Hostile, Benevolent, Implicit: How Different Shades of Sexism Impact Gendered Policy Attitudes

Gothreau, Claire; Arceneaux, Kevin; Friesen, Amanda

doi:10.3389/fpos.2022.817309

ORIGINAL RESEARCH article

Front. Polit. Sci., 04 July 2022

Sec. Political Participation

Volume 4 - 2022 | https://doi.org/10.3389/fpos.2022.817309

This article is part of the Research Topic Mind the Backlash: Gender Discrimination and Sexism in Contemporary Societies View all 11 articles

Hostile, Benevolent, Implicit: How Different Shades of Sexism Impact Gendered Policy Attitudes

$\nClaire Gothreau$ Claire Gothreau¹^*

Kevin Arceneaux²

Amanda Friesen³

¹Center for Philosophical-Experimental Study of Discrimination, Aarhus University, Aarhus, Denmark
²Center for Political Research (CEVIPOF), Sciences Po, Paris, France
³Department of Political Science, Western University, London, ON, Canada

Advances in gender equality and progressive policies are often stymied by cultural sexist systems and individual-level sexist attitudes. These attitudes are pervasive but vary in type—from benevolent to hostile and implicit to explicit. Understanding the types of sexism and their foundations are important for identifying connections to specific social and political attitudes and behaviors. The current study examines the impact of various manifestations of sexism on attitudes regarding policies and public opinion issues that involve gender equality or have gendered implications. More specifically, we look at attitudes on reproductive rights, support for the #MeToo Movement, equal pay, and paid leave policies. In Study 1 we use data from a high-quality web panel (n = 1,400) to look at the relationship between hostile, benevolent, and implicit sexism, and reproductive rights attitudes, as well as support for the #MeToo Movement. In Study 2 we use data from the American National Election Study (n = 4,270) to examine the relationship between hostile and modern sexism and attitudes on abortion, equal pay, and paid family leave. Overall, these results reveal a complicated relationship between different conceptualizations of sexism and gendered attitudes, underscoring the need to consider how different forms of sexism shape broader social and political views, from both a normative perspective for societal change and a measurement approach for research precision.

Introduction

Politics can be a masculine enterprise, both historically and presently around the world. For many decades, feminist scholars and activists have identified and criticized the gendered structures and attitudes that lead to sexist policies and exclusion of women from political spaces. After the U.S. election of Donald Trump, an election that featured the first major party woman nominee and a candidate that frequently made sexist remarks, and the advent of social movements like #MeToo, more researchers began exploring the role of sexism and gender attitudes in American politics. Sexism batteries became more commonly included in large surveys like the American National Election Study (ANES) and the Cooperative Congressional Election Study (CCES). Prior to 2016, these surveys had inconsistently included survey questions tapping sexist attitudes (Schaffner, 2021). The focus of much of the empirical research was the impact of sexism on vote choice. Even when controlling for partisanship, sexism is a powerful predictor of vote choice (Valentino et al., 2018), and hostile sexism, in particular, is connected to Trump support in 2016 (Ratliff et al., 2017; Cassese and Holman, 2019). We know less about the implications of different forms of sexism on public opinion and policy attitudes, particularly those that are explicitly and implicitly gendered. Moving beyond the electoral context and candidate support, we consider how different manifestations of sexism impact political attitudes and demonstrate that parsing out benevolent, hostile, modern, and implicit sexism may help us better understand why the connection between gender attitudes and issues like abortion have been mixed (Strickler and Danigelis, 2002; Jelen, 2015).

We build on an area of research that conceptualizes sexism and the opposition to gender equality as a way of justifying male dominance and maintaining existing gender relations (Jost and Kay, 2005; Cassese and Holman, 2019). All forms of sexism contribute to the maintenance of the gender status quo, but variation in these types of attitudes result in varying support for gender-related policies. The Ambivalent Sexism Inventory delineates hostile and benevolent sexism as distinct forms of prejudice (Glick and Fiske, 1996). Additionally, we explore the role of implicit sexism, prejudiced attitudes held at the nonconscious level (Jost et al., 2004). Because of the social desirability some people may exhibit in the presentation of sexist survey items, implicit tests of gender stereotypes can influence people's attitudes toward female candidates above and beyond their explicitly stated gender preferences (Mo, 2015).

The connection between sexism and gender attitudes and policy positions seems straightforward. We would expect that those who hold sexist attitudes would be less likely to support progressive policies with expressly gendered implications. In some policy areas, this direct connection appears to exist. However, in the gender-salient domain of reproductive rights, the association between sexism and gender attitudes is less clear and has only been somewhat elucidated by separating hostile from benevolent forms of sexism (Begun and Walls, 2015; Huang et al., 2016; Hodson and MacInnis, 2017; Petterson and Sutton, 2018).

Using an original survey and data from the ANES, we test whether and when hostile, benevolent, modern, and implicit sexism predict attitudes toward gendered public opinion issues and policy attitudes. We argue that all forms of sexism contribute to the subjugation of women in society. However, there are important nuances in different manifestations of sexism that have implications for public opinion and policy attitudes. Our results across the studies are considerable to unpack but the biggest takeaway is that sexism is not a uniformly negative predictor of progressive gender attitudes. We find that benevolent sexism was positively related to support for the #MeToo Movement whereas hostile sexism was a strong negative predictor. This reflects the fact that hostile sexism uncovers antipathy toward women while benevolent sexism taps the idea that women are morally superior and purer than men and should therefore be protected. We find that hostile sexism predicts less support for abortion and birth control access, as well as funding for Planned Parenthood. In our second study, we replicate these findings on abortion but find that modern sexism, not hostile, is related to less support for equal pay and paid leave policies. We also argue that sexism researchers should consider that the relationship between sexism and different political outcomes may be conditional on gender. We see in our data that for men, benevolent sexism does not always predict less progressive gendered policy attitudes, but it does for women. Although our data cannot speak to the exact mechanisms that connect different forms of sexism to policy and public opinion attitudes, we show that this connection does exist but is conditional on the type of sexism measured. Our results also underscore the need for more research aimed at understanding the antecedents and consequences of different forms of sexist attitudes.

Theory

How We Measure Sexism

Though empirical research on sexism and political outcomes like vote choice has proliferated in recent years, particularly after the U.S. election of Donald Trump, feminist activists and theorists have long discussed the role of sexist institutions and attitudes in stymying gender equality in education, pay, healthcare, and in politics. Feminist theorists have highlighted the ways in which sexism exists in institutionally structured settings, such as when women are paid less than men for the same labor, but also sexism in interpersonal interactions and even in the private sphere of the home (Okin, 1989; Nussbaum, 1998; Swim et al., 2001). Both forms of sexism reinforce existing gendered systems of dominance and subordination. Gender inequalities are often the result of sexism, but sexism also constitutes tacit beliefs and attitudes that individuals hold. Research in the social sciences in the 1980s and 90s began to try and measure these attitudes. We constructed Table 1 to define the main types of sexism measures used in social science research.

TABLE 1

Table 1. Measures of sexism used in social science research.

According to Glick and Fiske (1996), sexism is an ambivalent form of prejudice in which antipathy toward women who seek to undermine male dominance coexists with the idealization of women who occupy the roles carved out for them in the patriarchal system—wives, mothers, and homemakers in need of male protection (Glick and Fiske, 2001). Glick and Fiske (1996) introduced the Ambivalent Sexism Inventory that captured the way in which negative gender attitudes can be actively hostile but also paternalistic and patronizing. This measure reflects the fact that cultural representations of women, throughout history, have not always been strictly negative (Eagly and Mladinic, 1994). Women are represented as caregivers and housewives whose role is primarily within the domestic sphere. However, women are also subject to negative stereotypes and bias, particularly when they step outside of domestic roles. The Ambivalent Sexism Inventory reflects the duality of these cultural representations and stereotypes. The first dimension, hostile sexism, defines women as a group in competition with men, vying for social dominance. Someone who holds strong hostile sexist attitudes believes that women are inferior to men and, thus, incapable, and unworthy of power. As a result, this person is hostile toward women who do not accept their assigned roles in the patriarchy and perceive calls for gender equality as a ploy to usurp men's power and assert dominance over men. In contrast, the second dimension, benevolent sexism, adopts a more positive, but ultimately patronizing and paternalistic view of women. It shares with hostile sexism the notion that women are not capable of wielding power, and because of this, they require protection by men. As separate dimensions of sexism, individuals can be low on both (non-sexists) and high on both (ambivalent sexists), but they can also be high on one dimension and low on another. Hostile sexists are those who only see women as a threat to men's power, while benevolent sexist tend to simply see women as fragile, precious, and possessing moral superiority (Glick and Fiske, 1996).

The modern sexism measure was specifically constructed to capture attitudes that deny the existence of systematic discrimination against women (Swim et al., 1995). The development of this measure coincided with discussions in popular culture and feminist discourse about backlash against modern feminism (Banet-Weiser et al., 2019). Many critics of feminist movements hold the belief that we live in a post-feminist world in which equality has already been achieved (Anderson, 2015). These beliefs resulted in a cultural backlash against many of the advances achieved by feminism in the 1970s (Faludi, 1991). Contrary to traditional sexism, which openly endorses the idea that women are inferior to men, modern sexism is a subtler form of prejudice that involves a resentment toward demands for gender inequality. Individuals who hold modern sexist attitudes often feel negatively about the shifting roles of women in society. In analyses of the comparability of different sexism measures, the items on the modern sexism scale have been found to load together with hostile sexism items, although modern sexism items tapping antagonism and resentment more closely mapped onto hostile sexism than the items tapping denial of gender discrimination (Schaffner, 2021). A less frequently used scale is the Old-Fashioned Sexism Scale, also constructed by Swim et al. (1995). Modeled after the Old-Fashioned Racism scale, this scale measures blatant expressions of sexism like believing that men are smarter and more logical than women.

These varying patterns of sexist attitudes often have different attitudinal and behavioral implications. For example, hostile sexism is correlated with negative attitudes toward women in managerial positions whereas benevolent sexism is not (Masser and Abrams, 2004; Eagly and Carlie, 2007). Hostile sexists are more likely to condone violence toward women, including rape (Begany and Milburn, 2002; Masser et al., 2006), whereas benevolent sexists react negatively toward overtly crude, hostile treatment of women (Cassese and Holman, 2019). This is not surprising given that some studies have found only a weak positive correlation or no correlation between these two measures (Glick and Fiske, 2011). However, the mixture of negatively putatively positive stereotypes that make up hostile and benevolent sexism create “complementary gender stereotypes” that offer a justification for gender inequality (Jost and Kay, 2005). In addition, individuals need not be fully aware that they hold sexist stereotypes. Notions that women's roles are confined to being homemakers and mothers can be internalized and held at the nonconscious “implicit” level (Jost et al., 2004). When people formulate an attitude or a behavioral intention, their minds first draw on a network of nonconscious processes that serve as a starting point for conscious thought (Bargh and Chartrand, 1999; Lodge and Taber, 2013). Sometimes these intuitions are incorporated into people's attitudes and behavior without much consideration and guide people's political decisions outside of people's awareness (Arceneaux and Vander Wielen, 2017).

Sexism in American Politics

Much of the literature on sexism in American politics has focused on the ways in which sexist attitudes and stereotypes impact women political candidates (Sanbonmatsu, 2002; Bauer, 2015; Mo, 2015; Cassese and Holman, 2017). Though our main focus is on attitudes toward gendered policies, understanding the prevalence of sexism aimed at women in public life is helpful for investigating how else this prejudice is likely to spill over into policy preferences. Because partisanship is the strongest lever in American political behavior, there is a complicated relationship between sexism and candidate evaluation and vote choice, but ultimately research shows that when women run for office, particularly at higher levels, they face gender bias (Lawless, 2004; Paul and Smith, 2008). Vote choice chiefly comes down to incumbency and partisanship (Dolan, 2014), but gender stereotyping and sexism still play a role in electoral politics (Schneider and Bos, 2014) and often lead voters to have different standards of evaluation for men and women politicians (Barnes and Beaulieu, 2014; Barnes et al., 2020). Cassese and Barnes (2018) find that despite the blatant sexism present in the 2016 presidential race, many white women endorsed sexist beliefs, and these beliefs informed their vote choice. Both modern and traditional sexism were significant predictors of an individual's presidential vote in 2016 (Knuckey, 2019), and both hostile and benevolent sexists punish women politicians involved in sex scandals more than non-sexists (Barnes et al., 2020). Relatedly, concerns about gender discrimination predict support for a woman president (Huddy and Carey, 2009), and denials of this discrimination are associated with opposition to women politicians like Hillary Clinton (Sulfaro, 2007; Tesler and Sears, 2010; McThomas and Tesler, 2016) and the gender gap in partisanship (Simas and Bumgardner, 2017).

Indeed, women in the electorate are not immune from the influence of sexism in their politics. Personal experiences of sexism and sexual harassment can actually motivate political engagement (Bankert, 2020). Similarly, Hansen and Dolan (2020) find that women who reported being sexual harassed at work were more likely to mobilize. The broader #MeToo Movement, in which issues of sexual harassment and assault were brought to the forefront of American politics, also may have influenced increased participation among women (Dittmar, 2020). Sexism, in women's public and private lives, has also contributed to the gender gap in political interest and engagement (Carroll, 1989; Burns et al., 2001).

Importantly for our purposes, the type of sexism exposure matters. Experiencing hostile sexism can motivate engagement in collective action whereas benevolent sexism seems to decrease this interest (Becker and Wright, 2011). For those who hold sexist attitudes, hostile sexists who were exposed to sexist attacks against Hillary Clinton showed increased support for Trump and decreased support for Clinton, while benevolent sexists exposed to the same attack responded with increased support for Clinton (Cassese and Holman, 2019).

The Impact of Sexism on Policy Attitudes

Sexism not only impacts outcomes like vote choice and evaluations of political candidates, but it can impact attitudes, particularly political opinions that are gender salient. For example, modern sexism is associated with a denial of discrimination against women and a lack of support for policies designed to help women in the domains of education and the workplace (Swim et al., 1995). Hideg et al. (2016) find that benevolent sexism is associated with more support for employment equity policies supporting women, but this support did not extend into more stereotypically masculine workplace settings. Hostile sexists are less likely to support the adoption of gender quotas to increase women's representation in politics, whereas benevolent sexists are more likely to support these policies even though they do not support gender equality generally (Beauregard and Sheppard, 2021). Hostile sexism predicts victim-blaming attributions for the gender gap in income inequality (Connor and Fiske, 2019) as well as opposition to breastfeeding in public (Huang et al., 2020) and tolerance for sexual harassment (Russell and Trigg, 2004). Modern sexist attitudes are related to the belief that sexual harassment is not pervasive, the notion that the #MeToo Movement has gone too far, and opposition to workplace harassment training (Archer and Kam, 2021). Recent research even shows that sexism can impact compliance with public health measures, with higher levels of benevolent sexism actually increasing compliance (Chen and Farhart, 2020). This work is all in contrast to earlier research that failed to find definitive connections between sexism and gender-salient policy attitudes (Twenge, 1997), likely because measures of sexism now capture more subtle forms of gender-based prejudices.

Within the domain of reproductive rights, one of the most gender-salient policy areas, the connection between gender attitudes, sexism, and support for access to abortion and birth control is inconsistent (Jelen and Wilcox, 2003; Patel and Johns, 2009; Barkan, 2014). Some scholars find a positive correlation between opposition to abortion and both forms of ambivalent sexism, hostile and benevolent (Hodson and MacInnis, 2017), while others find only evidence for a correlation between abortion attitudes and benevolent sexism (Huang et al., 2016) or hostile sexism (Petterson and Sutton, 2018). These inconclusive findings may arise from the fact that most of these studies come from small convenience samples collected on college campuses, and they all focus on a relatively limited definition of reproductive rights—namely, abortion. People's attitudes about abortion tend to be relatively crystalized and heavily linked to moral absolutes (Wilcox and Norrander, 2002; Jelen and Wilcox, 2003; Mooney and Schuldt, 2008; Jelen, 2014; Ryan, 2014), whereas broader policy attitudes about women's reproductive rights, such as access to birth control, may be more malleable (Arceneaux and Kolodny, 2009).

Furthermore, gender identity and sexist attitudes may not supersede other identities like race and ethnicity. Women feel closer links to men of their race than their women peers of other groups (Junn, 1997; Gay and Tate, 1998). Compared to other groups, women's levels of group consciousness tend to be lower (Clayton and Crosby, 1992), which in part explains why they lack the political cohesion that other historically marginalized groups display (Cassese and Barnes, 2018). In U.S. politics, accounting for racial identity demonstrates that the supposed “gender gap” in women preferring Democratic to Republican candidates disappears, with white women selecting Republican presidents in an overwhelming majority of previous elections (Junn and Masuoka, 2020). White women are more likely to vote and prefer policies connected to their race and partisanship over their gender (Cassese and Barnes, 2018), and Black women also politically engage in ways more consistent with linked fate toward their racial rather than gender group (Stout and Tate, 2013). Thus, an intersectional lens is necessary to fully understand the experiences and preferences of women, particularly of Black American women (Crenshaw, 1989; Hancock, 2007; Brown, 2014). Indeed, attitudes on reproductive rights in the U.S. differ across racial/ethnic groups as well as religious affiliations (Smith, 2013; Jelen, 2014; Lizotte, 2015; Holman et al., 2020). The rich body of literature on intersectionality is necessary to understand how sexism operates in society, but there is less work on empirically connecting measures of sexism to intersectionality (see Junn and Masuoka, 2020 for discussion on how variation in socioeconomic and religious indicators matter more for the white woman vote). One of the challenges is the small sample problem—that most “representative” samples of American adults do not include enough participants from minority racial/ethnic groups to conduct meaningful analyses. Of course, this is the limitation of quantitative work that attempts to wedge intersectional theory into a model (Hancock, 2007). Nevertheless, there has been some recent research demonstrating that Black men and women are more likely to endorse benevolent sexism attitudes, as compared to white men and women (Davis et al., 2022), but we do not know how this plays out in public opinion of policy issues.

Previous literature in psychology and sociology has shed light on the relationship between sexism and a range of attitudes. Political science research, particularly since the 2016 U.S. election, has increasingly considered the role of sexism in shaping both vote choice and public opinion with a notable uptick in the number of articles in political science journals focusing on sexism (Schaffner, 2021). However, the scales used across studies often differ with some scholars using the full ASI scale (Cassese and Holman, 2019), while other studies rely on only the hostile sexism items from the ASI (Schaffner et al., 2018; Valentino et al., 2018). Furthermore, much of this literature describes the role of sexism in shaping vote choice. Our contribution is three-fold: First, we expand the dependent variables to consider how sexism may correlate with public opinion. We look at not only abortion, but also birth control, support for #Me Too, equal pay, and paid leave. Second, we do not assume that there is one pattern of sexist attitudes that can shape social and political views and instead look at how different measures of sexism predict gendered political attitudes. We utilize the measures of sexism that are most commonly used in political science: hostile, benevolent, and modern as well as an implicit measure of sexism to capture sexist attitudes that may exist outside of one's conscious awareness. Finally, we improve on previous research by leveraging two large representative samples.

Expectations

All manifestations of sexism, in some way, contribute to the maintenance of the gender status quo. However, sexism takes many different forms, and there are reasons to expect that different sexist attitudes may have differential impacts on public opinion and policy attitudes, particularly those that are expressly gendered. Hostile sexism is perhaps the least subtle form of sexism as it involves open hostility and resentment toward women and gender equality. Modern sexism, which also involves antagonistic attitudes toward women and demands for equality, is closely related to hostile sexism and sometimes characterized as comparable scales (Valentino et al., 2018; Schaffner, 2021). Therefore, we expect:

H₁: Hostile and modern sexism will have a significant and negative effect on support for gendered policy attitudes.

A fundamental feature of hostile sexism is the desire to maintain men's power (Cross et al., 2019). Hostile sexists express antagonistic and aggressive views about women and gender equality, positing that women are constantly vying for the social advantages, resources, and privileges that men have (Glick et al., 2000). Research has shown that hostile sexism is associated with negative evaluations of women in managerial positions, feminists, and in other roles that are deemed “non-traditional” (Glick et al., 1997; Masser and Abrams, 2004). We expect hostile sexism to be negatively correlated with expanded reproductive rights, including increased access to birth control and abortion, because the ability for women to have more control over their reproductive lives directly contradicts hostile sexist beliefs about women's subordinate status in society and affirms the belief that women are out to compete against men and vie for social dominance. Similarly, we theorize that hostile sexism will be negatively related to policies like paid leave and equal pay because of the antagonistic views about women that undergird hostile sexism.

Modern sexism is correlated with hostile sexism, though notably the modern sexism items that tap antagonistic and resentful attitudes toward women more closely map onto hostile sexism than the items focused on the denial of gender discrimination (Schaffner, 2021). We also expect that modern sexism will be associated with lower levels of support for reproductive rights, paid leave, and equal pay. Although these issues differ, they all involve pushing back against gender discrimination in some facet and the assumption that women face unequal conditions. Modern sexists do not believe that any gender inequalities stem from systemic discrimination and therefore would be unlikely to believe women deserve “special treatment” in the form of expanded reproductive rights or government intervention into ensuring equal pay and paid leave.

On the other hand, while benevolent sexist attitudes still ultimately uphold the gender status quo and male dominance, it is possible that these attitudes create cross pressures and competing considerations as it relates to gendered policy attitudes. For example, Hideg et al. (2016) find that benevolent sexism was associated with more support for employment equity policies for women, but this support disappeared when the workplace domain was stereotypically masculine, and those with high levels of benevolent sexist attitudes are more likely to support gender quota policies to increase women's presence in politics than those with low levels of benevolent sexism (Beauregard and Sheppard, 2021). This support stemmed from the belief that women need the help and protection of gender quotas to achieve success in politics and not from a belief in gender equality. Overall, benevolent sexism is associated with support for gender-based affirmative action in the workplace, but this association is based in the belief that women need assistance to be successful (Sibley and Perry, 2010). With certain topics, the desire to “protect” women may clash with the desire to maintain male dominance and uphold traditional gender roles. Therefore, we expect:

H₂: Benevolent sexism will have a significant and negative effect on support for gendered policy attitudes.

Ultimately, we still expect benevolent sexism to be negatively related to support for abortion, birth control access, Planned Parenthood funding, and the #MeToo Movement, even though cross pressures may exist. Though benevolent sexist attitudes are putatively positive in tone, these attitudes still serve to restrict women to traditional roles like caregivers and homemakers. Furthermore, a major assumption of benevolent sexists is that women are purer than men and morally superior (Glick and Fiske, 1996). This ideology also idealized women as nurturing mothers which may lead to less support for reproductive rights, as past research has shown (Huang et al., 2014).

We also explore the connection between implicit sexism and various gendered policy attitudes. Measured at the unconscious level, this type of bias occurs outside of our awareness and reflects the automatic associations we have attached to objects/words. Even those who do not report hostile or benevolent sexist attitudes can still be implicitly sexist, and when we only measure explicit attitudes, we risk missing a dimension of people's gender attitudes. We test whether this more subtle and inadvertent form of sexism can impact gendered attitudes. There is little research connecting implicit sexism to policy attitudes, but we know that implicit gender attitudes can impact support for women candidates (Mo, 2015), and other work has used implicit measures to elucidate the connection between implicit bias and attitudes toward immigration (Malhotra et al., 2012; Kroll, 2013), as well as implicit racism and support for voter ID laws (Banks and Hicks, 2016). We expect that the connection between implicit sexism and attitudes that are expressly gendered will be conditional on the strength of an issue attitude. For example, abortion is one of the few policy areas in which people have consistent attitudes, and these attitudes across the American population have been stable over time (Wilcox and Norrander, 2002; Jelen and Wilcox, 2003). It is unlikely that implicit sexism would impact relatively stable abortion attitudes that are more likely to be informed by explicit manifestations of sexism. Furthermore, abortion attitudes are closely linked to partisan identification (Killian, 2008; Levendusky, 2009). However, implicit sexism may impact attitudes on issues where opinions might be less crystallized like access to birth control.

H₃: Implicit sexism will have a significant and negative effect on support for gendered policy attitudes with the exception of abortion.

In some ways, we would logically expect that women would be more likely to support gender-salient policies. Women tend to express fewer sexist attitudes than men (Cowie et al., 2019), and these policies are more likely to directly impact their lives. However, women also are capable of holding sexist and gender system-justifying attitudes. Although women tend to express fewer sexist attitudes than men, some women buy into hostile and benevolent stereotypes as a way to “…justify and maintain the status quo” (Jost and Kay, 2005, p. 498). As noted above, gender identity is less politically influential relative to identities like race and ethnicity. Furthermore, past research has found a lack of gender differences in support for public opinion issues that have a disproportionate impact on women (Sapiro, 2003; Lizotte, 2015). Given this mixed evidence, we remain agnostic about whether the effect of sexist attitudes on gender-salient public opinion issues will be conditional on gender.

Materials and Methods

Study 1

Sample

To explore the relationship between sexism and gender salient policy attitudes, we recruited 1,400 respondents via the survey platform Prolific to take part in a “Gender Identity and Political Attitudes” survey in the Winter of 2018. Unlike publicly available survey data, our survey included multiple measures of explicit sexism, a measure of implicit sexism, as well as various measures of reproductive rights attitudes. Participants were paid $1.50 to take part in the 10-minute long study. The benefit of using Prolific is that they have algorithms in place to fairly allocate study spaces, decreasing the issue of non-naïve participants (Chandler et al., 2014). The sample was 50.1% men and 48.3% women. The mean age was 35.6, 74.4% of the sample were white, 5.4% were Black, 3.1% were Latino/a or Hispanic, 6.9% were Asian, 0.5% were Native American, 9.7% identified as multiracial and the median income was “Between $50,000 and $64,999.” Full demographic information can be found in the Appendix.

Measures

Participants first consented to the study and then responded to demographic questions, a 20-item Big Five personality battery (Donnellan et al., 2006), a Social Dominance Orientation battery (Sidanius and Pratto, 1999), as well as the Bem Sex Role Inventory. The Bem Sex Role Inventory is a commonly used measure of gender expression and gender roles. All participants completed the Gender-Career Implicit Association Test (IAT). The purpose of the IAT is to measure implicit gender attitudes in a way that is not subject to social desirability bias. In the congruent task, participants had to match up common male names with words related to work and careers and match up common female names with words related to family and home life. In the incongruent task, participants had to match male names with words related to family and home life and female names with words related to work and careers. The resulting D-Score measure is computed based on the difference in performance speeds between the two classification tasks. To compute the D-Score, we used the improved IAT algorithm specified in Greenwald et al. (2003) and the IAT package in R. Participants completed five items from both the Hostile and Benevolent sexism scales in the Ambivalent Sexism Inventory (Glick and Fiske, 1996). Alpha values indicate a high reliability for both the hostile subscale (Cronbach's Alpha = 0.92) and the benevolent subscale (Cronbach's Alpha = 0.84). All items were coded such that higher values indicated a more sexist response. A mean composite score was generated for each subscale by averaged responses across the items that ranged from 1 (not sexist) to 5 (sexist). The hostile and benevolent scales were moderately correlated with each other (r= 0.47), while the hostile and implicit sexism (D-score) (r = 0.05) and benevolent and implicit sexism (D-score) (r = 0.05) were not correlated.

To measure abortion attitudes, participates were asked “Under the following conditions, do you think pregnant women should be allowed to obtain a legal abortion…” The nine conditions ranged from “If the pregnancy was caused by rape” to “If the pregnancy was caused by a casual encounter.” The full battery of conditions can be found in the Appendix. All items were coded such that higher values indicated more support for abortion access. A mean composite score was generated by averaging response across all nine items (Cronbach's Alpha = 0.94). To tap birth control attitudes, participants were asked how much they agreed or disagreed with the following statements: (1) The government should make it easier for women to obtain birth control, (2) Single women should not be able to obtain birth control (reverse-coded), and (3) Men should have no say in a woman's decision about birth control. Items were coded such that higher values indicated more support for birth control access and a mean composite score was generated by average responses across the three items (Cronbach's Alpha = 0.51). Finally, participants were asked on a five-point scale how much they agreed with the federal government cutting off funding for Planned Parenthood (explained in the survey as a non-profit organization that provided sexual health care) with higher values indicating more support for Planned Parenthood funding. They were also asked how much they approved of the #MeToo Movement with higher values indicating more support.

Controls

Several different control variables were measured to account for other factors that may influence the outcomes in which we are interested. We controlled for age, gender, income, education, religiosity, race, and ideology as we may expect younger people, women, those who have higher incomes and are more educated, as well as liberals to be more supportive of various gendered public opinion issues. We controlled for religiosity with an item measuring the frequency of religious service attendance (Lizotte, 2015). Race was coded as a dummy variable with white and non-white as the categories. All variables were recoded to range from 0 to 1 for ease of interpretation of the unstandardized coefficients.

Results (Study 1)

To look at the relationship between sexism and gendered public opinion and policy attitudes, a total of four OLS regression models were estimated with results found in Table 2 (Hlavac, 2022). For all models, we looked at the variance inflation factor (VIF) to detect multicollinearity given that the hostile and benevolent sexism measures are moderately correlated with each other. We found no evidence of significant multicollinearity in any of the models. The key independent variables were the three sexism measures. We looked at the impact of these variables as well as a set of controls on abortion and birth control attitudes, support for Planned Parenthood funding, and approval of the #MeToo Movement. Our first hypothesis was that hostile sexism would have a significant and negative effect on support for gendered public opinion and policy attitudes. Indeed, there was a significant, negative correlation between hostile sexism and less support for abortion access, birth control access, funding for Planned Parenthood, and the #MeToo Movement. Figure 1 shows the marginal effect of hostile sexism on the dependent variables with continuous control variables set to their means and factors set to their reference categories. This is consistent with research that finds that hostile sexism predicts a variety of anti-egalitarian outcomes (Sakall, 2001; Murphy et al., 2011; Patev et al., 2019).

TABLE 2

Table 2. Study 1 regression results.

FIGURE 1

Figure 1. Effects of hostile sexism.

Hypothesis 2 predicted that benevolent sexism would be negatively related to support for gendered public opinion and policy attitudes. We find partial support for this hypothesis. Benevolent sexism is negatively related to abortion access and support for Planned Parenthood (p < 0.1), and the effects sizes are more modest than the effects of hostile sexism on these outcomes. Benevolent sexism was not significantly related to birth control attitudes, and contrary to theoretical expectations, benevolent sexism was positively correlated with support for the #MeToo Movement.

Finally, our third hypothesis predicted that implicit sexism, as measured with the implicit association test, would be negatively associated with support for gendered policy attitudes except for abortion. We only find partial support for this hypothesis as well. As expected, implicit sexism did not significantly predict abortion support. Implicit sexism was significantly related to Planned Parenthood access, with those higher in implicit sexism being less likely to support funding of Planned Parenthood. There was no significant relationship between implicit sexism and birth control attitudes or approval of the #MeToo Movement. Figure 2 shows the marginal effect of benevolent sexism on support for abortion, Planned Parenthood, and #MeToo with continuous control variables set to their means and factors set to their reference categories. Figure 3 displays the marginal effect of implicit sexism on support for Planned Parenthood. With respect to the control variables, age was negatively related to abortion attitudes and #MeToo support, being a woman was a positive and significant predictor of birth control attitudes, support for Planned Parenthood, and approval of the #MeToo Movement, and church attendance was a negative predictor of support for abortion, birth control access, and Planned Parenthood. Interestingly, church attendance was positively related to approval of the #MeToo Movement. Income was positively related to abortion attitudes and approval of #MeToo, while education was negatively related to birth control and Planned Parenthood support. Unsurprisingly, ideology was a significant negative predictor of all four dependent variables, with conservatives less likely to support abortion and birth control access, funding for Planned Parenthood, and they were less likely to approve of the #MeToo Movement.

FIGURE 2

Figure 2. Effects of benevolent sexism.

FIGURE 3

Figure 3. Effects of implicit sexism.

We also wanted to test whether the relationship between various measures of sexism and gendered policy attitudes are conditional on gender. We ran the same regression models described above but included interactions between gender and the three sexism scales. Full regression results can be found in Table 3. Again, the key independent variables were the three sexism measures. Regression results show that hostile sexism was significantly and negatively related to all four dependent variables. The interaction between gender and hostile sexism was positive and statistically significant for the birth control and Planned Parenthood models, indicating that the negative effect of hostile sexism is weaker for women as compared to men. The interaction between gender and benevolent sexism was negative and statistically significant when looking at abortion, birth control, and #MeToo attitudes. Figure 4 depicts how gender moderates the relationship between benevolent sexism and these attitudes. We see that for men, levels of benevolent sexism have virtually no effect on support for abortion or birth control access. However, there is a significant and negative relationship between benevolent sexism and these attitudes for women. When it comes to #MeToo support, benevolent sexism is a significant and positive predictor for both men and women, although it appears to be a slightly weaker relationship for women.

TABLE 3

Table 3. Study 1 regression results with gender interaction.

FIGURE 4

Figure 4. Effects of benevolent sexism by gender.

Study 2

Sample

While Study 1 allowed us to simultaneously estimate the effects of different measures, including implicit sexism, on attitudes toward reproductive rights, its main limitation is that it does not use a probability sample. Consequently, in Study 2, we use the nationally representative 2016 American National Election Study (n = 4,270) to replicate (partially) the findings on abortion attitudes in Study 1, and to extend our analysis to look at attitudes about equal pay and paid family leave. Reproductive rights issues may be a particular type of policy that taps into forms of sexism, but general equality of the sexes is at the heart of most sexist attitudes and debates. Thus, we are interested in better understanding what forms of sexism predict gendered workplace-related policies like equal pay and paid family leave (McBride and Parry, 2016). The ANES relies on a probability sample of eligible voters in the United States. The sample was 47.1% men and 52.9% women, and 71.7% white, 9.4% Black, 3.5% Asian, Native Hawaiian, or Pacific Islander, 10.6% Hispanic, 0.6% Native American or Alaska Native, and 4.2% other or multi-racial. The median income was between $50,000 and $64,999. Full demographic information can be found in the Appendix.

Measures

The 2016 ANES uses several different measures to tap gender attitudes, including an abbreviated version of the hostile sexism subscale of the Ambivalent Sexism Inventory (Glick and Fiske, 1996). The benevolent sexism subscale was not included on the ANES. Respondents were asked how strongly they agreed or disagreed with the following statements: (1) Many women interpret innocent remarks or acts as being sexist; (2) Most women fail to appreciate fully all that men do for them; (3) Women seek to gain power by getting control over men; and (4) Once a woman gets a man to commit to her, she tries to put him on a tight leash. All items were coded such that higher values indicated a more sexist response (Cronbach's Alpha = 0.79). The last two items were also measured in Study 1. The Modern Sexism Scale (Swim et al., 1995) was included, which consists of three items: (1) How much attention should the media pay to discrimination against women?, (2) When women demand equality these days, how often are they actually seeking special favors?, and (3) When women complain about discrimination, how often do they cause more problems than they solve? All items were coded such that higher values indicated a more sexist response (Cronbach's Alpha = 0.65). A mean composite score was generated for each subscale (hostile and modern) by averaging responses across the items. The hostile and modern scales were moderately correlated with each other (r = 0.42).

Finally, the ANES included the following question on abortion attitudes: “There has been some discussion about abortion during recent years. Which one of the opinions on this page best agrees with your view?” Response categories were, (1) By law, abortion should never be permitted; (2) By law, only in case of rape, incest, or a woman's life in danger; (3) By law, for reasons other than rape, incest, or woman's life in danger if need established; (4) By law, abortion as a matter of personal choice. Response categories were coded such that higher scores indicated more liberal abortion attitudes. We also analyzed an item asking about equal pay, which was “Do you favor, oppose, or neither favor nor oppose requiring employers to pay women and men the same amount for the same work?” Our final dependent variable was an item regarding paid leave—“Do you favor/oppose, or neither favor nor oppose requiring employers to offer paid leave to parents of new children?” Response categories were coded such that higher values indicated more support.

Controls

We controlled for age, gender, race, religiosity, income, education, and ideology as we may expect younger people, women, those who have higher incomes and are more educated, as well as liberals to be more supportive of various gendered public opinion issues. All variables were recoded to range from 0 to 1 for ease of interpretation of the unstandardized coefficients.

Results (Study 2)

To look at the relationship between sexism and gendered public opinion and policy attitudes, we estimated three OLS regression models, with results displayed in Table 4. For all models, we looked at the variance inflation factor (VIF) to detect multicollinearity given that the sexism measures are moderately correlated with each other. We found no evidence of multicollinearity in any of the models. The key independent variables were the two sexism measures. We looked at the impact of these variables, as well as a set of controls on abortion attitudes, equal pay, and paid leave. Our first hypothesis was that both hostile and modern sexism would have a significant and negative effect on support for gendered public opinion and policy attitudes. We find mixed support for this hypothesis. Consistent with our results from Study 1 and Hypothesis 1, hostile sexism was associated with less support for abortion access as well as equal pay. However, it was not a significant predictor of support for paid leave policies. Figure 5 shows the marginal effect of hostile sexism on the dependent variables with continuous control variables set to their means and factors set to their reference categories. Again, consistent with our first hypothesis, modern sexism is related to reduced support for abortion, equal pay, and paid leave policies as displayed in Figure 6. With respect to the control variables, age had a positive and significant effect on support for abortion and equal pay, but a negative and significant effect on support for paid leave. Women were significantly more supportive of equal pay and paid leave policies and less supportive of abortion, church attendance was negatively related to abortion support, education was positively related to abortion support, and conservatism was associated with less support for all three dependent variables. Income was a positive and significant predictor of support for abortion.

TABLE 4

Table 4. Study 2 regression results.

FIGURE 5

Figure 5. Effects of hostile sexism.

FIGURE 6

Figure 6. Effects of modern sexism.

Again, we were interested in whether the relationship between sexism and gendered policy attitudes is conditional on gender. Using the same analysis strategy as we used in Study 1, we estimated three regression models, interacting gender with both the hostile and modern sexism scales. Full regression results can be found below in Table 5. Contrary to our results in Study 1 in which the negative effect of hostile sexism on abortion support was weaker for women as compared to men, we do not see a significant interaction between gender and abortion support in this data. There was also no significant interaction between gender and hostile sexism when it came to paid leave and equal pay. Gender does appear to moderate the relationship between modern sexism and support for equal pay. More specifically, modern sexism was a weaker predictor of equal pay support for women as compared to men.

TABLE 5

Table 5. Study 1 regression results with gender interaction.

Discussion

We know that the influence of sexism on candidate evaluations and vote choice has increased in the past decade (Cassese and Barnes, 2018; Valentino et al., 2018). Adding to this literature, we demonstrate that sexism may have an impact beyond the electoral context to inform a myriad of political attitudes, particularly those attitudes that have expressly gendered implications. Scholars have used various forms of sexism scales to predict political phenomena, creating a mixed pattern of findings that are difficult to compare and unpack. By simultaneously estimating the associations between multiple measures of sexism, including an implicit association test, and political attitudes, we contribute to efforts to understand the mechanisms at work underlying sexist attitudes and suggest that sexism indeed comes in many shades that have implications for particular policies. Our contribution underscores the need for scholars who are using sexism batteries to carefully consider the scales they choose and the accompanying underlying attitudes about women. In other words, are the items measuring a form of sexism motivated by antipathy toward women, opposition to gender equality, a motivation to maintain existing gender relations, or ideas about the moral superiority and purity of women?

Including the implicit sexism measure alongside explicit measures of sexist attitudes in our statistical models allows us to assess whether unconscious sexism is related to attitudes toward reproductive rights as well as address the concern of social desirability bias that could be present in people's responses to explicit sexism scales. Furthermore, by using the implicit test around gender roles like workplaces and the home, we contribute to our understanding of how separate spheres ideology impacts attitudes beyond workplace and domestic space equality (Miller and Borgida, 2016). Specifically, individuals who are more likely to associate women with the home and men with work likely hold a set of attitudes that suggest women should be mothers and, thus, be less supportive of women's reproductive freedom. It's also possible individuals may connect birth control to sexual behaviors like sex outside of marriage or multiple partners and oppose Planned Parenthood to further restrict women's sex lives (Friesen et al., 2017). Future research should consider measuring support for condom access or erectile dysfunction medication to further elicit the role of sexism in these domains related to men's sexuality.

Our findings suggest that often the relationship between sexism and support for gendered policy is fairly straightforward. In Study 1, we found that hostile sexism, regardless of gender, was related to less support for abortion access, birth control access, funding for Planned Parenthood, and support for the #MeToo Movement. This suggests that antipathy toward women and beliefs that women are in competition with men vying for social dominance, the hallmarks of hostile sexism, are associated with support for restricting reproductive rights. In our second study, we replicated the finding that hostile sexism is negatively related to support for abortion. However, hostile sexism was only marginally related to equal pay and was not associated with support for paid leave while modern sexism was a negative predictor of all three. Previous research has found that modern and hostile sexism are closely related but that the modern sexism items tapping antagonism and resentment more closely mapped onto hostile sexism than the items tapping denial of discrimination. Because the modern sexism items on the ANES were more focused on the denial of discrimination rather than antipathy toward women, this suggests that—at least with respect to equal pay and paid leave policies—the mechanism driving these attitudes is more about a denial that women face unequal conditions rather than overt hostility toward women.

We were also interested in the relationship between benevolent sexism, the other sub-scale in the Ambivalent Sexism Inventory, and gendered attitudes. We found that benevolent sexists were less likely to support abortion and Planned Parenthood. However, the significant interaction between gender and benevolent sexism suggests that it is a stronger predictor for women as compared to men. Benevolent sexist women were more likely to show decreased support for abortion, birth control access, and support for the #MeToo Movement. More research needs to be done to understand how gender moderates the relationship between benevolent sexism and policy attitudes as our findings suggest that women and men may have different motivations for adopting benevolent sexist attitudes. One possibility of many is that benevolent sexist women may oppose abortion and birth control because they conflict with heteronormative and gendered notions of motherhood. In contrast, benevolent sexist men may not oppose abortion and birth control, per se, especially since these could offer men ways to “protect” women in their lives (e.g., a father encouraging his teenage daughter to obtain an abortion). Because there is some evidence of racial differences in benevolent sexism (Davis et al., 2022), scholars should be challenged to acquire larger samples of racial/ethnic groups to more adequately model intersectional effects.

Finally, we found that benevolent sexism was related to increased support for the #MeToo Movement. Although this was not an expected finding, it squares with the protective nature of benevolent sexism and is consistent with research that has found a positive relationship between benevolent sexism and support for gender quotas and even compliance with public health measures (Chen and Farhart, 2020; Beauregard and Sheppard, 2021). This connection raises important questions for how men view this movement and the strategies that political leaders and activists should take in pushing for more progressive policies. Understanding how various forms of sexism relate to policy attitudes can be helpful for social movement organizations (SMOs) hoping to appeal to and mobilize men on their behalf. For example, SMOs targeted men in Ireland's Repeal the 8^th referendum, the amendment in their constitution that prohibited abortion for any reason. This framing in messaging on social media included themes like “She lives on your street” or “Grandfathers for Yes” (Hunt and Friesen, 2021). Nearly all of the messages in the anti-abortion and pro-choice tweets aimed at men took on a benevolent sexism theme. The 8^th was repealed, and men did turn out on behalf of a “women's” issue but appealing to protective tropes can undermine broader gender equality goals. In the example of appealing to men in anti-sex trafficking movements, messages like “real men don't buy girls” could do more harm than good (Steele and Shores, 2015). By treating women as weak humans who need protection from men, benevolent sexist framing undermines the ultimate goal of achieving gender equality.

With the U.S. Supreme Court poised to overturn Roe v. Wade and severely restrict abortion rights for millions of women, this research also has implications for how we understand both abortion attitudes and attitudes about the potential rollback of abortion rights. Although our findings are only correlational in nature, our results suggest that both antipathy toward women and opposition to measures to address gender inequality predict negative support for abortion. This is a useful insight into the motivations that drive anti-abortion attitudes in light of the fact that most pro-life activist groups emphasize the desire to protect the sanctity of fetal life. Our research indicates that all different forms of sexist intuitions, hostile, benevolent, and modern, drive anti-choice attitudes.

Conclusion

Overall, our results reveal a complicated relationship between different conceptualizations of sexism and gendered attitudes, underscoring the need to consider how different forms of sexism shape broader social and political views, from both a normative perspective for societal change and a measurement approach for research precision. Because of the observational nature of our data, there are many limitations to our findings. We can only speculate about the exact mechanisms that connect sexism to gendered policy attitudes. Furthermore, both of our samples were predominantly white. This limits the generalizability of our findings to the broader population, and future research should explore the interaction between gender and race as it relates to sexism and political attitudes. For example, gender, race, and religiosity interact to shape abortion support, demonstrating the importance of these intersectional dynamics (Holman et al., 2020). More work needs to be done to fully understand the complexities of gender, race, and sexism in shaping political attitudes.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics Statement

The studies involving human participants were reviewed and approved by Temple University Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.

Author Contributions

CG and KA collaborated in designing and conducting this study. CG has taken the lead role in developing the article and conducting the statistical analysis. KA and AF contributed to the theoretical framework and to the discussion. All authors contributed to the article and approved the submitted version.

Funding

The data collection utilized in Study 1 was funded by the Behavioral Foundations Lab at Temple University.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Acknowledgments

We would like to thank Erin Cassese for helpful comments on this article.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpos.2022.817309/full#supplementary-material

References

Anderson, K. J. (2015). Modern Misogyny. Oxford: Oxford University Press.

Arceneaux, K., and Kolodny, R. (2009). The effect of grassroots campaigning on issue preferences and issue salience. J. Elect. Public Opinion Part. 19, 235–249. doi: 10.1080/17457280903072916