ASI (Glick and Fiske 1996) was used to examine hostile (antipathy towards women) and benevolent (attitudes of cherishing, but subjugating women) sexism. Benevolent sexism is further divided into three sub-scales: protective paternalism, complementary gender differentiation, and heterosexual sexual intimacy (Glick and Fiske 1996). Only the first two were used for this study.

Sexism is ubiquitous—found in men and women, different cultures and countries, etc. (e.g. Glick et al. 2000, 2001a). The purpose of focusing on gay men for this study was to analyze how ambivalent sexism may occur. Gender as a variable influences ambivalent sexism (e.g. Glick et al. 1997; Hart et al. 2013; Lee et al. 2010), but more research is needed to understand the impact of sexuality intersecting with gender. We understand intersectionality as the emphasis on “interlocking effects of race, class, gender, and sexuality, highlighting the ways in which categories of identity and structures of inequality are mutually constituted and defy separation into discrete categories of analysis” (Dill and Kohlman 2012, p. 154). As Shields (2008) notes, intersectionality is “an urgent issue” in empirical research on gender. We posit that intersectionality is also important for research on sexuality—particularly when studying areas largely dependent on heteronormativity like ambivalent sexism.

Some pop culture narratives assume a natural alliance between women and gay men (Frith et al. 2010; Kates 2000); which of course is true at times. Nevertheless, as Cole (2008, 2009) elucidates, part of the importance of intersectionality is to address social structures, which privilege or disadvantage marginalized groups. In other words, sexism cannot be fully eradicated until it is understood how it exists in all circumstances. The same is true for homophobia, racism, xenophobia, transphobia, and so forth. It consequently is problematic to assume that men who are gay are automatically not hostile or benevolent sexists.

Examining possible predictor variables significant for ASI also contributes to understanding intersectionality. Attitudes and beliefs commonly accepted via affiliations (such as religions or political parties) can differentiate individuals belonging to the same social groups. For instance, even though more US women vote for the Democratic Party overall (Tyson and Maniam 2016), the vast majority of white evangelical Christian women vote for the Republican Party (Smith and Martinez 2016). Consequently, by explicating influences such as religion, politics, nationalism, anti-immigration attitudes, and news consumption, researchers can avoid overgeneralizing populations.

Some predictor variables in this study performed as anticipated and some did not. Though ample previous research shows a connection between conservative ideology and ambivalent sexism (Christopher and Mull 2006; Feather and McKee 2012; Glick et al. 1997; Sibley et al. 2007; Viki et al. 2003), this study found no significant correlation. There was also less significance than expected for consuming conservative news. It is nonetheless important to note, as there is little previous research connecting consuming conservative news and ambivalent sexism.

Of all the predictor variables, religiosity had the biggest impact in the US and UK. The role of religiosity, especially for benevolent sexism, supports previous findings (Glick et al. 2002; Mikołajczak and Pietrzak 2014). It is also consistent with previous research which shows religiosity specifically increased ambivalent sexism in men (Maltby et al. 2010). Future research should examine how this affects the hierarchy of masculinities as noted in hegemonic masculinity (Connell and Messerschmidt 2005). For example, internalized heterosexism found in gay men and lesbians is linked to their level of sexism towards women (Murgo et al. 2017; Warriner et al. 2013).

Nationalism and anti-immigration attitudes were significant predictor variables for hostile sexism in the US and UK. It would thus appear that having aggressive attitudes towards women defined in hostile sexism make it more likely for the participants in this study to also have aggressive attitudes towards immigrants and the need to protect one’s country. This builds on previous literature which connects anti-immigration and nationalism attitudes to sexism (Sarrasin et al. 2015; Thobani 2000). One explanation for the connection is how hostile sexism, anti-immigration, and nationalism represent a belief in a patriarchal status quo (Connell and Connell 2000), and anger towards what endangers current social hierarchies. Even though gay men also face retribution for disrupting the status quo (Connell and Messerschmidt 2005), by focusing on other external disruptions (eliminating sexism and xenophobia), this study illustrates how gay men can exhibit sexism and anti-immigration attitudes in recompense for their loss of status in the hegemonic hierarchy. Even if gay men are not as high on the hierarchy as straight men, they attempt to be higher on the hierarchy than women. One group can be marginalized and still have privilege or negative attitudes towards other marginalized groups.

Noting that conservativism wasn’t a significant factor in this study, but predictor variables often associated with conservativism such as nationalism, anti-immigration attitudes, and religiosity were significant factors, shows individuals may harbor conservative ideologies without identifying as conservatives. Additionally, it illustrates there are specific topics, which transcend political ideologies (Huang et al. 2016). Overall, intersectionality is complex and multiple factors are needed in identifying problematic attitudes of marginalized groups. By identifying the significance of religiosity in benevolent sexism and nationalism and anti-immigration attitudes in hostile sexism, this study purports that gay men in this sample are most likely to exhibit ambivalent sexism in conjunction with other predictor variables rather than common demographic characteristics.

In relation to ambivalent sexism more broadly, the findings of the predictor variables call on the need to consider ambivalent sexism within studies of populism, which are strongly nationalistic and anti-immigration (Corbett 2016; Major et al. 2018). For instance, ambivalent sexism was a significant predictor variable of voting for Trump in 2016 (Bock et al. 2017; Ratliff et al. 2019).

Limitations and Future Research

This study’s limitations include the use of one quantitative method. The intent was to be able to statistically measure attitudes of sexism and possible predictor variables. Future research should include qualitative inquiry such as in-depth interviews to further uncover patterns found here. A second limitation is the study’s modest sample size. The findings, nonetheless, contribute to understanding of the marginalization of one marginalized population over another. Future research may include larger sample sizes, as well as non-western sampled populations. A focus on race and ethnicity would also be useful as the majority of participants were white non-Hispanic. Furthermore, since general populations have been well-studied in regards to ASI, our sample focused on gay men only. Future research could include a more general population in order to make further comparisons of sexuality and other factors.

The intent of this study was to illustrate how harmful attitudes subjugate populations, even by other subjugated populates. Future research should also extend to other harmful attitudes such as homophobia, racism, xenophobia and so forth found in other marginalized groups.

Conclusion

The current political and social landscapes in several countries around the globe are starting to scratch the surface of sexism and sexual abuse. The #metoo (metoomvmt.org) and #timesup (timesupnow.com) movements for instance, have brought much needed attention to sexual abuse and gendered inequalities present in women’s public and private lives. This study contributes to these movements by highlighting often overlooked ambivalent sexism of a marginalized group, in this case gay men. These findings also have practical uses for political and activist movements, as the goal of discussing intersectionality is to form better alliances (Cole 2009). We also suggest this study could have positive impact of gay men’s self-awareness of possible proclivities of ambivalent sexism, specifically religiosity in benevolent sexism, and nationalism and anti-immigrant attitudes in hostile sexism.

The benefits of reducing ambivalent sexism in gay men are many. Besides helping to eliminate sexism, it can reduce internalized heterosexism (Murgo et al. 2017; Warriner et al. 2013), and the influence hostile sexism has on intimate partner choices in gay men (Zheng et al. 2017; Zheng and Zheng 2015). Furthermore, it helps to eliminate power structures which disadvantage women in all parts of society (Connell and Messerschmidt 2005) and build better alliances between marginalized groups (Cole 2008; Montgomery and Stewart 2012).