Key PointsQuestion
What is the association between gender nonconformity (ie, gender expression that differs from societal expectations for feminine or masculine appearance and behavior) and indicators of mental distress and substance use among adolescents?
Findings
In this cross-sectional study of 6082 high school students, gender nonconformity was associated with feeling sad and hopeless, as well as suicidal thoughts and/or behaviors, among female and male students. In addition, gender nonconformity was strongly associated with substance use among male students.
Meaning
These findings underscore the importance of implementing school-based programs to prevent substance use and promote mental health that are inclusive of gender-diverse students.
Importance
The cultural roles and expectations attributed to individuals based on their sex often shape health behaviors and outcomes. Gender nonconformity (GNC) (ie, gender expression that differs from societal expectations for feminine or masculine appearance and behavior) is an underresearched area of adolescent health that is often linked to negative health outcomes.
Objective
To examine the associations of GNC with mental distress and substance use among high school students.
Design, Setting, and Participants
Cross-sectional study based on data from the Youth Risk Behavior Survey (YRBS) conducted in 2015. The setting was 3 large urban US school districts (2 in California and 1 in Florida). Participants were a racially/ethnically diverse population-based sample of 6082 high school students representative of all public school students in grades 9 through 12 attending these 3 school districts.
Main Outcomes and Measures
Sex-stratified adjusted prevalence ratios (APRs) (adjusted for race/ethnicity, grade, and sexual identity) for high gender-nonconforming students (very/mostly/somewhat feminine male students or very/mostly/somewhat masculine female students) and moderate gender-nonconforming students (equally feminine and masculine students) relative to a referent group of low gender-nonconforming students (very/mostly/somewhat masculine male students or very/mostly/somewhat feminine female students).
Results
Among 6082 high school students, 881 (15.9%) were white, 891 (19.1%) black, 3163 (55.1%) Hispanic, and 1008 (9.9%) other race/ethnicity. Among female students (2919 [50.0% of the study population]), moderate GNC was significantly associated with feeling sad and hopeless (APR, 1.22; 95% CI, 1.05-1.41), seriously considering attempting suicide (APR, 1.41; 95% CI, 1.14-1.74), and making a suicide plan (APR, 1.52; 95% CI, 1.22-1.89); however, substance use was not associated with GNC. Among male students (3139 [50.0% of the study population]), moderate GNC was associated with feeling sad and hopeless (APR, 1.55; 95% CI, 1.25-1.92); high GNC was associated with seriously considering attempting suicide (APR, 1.72; 95% CI, 1.16-2.56), making a suicide plan (APR, 1.79; 95% CI, 1.17-2.73), and attempting suicide (APR, 2.78; 95% CI, 1.75-4.40), as well as nonmedical use of prescription drugs (APR, 1.81; 95% CI, 1.23-2.67), cocaine use (APR, 2.84; 95% CI, 1.80-4.47), methamphetamine use (APR, 4.52; 95% CI, 2.68-7.61), heroin use (APR, 4.59; 95% CI, 2.48-8.47), and injection drug use (APR, 8.05; 95% CI, 4.41-14.70).
Conclusions and Relevance
This study suggests mental distress is associated with GNC among female and male students. Substance use also appeared to be strongly associated with GNC among male students. These findings underscore and suggest the importance of implementing school-based programs to prevent substance use and promote student mental health that are inclusive of gender diversity in students.
Gender (ie, the cultural roles and expectations attributed to women and men based on their sex) often shapes health behaviors and outcomes. For example, research supports that masculinity in men predicts heavy alcohol use1 and violence perpetration,2 and adherence to feminine norms by women is linked to less use of condoms3 and increased depression.4 One aspect of gender increasingly linked to negative health outcomes is gender nonconformity (GNC). Gender nonconformity describes when an individual’s gender expression does not align with societal expectations of their gender (Table 1).5,6 Gender-nonconforming individuals often experience increased social stress resulting from exposure to prejudice, discrimination, harassment, and violence.5-8 This appears to be especially true for adolescents; gender-nonconforming youths experience elevated rates of harassment and violence by peers,9,10 as well as increased exposure to childhood abuse.11
Stigma, discrimination, harassment, and other social stressors have well-documented effects on health.12,13 Research on racial/ethnic and sexuality-related health disparities points to the role of social and minority stress (or the repeated exposure to discrimination and harassment experienced by marginalized groups) as being particularly detrimental to mental health.14,15 Discrimination has been linked to depression among black16,17 and Latino18,19 populations, as well as suicidality among sexual minority youths.20 Exposure to bullying, harassment, and other social stressors also has been linked with elevated rates of substance use among both sexual minority and sexual nonminority youths.21 In addition, gender minority youths report increased substance use and minority stress experiences.22 The fact that gender-nonconforming adolescents experience discrimination, harassment, and other forms of social stress at rates above and beyond those of gender-conforming youths8,9,11 suggests that they may be at risk for poor mental health and substance use.
Evidence suggests that GNC may be linked to mental health and substance use disparities. Gender-nonconforming, sexual minority adults demonstrate elevated rates of psychological distress and anxiety compared with sexual minority adults who are gender conforming.8 Gender nonconformity in sexual minority women has been associated with increased exposure to violence and minority stress, which in turn are associated with mental health problems and elevated rates of substance use, including increased use of alcohol, tobacco, and other drugs.23 A few studies have evaluated the association between retrospective reports of childhood or adolescent GNC and mental health among young adults. Childhood GNC is associated with depressive symptoms in adolescence and young adulthood,24 and adolescent GNC appears to be associated with both more depression and less life satisfaction during young adulthood.25 Among adolescents, evidence suggests a link between GNC, homophobic bullying, and mental distress; however, gender-nonconforming boys experience more bullying and thus related mental distress than adolescent girls.26 Another study27 found tobacco use to be not only associated with high levels of self-rated masculinity in young men but also associated with childhood GNC, indicating that the association between substance use and GNC may not be linear.
Little is known about the associations between GNC, mental health, and substance use concurrently during adolescence. While it is clear that many lesbian, gay, bisexual, and transgender (LGBT) youths experience increased mental distress and substance use compared with heterosexual and cisgender youths, more information is needed about the associations of GNC with mental health and substance use among the general population of adolescents independent of sexual orientation. This information would enable practitioners to better understand the utility of considering the diversity of adolescent gender expression when designing mental health and substance use prevention programs for high school students. To address this gap, data from 3 racially/ethnically diverse population-based samples of US high school students from 3 large urban school districts were analyzed to describe the spectrum of gender expression, as well as the strengths of association between nonconforming gender expression (ie, GNC) and mental distress and substance use among students.
As part of the Centers for Disease Control and Prevention’s Youth Risk Behavior Surveillance System, approximately 20 large urban school districts conduct a Youth Risk Behavior Survey (YRBS) biennially using a 2-stage cluster sample design to produce a representative sample of high school students in grades 9 through 12 within each jurisdiction. In 2015, three large urban US school districts included an optional question on their YRBS questionnaire assessing gender expression among students. Data from these 3 school districts (2 in California and 1 in Florida) were combined into a single data set (6082 students). Questionnaires were administered in the classroom during a regular class period. Responses were recorded directly on computer-scannable answer sheets. Student participation in the survey was anonymous and voluntary, and local procedures were used to obtain parental consent. Each of these school districts reviewed and approved the YRBS using their local procedures. Overall response rates ranged between 70% and 90%, and sample sizes ranged between 1000 and 3000 students. The YRBS has been reviewed and approved by an institutional review board at the Centers for Disease Control and Prevention.
Using a validated measure,28,29 gender expression was assessed with the following statement and question: “A person’s appearance, style, dress, or the way they walk or talk may affect how people describe them. How do you think people at school would describe you?” Response options were “very feminine,” “mostly feminine,” “somewhat feminine,” “equally feminine and masculine,” “somewhat masculine,” “mostly masculine,” and “very masculine.” Based on a student’s response to this and to the question “What is your sex?” (response options were “female” or “male”), a 7-point GNC scale was created. Students were categorized from most gender conforming (1, indicating very feminine female students and very masculine male students) to most gender nonconforming (7, indicating very masculine female students and very feminine male students) (eTable in the Supplement). Because there were small numbers of somewhat, mostly, or very masculine female students, the following 3-level GNC variable also was created: (1) high GNC (somewhat, mostly, and very masculine female students and somewhat, mostly, and very feminine male students), (2) moderate GNC (equally feminine and masculine female students and male students), and (3) low GNC (somewhat, mostly, and very feminine female students and somewhat, mostly, and very masculine male students). Sexual identity was assessed with the question “Which of the following best describes you?” Response options were “heterosexual (straight),” “lesbian or gay,” “bisexual,” and “not sure.” Demographic characteristics assessed included sex (female or male), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, or other), and grade (9, 10, 11, or 12). The YRBS also measures a wide range of substance use behaviors, from tobacco and alcohol use to heroin and injection drug use (IDU), as well as various indicators of mental distress, from feeling sad and hopeless to suicidal thoughts and/or attempts (Table 2). All questions used in this analysis were asked at all 3 sites with one exception: 1 of the 3 sites did not ask the question “During the past 12 months, did you make a plan about how you would attempt suicide?”
In 2017, data were analyzed using statistical software (SUDAAN, version 10.0.1; RTI International) to account for the cluster sample design. Prevalence estimates with 95% CIs were calculated using Taylor series linearization. Prevalence estimates were not provided in tables when there were fewer than 30 observations in the denominator. Differences in the distribution of GNC by demographic characteristics and sexual identity were tested using χ2 statistics. Because gender expression varies by demographic characteristics and sexual orientation and because associations between GNC and other health risk behaviors tend to vary by sex, all associations of GNC with mental distress and substance use were stratified by sex and adjusted for race/ethnicity, grade, and sexual identity. In initial analyses of association, GNC was treated as a continuous variable ranging from 1 (most gender conforming) to 7 (most gender nonconforming); logistic regression models were used to examine both linear and nonlinear (quadratic) associations between GNC and mental distress and substance use. In subsequent categorical analyses, logistic regression models were used to calculate sex-stratified adjusted prevalence ratios (APRs) with 95% CIs that compared the prevalence of mental distress indicators and substance use among high gender-nonconforming students and moderate gender-nonconforming students relative to a referent group of low gender-nonconforming (ie, most gender conforming) students. Differences in the prevalence of mental distress and substance use between moderate and high gender-nonconforming students were tested by linear contrasts using t tests. Statistical tests were considered to be significant if 2-sided P < .05 or if the 95% CI did not include 1.0.
Among demographic characteristics, the weighted percentage (unweighted number) of students in the combined sample was 50.0% (95% CI, 47.9%-52.1%) (n = 2919) female and 50.0% (95% CI, 47.9%-52.1%) (n = 3139) male. Races/ethnicities were 15.9% (95% CI, 13.5%-18.7%) (n = 881) white non-Hispanic, 19.1% (95% CI, 16.6%-21.9%) (n = 891) black non-Hispanic, 55.1% (95% CI, 51.2%-58.8%) (n = 3163) Hispanic (of any race), and 9.9% (95% CI, 8.4%-11.7%) (n = 1008) other race/ethnicity. Grade levels were 28.5% (95% CI, 22.9%-34.9%) (n = 1581) in 9th grade, 25.0% (95% CI, 19.3%-31.7%) (n = 1662) in 10th grade, 23.5% (95% CI, 18.4%-29.6%) (n = 1438) in 11th grade, and 22.9% (95% CI, 17.8%-28.9%) (n = 1322) in 12th grade. For sexual identity, 87.6% (95% CI, 86.4%-88.7%) (n = 5114) of students identified as heterosexual (straight), 2.5% (95% CI, 2.0%-3.2%) (n = 151) identified as lesbian or gay, 5.6% (95% CI, 4.8%-6.6%) (n = 350) identified as bisexual, and 4.3% (95% CI, 3.7%-5.0%) (n = 264) responded “not sure” to the sexual identity question.
Gender nonconformity varied by sex, race/ethnicity, grade, and sexual identity (Table 3). High GNC was more prevalent among male students (13.0%) than female students (4.0%); was more prevalent among lesbian or gay (41.7%), bisexual (15.7%), and “not sure” (14.3%) students than heterosexual students (6.4%); and decreased as grade level increased.
In initial analyses, GNC was treated as a continuous variable, and indicators of mental distress were strongly associated with GNC (Table 4). Among female and male students, feeling sad and hopeless demonstrated both a linear and quadratic association with GNC. While feelings of sadness and hopelessness increased with GNC, this increase was nonlinear, with the greatest prevalence of sadness and hopelessness occurring near the middle of the GNC scale. Suicidal thoughts also were associated with GNC. Among female and male students, both seriously considering attempting suicide and making a suicide plan demonstrated a linear increase with increasing GNC. Among male students (but not female students) suicide attempts also demonstrated a linear increase with GNC.
Substance use was associated with GNC among male students (but not female students) (Table 4). Among male students, the prevalence of nonmedical use of prescription drugs and the use of cocaine, methamphetamine, heroin, and injection drug use all demonstrated linear increases with increasing GNC. Among female students, no category of substance use was associated with GNC.
In categorical analyses, among female students, feelings of sadness and hopelessness (APR, 1.22; 95% CI, 1.05-1.41), seriously considering attempting suicide (APR, 1.41; 95% CI, 1.14-1.74), and making a suicide plan (APR, 1.52; 95% CI, 1.22-1.89) were more likely among students with moderate GNC compared with those with low GNC (Table 5). Among male students, feelings of sadness and hopelessness also were more likely among students with moderate (APR, 1.55; 95% CI, 1.25-1.92) rather than low GNC; however, seriously considering attempting suicide (APR, 1.72; 95% CI, 1.16-2.56), making a suicide plan (APR, 1.79; 95% CI, 1.17-2.73), and attempting suicide (APR, 2.78; 95% CI, 1.75-4.40) were more likely among students with high GNC than low GNC.
In categorical analyses, among female students, alcohol was the only substance for which use was associated with GNC; compared with students with moderate GNC, students with high GNC were more likely to drink alcohol (Table 5). Among male students, nonmedical use of prescription drugs (APR, 1.81; 95% CI, 1.23-2.67), cocaine use (APR, 2.84; 95% CI, 1.80-4.47), methamphetamine use (APR, 4.52; 95% CI, 2.68-7.61), heroin use (APR, 4.59; 95% CI, 2.48-8.47), and IDU (APR, 8.05; 95% CI, 4.41-14.70) were more likely among students with high GNC compared with students with low GNC; IDU also was more likely among male students with moderate GNC compared with those with low GNC. Marijuana use (APR, 0.67; 95% CI, 0.48-0.95) was less likely among male students with moderate GNC than low GNC.
Within this study population, 1 in 5 students reported either moderate (11.9%) or high (8.4%) levels of GNC. Consistent with previous research,9,22,25,26 the prevalence of GNC varied by demographic categories, including sex, race/ethnicity, grade, and sexual identity. In particular, a greater prevalence of high GNC was reported by lesbian or gay students (41.7%) and bisexual students (15.7%) compared with heterosexual students (6.4%). To our knowledge, this study is the first to examine associations of GNC with mental distress and substance use independent of associations with sexual identity among a racially/ethnically diverse population-based sample of high school students in 3 large urban US school districts.
As previous research26 would suggest, while there were some similarities in the associations between GNC and mental distress among female and male students, there were differences as well. Feelings of sadness and hopelessness demonstrated a nonlinear increase with GNC among both female and male students. In categorical analyses, the greatest prevalence of feeling sad and hopeless was found among moderately gender-nonconforming female and male students. These findings may underscore unique challenges faced by students in the middle of the gender conformity spectrum. Although there is scant research on this group, youths experiencing moderate GNC may still experience significant distress related to experiences of discrimination and stigmatization.30 In contrast, associations between GNC and suicidal thoughts and attempts varied by sex. Among female students, suicidal thoughts and plans (but not attempts) demonstrated a linear increase with GNC; in categorical analyses, the prevalence of suicidal thoughts and plans appeared to peak among moderately gender-nonconforming female students. Among male students, suicidal thoughts, plans, and attempts all demonstrated a linear increase with GNC; in categorical analyses, the greatest prevalence occurred among male students expressing high levels of GNC. These patterns may echo the sex differences in suicide reporting and suicidal behaviors observed in adolescents more broadly. In Western contexts, reported rates of suicide attempts are typically higher in girls and women, while suicide completions are typically higher in boys and men.31,32 Female students reporting high GNC may be less (or no more) likely to report suicidal thoughts than more gender-conforming female students in part because of the association between masculinity and a reticence to report mental health symptoms. Conversely, male students reporting greater GNC may be more likely to disclose suicidal thoughts and attempts than male students who conform more strongly to conventional masculinity norms in part because they are not inhibited by masculine conventions that equate sharing mental distress with weakness.
Associations between GNC and substance use also varied by sex. Among female students, the prevalence of alcohol use was greater among high gender-nonconforming compared with moderate gender-nonconforming students; other categories of substance use did not vary by GNC. However, among male students, the prevalence of nonmedical use of prescription drugs, cocaine use, methamphetamine use, heroin use, and IDU each demonstrated a linear increase with GNC, with the greatest prevalence estimates found among high gender-nonconforming male students. A recent study21 of US high school students found that controlling for the increased social stress experienced by sexual minority youths reduced and in some cases eliminated the disparities in substance use by sexual orientation. This finding suggests that at least some of the increased substance use found among sexual minority youths may occur as a coping mechanism in response to experiences of social and minority stress.21 If social stress has a similar role with respect to GNC, our findings are consistent with previous research suggesting that GNC may be met with more overt harassment of gender-nonconforming male youths compared with gender-nonconforming female youths.26,33,34 Additional research is needed to better understand the links between gender norms, social stressors, and substance use behaviors across the spectrum of gender conformity. Also, given that transgender youths often have risk factors and experiences that overlap heavily with those of cisgender, gender-nonconforming youths, it is important to investigate whether associations of mental distress and substance use with self-reported transgender identity are similar to our findings with gender-nonconforming youths. Unfortunately, data on gender identity were not available in our study, and transgender students could not be identified. Finally, it is unknown whether our results would vary among a nationally representative population that would include both rural and suburban youths, as well as fewer racial/ethnic minority youths.
Our study has some limitations. First, school-based YRBS data apply only to youths who attend school, and sexual minority and gender minority youths may be disproportionately represented among high school dropouts and other youths who are absent from or do not attend school.35 Second, it is possible that students who are willing to report a stigmatized gender expression (ie, being gender nonconforming) may also be more willing to report stigmatized health behaviors and mental health symptoms, which could inflate observed estimates of the association between GNC and outcomes studied herein. Third, the extent of underreporting or overreporting of self-reported behaviors cannot be determined; however, the YRBS questionnaire items generally demonstrate good test-retest reliability.36 Fourth, the data are cross-sectional and provide only an indication of association, not causality. Furthermore, given the developmental context surrounding gender, our cross-sectional data do not allow us to examine how changes over time in either the individual’s or greater culture’s ideas about gender may affect the outcomes of interest. Fifth, small sample sizes at the most nonconforming end of the spectrum weakened our ability to examine associations across the full range of gender expression among female students.
Developing support systems within schools for gender-nonconforming students may be an important avenue to improving mental health and reducing substance use in this population. Providing safe spaces and school staff contacts who are knowledgeable and supportive of gender-nonconforming youths may also help to buffer stresses of the school environment. School staff may benefit from professional development on gender diversity to support students who are gender nonconforming more competently.37 Health education that is inclusive of discussions about gender and the variety of ways that it is expressed may be useful to decrease stigma for gender-nonconforming youths.38,39 In addition, given that verbal and physical harassment exacerbate both mental distress and substance use, instructing staff and students in bystander interventions may aid in improving the climate for gender-nonconforming youths. Based on previous research,40,41 such interventions may have ripple effects in schools among sexual and gender minority youths: when students witness teachers and other students intervening on behalf of students being bullied, they in turn are more likely to intervene.
Accepted for Publication: May 30, 2018.
Corresponding Author: Richard Lowry, MD, MS, Division of Adolescent and School Health, National Center for HIV/AIDS, Viral Hepatitis, Sexually Transmitted Disease (STD), and Tuberculosis (TB) Prevention, Centers for Disease Control and Prevention, 1600 Clifton Rd NE, Mail Stop E-75, Atlanta, GA 30329 (rlowry@cdc.gov).
Published Online: September 24, 2018. doi:10.1001/jamapediatrics.2018.2140
Author Contributions: Dr Lowry had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Lowry, Johns, Robin.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Lowry, Kann.
Administrative, technical, or material support: Gordon, Austin, Robin, Kann.
Supervision: Austin, Robin, Kann.
Conflict of Interest Disclosures: Dr Austin reported being supported by training grants T71-MC-00009 and T76-MC-00001 from the Maternal and Child Health Bureau, Health Resources and Services Administration, US Department of Health and Human Services. No other disclosures were reported.
Disclaimer: The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
1.Wells
S, Flynn
A, Tremblay
PF, Dumas
T, Miller
P, Graham
K. Linking masculinity to negative drinking consequences: the mediating roles of heavy episodic drinking and alcohol expectancies.
J Stud Alcohol Drugs. 2014;75(3):510-519. doi:
10.15288/jsad.2014.75.510PubMedGoogle ScholarCrossref 2.Santana
MC, Raj
A, Decker
MR, La Marche
A, Silverman
JG. Masculine gender roles associated with increased sexual risk and intimate partner violence perpetration among young adult men.
J Urban Health. 2006;83(4):575-585. doi:
10.1007/s11524-006-9061-6PubMedGoogle ScholarCrossref 6.Institute of Medicine. The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: National Academies Press; 2011.
8.Puckett
JA, Maroney
MR, Levitt
HM, Horne
SG. Relations between gender expression, minority stress, and mental health in cisgender sexual minority women and men.
Psychol Sex Orientat Gend Divers. 2016;3(4):489-498. doi:
10.1037/sgd0000201Google ScholarCrossref 9.Gordon
AR, Conron
KJ, Calzo
JP, Reisner SL, Bryn Austin
S. Nonconforming gender expression is a predictor of bullying and violence victimization among high school students in four U.S. school districts.
J Adolesc Health. 2016;58(2)(suppl):S1-S2. doi:
10.1016/j.jadohealth.2015.10.019Google Scholar 14.Meyer
IH, Frost
DM. Minority stress and the health of sexual minorities. In: Patterson
CJ, D’Augelli
AR, eds. Handbook of Psychology and Sexual Orientation. New York, NY: Oxford University Press; 2013:252-266.
17.Schulz
AJ, Gravlee
CC, Williams
DR, Israel
BA, Mentz
G, Rowe
Z. Discrimination, symptoms of depression, and self-rated health among African American women in Detroit: results from a longitudinal analysis.
Am J Public Health. 2006;96(7):1265-1270. doi:
10.2105/AJPH.2005.064543PubMedGoogle ScholarCrossref 24.Roberts
AL, Rosario
M, Slopen
N, Calzo
JP, Austin
SB. Childhood gender nonconformity, bullying victimization, and depressive symptoms across adolescence and early adulthood: an 11-year longitudinal study.
J Am Acad Child Adolesc Psychiatry. 2013;52(2):143-152. doi:
10.1016/j.jaac.2012.11.006PubMedGoogle ScholarCrossref 25.Toomey
RB, Ryan
C, Diaz
RM, Card
NA, Russell
ST. Gender-nonconforming lesbian, gay, bisexual, and transgender youth: school victimization and young adult psychosocial adjustment.
Dev Psychol. 2010;46(6):1580-1589. doi:
10.1037/a0020705PubMedGoogle ScholarCrossref 29.Greytak
EA, Gill
AM, Conron
KJ. Identifying transgender and other gender minority respondents on population-based surveys: special considerations for adolescents. In: Herman
JL, ed. Best Practices for Asking Questions to Identify Transgender and Other Gender Minority Respondents on Population-Based Surveys. Los Angeles, CA: Williams Institute; 2014:29-34.
31.Canetto
SS, Sakinofsky
I. The gender paradox in suicide.
Suicide Life Threat Behav. 1998;28(1):1-23.
PubMedGoogle Scholar 34.Sirin
SR, McCreary
DR, Mahalik
JR. Differential reactions to men and women’s gender role transgressions: perceptions of social status, sexual orientation, and value dissimilarity.
J Men’s Stud. 2004;12(2):119-132. doi:
10.3149/jms.1202.119Google ScholarCrossref