Ambivalent sexism among Christian and Muslim youth. The gendered pathway of perceived pressure for religious conformity Open Access

Contemporary European societies have removed formal barriers to grant men and women equal access to education, the labor market and politics. Despite enduring policy efforts (sometimes including gender quota), persistent gender inequalities in terms of employment rate, labor market position, payment, participation in decision-making positions, workshare in the household and childcare remain across European countries (The World Bank 2012; European Union 2017). Traditional gender norms embedded in attitudes are thought to play a substantial role in the reproduction of gender inequality because they carry (role) expectations that guide everyday practices (Ridgeway and Correll 2004). Research in this area has shown a complex interplay between more blatant and benevolent attitudes. Indeed, Glick and Fiske's (1996) Ambivalent Sexism Theory shows that in societies that strive for equality and tolerance, socially less accepted negative forms of sexism are supplemented by a seemingly positive and subtle form of sexism. Importantly, this subtle form of sexism is often not perceived as sexist and is therefore endorsed by both men and women (Connelly and Heesacker 2012; Cross and Overall 2018; Hammond et al.2014). While hostile sexism is directed towards wo(men) who challenge traditional gender roles, benevolent sexism is directed towards wo(men) who comply to traditional gender roles. That way, benevolent and hostile sexism are complementary justifications that reinforce and justify practices that reproduce gender inequality (Jackman 1994; Glick and Fiske 1996, 2001).

Some interpretations of Abrahamic religions work in similar ways by rewarding compliance to traditional gender roles and punishing people who deviate from them, which reproduces inequality (Jost et al.2014; Sumerau et al.2016). Consequently, a growing body of research across different countries and religious groups studied whether the importance individuals attach to their religious beliefs in everyday life predicts the endorsement of benevolent sexism and hostile sexism (Burn and Busso 2005; Gaunt 2012; Glick et al.2002, 2016; Hannover et al.2018; Maltby et al.2010; Mikołajczak and Pietrzak 2014; Stevenson 2014; Taşdemir and Sakallı-Uğurlu 2010), but also how traditional gender beliefs such as sexism mediate the association between religiosity and negative attitudes towards homosexuals (Roggemans et al.2015; Piumatti 2017). Therefore, understanding religious people's endorsement of ambivalent sexism will help to advance our understanding of prejudice towards anyone who deviates from traditional conceptions of how men and women should behave (including homosexuals). This paper aims to contribute to the emerging literature on the relationship between religion and ambivalent sexism in three ways.

First, although adolescence is an important period of attitude formation (Moshman 2005), previous studies on the relationship between ambivalent sexism and religion primarily focused on university students or adults. Because research among adolescents remains scarce, we focus on adolescents. Moreover, from adolescence onwards, hostile sexism starts to be complemented by benevolent sexism due to an increased mutual dependence regarding cross-gender friendships (Poulin and Pedersen 2007) and romantic relationships (Rudman and Glick 2008).

Second, research in this area has only devoted scant attention to the increasing ethnic and cultural diversity in European societies. The endorsement of ambivalent sexism among religious minority groups has rarely been studied (but see, Hannover et al.2018). Therefore, in this study, we compare a religious minority group (Muslims) to a religious ‘majority’ group (Christians) in Belgium. We also pay justice to the fact that religious groups are often not ethnically homogenous by studying young people from Turkish and Moroccan descent (the two largest groups affiliating to Islam in Belgium).

Finally, most studies solely focused on religious experience on the individual level (e.g. religious affiliation, subjective importance of one's religious beliefs) to explain the relationship between religion and ambivalent sexism. Yet identifying as religious, also implies being part of a larger religious group in which religious members can exert pressure to conform to religious norms. Indeed, religious groups are first and foremost social groups with their own (gendered) group norms that guide beliefs, attitudes, values and eventually behavior. Through communication processes within religious groups, social control may be exerted by other members to conform to religious norms. Especially during adolescence group-level dynamics become apparent (King et al.2002). Until now, research has neglected such (gendered) collective religious pathways through which sexist attitudes are guided.

To sum up, we investigate how the endorsement of ambivalent sexism is related to adolescents’ ethnic background (native, Moroccan or Turkish descent), religious affiliation (Christian or Muslim), religiosity (subjective importance of one's belief in everyday life), and perceived pressure for religious conformity (the expectation from significant others to follow religious rules) by performing multivariate multilevel regression analyses. We rely on data from two independent samples gathered in 2013 and 2018 by the Youth Research Platform among Dutch-speaking religious boys (N = 1637) and girls (N = 2058) between 14 and 18 years old.

Sexist attitudes are based on stereotypical descriptions of how men and women generally behave (descriptive stereotypes) and should behave (prescriptive stereotypes) (Burgess and Borgida 1999). These gender stereotypes are built on an essentialist view in which men are described and expected to be agentic (i.e. dominant, instrumental, assertive and rational) and women as communal (i.e. caring, nurturing, social and emotional) (Prentice and Carranza 2002). To understand why gender stereotypes contribute to gender inequalities, we turn to the Ambivalent Sexism theory (Glick and Fiske 1996). This theory points to two important mechanisms that render gender inequalities persistent in contemporary societies (Ridgeway 2011).

First, gender inequality is thought to persist because gender stereotypes are constantly shaped and reinforced through daily interactions, partly because people regulate each other's behavior (West and Zimmerman 1987). That is, people who deviate from gender norms are often socially punished in subtle or explicit ways (Rudman and Fairchild 2004). Conversely, people who conform to gender norms tend to be socially rewarded (Mayeux and Kleiser 2020). This system of numerous social rewards and punishments is also described in Glick and Fiske's (1996) Ambivalent Sexism Theory in which two – seemingly contradictory – forms of sexism towards women are discussed, namely hostile and benevolent sexism. Hostile sexism refers to an envious form of prejudice that negatively stereotypes women who do not comply to traditional gender roles (e.g. career women, childless women and feminists) as controlling and manipulative (Becker 2010; Sibley and Wilson 2004). Benevolent sexism, on the contrary, is a form of prejudice that positively stereotypes women who comply to traditional gender roles as caring, nurturing and even morally superior to men. However, this form of sexism indirectly describes and prescribes women's dependence to men. More specifically, benevolent sexism is grounded on three beliefs: men and women hold complementary gender traits (gender differentiation), men and women need each other for intimacy (heterosexual intimacy) and because women are viewed as beautiful yet vulnerable they need protection by men (protective paternalism).

Second, precisely due to benevolent sexism's chivalrous and positive undertone, it is less likely to be perceived as negative by women and men (Connelly and Heesacker 2012; Cross and Overall 2018; Hammond et al.2014). Importantly, benevolent sexism plays a significant role in justifying and consolidating gender inequality, because it provides an important incentive for women to not only conform to traditional gender roles but also to actively endorse benevolent sexism (Glick and Fiske 2001). Previous studies have consistently shown that women endorse hostile sexism to a lesser extent than men (de Lemus et al.2010; Ferragut et al.2017; Glick and Fiske 2011; Mastari et al.2019). Regarding the endorsement of benevolent sexism, the observed patterns are more mixed. Some studies found no gender differences (Gaunt 2012; Glick et al.2000, 2004), others found higher endorsement among men/boys (Glick and Fiske 2001; Ferragut et al.2013), while in some studies women/girls endorsed benevolent sexism to a higher extent than men/boys (Sibley and Wilson 2004; Vandenbossche et al.2017). Importantly, women's endorsement of benevolent sexism reduces their resistance towards an unequal gender system (Connelly and Heesacker 2012; Cross and Overall 2018; Hammond et al.2014). Moreover, by agreeing to benevolent sexism women legitimize men's dominant position and justify a system of rewards and punishments which helps to maintain gender inequality (Jost and Kay 2005; Sibley et al.2007).

Jost and colleagues (2014) hold that religion is an important belief system that helps to accept and justify sexist beliefs through moral and religious values. Similar to the workings of benevolent sexism, Abrahamic religions (such as Christianism, Islam and Judaism) encourage traditional gender roles as ‘God-given’ which (religiously) justifies them (Glick et al.2002; Whitley 2009; Fitzpatrick Bettencourt et al.2011; Haggard et al.2019; Howard et al.2020). Religious views on gender roles bear the potential that religious people express hostility towards others who don't conform to group norms (i.e. traditional gender norms, heteronormative norms which proscribes anti-gay sentiment, etc.) (Duck and Hunsberger 1999), as is the case for hostile sexism. In discussing the relationship between ambivalent sexism and various religious aspects, we distinguish between dynamics on the individual and group level.

Research on the endorsement of ambivalent sexism among religious people mostly focused on whether the subjective importance of one's belief in everyday life (further: religiosity) predicts ambivalent sexism. We first discuss previous research regarding the endorsement of benevolent sexism, followed by hostile sexism.

First, higher levels of religiosity predicted benevolent sexism towards men and women (but not hostile sexism) among adult Christian men and women in Spain (Glick et al.2002). In another study, intrinsic religiosity, extrinsic religiosity and scriptural literalism were all related to benevolent sexism (but not hostile sexism) among (young) adult Christian men and women in the United States (Burn and Busso 2005). In a study among Christian Orthodox young adults in the United States, no relationship was found between ambivalent sexism towards women and religiosity among women. For men, higher religiosity was only associated with one subfactor of benevolent sexism (towards women), i.e. protective paternalism (Maltby et al.2010). In another study conducted in Poland, higher levels of religiosity were only related to the endorsement of benevolent sexism (towards women) among adult Christian women, whereas no relationship was observed for men (Mikołajczak and Pietrzak 2014). Next, among Jewish men and women in Israel, higher levels of religiosity predicted benevolent sexism towards women and men (Gaunt 2012). Interestingly, religiosity was negatively related to men's hostile sexism towards both women and men, while no relationship was found between women's hostile sexism (towards men and women) and religiosity. Finally, among religious people in a Belgian student sample, religiosity was both related to benevolent and hostile sexism towards women (Van Assche et al.2019).

Second, in some studies, religiosity was related to a stronger endorsement of hostile sexism. In a study among Mormon adults in the United States, for example, higher levels of religiosity and fundamentalism were related to the endorsement of hostile and benevolent sexism towards women (Stevenson 2014). Similarly, religiosity was also related to the endorsement of hostile and benevolent sexism towards women among Muslim in Turkey (Taşdemir and Sakallı-Uğurlu 2010). In another study in Turkey among Muslims, Glick et al. (2016) found that hostile and benevolent sexism towards women correlated with higher levels of religiosity. In turn, men's religiosity and the endorsement of hostile sexism predicted their endorsement of honor beliefs, which are beliefs about protecting the family honor by respecting traditional gender roles. For women, such honor beliefs were also predicted by religiosity and, contrary to men, to the endorsement of benevolent sexism. Finally, a study in Germany by Hannover et al. (2018) among young Muslim adults revealed that religiosity predicted higher levels of benevolent and hostile sexism towards women, although this was mediated by fundamentalism. The latter partly explained why Muslims supported ambivalent sexism to a higher extent than Christians.

Gender differences have been observed in virtually all studies and some authors explicitly indicate that gender moderates the relationship between religion and ambivalent sexism (Maltby et al.2010; Mikołajczak and Pietrzak 2014; Stevenson 2014). Therefore, to understand through which gendered ways religion and ambivalent sexism towards women are related to each other, we study this for boys and girls separately.

The above-mentioned studies have found mixed results across different religious affiliations. Belgium is an excellent case study to investigate differences in ambivalent sexism between two religious affiliations: Christianity and Islam. Although religious involvement among the native Belgian population thoroughly weakened from the 1960s onwards, today a considerable group still identifies as Christian or Catholic and continues to participate in religious traditions such as baptizing, communions and marrying before the Church (Conway and Spruyt 2018; Dobbelaere and Billiet 2010; Van Droogenbroeck et al.2016). Around the same time of the religious turning point in the 1960s, Belgium started recruiting foreign guest workers with different religious backgrounds. The largest influx came from Morocco and Turkey, in which Islam is the most practiced religion (Crul and Vermeulen 2003). In this study, we are therefore interested whether affiliating to Islam or Christianity relates differently to ambivalent sexism. In line with previous research, we also expect that higher levels of religiosity predict ambivalent sexism among religious adolescents.

While previous research has most often focused on individual aspects to understand the relationship between ambivalent sexism among religious people, we additionally concentrate on group level processes which potentially shape gender attitudes. More specifically, we investigate whether perceived pressure for religious conformity and belonging to an ethnic group relates to higher levels of ambivalent sexism.

To understand how identifying with a certain group can affect attitudes and behavior, we rely on self-categorization theory (Turner et al.1987). As a psychological process, categorizing oneself to a specific group generates a sense of group belonging and internal cohesion, which can go hand in hand with normative pressures from other group members regarding attitudes and behavior (Tajfel and Turner 1979). In-group norms are established and reinforced by direct and indirect verbal and non-verbal communication of what other people say or do (Hogg and Reid 2006). In terms of religion, self-categorizing as being an adherent of a religion, also means perceiving oneself as belonging to a religious group. Within religious groups, members can be prone to pressure for religious conformity by other members (Hogg et al.2010). Especially during adolescence, a period during which young people explore their (social) identity, parents and peers hold a potential power to influence their behavior (King et al.2002). In a study among adolescents in Flanders (Belgium), Van Droogenbroeck and Spruyt (2019) found that perceived pressure for religious conformity predicted higher levels of anti-gay sentiment among Muslims and Christians. Similarly, we argue that adolescents who experience pressure to conform to religious rules by parents and peers (both Islamic and Christian), support benevolent and hostile sexism to a higher extent. When parents and/or friends exert pressure to conform to religious rules, young people can show their religious dedication by endorsing benevolent sexism in which traditional/religious gender roles are respected and/or by endorsing hostile sexism towards people who don't respect traditional/religious gender roles. By behaving accordingly, group members secure their place within the group (Klein et al.2007). Moreover, texts in the Quran and Bible describe how women should behave, while men's role is to supervise whether women are obedient to God's or Allah's will (see Anwar 2006; Burn and Busso 2005). Because religious texts are more centered around girls’ or women's behavior, we expect that perceived pressure to conform to religious rules predicts girls’ ambivalent sexism more strongly than boys’.

Next, we hypothesize that differences are likely to occur in the endorsement of ambivalent sexism according to adolescents’ ethnic background. Previous research indicates that people of Turkish descent hold more traditional beliefs than people of Moroccan descent (Lesthaeghe et al.2000). This is partly due to more effective cultural transmissions among the Turkish group and the higher social control to preserve cultural traditions (Güngör et al.2011). That is, Turkish diaspora hold a strong and unified Turkish identity (Manço and Manço 1994) in contrast to the Moroccan diaspora – a heterogeneous group of Arabs and Berbers (Maddy-Weitzman 2001). Consequently, we expect that young religious people of Turkish descent support ambivalent sexism to a higher extent than those of Moroccan descent. Moreover, because cultural values and traditions are especially transmitted through the female family line (Phinney 1990), we expect to find a stronger relationship between ambivalent sexism and ethnic background for girls compared to boys.

We rely on data from the Youth School Monitor gathered in 2013 and 2018 by the Flemish Youth Research Platform (http://www.jeugdonderzoeksplatform.be/en) among two independent representative samples of Flemish secondary schools throughout Belgium (N₂₀₁₃= 98; N₂₀₁₈= 100). The data can be accessed via http://osf.io/376bh/. The survey consisted of various questions on topics such as general well-being, leisure time, socio-political attitudes, well-being at school, perpetration and victimhood. The cross-sectional data from 2013 (N = 4688) and 2018 (N = 5858) were pooled to achieve a sufficiently high number (and associated statistical power) of religious young people within both datasets. Pupils were between 14 and 18 years old (M_age= 17, SD = 1.60). For our analyses, we are only interested in religious people affiliating to Christianity or Islam (see further). Of the total sample of 10546 respondents, 33.6% (n = 3541) identified as being Christian, while 25% (n = 2643) identified as Muslim. We arrived at a final sample of 3695 religious adolescents (n_boys= 1637; n_girls= 2058) in our analyses because the monitor of the Youth Research Platform uses rotating modules. Although religious affiliation was collected among all respondents some items such as perceived pressure for religious conformity were not presented to all respondents. However, the rotating modules are completely randomized, thus selection bias is non-existent. Since respondents were nested in schools, we analyzed our data by means of multilevel regression analysis and estimated a series of nested models.

To capture benevolent and hostile sexism, respondents rated 8 items on a 5-point Likert scale (from 1 ‘Strongly disagree’ to 5 ‘Strongly agree’). Because the survey covered various subjects (see above), we worked with a shortened scale of the Ambivalent Sexism Inventory (Glick and Fiske 1996) which has been used in other studies (Mastari et al.2019; Vandenbossche et al.2017) to reduce respondents’ time and mental load. Hostile sexism was measured by four items (e.g. ‘Women seek special favours under guise of equality’; Cronbach's Alpha = .71). Benevolent sexism was also measured by four items and all three facets of benevolent sexism were covered (e.g. ‘Women have a quality of purity few men possess’; Cronbach's Alpha = .65). For both benevolent and hostile sexism (r = .31; p < .01), we constructed a summation scale ranging from 0 (low sexism) to 10 (high sexism).

Ethnic background was determined by a combination of the country of birth of the grandmother on mother's side, mother, father and the respondent (n_native= 2359; n_Moroccan= 866; n_Turkish= 470). Religious adolescents of native descent functioned as the reference group.

Respondents were asked how they would describe themselves religiously or spiritually. Options were: Catholic, Protestant, Christian (but not Catholic or Protestant), Muslim, Jewish, secular or non-religious. Christians, Catholics and Protestants were taken together as the reference group of Christians (n = 2164) next to Muslims (n = 1531). Religious affiliation was entered as a dummy variable in the analyses.

Religiosity was captured with one item ‘How important is religion to you?’ rated on an 11-point Likert scale ranging from 0 ‘Not important at all’ to 10 ‘Very important’.

Perceived pressure for religious conformity measured the perceived expectations from parents and friends to conform to religious rules. The concept was measured by two items ‘My parents expect me to live by the rules of my religion’, ‘My friends call me out if I do something against my religion’. Respondents rated items on a Likert scale ranging from 1 ‘Strongly disagree’ to 5 ‘Strongly agree’. Items for perceived pressure for religious conformity were summated to a scale ranging from 1 to 5 (Cronbach's Alpha = .79).

Year of data collection was entered as a control variable (n₂₀₁₃= 1935; n₂₀₁₈= 1760).

Because previous research showed that adolescents in Flanders enrolled in the technical and vocational educational track are more likely to endorse ambivalent sexism compared to adolescents in the general and arts track (see Vandenbossche et al.2017) and we know that people with a Turkish or Moroccan background are overrepresented in technical and vocational education, we also controlled for the educational track pupils were enrolled in. A dummy variable was constructed in which the technical and vocational track were collapsed (n_{technical and vocational track}= 2065) and the general and arts track acted as the reference category (n_{general and arts track}= 1630).

Table 1 presents the descriptive statistics for boys and girls separately. Adolescents enrolled in the vocational or technical track were slightly overrepresented in our data among boys (58.40%) and girls (53.89%). Boys were significantly more represented in the technical or vocational track compared to girls. 22.79% of boys and 23.96% of girls were of Moroccan descent, while 12.89% of boys and 12.59% of girls were of Turkish descent. 42.03% of boys and 40.96% of girls affiliated to Islam, the majority thus affiliated to Christianity. Girls rated their religiosity significantly higher than boys, whereas boys reported experiencing more pressure for religious conformity. It should be noted that the gender differences in religiosity and pressure for religious conformity were rather small. Regarding ambivalent sexism, girls endorsed benevolent sexism significantly stronger than boys, whereas boys endorsed hostile sexism significantly stronger than girls.

In the following we discuss results for the multivariate multilevel regression analyses to predict benevolent sexism (Table 2) and hostile sexism (Table 3) for boys (models 1a to 4a) and girls separately (models 1b to 4b). The final model was built in four steps. This paper aims to control for differences in the endorsement of ambivalent sexism within religious groups by controlling for educational track and ethnic background. Therefore, adolescents’ non-religious socio-demographic variables – educational track and ethnic background – were entered in the first step. Next, we added the religious variables one by one: religious affiliation in the second model, self-rated religiosity in the third model and in the fourth and final model perceived pressure for religious conformity was added.

Table 2 presents the results of the multilevel analysis on benevolent sexism (BS) among boys (model 1a to 4a) and girls (model 1b to 4b) separately. The first model indicated that boys and girls enrolled in technical or vocational education were more likely to endorse BS compared to boys and girls enrolled in general or arts education (boys: b_education= .33**; girls: b_education= .36***). No difference was found in model 1a between boys of native and Moroccan descent (b_Moroccan= −.05). Boys of Turkish descent were more likely to support BS compared to boys of native descent (b_Turkish= .24*). Among girls, model 1b showed that BS was significantly related to ethnic background. More specifically girls of Moroccan descent (b_Moroccan= .33***), but especially girls of Turkish descent (b_Turkish= .65***) reported higher levels of BS compared to girls of native descent. In the second model, religious affiliation was added. The initial differences in benevolent sexism between boys of Turkish and native descent on the one hand, (model 2a) and girls of Moroccan and native descent on the other hand were no longer observed (model 2b). The difference in BS between girls of native and girls of Turkish descent remained significant (b_Turkish= .51***). For both boys and girls, no differences in BS were found between affiliating to Islam or Christianity. Next, model 3 revealed that higher levels of religiosity were weakly related to BS for boys (b_religiosity= .04**), but not for girls (b_religiosity= .02). Finally, perceived pressure for religious conformity was added in model 4. Perceived pressure for religious conformity was not related to BS for boys (b_{religious pressure}= .06). Girls’ support for BS, however, was predicted by perceived pressure for religious conformity (b_{religious pressure}= .21***). By entering perceived pressure for religious conformity in model 4b the difference in BS between girls of Turkish and native descent further decreased (b_Turkish= .39**). For boys, entering perceived pressure for religious conformity resulted in a significant difference between boys of native and Moroccan descent. Boys of Moroccan descent were less likely to endorse BS compared to boys of native descent (b_Moroccan = −.29*). Throughout all the models, the relation between BS and educational track for boys and girls remained significant.

Table 3 presents the results of the multilevel regression analyses for hostile sexism (HS) for boys (model 1a to 4a) and girls (model 1b to 4b) separately. The first model indicated that boys enrolled in technical or vocational track were more likely to endorse HS compared to boys enrolled in general or arts track (b_education= .18*). No differences in HS were found in the first model between boys of native and Moroccan descent (b_Moroccan= −.08), although boys of Turkish descent reported slightly higher HS than boys of native descent (b_Turkish= .24*). For girls, the first model showed a stronger endorsement of HS among girls enrolled in the technical or vocational track compared to girls enrolled in the general or arts track (b_education= .52***). In the first model, differences were observed in the support for HS according to girls’ ethnic background. More specifically, girls of Moroccan (b_Moroccan= .40***) and especially girls of Turkish descent (b_Turkish= .72***) were more likely to support HS compared to girls of native descent.

Ethnically and religiously diverse European societies today find themselves in a peculiar position with regard to the pursuit of gender equality. Although growing political efforts have been invested to grant men and women equal access to education, the labor market and politics, gender inequalities continue to persist in both public and private spheres (The World Bank 2012; European Union 2017). Traditional gender beliefs, such as ambivalent sexism, are thought to play a key role in the perpetuation of gender inequalities. Previous studies showed that Abrahamic religions such as Christianity and Islam are associated with the endorsement of traditional gender beliefs as present in ambivalent sexism (Burn and Busso 2005; Glick et al.2002, 2016; Hannover et al.2018; Maltby et al.2010; Mikołajczak and Pietrzak 2014; Stevenson 2014; Taşdemir and Sakallı-Uğurlu 2010). Belgium, like most European countries has a growing group of youth with diverse ethnic and religious backgrounds in the cities. Studying the relationship between ambivalent sexism and religion among youth within ethnically and religiously diverse European societies is important to understand the potential tension between striving for gender equality and the presence of religion. Our findings lead to three general conclusions.

So far, studies on religion and ambivalent sexism have mostly focused on one religion and individual aspects of religious experience (e.g. Gaunt 2012; Glick et al.2002; Mikołajczak and Pietrzak 2014). This article showed that initial differences between Muslim and Christian youth in the endorsement of hostile sexism disappeared after adding both individual and group level aspects of adolescents’ religious experience: religiosity and perceived pressure for religious conformity to the analyses. The support for ambivalent sexism among Christian and Muslim youth did not seem to be a consequence of substantive differences between both faiths, but mainly follows from how important religion is to adolescents (individual level) and how much pressure they experience to conform to religious norms (group level). Thus, both Islam and Christianity seem to be belief systems that foster the endorsement of sexist beliefs through gendered religious values.

In line with previous research we found that boys’ religiosity predicted benevolent sexism, but not hostile sexism (Burn and Busso 2005; Gaunt 2012; Glick et al.2002; Maltby et al.2010; Van Assche et al.2019). Surprisingly, among religious boys, the Moroccan or Turkish culture did not predict the support for ambivalent sexism. Among religious girls, being of Turkish descent predicted a stronger endorsement of benevolent sexism. From a cultural point of view, it is possible that girls of Turkish descent aim to preserve traditional gender conceptions for the sake of cultural traditions which are often transmitted to girls (Güngör 2007).

Girls’ endorsement of ambivalent sexism was also strongly predicted by perceived pressure for religious conformity, rather than their own religiosity. This raises the question whether religious girls ‘do gender’ in function of their religion and, importantly, their religious environment, rather than their own personal religious experience. Although girls do not lack total agency in their religious experience, our findings suggest that religious girls’ gendered attitudes stem from normative expectations expressed by their religious environment. Some authors indeed argue that religious people ‘do gender’ in function of religious values (Avishai 2008). Among boys, we found no relationship between perceived pressure for religious conformity and ambivalent sexism. Thus, although boys also experience pressure for religious conformity, it does not relate to their attitudes towards women. To understand whether the lack of an association between pressure for religious conformity and ambivalent sexist attitudes towards women among boys was the result of studying attitudes directed towards women (and not men), we ran the same analysis with ambivalent sexism towards men (table A2 in Appendix, see Supplementary). Interestingly, these analyses showed that perceived pressure for religious conformity predicted a higher endorsement of ambivalent sexism towards men. Thus, pressure for religious conformity directed towards oneself and by extent to the own gender group, translated into the internalization of beliefs about traditional gender roles. This reveals the gendered dynamics within religious groups. When significant others grant importance to the compliance of religious rules which are inherently gendered, it seems religious adolescents construct gendered world views in which their own gender group should comply to traditional (religious) roles. This (1) urges for an approach in which religious identities and gender identities are simultaneously studied and (2) highlights the importance of group-level variables such as the perceived pressure for religious conformity.

Although this study contributes to an emerging literature by considering perceived pressure for religious conformity and ethnic background it suffers from a few limitations. Further research should use data that allow to determine causality and include more indicators that tap into individuals’ religious experience. More specifically, the level of fundamentalism as opposed to a religious quest orientation seems interesting. Previous research already showed that ambivalent sexism is related to fundamentalism (Gaunt 2012; Hannover et al.2018; Stevenson 2014). It would be interesting to study whether perceived pressure for religious conformity moderates the relationship between fundamentalism and ambivalent sexism. In addition, rather than including ethnic background as a demographic variable, future research should consider ethnic group dynamics more thoroughly by considering one's ethnic identity (or ethnic group belongingness) (Phinney 1990) and pressures for ethnic conformity which can lead to affirming cultural traditions more strongly (Van Kerckem et al.2014). That way, individual and group-level aspects of one's religious and ethnic background are considered simultaneously.

The limitations signal interesting opportunities for further research. They, however, do not disqualify this paper's general message: in contemporary European societies where ethnic and religious diversity is on the rise and persistent gender inequalities remain, it is crucial to study the dynamics of belonging to a religious group and study different ethnic-cultural groups simultaneously. Starting from these principles, this study showed how individual-level and group-level religious predictors of ambivalent sexism are differently related for religious boys and girls.

No potential conflict of interest was reported by the author(s).

Gil Keppens ([email protected]) is a postdoc researcher at Research Group TOR at the Vrije Universiteit Brussel (Belgium). He obtained his Master's degree (2012) and PhD (2018) at the Department of Sociology of Vrije Universiteit Brussel where he gained experience in the design, data-collection, analysis, and valorization of qualitative and large-scale quantitative research. Gil is actively engaged in truancy research and founding member of the International Network for School Attendance. His main research interests include youth research, sociology of education and educational policy.

Bram Spruyt ([email protected]) is associate Professor of Sociology at the Vrije Universiteit Brussel (Belgium) and one of the supervisors of the Flemish Youth Research Platform. He teaches Sociology, Cultural sociology, Researching culture. His main research interests include public opinion research, youth research, sociology of education and cultural sociology. Bram Spruyt is member of the Young Academy of Belgium.

Filip Van Droogenbroeck ([email protected]) is assistant Professor in data analytics in the Business Technology & Operations department at the Faculty of Social Sciences and Solvay Business School of the Vrije Universiteit Brussel (VUB, Brussels, Belgium). Formerly he was research professor in Sociology (research group TOR) at the VUB. His main research interests include sociology of youth, sociology of education, multi-level research methods in organizational research, and the interaction between artificial intelligence and the social sciences.

Laora Mastari ([email protected]) is researcher at the Vrije Universiteit Brussel (Belgium). Contact details: Vrije Universiteit Brussel, Sociology Department, Pleinlaan 5, 1050 Brussels (Belgium) +326148151. She obtained her master's degree in Sociology at the Vrije Universiteit Brussel in 2016. From November 2016 to September 2021 she worked as a researcher for the Youth Research Platform while simultaneously writing her doctoral dissertation on gender conformity among young people in Flanders.