Since Bourdieu’s (1984) seminal work, research has focused on inequalities in highbrow cultural tastes, that is, involvement in the arts—art, music, and theatre—and literature. Bourdieu (1984, 1986) argues that highbrow cultural tastes function as a form of cultural capital that is transposable into economic and social advantages (see for instance Lizardo 2006a). Indeed, leisure-time cultural activities in childhood and adolescence yield important benefits, such as the acquisition of specific cognitive and non-cognitive skills that lead to higher grades in school and access to elite colleges, as well as generate better career opportunities (Covay and Carbonaro 2010; DiMaggio 1982; Dumais 2002; Kaufman and Gabler 2004; Lareau 2003).

Although researchers heavily focus on the relationship between tastes and social class inequalities, gender differences in cultural tastes and the mechanisms behind these differences remain relatively understudied, in particular among adolescents (DiMaggio 2004; Schmutz et al. 2016; Siongers and Lievens 2014). In fact, the emphasis on highbrow tastes as a manifestation of cultural capital makes the consistent empirical finding that women are more involved in highbrow culture than men are puzzling (Bihagen and Katz-Gerro 2000; Christin 2012; Lizardo 2006b; Purhonen et al. 2011). Even though women are more likely to have highbrow tastes and tend to do better in school and thus appear to be advantaged in terms of cultural capital, they are not always able to translate this into an advantaged socioeconomic position (Bihagen and Katz-Gerro 2000; Christin 2012). Nevertheless, current explanations of the gender gap in cultural tastes often endorse a cultural capital perspective, overlooking that highbrow cultural tastes not only function as a marker of cultural capital and underlying social differences, but also reflect unequal expectations for (young) men/boys and (young) women/girls. These expectations mirror society-wide cultural norms that define highbrow culture as belonging to the feminine sphere, even among young people (Bihagen and Katz-Gerro 2000; Christin 2012; Nosek et al. 2002; Nosek and Smyth 2011; Tepper 2000).

In the present paper, we propose an alternative perspective on the puzzle of women’s highbrow cultural taste by addressing two lacunae in existing research. A first hiatus is that empirical research on gender differences in cultural preferences mainly studies adults, not children or adolescents, and often uses explanations that apply to adults’ life worlds, for instance class- and work-related explanations. Thus, even though childhood and adolescence are crucial periods in the development of cultural preferences (Bourdieu 1984), empirical research tends to overlook the finding that there is already a substantial gender gap in cultural tastes before children and adolescents reach adulthood, irrespective of social class (Dumais 2002; Kaufman and Gabler 2004; Schmutz et al. 2016; Siongers and Lievens 2014; Willekens and Lievens 2014). This calls for research that tries to clarify how young people’s interests in arts and literature become gendered, keeping social inequalities constant.

A second shortcoming is that many existing studies apply a gender-role socialization approach to gender differences and invoke a dichotomous and unidimensional vision on gender. Such an approach neglects current perspectives in gender studies that treat gender as a multidimensional and fluid concept that pertains to all aspects of social life (Egan and Perry 2001; Risman 2004; Thorne 1997). So, there should be more attention to the dynamic identity-related and interactional mechanisms underlying the gender gap in cultural interests in quantitative research. It is during childhood and adolescence that young men and women become gendered: they learn from parents and peers what kind of behavior is expected from a male or female and they learn to identify the self as masculine or feminine and behave accordingly (Egan and Perry 2001; Leaper and Friedman 2007). The dynamic view on gender differences that we endorse argues that socialization does not result into a prototypical man versus woman, but leads to a diversity in terms of gender typicality and gender conformity within each gender.

The present paper adds to existing research by relating gender differences in highbrow cultural taste to adolescents’ gender identity and experienced cultural expectations of gender-conforming behavior, keeping social inequalities constant (Egan and Perry 2001; West and Fenstermaker 1995). Using survey data from a representative sample of over 5000 Flemish secondary school students (7th grade), we analyze whether (a) female adolescents show higher interest in highbrow cultural activities than do male adolescents and (b) whether this difference is associated with gender identity and with (internalized and peer) pressure for gender-conforming behavior, particularly for boys.

Gender and Highbrow Cultural Preferences

Current research in various Western countries shows that adult women are more likely to express a taste for what is traditionally considered high culture: they are more likely than are men to be interested in and to participate in high-status cultural activities such as going to the opera, attending a play, and visiting an (arts) museum (Bihagen and Katz-Gerro 2000; Christin 2012; Falk and Katz-Gerro 2016; Lizardo 2006b; Purhonen et al. 2011). Research on adolescents is scarce, but shows a similar gender gap among youth (Schmutz et al. 2016; Siongers and Lievens 2014, Willekens and Lievens 2014).

Based on research on adults, there are two large groups of partially overlapping explanations for the gender gap in highbrow culture consumption. A first group of explanations focuses on cultural capital, employment, and social class. It is argued that the gap originates in gendered educational and work-related choices and contexts and in differential involvement in the labor force (Bihagen and Katz-Gerro 2000; Collins 1988; DiMaggio 2004). For instance, Christin (2012) shows that women’s overrepresentation in the cultural and education sectors explains part of the gender differences in the United States. Relatedly, Lizardo (2006b) indicates that the gender gap is much smaller in occupational fields where the proportion of cultural capital (relative to the proportion of economic capital) is higher. Moreover, DiMaggio and Mohr (1985) argue that women invest in cultural capital to have a better position on the marriage market. Bourdieu (1984, 2001) and Collins (1988) suggest that (middle-class) women consume more highbrow culture because they are responsible for the family’s public image and for the cultural reproduction, that is, the “cultural housekeeping” (see Lovell 2001, p. 23) within the family.

However, these explanations overlook that highbrow cultural tastes are not necessarily a marker of cultural capital related to social inequality, but may also originate in gender inequality. Indeed, activities related to the arts and literature have strong feminine connotations, and there is substantial evidence for this gender-typing of highbrow cultural activities. For instance, psychological research explaining engagement in different school topics using Implicit Association Tests indicates that children and (young) adults consider math and science as “male” subjects, but arts and reading have a feminine connotation (Cvencek et al. 2011; Heyder and Kessels 2013; Nosek et al. 2002; Nosek and Smyth 2011). Martino (1999) shows that the academic and emotional nature of cultural practices, like reading books and playing theatre, is at odds with the conceptions of masculinity in an Australian high school. Other work on ballroom dancing in the United States, for example, suggests that the artistic, creative, and aesthetic aspects of the activity give it a feminine connotation (Leib and Bulman 2009).

A second group of explanations of the gender gap in tastes focuses on gendered socialization and separate spheres ideologies. These arguments pay more attention to the feminine connotations of the arts and literature. Tepper (2000) for example, suggests that differential cultural taste is explained by society-wide cultural norms, originating in Victorian separate spheres ideology, that define highbrow cultural activities as related to the feminine sphere because these activities are passive, private, non-competitive, and academic. The gendered connotations to cultural activities are expected to lead to gender-specific early socialization in the arts and literature within the family (Bihagen and Katz-Gerro 2000; Christin 2012; Katz-Gerro and Jaeger 2015; Tepper 2000). For instance, using adults’ retrospective accounts, Christin (2012) shows that a quarter of the gender differences in cultural participation among American adults is explained by differential participation in the arts as a child, and she argues that parents may stimulate their sons and daughters to like those practices that are deemed acceptable for boys and girls. Even though Katz-Gerro and Jaeger (2015) show that there is little evidence that parents engage in gender-specific cultural socialization in Denmark, other socializing agents, such as peers, may be responsible for the gender differences they find.

The gendered socialization and separate spheres explanations share the assumption of the gender-role socialization approach to the study of gender differences, tracing differences between men and women back to childhood gender-role socialization. This approach—common in Bourdieu’s days (Bourdieu 2001; Silva 2005)—has received much critique (Lopata and Thorne 1978; Risman and Davis 2013). First of all, gender-role socialization approaches overemphasize the importance of early childhood and socialization in the family as the site of gender socialization and overlook that peers play an important part in youths’ gender socialization as well (see Harris 1995; Leaper and Friedman 2007). Moreover, they have a static vision on gender that leaves little agency and possibilities for resistance to the individual, as if (young) men and women always conform to pressures from others to adhere to stereotypical gender beliefs. Lastly, gender-role socialization perspectives encumber current explanations with a dualistic vision of gender, which often results in analyzing gender differences by including an independent variable “man/woman” in the analysis and drawing conclusions based on “average” women and men. We argue that treating gender as a multidimensional and fluid concept and thus explicitly considering gender identity and gender conformity as variable processes holds most promise to solve the puzzle of gender differences in cultural preferences (Egan and Perry 2001; Risman 2004; Thorne 1997).

Gender Identity and Interactional Cultural Expectations

As a response to the critiques on the gender-role socialization perspective, social psychology has focused on gender identity at the individual level and sociology on interactional expectations (Risman 2004; Risman and Davis 2013). Gender identity is the extent to which a person perceives and identifies the self to be masculine or feminine in relation to what is considered masculine or feminine in a specific context (Egan and Perry 2001; Tobin et al. 2010; Vantieghem et al. 2014b; Wood and Eagly 2009). According to gender identity theory, differences between males and females originate in masculine and feminine gender identities (Egan and Perry 2001). Since the downfall of the gender-role socialization perspective, gender identity is understood to be multidimensional; it relates to different fields of life (not only psychological traits for instance; see Vantieghem et al. 2014b, for an overview of the evolution of the concept gender identity).

The multidimensional approach to gender identity of Egan and Perry (2001) has been highly influential in the field. According to Egan and Perry, gender identity has several facets, two of which are the focus of this article because qualitative research has shown them to influence cultural tastes (Cann 2014, 2015; Pascoe 2007): gender typicality and pressure for gender conformity. Gender typicality is the extent to which one feels to be a typical member of one’s gender category, of course in reference to the social ideas of what a woman or a man is. Pressure for gender conformity refers to the pressure people experience from themselves and from others to conform to gender stereotypes.

The idea of gender conformity can also be traced back to the sociological “doing gender” perspective, where it refers to cultural expectations on the interactional level (West and Fenstermaker 1995; West and Zimmerman 1987, 2009). This symbolic-interactionist theory argues that people display gendered behavior and use gendered symbols in everyday social interactions in the production and justification of their gender. Gender is theorized to be a process, something one continually does (rather than is) as a constant proof that one lives up to the cultural expectations of what women and men should be and do. As a consequence, gender is not a natural, essentialist attribute of an individual but something that originates in social interactions. One is only considered a competent member of society when producing gender in an acceptable way, and when not performing gender as expected, an individual faces punishment (e.g., ostracism and bullying). However, the theory allows for agency because the individual still has the choice to consciously resist these gendered expectations and “undo” his/her gender, which is of course easier in contexts where there is less pressure to conform to gender stereotypes.

Currently, there is a growing body of research in educational sociology using gender identity theory and doing gender perspectives to explain girls’ and female adolescents’ advantage in certain educational outcomes and to explain (fe)male adolescents’ gendered subject interest and motivation (Leaper and Van 2008; Leaper et al. 2012; Vantieghem et al. 2014b). Furthermore, there is some qualitative research linking leisure-time activities to the formation of gender identity. Cann (2014, 2015), for example, shows how gender is experienced, negotiated, and (re)produced in pop-cultural tastes regarding sports, music, and reading among British adolescents. Cherland (1994) finds that fiction reading among Canadian girls—much approved by their parents—is part of the production of a feminine identity. The constructed gender identities in cultural activities are often complex, fluid, and multi-faceted as research on ballroom dancing by American adults shows (Leib and Bulman 2009). Recent work by Schmutz et al. (2016) indicates that the gender gap in highbrow tastes of adolescents differs across school contexts and is larger in rural schools than in urban schools, which is assumed to reflect gender conservative versus progressive attitudes in those schools.

Moreover, it appears to be more important for boys and young men not to be feminine or engage in feminine activities than it is for girls and young women to perform the “right” gender through feminine cultural activities (Cann 2014, 2015; Cherland 1994; Reay 2002). It is theorized that because of the devaluation of the feminine, it is more acceptable for (certain) girls and female adolescents to do or like things that boys and male adolescents usually do or like than vice versa, because those girls and young women engage in behavior that is socially valued (e.g., pursuing a degree in engineering, rather than in humanities, which will lead to a better paid job) (Cann 2014, 2015; Kane 2006). Boys and young men however, have much more status and prestige to lose when showing gender-incongruent behavior and have to resist being associated with devalued feminine attributes. As a consequence, young men define masculinity in terms of it not being feminine and try to distance themselves from femininity (Martino 1999).

Whereas for girls and young women it is possible to express their femininity in other ways than via highbrow cultural preferences, for boys and young men being associated with feminine activities is suspect per definition (Tepper 2000). Boys and (young) men interested in reading, ballroom dancing or taking drama classes are considered sissies, poofters or fags (Leib and Bulman 2009; Martino 1999; Pascoe 2007). These homonegative slurs are theorized to be important means by which boys’ masculinity and gender conformity is policed and judged by their peers. Gender non-congruent boys have fewer friends and are more likely to be bullied by peers (Young and Sweeting 2004). Peers are considered to be an important socializing factor in the formation of identity among adolescents, sometimes even more so than parents are (Harris 1995).

The Present Study

Turning to the present study involving male and female adolescents, we expect young women to be more interested in highbrow culture than young men will be (Hypothesis 1). We expect this difference in interest between male and female adolescents to be related to both individual gender identity (as measured by gender typicality) and gendered interactional cultural expectations (as measured by gender conformity pressure). Specifically, we hypothesize that higher gender typicality will be related to lower highbrow cultural interest for male adolescents, but to higher interest for female adolescents (Hypothesis 2a). The association between gender typicality and highbrow cultural interest is expected to be stronger for male adolescents than for female adolescents (Hypothesis 2b). Moreover, we predict that higher pressure for gender conformity will be related to lower interest in highbrow culture for male adolescents, but to higher interest for female adolescents (Hypothesis 3a). The association between gender conformity pressure and highbrow cultural interest is hypothesized to be stronger for male adolescents than for female adolescents (Hypothesis 3b).

Method

Participants

We employ data from 5227 Flemish secondary school students (2773 young men; 2454 young women) in the first year of secondary education (7th grade in the American educational system). The study of 7th graders is useful for our purpose because the onset of biological puberty in early adolescence makes gender differences particularly salient and pressures to conform to gender norms become stronger (see research on the gender intensification hypothesis, such as Galambos et al. 1990; Hill and Lynch 1983). Moreover, these adolescents are old enough to expect that tastes formed in this life phase will remain more or less stable and continue in adulthood (see for instance Nagel and Verboord 2012). Most of them (n = 4190, 80.2%) were 12 years old (M age = 12.18, SD = .48, range = 10–15), meaning that the majority of the students were on track, whereas 949 (18.2%) of the students were older than 12 and had school delay. A small group of 88 (1.7%) students was younger than 12, meaning that they were allowed to skip one or two grades during their school career. The large majority (n = 4833, 92.5%) of the students were in a general, academic track, whereas 394 (7.5%) were in a vocational track. About 19% (n = 1015) of the students had a working-class background; a majority (n = 3028, 57.9%) had a middle-class background; and the remaining 22.7% (n = 1184) had a service-class background. A complete description of the sample (total as well as for boys and girls separately) is shown in Table 1.

Table 1 Descriptive statistics for the total sample and separately for young men and young women
Full size table

Procedure

Data were gathered from September to December 2012 as part of the project ‘Teaching in the bed of Procrustes’; detailed methodological information is also available in Vantieghem (2015, pp. 63-73). Based on an inventory of all Flemish schools that offer 7th grade education, certain criteria were applied to define subpopulations. Schools were selected to represent all geographical regions in Flanders, the northern Dutch-speaking part of Belgium. Selection was also based on school denomination (public versus Catholic), and a proportional representation of schools in cities and rural areas was assured. Within these parameters, three random samples were drawn. When a school in the first sample refused, a matched school in the next random sample was contacted. Thus, the selected schools represent the Flemish educational context on important dimensions, while the randomness within these subgroups was maintained. In total, 58 (47%) schools participated in the study (124 were contacted). Comparison of the sample to the Flemish school population indicated that the schools are representative and that no systematic biases occurred. A waiver of parental consent and the use of child assent were approved by the school and the Belgian Commission for the Protection of Privacy, based in the minimal risk and confidential nature of the study (see Vantieghem, 2015, p. 66).

All seventh graders in the participating schools were asked to complete the paper-and-pencil questionnaire, administered by a researcher, in their classrooms in the time span of one course (50 min). A researcher was present the entire time to explain the purpose and procedure of the survey and to answer questions. Students were assured that the survey was confidential: the unique code on the questionnaire was necessary to link their answers to a follow-up survey, but the datasets for analysis contain only anonymous data. In total, 6380 students filled out the questionnaire. However, students with missing data on one of the variables were excluded, so the analyses are based on 5227 (82%) students. For the scales used in the analysis, having missing answers on more than 25% of the constituent items resulted in a missing on the scale.

Dependent Variable: Interest in Highbrow Culture

As an indicator of highbrow cultural taste, we use a scale measuring interest in a variety of highbrow cultural activities. We choose an indicator of cultural taste preferences, and not of cultural participation, for the following reasons. According to Peterson, “respondents’ self-reports of their preferences [compared to cultural participation] seem a more direct measure of the way they use art in shaping identity and symbolically announcing their place in the world” (Peterson 2005, p. 265). Because identity processes are central in our approach, studying cultural taste preferences as measured by interests makes sense. Although cultural participation may appear more interesting at first sight in relation to the “doing gender” perspective, one has to keep in mind that displaying specific attitudes toward the arts and literature is as much a part of the performance of gender as participation in those realms (see Cherland 1994). After all, the cultural participation of early adolescents is frequently mediated by parents or teachers and compulsory, and thus not always an expression of individual preference.

In the survey, students were asked how interested they are in 24 leisure activities: watching sport as a spectator or fan, doing sports, making music, doing drama/word courses, painting/drawing/clay modelling, cooking, going to the movies, listening to music, attending a concert, attending a play or dance performance, visiting a library, visiting an (art) museum, watching TV, playing video games, using social media (Facebook, Twitter, …), surfing on the internet, collecting things (stamps, comic books, …), reading, playing board games (e.g., Monopoly), shopping, going out (to parties, …), meeting friends (at home or somewhere else), spending time with family, youth organization (e.g., Scouting). Each was rated using a 4-point scale from 0 (not interested) to 3 (very interested).

Categorical Principal Component Analyses (CatPCA), which were performed on the wide range of (highbrow and other) cultural activities included in the questionnaire, showed that the following cultural activities have high loadings on the first dimension: making music, doing drama/word courses, painting/drawing/clay modelling, attending a concert, attending a play or dance performance, visiting a library, visiting an (art) museum, and reading. These activities are all related to the arts and literature and were used to measure Interest in Highbrow Culture. A mean scale score with high internal consistency (Cronbach’s α = .77) was computed. Higher scores on this scale mean higher interest in highbrow cultural activities. On average, young men scored almost .5 points lower on this scale (see Table 1). Six activities loaded on the second dimension: watching sport as a spectator or fan, doing sports, going to the movies, listening to music, watching TV and playing video games. The composite measure based on these activities indicates pop-cultural taste and is used as a control in the analysis.

Independent Variables: Gender, Typicality, and Conformity

The central explanatory variables are gender of the student, gender typicality, and pressure for gender conformity. As can be seen in Table 1, 53% of the adolescents in the sample are boys (the reference group). The indicators of gender typicality and gender conformity pressure are based on the Self-Perception Profile of Egan and Perry (2001), which is increasingly used for research on gender identity in children (Perry and Pauletti 2011) and applied in research on gendered educational outcomes (Leaper et al. 2012; Leaper and Van 2008; Vantieghem and Van Houtte 2015; Vantieghem et al. 2014a). The Dutch translation and answer formats of the measures were based on the study of Bos and Sandfort (2010).

To measure gender identity, we used the gender typicality subscale of this gender identity questionnaire. Gender Typicality measures the extent to which males and females feel they are typical for their gender (in general and with regard to important dimensions) and is a mean scale based on six items using a 5-point Likert scale ranging from 0 (completely disagree) to 4 (completely agree). Sample items are: “I feel that I am a good example of a typical boy/girl,” “I feel that the things I am good at are similar to those of most boys/girls,” and “I feel that my personality is similar to that of most boys/girls,” with each designation matching the gender of the respondent. The higher a student scores on the measure, the stronger his/her feeling of gender typicality. The scale has good internal consistency reliability (α = .78). On average pupils had a score of 2.48, ranging from 0 to 4, and young men reported higher typicality than did young women (see Table 1).

The pressure for gender conformity scale indicates the extent to which a pupil experiences pressure for gender-conforming behavior. Four items measure pressure from peers and four items measure pressure experienced from oneself, each using a 4-point Likert-type response scale from 0 (completely disagree) to 3 (completely agree). Sample items are: “I think it is important to act just like other girls/boys” (internalized pressure) and “The boys (girls) I know would be upset if I wanted to learn an activity that only girls (boys) usually do” (pressure from peers), again matching items to the respondent’s gender. Higher scores mean higher perceived pressure for gender conformity. The measure is a mean scale with good internal consistency reliability (α = .82), with all items loading highly on the first dimension of an oblique factor analysis, which indicates that the questions measuring internalized pressure and measuring pressure from peers do not have a different underlying construct. This overlap is not surprising because, like pressure from peers, pressure from oneself originates in social interaction. On average, male adolescents scored considerably higher on the scale than did female adolescents (see Table 1), which means that young men perceive more pressure from others and themselves to conform to gender stereotypes. Also young women experience a considerable amount of pressure (see Table 1). Gender Typicality and Gender Conformity Pressure are correlated (r = .22, p < .001), but they can still be meaningfully separated.

Controls

We take into account the following demographics and indicators that are associated with the independent variables and that might influence the effects of these variables on the dependent variable (see Table 1). Age Group indicates whether the pupil is 12-years-old or younger (reference category) or whether (s)he is 13 or older, meaning s/he has school delay. We also account for the student’s educational track: general (reference) or vocational. We include a categorical indicator of the socio-economic background of the student. Following the EGP-classification (Erikson and Goldthorpe 1992; Evans 1992), this variable is based on the current or last occupation of the father and mother and distinguishes working-class youth (reference category) from middle class and service class youth. The adolescent’s social class is the highest score of both parents (when applicable). We add Ethnicity to the model, which is based on the country of birth of the adolescent’s parents. The variable distinguishes students from Belgian or Western-European origin (reference group) from students with either one or both parents from non-Western European origin.

We also consider the adolescent’s results on a reading test as a proxy for academic attainment because we know there is an association between cultural preferences and academic attainment (Dumais 2002; DiMaggio 1982), as well as between gender typicality and gender conformity pressure and educational outcomes (Vantieghem and Van Houtte 2015; Vantieghem et al. 2014a). In this test, students answered questions on content and vocabulary of four texts they had to read. As mentioned, we also include a mean scale of interest in pop-cultural activities. Preliminary analyses showed that many adolescents reporting high interest in highbrow cultural activities were also interested in these other cultural activities. Because research shows that these pop-cultural tastes are influenced by identity and pressure-related processes (Cann 2014, 2015), we control for interest in these other common cultural activities.

Lastly, we control for the Egalitarian Gender Role Ideology of each adolescent. This variable is a mean scale of 15 variables indicating egalitarian and traditional beliefs about female and male roles (α = .79), developed by Vermeersch et al. (2010). It includes questions such as “The husband should be the decision-maker of the family”; “There is something wrong with girls who talk dirty”; and “Women should first consider their children, and only then their career” (all reverse-scored). The higher a pupil scores, the more egalitarian the ideas about gender roles the pupil holds. We take gender beliefs into account because research has shown that traditional gender beliefs are associated with worse outcomes on academic indicators for both males and females (Davis and Pearce 2007; Scott 2004) and because egalitarian gender-role attitudes are heavily correlated with pressure for gender-conforming behavior (r = −.50, p < .001). Adolescents growing up in contexts where traditional ideas about gender roles prevail experience higher pressure for gender-conforming behavior (both internalized and from peers) and vice versa.

Results

We used multilevel models to account for the nested structure of the data given that pupils are clustered in schools (Hox 2010). First, we ran an intercept-only model which separates variance at the individual level from variance on the school level (shown in Table 2, null model). This model served as the baseline model for the following models. Without any controls, about 10% of the variation was on the school level as indicated by the Intra-class Correlation Coefficient (ICC) or the Variance Partitioning Coefficient (VPC): ICC = .038/(.038 + .335) = .102.

Table 2 Multilevel regression models predicting highbrow cultural tastes
Full size table

To test Hypothesis 1, which predicted that young women are more interested in highbrow culture than young men are, we added the variable Gender, which had a significant effect, in Model 1 (see Table 2). On average, young women scored .42 higher on the highbrow interest scale than young men did. Thus, Hypothesis 1 was supported. Using the variance components on the individual level (е0j), we calculated that adding Gender in Model 1 explained almost 11% of the variation in highbrow cultural taste among these adolescents compared to the null model [level 1 explained variance = 1-(.299/.335) = .107]. This indicates that gender has a substantial effect on highbrow cultural interest. As shown in Model 2, the coefficient of Gender was only slightly reduced when all control variables were introduced.

To evaluate Hypothesis 2a (which predicted that gender typicality has a positive association with highbrow cultural taste for young women, but a negative association for young men) and Hypothesis 2b (which expects the association for male adolescents to be stronger), we added Gender Typicality, Gender Conformity Pressure, and an interaction term between students’ Gender and Gender Typicality in Model 3 (see Table 2). To make the interpretation easier, the relationship between Gender, Gender Typicality and Highbrow Cultural Taste for young men and women is also diagrammed in Fig. 1. In line with Hypothesis 2a, we found that the more typical a male adolescent considers himself to be, the lower his interest in highbrow culture, but the more gender congruent a female adolescent is, the higher her interest in highbrow cultural activities. As can be seen in Fig. 1, the predicted gap in highbrow interest between atypical male adolescents and atypical female adolescents (on the left-hand side of the figure, Δa = .244) is only half of the gap in highbrow cultural taste between gender typical male and female adolescents (on the right-hand side, Δb = .497). Thus, Gender Typicality is strongly related to the gender gap in cultural taste between young men and women. As illustrated in Fig. 1, the negative effect of Gender Typicality on Highbrow Cultural Taste for young men is much more pronounced than the positive effect is for girls: the line depicting the relationship between Gender Typicality and Highbrow Cultural Tastes for young women is rather flat, whereas the line for young men is much steeper. Thus, Hypothesis 2b was supported.

Fig. 1
figure 1

The association among gender, gender typicality and highbrow cultural taste

Full size image

To evaluate Hypothesis 3a (which expected a negative association between Gender Conformity Pressure and Highbrow Cultural Interest for young men and a positive association for young women) and Hypothesis 3b (which expected the association for young men to be stronger), we included Gender Typicality, Gender Conformity Pressure, and an interaction term between students’ Gender and Gender Conformity Pressure in Model 4 (see Table 2). The relationship among Gender, Gender Conformity Pressure, and Highbrow Cultural Taste for male and female adolescents is illustrated in Fig. 2. Gender Conformity Pressure had a negative effect for male adolescents and a (very small) positive effect for female adolescents, which confirms Hypothesis 3a. The more pressure for gender conformity a young man experiences, the lower his interest in highbrow culture. For young women, higher gender conformity pressures were associated with only slightly higher interest in highbrow culture. As illustrated in Fig. 2, the gap between male and female adolescents who do not experience pressure for gender conformity is 61% of the difference in highbrow interest between male and female adolescents who experience a lot of pressure for gender conformity (Δa = .308 versus Δb = .506). So, pressure for gender conformity influences the highbrow cultural taste gap between young men and women substantially. As illustrated in Fig. 2, the negative effect of Gender Conformity Pressure on Highbrow Cultural Taste for young men is much more pronounced than the positive effect is for young women: the line depicting the relationship between Gender Conformity Pressure and Highbrow Cultural Tastes for young women is almost flat, whereas the line for young men is much steeper. Thus, Hypothesis 3b was supported.

Fig. 2
figure 2

The association among gender, gender conformity pressure and highbrow cultural taste

Full size image

Discussion

In the present paper, we studied gender differences in highbrow cultural taste among a representative sample of Flemish 7th graders. As expected, compared to male adolescents, female adolescents report considerably higher interest in highbrow cultural activities related to arts, music, theatre, and literature, which are known to have feminine connotations (Christin 2012; Nosek and Smyth 2011; Tepper 2000). Interestingly, this gender gap in tastes is strongly related to the gender typicality and gender conformity pressures experienced by these early adolescents. Higher gender typicality and higher pressure for gender conforming behavior are associated with slightly higher highbrow cultural interest for young women, but with much lower highbrow cultural interest for young men. We conclude that gender typicality and gender conformity pressure are important mechanisms reinforcing the gender gap in highbrow cultural tastes. Thus, our results are in line with existing qualitative research on children’s and adolescents’ cultural preferences which already suggested that they construct their gender identity and do their gender through gender-typed cultural activities such as reading (specific) books and listening to certain kinds of (popular) music (Cann 2014, 2015; Cherland 1994).

Moreover, the stronger effects of gender typicality and gender conformity pressure for male adolescents indicate that it is more difficult for young men to like an activity perceived as feminine than it is for young women to dislike a feminine activity. Female adolescents probably have other ways to do their gender appropriately whereas for male adolescents being associated with femininity is suspicious per se. Whereas existing research mostly focuses on why women express high interest in highbrow culture, our results show that from a gender perspective another question is equally important: Why do (young) men have so little interest in highbrow cultural activities? Although this re-framed question has been ignored by most existing research, our study shows that gendered identity-related and interactional processes make it particularly difficult for growing-up boys and young men to engage in activities that are considered feminine. Cultural conceptions of what (young) men are and should like are very much linked to their highbrow cultural tastes, which prevents them from exploring their talents in arts, music, theatre, and literature and from learning the specific skills (creativity, self-expression, etc.) that these activities have to offer.

Limitations

An important limitation of our study is related to the cross-sectional nature of the data; we were only able to uncover the associations among the variables, not their causal direction. For example, there is research that suggests that identity processes and pressures influence the interest in cultural activities of (fe)male adolescents (Martino 1999; Pascoe 2007). Other work indicates that participation in gender-typed leisure activities in childhood may also affect the identity development later in adolescence (McHale et al. 2004). When it comes to identity processes, it is possible that (a) feeling a(n) (a)typical young (wo)man leads to (a)stereotypical cultural tastes, that (b) having (a)stereotypical cultural tastes causes feelings of gender (a)typicality or that (c) gender typicality and gendered cultural tastes have reciprocal effects. Further research looking into this causality question is needed.

We further argue that the longitudinal data needed for this analysis has to comply with certain conditions and standards that at present are not yet met by existing data. First, there have to be many measurement points because as a consequence of regression to the mean, adolescents with extreme scores on the measures may score values closer to the mean in subsequent measurements, and this natural variation may be mistaken for real change. Second, there has to be enough time between measuring points for gender identity to vary because gender identity tends to be stable for shorter time periods (Egan and Perry 2001). However and third, the measuring points should be close enough in time to minimize biological changes during adolescence (for instance development of secondary sex characteristics between measurements), which may change the meaning and assessment of gender identity independent of changing tastes. Fourth, the first measurement needs to be as early in life as possible because gender identification is an ongoing process starting very early in life (Leaper and Friedman 2007). We expect this kind of longitudinal research to find feedback processes: although specific gender identities and conformity pressures make it more likely that a child or an adolescent is interested in highbrow culture, an interest in culture, and particularly participation in highbrow activities, provide a context that can influence the adolescent’s identity development and the pressures for conformity s/he experiences. Most likely, the processes described in our paper are the result of a complex interplay in which tastes, on the one hand, and identity and pressure, on the other, are both causes and consequences. This intertwining of gender and cultural tastes underscores the relevance of the research question.

Future Research Directions

Our approach to the study of the gender gap in highbrow cultural tastes among adolescents can be extended by looking into the relationship with the student’s social class, with the pupil’s school context, and with current perspectives on cultural capital. Additionally, we should investigate whether it can be applied to adults’ tastes or in other parts of the world. First, although we argued that existing explanations tend to focus too much on work- and class-related inequalities, the social background of the adolescent remains relevant. Although the gap between young men and women does not differ significantly according to social class in the final models we presented, working-class youth in the sample experienced more pressure for gender conforming behavior than did middle-class and service-class youth (who experience the least pressure; analyses not reported). In this way, social class does reinforce gender inequalities.

Although we contributed to existing research by unraveling the gendered processes at play, it is true that an adolescent does not have a gender or a class position, but rather both at the same time and these are intertwined in complex ways, as understood by proponents of an intersectional perspective (Anthias 2004; Crenshaw 1991; Yuval-Davis 2006). Qualitative educational research such as Willis’ (1977) classic “Learning to labour” suggests that there are specific types of young men and women with specific femininities and masculinities (in terms of the combinations of identity, pressure, gender role ideology, etc.) that are more prevalent in some social classes than in others (Jackson 2006; McRobbie 2000; Willis 1977). A good example is Willis’ working-class lads. An integration of our gender perspective and social class research will allow researchers to explore how mechanisms behind the gendering of cultural taste relate to the formation of the (class) habitus and will add to research on why masculinity is so important for lower class boys and young men (Reay 2002). Future research that studies how class and different manifestations of gender intersect in their relation with highbrow cultural taste will also benefit from looking into the effects of school characteristics: being an atypical young man with a certain social background may be something different in secondary schools with only vocational versus academic tracks, in urban versus rural schools (see for instance Schmutz et al. 2016), schools with low or high average gender conformity pressure and typicality, etc.

Second, now we know that gender identity and pressures for gender conforming behavior support the gendering of tastes in adolescence—a life stage in which gender differences intensify (Hill and Lynch 1983), but the question remains as to what extent these mechanisms are relevant to explain the gender gap in adulthood. It is important to stress that we consider our perspective complementary and not contradictory to existing explanations that have mainly focused on gendered early socialization, separate spheres arguments, and differential educational, work-related and class-related contexts. Our approach adds to existing separate spheres arguments because we uncover the mechanisms by which cultural beliefs of the femininity of highbrow culture may become part of an enduring set of dispositions. Furthermore, although Christin (2012) links the gender gap in cultural participation among adults to the gender gap in childhood participation, for instance, our perspective offers part of the explanation of why there is a gap between boys and girls in the first place.

Moreover, although current educational, work-related, and class-related explanations often mistake manifestations of gender inequality for social inequalities, our perspective encourages a look beyond the surface to the gendered nature of taste differences between men and women. For instance, Christin’s (2012) finding that the gender gap in tastes is explained by women’s overrepresentation in the cultural and education sectors and Lizardo’s (2006b) finding that the gender gap is largest in sectors with high economic capital (relative to cultural capital) may actually reflect gender identity processes and interactional pressures because gender typicality and experienced pressures for gender conformity can lead to stereotypical subject choices (Leaper and Van 2008; Leaper et al. 2012) and because occupational gender stereotypes and gender (in)congruence affect men’s interest in female-dominated occupations (Forsman and Barth 2016). Gender congruent women who experience a lot of pressure to conform to stereotypical gender beliefs and atypical men may choose to work in those high-cultural-capital, mostly female-dominated sectors, whereas gender congruent men experiencing pressure for gender conformity in adolescence and young adulthood self-select into those male-dominated sectors where economic capital dominates. Thus, the relationship between tastes and the different dimensions of gender among adults needs to be examined.

Although the association between the gender gap in cultural tastes and gender identity and pressure will most likely not be as strong among adults as among adolescents because other group identifications may be more relevant (e.g., class or ethnicity), it is to be expected that there will remain at least a modest association between these gendered mechanisms and the gender gap in tastes among adults. This finding would call for a revised perspective on cultural tastes as a marker of cultural capital because, if a substantial part of the gender gap is related to gender inequality rather than social inequality, this could contribute to our understanding of why women may not be able to translate cultural tastes into social advantages to the extent that men can. Thus, women’s cultural tastes may only partly function as capital. More research is needed on whether women have the same exchange rate for their highbrow cultural tastes, knowledge, and investments as men do, also considering the growing relevance of new or so-called emerging forms of cultural capital, which reduce the importance of highbrow culture (Prieur and Savage 2013, Roose 2015).

Third, research in other contexts should verify whether the processes underlying the gendering of tastes in Flanders are similar in other countries. We have reason to expect that the processes we uncovered are at least fairly similar in other Western societies. The gender gap in cultural involvement is present in most Western countries (for the US: Christin 2012; and for the EU countries: Falk and Katz-Gerro 2016). Additionally, British and Canadian qualitative research links gender identity to (pop-)cultural preferences (Cann 2014, 2015; Cherland 1994), and American research shows that gender identity and conformity pressures impact on stereotypical subject choices and engagement among students (Leaper et al. 2012; Leaper and Van 2008) and that (young) adults associate arts with femininity (Nosek and Smyth 2011). In this regard, it is important to consider that the meaning of highbrow culture and what is understood as masculine or feminine may differ slightly from country to country, highlighting the need for complementary qualitative research on this topic.

Practice Implications

Our work shows that gender identity processes and pressures for gender conforming behavior are important mechanisms reinforcing the gap in cultural tastes between young men and women. This suggests that changes in what kind of cultural behavior is expected from boys and girls and young men and women will go hand in hand with a decline in the gender differences in taste. Trying to create equal access to arts-, music-, theatre-, and literature-related leisure activities is important because leisure time is an important part of adolescents‘ lives and identities (Cann 2014, 2015; Frønes 2009). Leisure time cultural activities give them the opportunity to develop and grow as a person, to express creativity and emotions, to explore new talents, and to develop artistic abilities in which they can excel.

Our research suggests that the creation of safe environments where there is little pressure to conform to stereotypes of what typical (fe)male adolescents should do and like is important for young men to be able to explore their talents in the domains of arts, music, literature and theatre and to get access to these enriching experiences. Because education is an institution with an emancipatory role that has a large impact on the everyday lives of young people, it is important that schools can provide such a safe environment in which all gender expressions are welcomed and valued. One of the aims of the project “Teaching in the bed of Procrustes” (Vantieghem, 2015) was to provide guidelines and workshops for teachers to work towards a gender-progressive environment in their classrooms and schools, which has materialized in a practice-oriented book for teachers on which many of the suggestions proposed here are based (Van Maele et al. 2015, see also the doctoral dissertation of Vantieghem 2015, pp. 196–201).

A more gender-equal environment in schools can be achieved by avoiding school policies, curricula, and teacher-student interactions that reinforce traditional gender binaries. Because we know that bullying is an important means by which the behavior of youth is policed by their peers, it is important that schools have a developed anti-bullying policy that pays attention to gender-based bullying and creates recognition for different gender expressions. Of course, teachers are essential for the implementation of these policies and they should receive guidelines on how to recognize and deal with gender-based bullying. Moreover, using gender as means to organize groups and courses should be avoided. For instance, often sport-classes are gender-segregated, even though segregation in courses reinforces stereotypical behavior (Martin et al. 2014). Segregated classes or group work will not benefit atypical young men who experience a lot of pressure to conform to stereotypes.

Furthermore, schools and teachers should stimulate pupils, young men and women alike, to use arts and poetry to express themselves, for instance through workshops in which different talents (music, visual arts, dance, theatre, handicrafts, …) can be discovered. Lastly, when it comes to studying arts, literature, and poetry in class, it is important that teachers do not (consciously or not) reinforce stereotypes by having different expectations for young men and women or by focusing on topics that are closer to young men’s interests. Cherland (1994), for instance, indicates that teachers’ reading lists in elementary school tended to contain adventure books with male protagonists because they expect boys to avoid books regarded as being for girls. Although the intention behind the teachers’ decision may be to make boys more committed to their school work, it also accepts the stereotype that literary books are just not something for boys.

Conclusions

Gender differences in adolescents’ highbrow cultural tastes are important to study—not only because these preferences function as cultural capital later in life, but also because the gendered nature of these tastes shapes gendered experiences in childhood. Our study shows that young men’s experienced pressures for gender conforming behavior and gender typicality are strongly related to their low interests in arts-, theatre-, music- and literature-related activities. Thus, the gender gap in highbrow cultural tastes among adolescents is strongly associated with the gender identity and interactional pressures experienced by these adolescents. By focusing on youth and studying gender as a multidimensional and fluid concept, our paper adds to existing research by uncovering the mechanisms supporting the gendering of cultural preferences.