SCHOOL CULTURE AND THE WELL-BEING OF SAME-SEX-ATTRACTED YOUTH

Abstract

This study assesses how variations in heteronormative culture in high schools affect the well-being of same-sex-attracted youth. The authors focus on the stigmatization of same-sex attraction (rather than identity or behavior) to better understand how heteronormativity may marginalize a wide range of youth. Specifically, the authors use data from the National Longitudinal Study of Adolescent Health to examine how variation across schools in football participation, religious attendance, and urban locale affects same-sex-attracted adolescents' depressive symptoms, self-esteem, fighting, and academic failure. The results suggest that though same-sex-attracted youth are at greater risk for decreased well-being, these youth are at higher risk in nonurban schools and in schools where football and religion have a larger presence. Results vary for boys and girls: The urban locale of a school has a larger impact for boys, while school religiosity has a greater impact for girls.

Schools can be an unwelcoming environment for adolescents with nonheterosexual feelings, behaviors, and identities. In schools, as in the broader culture, heterosexuality is often assumed and institutionally enforced through rituals, daily interactions between students and teachers, and the curriculum (Chesir-Teran 2003). This heteronormativity can be explicit, including homophobic name-calling or verbal and physical harassment of students who deviate from normative gendered forms of sexuality (Kosciw, Diaz, and Greytak 2007). It can also be relatively subtle, perpetuated through pervasive heteronormative discourse and symbols of appropriate gender and sexual relations displayed in classrooms, peer groups, and extracurricular activities (Eder and Parker 1987; Nayak and Kehily 1996). Such displays of heterosexism are stigmatizing for same-sex-attracted youth and have negative effects on their well-being (D'Augelli 2002).

While research suggests that a better understanding of heteronormativity is needed to improve the well-being of sexual minority youth (Chesir-Teran 2003), little empirical work has been done to explore how dimensions of heteronormativity at an institutional level affect well-being at the individual level (Goodenow, Szalacha, and Westheimer 2006). In this study, we conceptualize heteronormativity within schools as the aggregation of staff and students' learned schemas or habitus (Bourdieu 2001; Ridgeway 2006), students' daily practices that reinforce normative expressions of gender and sexuality (Nayak and Kehily 1996; Pascoe 2007), and the visibility or school sponsorship of heteronormative institutions such as football and religion (Foley 1990; Osborne and Wagner 2007). Because of the central role schools and peers play in the lives of developing adolescents, it is important to understand schools as normative contexts that shape adolescents' well-being. This insight can further our understanding of how schools reinforce or deconstruct strongly embedded heteronormative patterns that marginalize individuals who deviate from hegemonic forms of sexuality.

THEORETICAL BACKGROUND

Heteronormativity and Stigma

In the United States, as in most other Western societies, heterosexuality is normative and upheld in relation to other “deviant” sexualities or sexual behaviors (Butler 1990; Foucault 1978). Heteronormativity denotes “the myriad ways in which heterosexuality is produced as a natural, unproblematic, taken-for-granted, ordinary phenomenon” (Kitzinger 2005, 478) and is premised on (and supported by) a “natural” binary division of the sexes and the privileging of other-sex desire and relationships (Kitzinger 2005; Richardson 1996). Heteronormativity involves the celebration of heterosexual relationships as well as the organization of culture, including institutions, around such relationships (Foucault 1978). It also involves the celebration of socially constructed gendered behaviors that highlight the differences between men and women as well as the sanctioning of gender transgressions that disrupt this “natural” binary (Neilson, Walden, and Kunkel 2000).

Heterosexual privilege, the social construction of gender difference, and gender inequality are tightly woven and intersecting phenomena (Stein 2008). Masculinity and femininity are sexualized, and hegemonic forms of masculinity and femininity demand heterosexual desires and behaviors (Nayak and Kehily 1996; Richardson 1996). Sexual practices deemed acceptable for young men and women are those that reaffirm men's sexual control of women and deny women's sexual desire (MacKinnon 1987). Research on schools and adolescent peer cultures has illustrated this intersection of gender and sexuality. For example, in schools, it is often the case that being called “gay” or a “fag” is a reference not so much to boys who are romantically attracted to boys or who self-identify as gay but to boys acting “nonmasculine” (Pascoe 2007; Plummer 2001). Similarly, the “dyke” label is usually assigned to girls who do not enact the cultural scripts assigned to women (Neilson, Walden, and Kunkel 2001). Such practices not only compel boys and girls to display “appropriately” gendered behavior but also marginalize and stigmatize same-sex sexual desire. Moreover, as boys and girls “do gender” to avoid such social sanctions, they continue to create and reinforce heteronormativity (Neilson, Walden, and Kunkel 2000) within their environments.

The naturalness of heterosexuality and other-sex attraction is institutionalized in marriage, often through religion (Rubin 1999; Weeks 1985) and is more subtly reinforced through a variety of taken-for-granted cultural discourses, practices, and symbols (Kitzinger 2005). In contexts where these discourses, practices, and symbols are widespread and institutionally supported, alternative sexual desires are marginalized and negatively sanctioned. Consequently, in heteronormative school contexts, same-sex-attracted adolescents may be stigmatized as their feelings conflict with strongly held and widely displayed normative expectations of appropriate gendered sexuality (D'Augelli 1998).

Stigma and the Well-being of Same-Sex-Attracted Youth

Amid widespread cultural discourses and practices that celebrate and enforce heterosexuality, youth who feel same-sex attractions may develop a sense of difference and internal distress. Feeling different at a time in life when the need to “fit in” is especially pronounced can be itself distressing (Crosnoe 2000), yet the stigma attached to same-sex attraction may lead same-sex-attracted youth to experience additional shame and discrimination (DiPlacido 1998). It is important that a stigma does not have to be visible or known to others to be distressing, nor does it have to be incorporated into one's identity (Goffman 1963). Consequently, adolescents with same-sex attractions may experience emotional distress, evaluate themselves negatively, or withdraw in an effort to avoid others' disapproval and rejection (Troiden 1989).

A large body of research demonstrates the consequences of stigma related to non-normative gendered sexualities. Sexual minority youth experience lower levels of social-emotional well-being than other youth (D'Augelli 2002). Research has also documented sexual minority youths' increased likelihood of being involved in fights, most often as the victims of physical harassment by peers (Kosciw, Diaz, and Greytak 2007; Russell, Franz, and Driscoll 2001) as well as the negative academic consequences of stigmatization faced by same-sex-attracted youth (Pearson, Muller, and Wilkinson 2007). Thus, because of the pervasiveness of heteronormativity in society, youth with same-sex attractions often experience feelings of difference and stigma and lower levels of well-being. However, while the impact of stigmatization on the well-being of same-sex-attracted youth has been well documented, research often neglects the diversity of school cultures and the ability of schools to reify or to subvert heteronormativity.

Organizational Culture and the Role of Schools

Schools are important socializing institutions in which young people struggle to define themselves in relation to others, particularly peers. This influence is not limited to friends: Recent research has documented the important influence of wider social networks, rather than friends, on individual behavior and identity (Giordano 2003). Because sexuality becomes increasingly central to identity and social relationships during adolescence, schools are critical social contexts in which dominant beliefs about sexuality are played out.

Social institutions such as schools often mirror the larger structure of society, including the norms and behaviors of acceptable sexuality (Epstein and Johnson 1998). Recent theoretical explanations of organizational culture (Hallett 2007) and empirical investigations of variation in heteronormativity across schools (Goodenow, Szalacha, and Westheimer 2006) suggest that the level of stigmatization experienced by same-sex-attracted youth may vary by the school's heteronormative climate, or the degree to which hegemonic forms of sexuality organize and stratify relational contexts (Ridgeway and Correll 2004).

The level of heteronormativity within a school is created and reinforced in part through the aggregation of staff and students' cultural schemas or habitus, which include taken-for-granted outlooks, beliefs, and experiences that are carried into and developed within schools (Bourdieu 2001; Hallett 2007; Ridgeway 2006). When a greater number of individuals within a school operate within heteronormative schemas, heteronormativity acquires more legitimacy and power, creating a relational context that limits available outlets for adolescent sexuality and stigmatizes same-sex desire. Individual heteronormative schemas may be reinforced through involvement in activities such as football and religion, which involve symbolic rituals that celebrate heterosexuality and gender differences, or in the context of less tolerant and less diverse communities. Indeed, existing research on adolescent attitudes toward homosexuality and gendered expectations of behavior has linked participation in hypermasculine team sports such as football (Osborne and Wagner 2007), greater religiosity (Olson, Cadge, and Harrison 2006), and residence in nonurban communities (Stein 2001) with strongly heteronormative outlooks. Furthermore, when a predominantly heteronormative institution, such as religion, becomes largely visible within a school, it may influence school culture regardless of the heteronormativity displayed by individual students. Moreover, when activities and rituals that promote heteronormativity are organized by the school itself, such as through football, heteronormativity may be institutionally legitimated and enforced.

Prevalence of Football

As a traditionally male-dominated domain, competitive sports are often an arena in which heterosexualized gender systems are displayed and symbolically enforced (Barron and Bradford 2007; Eder and Parker 1987; Foley 1990). In high schools, sports coaches and teammates often emphasize toughness, aggression, and stoicism, and the failure to meet these standards is associated with a lack of masculinity and heterosexuality (Messner 1992). In fact, previous research has shown that male adolescents who participate in core sports, including football, are significantly more likely than other boys to express both homophobic attitudes (Osborne and Wagner 2007) and anxiety about their performance of traditional masculinity (Nayak and Kehily 1996). When a greater proportion of the student body within a school plays football, these heteronormative attitudes may be more pervasive, dominating school culture and stigmatizing those who do not fit traditional definitions of masculinity and hegemonic male sexuality (Eder and Parker 1987). Furthermore, because high school football is a male-dominated arena in which hegemonic masculinity and male heterosexual prowess are emphasized, this aspect of school culture may have a greater impact on the well-being of same-sex-attracted boys relative to same-sex-attracted girls.

Organized high school football, as an institution, is accompanied by numerous symbolic rituals that naturalize heterosexual attractions and relationships and the social construction of gender difference. These rituals are observed by the entire student body and include pep rallies and homecoming events, where highly masculinized football players, feminized cheerleaders, and heterosexual courting rituals are overtly celebrated (Foley 1990). When a greater proportion of the student body within a school plays football, the school-sponsored rituals surrounding institutionalized football may take on more symbolic power and legitimacy, dominating school culture and stigmatizing same-sex-attracted youth.

Prevalence of Religiosity

Individuals more attached to religious institutions express higher levels of heteronormativity, exhibiting more homophobia and adherence to “traditional family values” than their less attached counterparts (Olson, Cadge, and Harrison 2006). The fundamental tenets of most of the Judeo-Christian religions predominant in U.S. society teach the restriction of sexual activity to married, heterosexual couples and label all other forms of sexuality as deviant (Greenberg and Bystryn 1982; Rubin 1999). Many major religions have created powerful rituals and beliefs that celebrate “appropriate” expressions of sexuality as a way to preserve heterosexual marriage and the traditional family unit (Greenberg and Bystryn 1982). Although heteronormative beliefs vary by religious affiliation, with conservative affiliations being the most heteronormative, it may be the collection of many students who attend religious services regularly, regardless of affiliation, that contributes to a school culture that is less tolerant of alternative expressions of sexuality (Olson, Cadge, and Harrison 2006).

While religious condemnation of alternative expressions of sexuality is directed at both men and women, religious institutions have generally been more concerned with the sexual desire and expression of women and girls. Girls are provided with cultural scripts that deny women's sexual desire and emphasize women's obligation to fulfill the emotional and sexual needs of their husbands (rather than their own) (Heath 2009; Weeks 1985). Likely as a result of these gendered messages about sexuality, girls' sexual behavior is more influenced by their religiosity (Rostosky et al. 2004). Consequently, being situated in a context in which peers are attached to religious institutions may create greater feelings of difference and stigma for girls who experience same-sex attractions than for similarly situated boys.

School Locale

While the prevalence of religiosity and football participation within high schools may signal more heteronormative contexts, the geographic location of schools may also shape the climate within schools. Community size is positively linked to nontraditional behaviors as well as the tolerance and acceptance of such behaviors (Wilson 1995), and individuals in smaller communities generally report higher levels of homophobia and less acceptance of sexual diversity than those in larger, more urban areas (Stein 2001). Consequently, individuals in urban areas may be more tolerant and accepting of diversity, relative to those living in suburban or rural areas; therefore, schools in urban areas may display lower levels of heteronormativity. Thus, we expect that sexual minority youth may have more difficulty adjusting in schools located in rural and suburban communities, as these smaller and less urban communities may reinforce a heteronormative culture.

In addition to a more traditional and conservative culture emphasizing family values prescribed by heteronormativity, nonurban areas are also less likely to have visible gay spaces, which may soften some of the stigma and shame attached to same-sex attraction encountered in schools as well as the isolation caused by feelings of difference (Flowers and Buston 2001). Research has shown that for boys, while a school culture emphasizing sports often stigmatizes same-sex attractions and identities, the “gay scene” outside of school often serves as a safe space for exploring non-normative sexuality and desire (Barron and Bradford 2007). In addition, it has been suggested that visible lesbian subcultures have been on the decline, even in urban areas such as San Francisco (Stein 2006), perhaps making the urban gay subculture a predominantly male scene. Thus, for same-sex-attracted youth, living in an urban environment may play more of a protective role in the lives of boys than of girls.

The purpose of the current study is to explore how school characteristics associated with heteronormativity shape the relative well-being of adolescents with same-sex romantic attractions. Specifically, we look at how the prevalence of football and level of religiosity within the school as well as the urban location of the school moderate the impact of same-sex attraction on social-emotional well-being and externalized distress. In addition, given that girls and boys are culturally scripted and expected to do sexuality and gender differently, we consider whether heteronormativity in schools affects boys and girls differently. Previous research suggests that girls and boys experience the formation of sexual identity (Savin-Williams and Diamond 2000) and the stigma attached to same-sex attraction (D'Augelli, Pilkington, and Hershberger 2002) differently. Moreover, different aspects of the school environment, as described above, may matter more or less for same-sex-attracted girls and boys. Though previous studies of sexual minority adolescents' experiences in schools suggest that sexual minority boys may be more affected by heteronormative school contexts (Pascoe 2007), exploring the effects of heteronormativity on same-sex-attracted rather than lesbian, gay, bisexual, and queer (LGBQ) identified youth may result in findings that diverge from previous research.

DATA AND METHOD

To examine how school culture affects the well-being of same-sex-attracted youth, we employed data from the National Longitudinal Study of Adolescent Health (Add Health) and its education component, the Adolescent Health and Academic Achievement Study (AHAA) (Muller et al. 2007). Add Health is a nationally representative, school-based study of seventh to twelfth grade students who were first interviewed during the 1994-95 school year. The Add Health sample of 20,745 students was drawn from a random sample of high schools in the United States that were stratified by region, urbanicity, size, type, racial composition, and grade span (Harris et al. 2003).

Add Health was the first nationally representative sample of youth that included any information on sexual attraction; therefore, it offered the first opportunity to assess the degree and extent of problems faced by sexual minority adolescents nationwide. In addition, the Add Health and AHAA data include information on several domains of adolescents' lives, including socioemotional well-being, experiences with fighting, and academic outcomes. Add Health contains large within-school samples that allow us to measure different aspects of school culture by aggregating individual responses to questions about sports and religion. We limit our sample to high school students (grades 9 to 12) who completed both an in-school and an in-home (wave I) interview, had a valid sampling weight, and were not missing information on same-sex attraction (n = 7,082 girls and 6,861 boys). Descriptive statistics for all individual-level and school-level variables are presented in Table 1.

TABLE 1.

Descriptive Statistics for All Analytic Variables

	Girls (N = 7,082)			Boys (N = 6,861)
Individual-level Variables	M or Proportion	SD	% Missing	M or Proportion	SD
Outcomes
Depression	0.66	0.40		0.55	0.33
Self-esteem	3.93	0.55		4.15	0.52
Fighting	0.21			0.42
Course failure	0.44			0.57
Independent variables
Romantic attraction
Other-sex attraction	0.87			0.83
Same-sex attraction	0.06			0.07
No attraction	0.08			0.10
Race/ethnicity
White	0.65			0.65
Latino/a	0.11			0.12
Black	0.16			0.15
Asian	0.05			0.05
Other	0.03			0.03
Parents' education	2.78	1.12	2.13	2.79	1.14
Age	16.68	1.29		16.84	1.36
Family structure
Single parent	0.23			0.22
Stepparent	0.17			0.17
Relationship with parents	3.42	0.58	2.73	3.50	0.50
Pubertal development
Younger	0.19			0.21
Older	0.41			0.40

School-level Variables (n = 73)	M or Proportion	SD	Min	Max
School size
Small	0.16
Medium	0.37
Large	0.47
School region
West	0.19
Midwest	0.23
South	0.38
Northeast	0.19
Urbanicity
Urban	0.29
Rural	0.19
Suburban	0.52
School religiosity (unstandardized)	2.65	0.46	1.51	3.90
Prevalence of football (unstandardized)	0.39	0.16	0.00	0.78
Proportion sexual minority	0.06	0.03	0.01	0.14

SOURCE: National Longitudinal Study of Adolescent Health.

NOTE: Descriptive statistics for individual-level variables are weighted.

Measures

Romantic attraction

Our measure of same-sex attraction was based on questions from the wave I in-home survey that asked the respondent whether he or she had ever had a romantic attraction to a female and to a male. We considered youth to be same-sex attracted if they ever reported having been attracted to a person of the same sex, regardless of whether they also reported other-sex attractions. Students who reported that they had never been romantically attracted to a male or female were included in analyses as a separate category. Other-sex-attracted adolescents served as a reference group in all models.

Because sexuality is fundamentally in development during adolescence, we believe romantic attraction is a more developmentally appropriate measure of sexual orientation than sexual behaviors, relationships, or identity. In fact, research has found that sexual minority teenagers consider the desire to be romantically involved a more important aspect of sexual orientation than identity or behavior (Savin-Williams 2005). Many adolescents who experience same-sex desires will never identify as gay, lesbian, or bisexual (Savin-Williams 2005); therefore, our romantic attraction measure likely reveals both self-identified LGBQ youths as well as those who do not, and may never, identify as such. A focus on same-sex attraction allows this study to explore the power of heteronormative contexts to constrain the sexual feelings and desires of all youth, not just those who identify as LGBQ. We should be clear, however, that we do not know how many of these youth incorporated these attractions into their sexual identity, made their attractions known to their peers, or displayed any gender nonconformity in their behaviors or dress.

Dependent variables

Depression was assessed using 19 items from the Center for Epidemiological Studies Depression scale that measures both malaise and mood aspects of emotional distress. Responses to the items were averaged to form a single continuous indicator of depressive symptoms (α = .86) that ranges from 0 to 3. Self-esteem is assessed with a single continuous indicator ranging from 1 (strongly disagree) to 5 (strongly agree); higher values represent higher levels of self-esteem (α = .85). It is based on seven in-home survey questions assessing how strongly the respondent agrees or disagrees that he or she has a lot of good qualities, is physically fit, has a lot to be proud of, likes himself or herself the way he or she is, feels he or she is doing everything just about right, feels socially accepted, and feels loved and wanted. Fighting is a dichotomous measure indicating whether or not a student reported being in a physical fight in the past year. This question asks only whether the respondent “got into a physical fight”; we are unable to distinguish whether adolescents were physically victimized or if they initiated the fight (possibly in response to harassment). Academic failure is a dichotomous measure indicating whether or not a student ever failed any type of course in high school and is constructed using data taken from students' high school transcripts.

School characteristics

School religiosity is an aggregation, at the school level, of individual religious attendance. During the in-home interview, students were asked how often they have, in the past year, attended religious services, and responses ranged from never to once a week. Responses were reverse coded so that higher values (1–4) represented more attendance and were then averaged within each school to produce a measure describing the average religious attendance of students within the school. Individuals who reported no religious affiliation were given the lowest value (1) on the religious attendance measure. Prevalence of football was measured as the proportion of male students who participated in school football during the 1994-95 school year (limited to boys, as very few girls participated in football). This measure was based on a question from the in-school survey that asked students to identify school-organized clubs, organizations, and teams they participated in or planned to participate in during the current school year. Finally, urban refers to schools located within urban areas, with schools located in rural or suburban areas serving as the reference group. Information on school locale comes from the 1994-95 school administrator survey.

Control variables

We control factors related to adolescent problem behaviors, including family relationships (Steinberg 2001) and pubertal development (Siegel et al. 1999). Relationship with parents was measured with five questions (coded 0 to 4) that asked respondents about their relationships with each resident parent (mother and father). We constructed separate indicators for mother (α = .84) and father (α = .88) by taking the mean of the five responses and took the higher score. Relative pubertal development indicates whether the respondent described himself or herself as looking younger, older, or about average compared to other students his or her age. We also controlled for students' background characteristics, including racial/ethnic identity, age, parents' highest level of education, and family structure. Finally, we included school-level controls for school size (small, medium, or large) and region (South, West, Midwest, or Northeast), which were taken from the school administrator survey. We used mean imputation to fill in missing values for relationship with parents and parents' education, and in models we included dummy variables indicating values were missing.

Analytical Plan

We employed multilevel modeling using the HLM6 software to examine the impact of school- and individual-level variables on adolescents' social-emotional well-being and problem behaviors. First, we estimate the effect of same-sex attraction on depression, self-esteem, fighting, and academic failure, controlling for individual-level factors (model 1). To examine whether this relationship varies according to the culture of the school, we estimate a series of cross-level interactions between same-sex attraction and school characteristics: school religiosity (model 2), the prevalence of football (model 3), and urban locale (model 4). All models include school-level controls for size and region (although coefficients are not shown in the table), and all analyses are performed separately for boys and girls. In our regression models, dichotomous variables are uncentered and ordinal and continuous variables are grand mean centered (individual values converted into deviations from the overall sample mean). The intercepts in these models can thus be interpreted as the average outcome for the hypothetical individual who is in the reference category for all dummy variables and average on all continuous variables. We include the student-level weight normalized at the individual level in all models.

RESULTS

Overall, our findings support previous literature suggesting that same-sex-attracted youth are at greater risk for lower social-emotional well-being and externalized signs of distress—fighting and failure. We find that this is especially the case for same-sex-attracted girls. We also find that our school culture variables representing heteronormativity, including school religiosity, the prevalence of football, and school locale, help to moderate this relationship. Below we present results from multilevel fixed-effects hierarchical linear models for boys and girls predicting depression, self-esteem, fighting, and failure.

Depression

As shown in model 1 of Table 2, both same-sex-attracted girls and same-sex-attracted boys report higher levels of depression compared to their other-sex-attracted peers, even after controlling for important socio-demographic background characteristics, parental relationships, and pubertal development. For girls, this relationship does not vary by the level of religious attendance within the school (model 2), the prevalence of football in the school (model 3), or school locale (model 4). However, for boys, the association between same-sex attraction and depression does vary by school locale. Results of model 4 suggest that, among boys, the negative effect of same-sex attraction on depression is weaker in urban schools than it is in rural and suburban schools, as shown by the statistically significant and negative cross-level interaction between same-sex attraction and urban school locale. In urban schools, the same-sex attraction effect is weaker (i.e., b = .11 + −.11(1) = 0.0) than it is in suburban and rural schools (i.e., b = .11 + −.11(0) = .11), suggesting that urban schools may protect against depression among same-sex-attracted boys. Thus, same-sex-attracted boys attending schools in urban areas experience less social-emotional distress associated with the stigmatization of same-sex attraction, as seen by their lower rates of depressive symptoms, than those attending schools in suburban or rural areas.

TABLE 2.

Results from Hierarchical Linear Models Predicting Depression for Girls and Boys

	Girls				Boys
	Model 1	Model 2	Model 3	Model 4	Model 1	Model 2	Model 3	Model 4
	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
Romantic attraction
None	.01	.01	.01	.01	.04*	.04*	.04*	.04*
Same sex	.14***	.14***	.14***	.12***	.08***	.08***	.07***	.11***
Cross-level interactions
Same-sex attraction by						.01
School religiosity		−.02
Prevalence of football			−.01				.03
Urban				.10				−.11*
School-level variables
School religiosity		.00				−.01
Prevalence of football			.00				.00
Urban locale				−.01				.02
Individual-level controls
Age	.00	.00	.00	.00	.02***	.02***	.02***	.02***
Race (white)
Black	.08***	.08***	.08***	.08***	.04*	.04*	.04*	.04
Latino	.07*	.06*	.07*	.07*	.06**	.06**	.06**	.06**
Asian	.09**	.09**	.09**	.09**	.08**	.08**	.08**	.08**
Other	.03	.03	.03	.03	.07	.07	.07	.07
Parents' education	−.03***	−.03***	−.03***	−.03***	−.02***	−.02***	−.02***	−.02***
Family structure (intact)
Single parent	.03	.03	.03	.03	.02	.02	.02	.02
Stepparent	.04	.04	.04	.04	.04**	.04**	.04**	.04**
Relationship with parents	−.19***	−.19***	−.19***	−.19***	−.17***	−.17***	−.17***	−.17***
Pubertal development (average)
Younger	.09***	.09***	.09***	.09***	.06***	.06***	.06***	.06***
Older	.07***	.07***	.07***	.07***	.01	.01	.01	.01
Intercept	.55***	.59***	.59***	.59***	.49***	.49***	.49***	.49***
N	7,082				6,860

SOURCE: National Longitudinal Study of Adolescent Health.

NOTE: Models 2, 3, and 4 control for school size and region.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

Self-esteem

Table 3 presents similar results for self-esteem. For girls, the main effect of same-sex attraction in all models is negative and statistically significant, suggesting that same-sex-attracted girls have lower self-esteem in adolescence than their other-sex-attracted counterparts. However, this relationship does not vary by school religiosity (model 2), prevalence of football (model 3), or locale (model 4).

TABLE 3.

Results from Hierarchical Linear Models Predicting Self-esteem for Girls and Boys

	Girls				Boys
	Model 1	Model 2	Model 3	Model 4	Model 1	Model 2	Model 3	Model 4
	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
Romantic attraction
None	.02	.02	.02	.02	.00	.00	.00	.00
Same sex	−.12**	−.12**	−.12**	−.12*	−.03	−.03	−.01	−.03
Cross-level interactions
Same-sex attraction by						−.01
School religiosity		.01
Prevalence of football			.01				−.11**
Urban				.00				.00
School-level variables
School religiosity		.01				.01
Prevalence of football			.02				.02
Urban locale				.06*				−.01
Individual-level controls
Age	.01	.01	.01	.01	−.01	−.01	−.01	−.01
Race (white)
Black	.09**	.09**	.09**	.09**	.12***	.12***	.12***	.13***
Latino	−.03	−.03	−.02	−.03	−.04	−.03	−.03	−.03
Asian	−.10*	−.09*	−.09*	−.10*	−.11**	−.10**	−.10**	−.10**
Other	.00	.00	.00	.00	−.09	−.09	−.08	−.09
Parents' education	.02**	.02**	.02**	.02**	.03***	.03***	.03***	.03***
Family structure (intact)
Single parent	.03	.03	.03	.03	.04	.04	.04	.04
Stepparent	−.01	−.01	−.01	−.01	−.04	−.04	−.03	−.04
Relationship with parents	.39***	.39***	.39***	.39***	.39***	.39***	.39***	.39***
Pubertal development (average)
Younger	−.01	−.01	−.01	−.01	−.03	−.03	−.03	−.03
Older	−.02	−.02	−.02	−.02	.05***	.05***	.05***	.05***
Intercept	3.94***	3.86***	3.85***	3.84***	4.13***	4.10***	4.10***	4.10***
N	7,080				6,860

SOURCE: National Longitudinal Study of Adolescent Health.

NOTE: Models 2, 3, and 4 control for school size and region.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

For boys, we see something different. As shown in model 1, there is not a statistically significant effect of same-sex attraction on self-esteem when variation in school culture is not considered. However, there is evidence that, for boys, the relationship between same-sex attraction and self-esteem is moderated by the prevalence of football within the school (model 3). The cross-level interaction between prevalence of football and same sex-attraction is negative and statistically significant, suggesting that the visibility of a hypermasculine sport such as football in schools may decrease levels of self-esteem for same-sex-attracted boys within these schools. Taken together, these results suggest that while the negative relationship between same-sex attraction and social-emotional well-being is, on average, stronger for girls than for boys, this relationship is more dependent on a football culture and urban locale for boys than it is for girls.

Fighting

Models predicting involvement in fights are shown in Table 4. For girls, we do see a main effect of same-sex attraction on fighting, such that, as shown in model 1, same-sex-attracted girls are 65 percent (odds ratio [OR] = exp(.50) = 1.65) more likely than other-sex-attracted girls to have been in a physical fight in the past year. In addition, we see that this effect varies by the average religiosity of the school, as shown by the significant and positive cross-level interaction between same-sex attraction and school religiosity (model 2). In schools with average religious attendance one standard deviation above the mean, same-sex-attracted girls are more likely to have been in a fight than are those in schools with average religious attendance one standard deviation below the mean (b = .55 + .38(1.0) = .93 vs. b = .55 + .38(−1.0) = .17).

TABLE 4.

Results from Hierarchical Linear Models Predicting Fighting for Girls and Boys

	Girls				Boys
	Model 1	Model 2	Model 3	Model 4	Model 1	Model 2	Model 3	Model 4
	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
Romantic attraction
None	−.19	−.19	−.19	−.19	−.61***	−.61***	−.61***	−.61***
Same sex	.50**	.55**	.49**	.43*	−.10	−.09	−.09	.09
Cross-level interactions
Same-sex attraction by						.10
School religiosity		.38*
Prevalence of football			−.01				−.06
Urban				.26				−.75*
School-level variables
School religiosity		−.09				−.12*
Prevalence of football			.00				.03
Urban locale				.03				.24*
Individual-level controls
Age	−.10***	−.10***	−.10***	−.10***	−.06*	−.06*	−.06*	−.06*
Race (white)
Black	.65***	.70***	.69***	.68***	.38***	.39***	.37***	.37***
Latino	.34*	.32*	.32	.31	.13	.13	.13	.11
Asian	.05	.02	.03	.03	−.13	−.12	−.13	−.13
Other	.12	.10	.10	.10	.23	.20	.21	.20
Parents' education	−.16***	−.17***	−.17***	−.17***	−.12***	−.11***	−.12***	−.12***
Family structure (intact)
Single parent	.39***	.38***	.38***	.39***	.18*	.18*	.18*	.18*
Stepparent	.29*	.29*	.29*	.29*	.13	.13	.13	.13
Relationship with parents	−.40***	−.40***	−.40***	−.40***	−.20**	−.20**	−.20**	−.20**
Pubertal development (average)
Younger	.21	.21	.22	.22	−.10	−.10	−.10	−.10
Older	.41***	.40***	.41***	.41***	.19*	.19*	.19*	.20*
Intercept	−1.98***	−2.03***	−2.01***	−2.00***	−.48***	−.62***	−.56***	−.61***
N	7,072				6,844

SOURCE: National Longitudinal Study of Adolescent Health.

NOTE: Models 2, 3, and 4 control for school size and region.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

For boys, experiencing same-sex attraction, on average, does not seem to increase the odds of getting into a fight. However, while the main effect of same-sex attraction on fighting is not statistically significant, the cross-level interaction between urban school locale and same-sex attraction is negative and statistically significant (model 4); this suggests that same-sex-attracted boys may be less likely to get into physical fights in urban schools than they are in rural or suburban schools. In fact, in urban schools, same-sex-attracted boys are less likely than their other-sex-attracted peers to have been in a fight (b = .09 + −.75(1) = −.66).

Academic Failure

Finally, results for analysis of course failure are presented in Table 5. For girls, we see that the effect of same-sex attraction on the likelihood of failure varies by school religiosity (model 2) and the prevalence of football (model 3). Same-sex-attracted girls are more likely to fail a course in high school when they are located within a more religious context (b = .44 + .70(1.0) = 1.14 vs. b = .44 + .70(−1.0) = −.26) and in schools with a larger football culture (b = .31 + .38(1.0) = .69 vs. b = .31 + .38(−1.0) = −.07). For girls, this finding is consistent with the finding for fighting, which suggests an important association between school culture and externalizing behaviors for same-sex-attracted girls in high schools.

TABLE 5.

Results from Hierarchical Linear Models Predicting Failure for Girls and Boys

	Girls				Boys
	Model 1	Model 2	Model 3	Model 4	Model 1	Model 2	Model 3	Model 4
	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
Romantic attraction
None	.26	.27	.26	.26	.20	.20	.20	.20
Same sex	.33	.44*	.31	.17	.47**	.47**	.50**	.67**
Cross-level interactions
Same-sex attraction by						−.02
School religiosity		.70*
Prevalence of football			.38*				−.20
Urban				.74				−.73*
School-level variables
School religiosity		−.10				−.16
Prevalence of football			−.08				−.11
Urban locale				.40				.36
Individual-level controls
Age	−.01	−.01	−.01	−.01	.00	.00	.00	.00
Race (white)
Black	.65***	.65***	.65***	.65***	.53*	.53*	.53*	.52*
Latino	.50**	.46**	.46**	.46**	.18	.15	.16	.14
Asian	−.61	−.62	−.62	−.62	−.25	−.27	−.27	−.28
Other	.39	.39	.38	.40	.74	.74	.77	.74
Parents' education	−.36***	−.36***	−.36***	−.36***	−.41***	−.41***	−.41***	−.41***
Family structure (intact)
Single parent	.58***	.58***	.57***	.58***	.31*	.30*	.31*	.30*
Stepparent	.46***	.46***	.46***	.46***	.50***	.49***	.50***	.49***
Relationship with parents	−.15	−.14	−.15	−.15	−.24*	−.24*	−.24*	−.24*
Pubertal development (average)
Younger	−.06	−.07	−.06	−.06	.01	.01	.01	.01
Older	.22*	.22*	.22*	.23*	.12	.12	.12	.12
Intercept	−.74***	−1.33***	−1.30***	−1.33***	.08	−.27	−.20	−.23
N	4,419				4,092

SOURCE: National Longitudinal Study of Adolescent Health.

NOTE: Models 2, 3, and 4 control for school size and region.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

Same-sex-attracted boys are more likely than other-sex-attracted boys to fail at least one course in high school. Specifically, as seen in model 1, same-sex-attracted boys are 60 percent (OR = exp(.47) = 1.60) more likely to fail a course in high school than are other-sex-attracted boys. As with findings for depression and fighting, the effect of same-sex attraction for boys appears to vary by school locale (model 4). In urban schools, the same-sex attraction effect is much weaker (i.e., b = .67 + −.73(1) = −.06) than it is in suburban and rural schools (i.e., b = .67 + −.73(0) = .67).

DISCUSSION

Our study demonstrates that school culture shapes the well-being of same-sex-attracted youth. While previous research has demonstrated that sexual minority youth are at greater risk for experiencing depressive symptoms (D'Augelli 2002), low self-esteem (D'Augelli 1998), physical conflict with peers (D'Augelli, Pilkington, and Hershberger 2002), and poor academic performance (Pearson, Muller, and Wilkinson 2007), our study suggests that these risks may depend in part on the level of heteronormativity found within the schools they attend. Specifically, the visibility of football and religion as well as the location of the school in an urban, suburban, or rural community influence the degree to which same-sex-attracted youth experience decreased well-being. Moreover, these different aspects of school culture appear to influence same-sex-attracted boys and girls differently: While the prevalence of football within schools affects the well-being of both girls and boys with same-sex attractions, school religiosity appears to play a greater negative moderating role for same-sex-attracted girls, while school locale, with urban settings exhibiting more tolerance, plays a greater role for same-sex-attracted boys.

The visibility of football within the school appears to negatively affect both boys and girls to some extent: Compared to their peers with only other-sex attractions, same-sex-attracted boys report lower levels of self-esteem and same-sex-attracted girls are more likely to fail a course when attending schools where more boys are involved in school-sponsored football. Research suggests that the heteronormative practices and discourses employed by adolescents who participate in football may reinforce hierarchies of masculinity and sexuality and stigmatize non-normative sexual desires. In addition, school sponsorship may have the power to legitimate patterns of discourse and behavior that create feelings of difference. Our findings provide some evidence that there is greater stigma attached to nonheterosexual feelings and behaviors in schools where football is more visible, for both boys and girls. While we expected that the prevalence of football in schools might negatively affect the well-being of same-sex-attracted boys more than same-sex-attracted girls because of football's emphasis on the performance of hegemonic masculinity, our results do not support this. It is possible that if we had focused on LGBQ identity or gender nonconformity rather than same-sex attraction, our findings might have been stronger for boys, who might be more targeted by peers in these contexts. However, the symbolic rituals associated with high school football reinforce compulsory heterosexuality for both boys and girls, so same-sex-attracted girls and boys may be equally affected by the visibility of football in school.

Findings regarding the impact of religiosity within the school suggest that religion has a unique impact on the well-being of same-sex-attracted girls. While same-sex-attracted girls are more likely than other-sex-attracted girls to be involved in fights and to fail a course in schools with more religious student bodies, same-sex-attracted boys are not affected by the religiosity of the schools they attend. Perhaps this is because of the greater policing of women's sexual desires and behaviors compared to men in many religious institutions. While same-sex relationships of both men and women have been targeted by conservative religious organizations, women and girls have often been the focus of discourses of sexual morality (Heath 2009; Weeks 1985). Coupled with the greater importance that girls place on religion (Miller and Hoffman 1995; Rostosky et al. 2004), it is not surprising that the visibility of religion in school may play a greater role in the well-being of same-sex-attracted girls. The apparent impact of levels of religiosity among students within schools suggests a powerful influence of institutionalized religion and its ability to pervade secular institutions. This is particularly important given that institutionalized religion has been, today and in the past, one of the largest voices against same-sex relationships. Its impact at the school level suggests how ubiquitous and potentially harmful these cultural messages can be for young people.

For the last school culture measure we investigated—urban locale—we again found differences between boys and girls. For boys, but not for girls, attending a school in an urban versus suburban or rural area appears to buffer the negative effect of same-sex attraction on depression, fighting, and failing a course in high school. Individuals from rural areas often report more negative feelings toward homosexuality and higher levels of heteronormativity, and the findings of this study reinforce previous findings that sexual diversity is often less tolerated and more stigmatized in nonurban contexts and the schools within them (Stein 2001). In addition, the mere visibility of alternative forms of sexuality often found in urban areas may lessen the stigma and associated feelings of isolation and difference experienced by adolescents with same-sex attractions (Flowers and Buston 2001). Same-sex-attracted youth in nonurban areas may feel excluded not only from the heterosexual culture within their own immediate community but also from an inaccessible gay subculture that exists in urban areas (Flowers and Buston 2001). The greater visibility of this gay male culture in many urban areas, relative to the decline of visible lesbian subcultures (Stein 2006), may help explain the finding of a moderating effect of urban location for boys but not for girls.

It is important to acknowledge the implications of our use of same-sex attraction as a measure of sexual minority as well as how this measure fits into our larger theoretical framework of understanding how heteronormativity within schools shapes the development of youth. Much of the research on heteronormativity within schools deals with the verbal and physical harassment aimed at gender transgressions or nonheterosexual identities (Savin-Williams 2005). By using same-sex attraction as a measure, we are tapping a sense of difference and stigma that is more closely tied to sexuality. While gender and sexuality are tightly intersecting statuses, recent advances in the sociology of sexualities and queer theory call for a separate examination of sexuality, one that acknowledges sexuality as an organizing principle of power that affects everyone and that avoids the pitfalls of defining sexuality “almost exclusively in relation to a discrete orientation or identity category” (Stein 2008, 116). More work in the sociology of sexuality needs to address, more broadly, how and in what contexts varieties of adolescent sexual attractions and desires become stigmatizing.

Previous research has suggested how heteronormativity is produced within schools and some potential interventions to reduce it (Chesir-Teran 2003). Much of this work has focused on specific programs designed to meet the needs of LGBQ youth, such as gay–straight alliances and safe school policies (Kosciw, Diaz, and Greytak 2007). While important, such interventions may ignore the subtle yet powerful maintenance of heteronormativity within schools that occurs through the reproduction of schemas and daily practices organized around heterosexual relationships, which are further legitimated by the school and by institutions embedded within the school, such as sports and religion. We suggest thinking beyond interventions specifically aimed at LGBQ-identified students that are based on assumptions about identities and that ignore a larger population of youth who may feel stigmatized by dominant forms of sexuality.

Qualitative comparisons of schools may be needed to fully understand why school locale and the prevalence of football and religion are associated with the marginalization of students with same-sex attraction and to uncover other mechanisms that reproduce heteronormativity within schools. More work should be done to explore the relational contexts of other spaces within schools, such as drama and fine arts programs, to determine the dominant discourses and practices that pervade these spaces and their possible role in reducing the stigma of same-sex attraction. By exploring how the local school context shapes the well-being of same-sex-attracted adolescents, we can reveal ways schools can help protect all youth from negative outcomes related to heteronormative cultures. Moreover, we may be able to inhibit the reproduction of inequalities, including those related to gender and sexuality, which often occurs through the construction and reconstruction of heteronormative culture through everyday interactions and the legitimization of these interactions within schools.

Biographies

Lindsey Wilkinson is an assistant professor of sociology at Portland State University and has a primary interest in how schools shape the well-being of adolescents. Her previous research has addressed the effect of English as a second language placement on the social and academic outcomes of language-minority youth.

Jennifer Pearson is an assistant professor of sociology at Wichita State University. Her primary research interests are in the areas of gender, sexuality, and education. Her current work explores how school context and school success shape young women's sense of sexual empowerment.