Abstract
Purpose.
To examine the association between gender expression, peer victimization (PV), and disordered weight control behaviors (DWCB) in four population-based samples of U.S. high school students.
Methods.
Analyses include data from 5,488 U.S. high school students from the 2013 Youth Risk Behavior Surveys in four jurisdictions (Broward County, FL; Chicago, IL; Los Angeles, CA; San Diego, CA). Participants were 56% Hispanic/Latino, 21% Black/African American, and 14% White. Two items asked about perceived gender expression; responses were classified into three groups: highly gender conforming (e.g., very masculine boys), moderately gender conforming, gender nonconforming (e.g., feminine boys). Sex-stratified multivariable logistic regression models were used to examine the association between gender expression, PV, and DWCB in previous month (fasting, using diet pills/liquids/powders, and purging [vomiting or using laxatives]), controlling for potential confounders.
Results.
Overall, 12% of respondents reported fasting, 6% reported diet pill use, and 5% reported purging, with significantly higher prevalence among gender nonconforming compared to gender conforming males (p<.001). In adjusted models, gender nonconforming males had greater odds of fasting (odds ratio [95% confidence interval]: 3.0 [2.0-4.7]), diet pill use (6.1 [3.7-9.9]), and purging (7.2 [3.6-14.8]), relative to moderately conforming males. No significant associations were found among females. Adding PV to models modestly attenuated the association between gender nonconformity and DWCB for males.
Conclusions.
In probability samples of U.S. high school students, we observed marked differences by gender expression in DWCB among males but not females. Gender expression-related stigma should be addressed within clinical and school-based interventions to prevent DWCB.
Keywords: gender expression, gender nonconformity, eating disorders, weight control, bullying, peer victimization
Disordered weight control behaviors (DWCB), such as skipping meals, using diet pills without a prescription, self-induced vomiting, and taking laxatives to lose or maintain weight, remain a significant public health concern among adolescents. Such behaviors typically occur prior to the onset of clinical eating disorders [1] and have been estimated to be the third most common cause of chronic illness in children and adolescents, following obesity and asthma [2,3]. Health outcomes associated with DWCB may be severe and include adverse physiological (e.g., esophagitis, gastric rupture, impairment of digestive functioning) and mental health (e.g., depression) effects [2,4].
A substantial body of research has documented inequities in the prevalence of DWCB among youth by sex, race/ethnicity, sexual orientation, and housing instability [5–7]. Specifically, female youth, youth of color, and sexual minority youth are more likely to engage in these behaviors than male youth, non-Hispanic white youth, and heterosexual youth, respectively. However, more recent research also points to substantial variation in risk within those groups. In particular, emerging research suggests that gender expression, or the degree to which an individual is perceived as conforming or not conforming to societal expectations for masculine or feminine appearance and behavior [8], may be another risk factor for engaging in DWCB [9,10].
One possible mechanism for the potential association between gender expression and DWCB is that individuals perceived to have a nonconforming gender expression are at heightened risk of experiencing peer bullying and peer violence victimization (hereafter, peer victimization [PV]) [11,12], including in school and online environments (“cyberbullying”) [13,14]. The minority stress framework [15] posits that exposure to stigma such as PV may elevate risk of engaging in DWCB as a coping response. DWCB may also be linked to gender conformity based on pressures to meet dominant cultural physical appearance ideals, such as those equating femininity with thinness among women and masculinity with lean muscularity among men [16,17]. These cultural appearance ideals have been shown to differentially shape behaviors and outcomes by sex, with a higher prevalence of disordered eating behaviors among girls/women in the U.S. compared to boys/men [2,18]; this has been linked to the disproportionate demand placed on girls/women to meet cultural femininity ideals of body shape and size [19]. Thus, we might expect both minority stress processes and cultural appearance ideal pressures to differentially impact the association between gender expression and DWCB for female compared to male youth. For example, it may be that highly gender conforming female youth (who may be more likely to internalize a “feminine thin ideal”) and nonconforming female youth (who may be more likely to experience minority stressors such as PV) are both at elevated risk of DWCB, while among male youth only those who are nonconforming are at elevated risk compared to their more gender conforming peers. However, no studies have examined perceived gender expression as a risk factor for DWCB in adolescents and how this association may vary by sex and experiences of PV.
This study seeks to address these gaps by analyzing associations between gender expression, sex, PV, and DWCB among youth using data from high school students participating in the 2013 U.S. Youth Risk Behavior Surveillance System (YRBSS) survey in four school districts. Our primary hypothesis is that youth who report a more nonconforming gender expression will have higher odds of engaging in DWCB than youth who are highly or moderately gender conforming. Our secondary hypothesis is the association between gender expression and DWCB will vary by sex; we hypothesize that greater nonconformity will be associated with increased odds of DWCB among boys; whereas among girls, we hypothesize that gender nonconforming and highly gender conforming will have greater odds of DWCB than their moderately conforming peers. Our third hypothesis, informed by the minority stress framework [11,15] as described above, is that experiences of PV will attenuate the association between gender expression and DWCB.
METHODS
Study population
This study draws on the 2013 YRBSS conducted by the U.S. Centers for Disease Control and Prevention. The YRBSS uses a two-stage cluster sample design to produce a representative sample of 9th through 12th grade students in the U.S. within each included jurisdiction [20], in compliance with local procedures on parental consent/notification. The Boston Children’s Hospital Institutional Review Board designated this as not human subjects research because all analyses used secondary, de-identified data.
The present study used YRBSS data from the four jurisdictions that elected to add a novel measure of perceived gender expression (described below): Broward County, FL; Chicago, IL; Los Angeles, CA; and San Diego, CA (n=6,000). Data from 2013 were used as this was the first year that gender expression was offered as an optional measure that jurisdictions could choose to include in their local YRBSS survey and the most recent year DWCB were systematically collected. Due to small sample size, participants who were 12 years of age were excluded. Individuals missing responses on key variables were excluded from the analysis (gender expression, 4.9%; sex, <1%; key covariates, 3.1%). In addition, participants missing information on each outcome were excluded from outcome-specific analyses (range: 9.4%-10.2% missing). For each outcome-specific analysis, participants excluded due to missing data were significantly more likely to be male (p<.0001), younger age (p=.04), Black or African American (p<.0001), and sexual minority (p<.0001), and were significantly less likely to be in San Diego compared to the other three jurisdictions (p<.0001). Final unweighted analytic samples ranged from n=5,389 to n=5,488.
Measures
Primary outcome
Disordered weight control behaviors (DWCB).
Participants were asked if they engaged in three types of behaviors to lose or to keep from gaining weight in the previous 30 days (coded yes/no): going “without eating for 24 hours or more (also called fasting),” taking “any diet pills, liquids, or powders without a doctor’s advice,” and vomiting or taking laxatives.
Primary exposure
Gender expression.
Perceived gender expression was assessed with a self-report question based on a two-item measure developed for adolescents [8]. This item is designed to elicit self-reported perceptions of others, due to central role that such perceptions play in processes of stigmatization and discrimination [21]. Participants were asked: “A person’s appearance, style, dress, or the way they walk or talk may affect how people describe them. How do you think other people at school would describe you?” with response options ranging on a seven-point scale from “very feminine” to “very masculine.” Responses were recoded based on participant sex (range: 1-7) and classified in three groups, similar to previous research [12,22]: (a) highly gender conforming (score=1, corresponding to “very masculine” for boys and “very feminine” for girls); (b) moderately gender conforming (score=2-4); and (c) gender nonconforming (score=5+).
Hypothesized mediator
Peer Victimization.
Two indicator variables were constructed to assess distinct forms of PV: bullying victimization and violence victimization. Peer bullying victimization was measured using two YRBSS items as a report of having been bullied on school property or electronically (i.e., cyber-bullying) in the previous year (0=no, 1=yes). Peer violence victimization was a composite of three YRBSS items (0=no to all, 1=yes to any): having been threatened or injured with a weapon on school property in past year; having been injured in a physical fight and needing treatment in past year; and skipping school due to feeling unsafe in the past month.
Sociodemographics
For self-reported sex, participants were asked “What is your sex?” (male or female). The 2013 YRBS did not include a question about gender identity; therefore, transgender students cannot be identified in this analysis. Age was assessed via self-report in years. Race/ethnicity was assessed using two survey questions: “Are you Hispanic or Latino?” (yes or no) and “What is your race?” (select all that apply). Responses were classified as Hispanic/Latino, Black/African American, White, Asian/Pacific Islander, and another race or multiracial. Sexual orientation identity was assessed with the question, “Which of the following best describes you?” with four response options (heterosexual (straight), gay or lesbian, bisexual, not sure).
Statistical Analysis
Descriptive analyses and hypothesis testing were carried out with SAS software v.9.4 procedures designed to account for complex sampling design and survey weights [23]. Chi-square tests were used to examine sex-stratified bivariate associations between DWCB, PV and gender expression in three categories (highly gender conforming, moderately gender conforming, gender nonconforming) with α-level=.05. Multivariable logistic regression models were used to estimate the odds of each of the three outcome variables (fasting, diet pill use, and purging) associated with gender expression group (reference = moderately gender conforming youth), adjusted for geographic region, age, racial/ethnic identity, and sexual orientation identity (Model 1) . We tested for effect modification by sex by examining models with sex-by-gender expression interaction terms. Statistically significant interaction terms were found for diet pill use (p=.001) and purging behaviors (p=.008) but not for fasting. Based on these findings in conjunction with the literature documenting differences in prevalence and psychopathology of disordered eating behaviors for girls vs. boys [18,24] all multivariable models were sex-stratified. To examine the potential role of exposure to PV, we ran separate models adding the two PV covariates (Model 2). Due to prior research showing high levels of PV among sexual minority youth in schools, we tested for collinearity between sexual orientation and the two PV indicators; we found no evidence of collinearity (VIF<3 and significant independent associations between sexual orientation and PV).
RESULTS
Results are from 5,488 participants (53.4% female; average age 16 years (SD=0.1)) across four school districts. Over half (55.8%) of the sample was Hispanic/Latino, followed by Black/African American (17.4%), White (13.5%), and Asian/Pacific Islander/Native Hawaiian (10.2%). Approximately one-third (33.2%) of the sample was classified as highly gender conforming, the majority (59.1%) as moderately gender conforming, and 7.7% were classified as gender nonconforming. Table 1 displays sociodemographic characteristics and weighted prevalence of PV and DWCB by sex. Figure 1 displays distribution of gender expression by sex and sexual orientation identity.
Table 1.
Sociodemographic characteristics and key outcomes among high school students in four U.S. school districts
| Total | Female | Male | |
|---|---|---|---|
|
| |||
| No. | 5,488 | 2,929 | 2,559 |
| Age in years, M (SE) | 16.0 (0.1) | 16.0 (0.1) | 16.0 (.01) |
| Race/Ethnicity, No. (%) | |||
| White | 813 (13.5) | 407 (13.9) | 406 (15.9) |
| Black/African American | 1084 (21.1) | 640 (21.9) | 444 (17.4) |
| Hispanic/Latino | 2800 (55.8) | 1491 (50.9) | 1309 (51.2) |
| Asian/Pacific Islander/Native Hawaiian | 506 (6.4) | 246 (8.4) | 260 (10.2) |
| Additional races | 285 (3.2) | 145 (5.0) | 140 (5.5) |
| Sexual Orientation Identity, No. (%) | |||
| Heterosexual | 4867 (88.7) | 2515 (85.9) | 2352 (91.9) |
| Gay/Lesbian | 107 (1.9) | 51 (1.7) | 56 (2.2) |
| Bisexual | 304 (5.5) | 241 (8.2) | 63 (2.5) |
| Not Sure | 210 (3.8) | 122 (4.2) | 88 (3.4) |
| Region, No. (%) | |||
| Broward Co., FL | 1314 (23.9) | 722 (24.7) | 592 (23.1) |
| Chicago, IL | 1383 (25.2) | 809 (27.6) | 574 (22.4) |
| Los Angeles, CA | 1479(26.9) | 759 (25.9) | 720 (28.1) |
| San Diego, CA | 1312 (23.9) | 639 (21.8) | 673 (26.3) |
| Gender Expressiona, No. (%) | |||
| Highly gender conforming | 1822 (33.2) | 1038 (35.4) | 784 (30.6) |
| Moderately gender conforming | 3242 (59.1) | 1782 (60.8) | 1460 (57.1) |
| Gender nonconforming | 424 (7.7) | 109 (3.7) | 315 (12.3) |
| Peer Victimization, No. (%) | |||
| Bullying victimization in past yearb (yes) | 1003 (18.3) | 612 (21.0) | 391 (15.3) |
| Peer violence victimizationc (yes) | 728 (13.3) | 388 (13.2) | 340 (13.3) |
| Disordered Weight Control Behaviorsd, No. (%) | |||
| Fasted at least 24 hours (yes) | 640 (11.8) | 437 (15.1) | 203 (8.1) |
| Used diet pills, liquids, powders (yes) | 310 (5.7) | 197 (6.8) | 113 (4.5) |
| Vomited or used laxatives (yes) | 282 (5.2) | 195 (6.7) | 87 (3.5) |
Note: All proportions weighted.
Gender expression (GE) groups: Highly gender conforming: GE score=1; Moderately gender conforming: GE score=2-4; Gender nonconforming: GE score=5-7.
Reported experiencing any bullying victimization on school property or electronic bullying victimization in past year.
Reported one or more of three indicators of school-based peer violence victimization: injured in a physical fight and needed treatment 1+ times in past year; threatened or injured with a weapon 1+ times in past year; or missed school due to feeling unsafe 1+ days in past 30 days.
Reported engaging in behavior to lose or maintain weight in past 30 days.
Figure 1.
Distribution of gender expression by sex and sexual orientation identity (U.S. YRBSS 2013, n=5,488)
Table 2 presents bivariate associations between gender expression, PV, and DWCB. Gender nonconforming male and female students had twice the prevalence of peer violence victimization compared to highly gender conforming or moderately gender conforming students (25.7% vs. 12.9% and 12.7%, respectively among females, p=0.03; 27.3% vs. 10.8% and 12.2%, respectively among males, p<.0001). Gender nonconforming male students had a significantly higher prevalence of peer bullying victimization and of engaging in each disordered weight control behavior over the past month than moderately gender conforming and highly gender conforming male students; patterns were similar but not statistically significant among female students.
Table 2.
Prevalence of peer victimization and disordered weight control behaviors by gender expression among high school students in four U.S. school districts
| Gender Expression Group |
||||
|---|---|---|---|---|
| Highly Gender Conforming | Moderately Gender Conforming | Gender Nonconforming | p-value | |
| Female, No. | 1,038 | 1,782 | 109 | |
| Peer Victimization, % (n) | ||||
| Bullying victimization | 17.9 (185) | 22.2 (395) | 29.4 (32) | 0.31 |
| Peer violence victimization | 12.9 (134) | 12.7 (226) | 25.7 (28) | 0.03 |
| Disordered Weight Control Behaviors, % (n) | ||||
| Fasted at least 24 hours | 14.8 (151) | 15.5 (274) | 11.0 (12) | 0.78 |
| Used diet pills, liquids, or powders | 6.3 (64) | 7.0 (124) | 8.3 (9) | 0.61 |
| Vomited or used laxatives | 6.3 (64) | 6.9 (121) | 9.2 (10) | 0.30 |
|
| ||||
| Male, No. | 784 | 1,460 | 315 | |
| Peer Victimization, % (n) | ||||
| Bullying victimization | 9.5 (74) | 16.0 (233) | 26.9 (84) | <0.001 |
| Peer violence victimization | 12.2 (96) | 10.8 (158) | 27.3 (86) | <.0001 |
| Disordered Weight Control Behaviors, % (n) | ||||
| Fasted at least 24 hours | 5.7 (44) | 6.5 (92) | 22.0 (67) | <.0001 |
| Used diet pills, liquids, or powders | 3.2 (25) | 2.6 (38) | 16.7 (50) | <.0001 |
| Vomited or used laxatives | 2.1 (16) | 1.8 (26) | 15.2 (45) | 0.001 |
Note: All proportions weighted. Bivariate associations tested for statistical significance with Chi-square tests; differences with p<.05 considered statistically significant.
Table 3 shows results from multivariable logistic regression models estimating associations between gender expression, PV, and DWCB. Among female students, there were no significant associations between gender expression and DWCB. Among male students, associations were significant in adjusted models for all three DWCB. Relative to moderately gender conforming male students, gender nonconforming males had significantly higher odds of fasting (OR=3.01 [1.97, 4.60]), diet pill use (OR=6.07 [3.70, 9.94]), and purging behaviors (OR=7.23 [3.55, 14.76]), controlling for potential confounders (Table 3, Model 1). Effect estimates also indicated potential elevated risk of purging behaviors among highly gender conforming male students, relative to moderately conforming male students, although confidence intervals covered the null. Table 3, Model 2 shows results when peer bullying victimization and peer violence victimization were added to the models. Modest attenuation of the effect estimates among male students was observed, particularly in the model including peer violence victimization.
Table 3.
Multivariable logistic regression estimates of association between gender expressiona and disordered weight control behaviorsb in four U.S. school districts
| Female (n=2,915) | Male (n=2,521) | |||||||
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| Model 1c | Model 2d | Model 1c | Model 2d | |||||
|
| ||||||||
| OR | (95% CI) | OR | (95% CI) | OR | (95% CI) | OR | (95% CI) | |
| Fasted at least 24 hours | ||||||||
| Highly gender conforming (Ref=Mod. conf.) | 1.13 | (0.85, 1.50) | 1.13 | (0.84, 1.51) | 0.94 | (0.58, 1.50) | 0.96 | (0.59, 1.55) |
| Gender nonconforming (Ref=Mod. conf.) | 0.68 | (0.32, 1.44) | 0.61 | (0.27, 1.38) | 3.07 | (2.00, 4.69) | 2.62 | (1.63, 4.20) |
| Bullying victimization (Ref=None) | 2.02 | (1.48, 2.75) | 1.46 | (0.94, 2.28) | ||||
| Peer violence victimization (Ref=None) | 1.99 | (1.45, 2.74) | 2.91 | (1.96, 4.32) | ||||
| Used diet pills, liquids, powders | ||||||||
| Highly gender conforming (Ref=Mod. conf.) | 1.05 | (0.72, 1.53) | 1.03 | (0.72, 1.49) | 1.03 | (0.49, 2.16) | 1.08 | (0.52, 2.26) |
| Gender nonconforming (Ref=Mod. conf.) | 1.15 | (0.50, 2.65) | 1.00 | (0.40, 2.50) | 6.07 | (3.70, 9.94) | 5.02 | (3.06, 8.23) |
| Bullying victimization (Ref=None) | 3.19 | (2.10, 4.85) | 1.73 | (0.85, 3.50) | ||||
| Peer violence victimization (Ref=None) | 1.83 | (1.32, 2.53) | 2.67 | (1.64, 4.34) | ||||
| Vomited or used laxatives | ||||||||
| Highly gender conforming (Ref=Mod. conf.) | 1.24 | (0.85, 1.80) | 1.25 | (0.85, 1.84) | 1.37 | (0.64, 2.93) | 1.41 | (0.63, 3.19) |
| Gender nonconforming (Ref=Mod. conf.) | 1.63 | (0.74, 3.58) | 1.51 | (0.62, 3.68) | 7.23 | (3.55, 14.76) | 6.89 | (3.29, 14.44) |
| Bullying victimization (Ref=None) | 2.49 | (1.73, 3.58) | 2.13 | (1.11, 4.11) | ||||
| Peer violence victimization (Ref=None) | 1.78 | (1.17, 2.72) | 1.16 | (0.57, 2.36) | ||||
OR=Odds Ratio; 95% CI=95% Confidence Interval; bold=statistically significant, α=.05. All models weighted and adjusted for cluster sampling design.
Gender expression (GE) groups: Highly gender conforming: GE score=1; Moderately gender conforming: GE score=2-4; Gender nonconforming: GE score=5-7.
Three disordered weight control behaviors were collected: fasting at least 24 hours, using diet pills, liquids, or powders without a doctor’s advice, or using laxatives or vomiting to lose or maintain weight in past 30 days (yes/no).
Model 1 adjusted for geographic region, age, racial/ethnic identity, and sexual orientation identity.
Model 2 adjusted for Model 1 covariates, plus bullying victimization (online or in-person) and peer violence victimization.
DISCUSSION
This study is among the first to examine associations between perceived gender expression and DWCB and to assess for variations in this association by sex and PV among a large population-based sample of adolescents. Study findings demonstrate that gender expression may be a risk indicator for DWCB, particularly among adolescent boys who are perceived to violate societal gender norms (boys who are perceived as more feminine).
Consistent with prior research and national surveillance estimates [25], prevalence of DWCB ranged from 5-12% and prevalence estimates were higher among female than male youth in our sample. Based on the most recently available high school enrollment data in these jurisdictions, ranging from 29,000 to 154,000 students enrolled [26–29], YRBSS estimates suggest there could be as many as 42,000 youth fasting, 20,000 using diet pills, and 18,500 using laxatives or vomiting to control weight each month in these four districts combined. Extrapolating to current U.S. Census estimates of youth ages 15-19 enrolled in school [30], there may be as many as 2.2 million youth engaging in at least one of these behaviors each month in the U.S.
Results from the current study support existing literature demonstrating increased risk of DWCB among gender nonconforming youth [9,10] and among youth who have experienced PV [31,32], and underscore that these elevated risks are independent of sexual orientation identity. Previous research on gender expression and DWCB has largely focused on sexual minority populations, which have higher likelihood of being perceived as gender nonconforming and also have higher risk of DWCB than heterosexual populations [6,33]. Studies of lesbian and bisexual women have identified gender nonconformity as a protective factor against body dissatisfaction (a risk factor for DWCB) via reduced thin-ideal internalization [34]; in contrast, findings on gender conformity have been mixed, with some studies identifying greater femininity as a risk factor for body dissatisfaction and DWCB and others finding no association [35]. Our findings extend this prior research to a population-based sample of youth and demonstrate that gender nonconformity is associated with DWCB among heterosexual and sexual minority youth.
We found support for our secondary hypothesis that the association between gender expression and DWCB varies by sex. The most gender nonconforming males had 3-7 times greater odds of DWCB than their moderately gender conforming peers. These striking disparities offer support for prior scholarship suggesting that greater alignment with femininity norms among men is associated with disordered eating behaviors [17,36]. In addition, minority stress pathways [15] may play a role in these disparities. Minority stressors, such as experiences of PV targeting those perceived to violate societal gender norms [11,14], may trigger behavioral stress responses, including DWCB. At the other end of the gender expression spectrum, we observed elevated though not statistically significant odds of purging among the most gender conforming relative to moderately conforming boys. Although earlier research suggested that conformity to masculinity norms was protective against restrictive eating behaviors for men [36], more recent research has called for greater attention to the pressure to conform to a “muscular ideal” as a driver of muscularity-oriented DWCB for boys and men [17,36]. Our findings suggest that with greater power we might have detected statistically significant elevated risk of purging for highly gender conforming (i.e., very masculine) boys. Future public health surveillance efforts should include questions about muscularity-oriented behaviors among youth, such as use of muscle-building supplements, anabolic steroids, and compulsive exercise.
In contrast to our findings among male youth and contrary to our hypothesis, our findings among female youth were inconclusive. For fasting and purging behaviors the odds ratios were of a moderate magnitude but no statistically significant associations were observed between gender expression and DWCB among females. This finding could be due to low statistical power given relatively small sample sizes, but it is also in line with prior research by VanKim et al. [37] who reported significant differences by gender expression in eating behaviors like skipping breakfast among male but not among female young people. However, our findings were contrary to our hypothesis that both gender conforming and nonconforming girls would have increased odds of DWCB. Previous research demonstrates that gender nonconformity is a risk indicator for stress-related behaviors among girls and boys [38] and that greater conformity to femininity norms is a risk factor for negative body image and DWCB among young women [39].
There are a few possible explanations for this null association. The Tripartite Influence Model posits that sociocultural appearance pressures from family, peers, and media disproportionately affect girls and women, driving body dissatisfaction and increasing risk of DWCB [19]. In our sample, female youth had significantly higher prevalence of each DWCB than male youth—and prevalence was similar across each level of gender expression, unlike for male youth. This suggests sociocultural appearance pressures may impact girls regardless of their gender expression (i.e., for feminine, androgynous, and masculine girls alike). A second potential explanation is that there may be contrasting gender expression-related pathways for subgroups of female youth, generating what appears to be a null effect. That is, sociocultural appearance pressures may be heightening DWCB risk for gender conforming girls, while exposure to minority stressors (e.g., PV) may be heightening DWCB risk for gender nonconforming girls [31]. Indeed, in our analysis, although DWCB were similar across gender expression groups, the most gender nonconforming female youth reported significantly higher levels of PV. This raises questions for future research about how best to tease apart gender-related pathways that drive DWCB risk. These findings, in concert with existing evidence that youth of color, sexual and gender minority youth, and other youth with marginalized social statuses are disproportionately burdened by DWCB [5–7], also point to the need for more intersectional research on DWCB among youth. In addition to investigating gender-related pathways, future research should explore potential interaction effects of gender expression with race/ethnicity, socioeconomic position, and other critical dimensions of social inequality.
We found partial support for our third hypothesis, that PV would attenuate the association between gender expression and DWCB. Peer bullying and peer violence victimization modestly attenuated the observed associations among boys but all associations remained significant and of high magnitude, suggesting PV may play some role but other explanations are needed. Future research should use instruments that can assess exposure to gender expression-related PV with greater nuance and should capture victimization by other types of perpetrators (school staff, family members) and in a range of settings [11]. Additional mechanisms by which perceived gender expression may affect DWCB should be explored, including exposure to gender-related appearance ideals via peer, family, conventional media, and social media.
Limitations
This study has several limitations to note. First, data were collected via self-report and may be subject to recall bias or social desirability bias in reporting. Second, like all cross-sectional surveys, these analyses cannot identify direction of association. Although for some people nonconforming gender expression can emerge early in childhood [40] and thus is likely to precede development of DWCB, we cannot exclude the possibility of the reverse direction: that participating in DWCB may affect how youth report their perceived gender expression. Third, the 2013 YRBSS did not ask about gender identity and thus we could not examine the experiences of transgender youth. Future research should extend this work to youth of all gender identities. Fourth, missing data were not evenly distributed across the participants, which could have introduced bias; however, overall item-level missingness was low (ranging from <1% to 5% on any single item). Finally, these data are drawn from probability samples in four geographic regions with distinct ethnic, economic, and sociopolitical characteristics that could impact perceived gender expression, PV in school settings, and youth’s DWCB; therefore, these findings may not be generalizable to other geographic contexts.
Conclusion
These findings offer novel insights into the role of perceived gender expression as a potential indicator for health risk behaviors among adolescents. From a clinical perspective, findings highlight the need for clinicians to be aware of these risks when obtaining history from adolescents and underscore the importance of the confidential social history as part of routine adolescent health care [41]. Results from this study also highlight the importance of asking adolescents about the type, frequency, and severity of weight control behaviors they may engage in and using positive framing around body image and weight during health maintenance visits [2]. Clinicians should consider exploring the adolescent’s self-perceived gender expression when assessing risk factors for PV and DWCB.
This study also has implications for public health researchers and educators. Our findings reinforce the importance of collecting public health surveillance data on DWCB, which are prevalent and place youth at risk of eating disorders and lifelong adverse health outcomes [4]. Unfortunately, since 2013 the YRBSS has not collected these data; we are hopeful that in the future monitoring of these important adolescent health behaviors will resume. Further, the elevated risk of PV among gender nonconforming male and female students suggests that school-based anti-bullying programs and adolescent health equity initiatives should address the role of societal gender norms and foster school environments that celebrate gender diversity.
IMPLICATIONS AND CONTRIBUTION.
This study’s findings highlight the need for: (a) clinicians to consider exploring the adolescent’s self-perceived gender expression when discussing possible risk factors for bullying/victimization and disordered weight control behaviors; and (b) school-based anti-bullying programs to include initiatives to foster school environments that celebrate gender diversity.
ACKNOWLEDGMENTS
The authors thank the 2013 Youth Risk Behavior Surveillance System participants and the Centers for Disease Control and Prevention’s Division of Adolescent and School Health for the data used in this study. S.B. Austin is supported by the Maternal and Child Health Bureau, Health Resources Services Administration, U.S. Department of Health & Human Services (T71-MC00009; T76-MC00001). M.L. Wang is supported by NIH (1R01DK120713). An earlier version of the analyses described in this manuscript was presented at the International Conference on Eating Disorders in May 2016 in San Francisco, CA. The authors have no financial or other conflicts of interest to disclose. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, Department of Health and Human Services, or the Centers for Disease Control and Prevention. We thank Rose Eiduson, MPH, of Boston Children’s Hospital, for her assistance with editing this manuscript. We thank all the students in the 2013 Youth Risk Behavior Survey.
Abbreviations
- DWCB
disordered weight control behaviors
- LGBT
lesbian, gay, bisexual, transgender
- PV
peer victimization
- YRBSS
Youth Risk Behavior Surveillance System
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- [1].Levinson JA, Sarda V, Sonneville K, et al. Diet Pill and Laxative Use for Weight Control and Subsequent Incident Eating Disorder in US Young Women: 2001–2016. Am J Public Health 2019;110:109–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Golden NH, Schneider M, Wood C, et al. Preventing Obesity and Eating Disorders in Adolescents. Pediatrics 2016;138. [DOI] [PubMed] [Google Scholar]
- [3].Fisher M, Golden NH, Katzman DK, et al. Eating disorders in adolescents: A background paper. J Adolesc Health 1995;16:420–437. [DOI] [PubMed] [Google Scholar]
- [4].Herzog W, Deter HC, Fiehn W, et al. Medical findings and predictors of long-term physical outcome in anorexia nervosa: a prospective, 12-year follow-up study. Psychol Med 1997;27:269–79. [DOI] [PubMed] [Google Scholar]
- [5].Austin SB, Spadano-Gasbarro J, Greaney ML, et al. Disordered Weight Control Behaviors in Early Adolescent Boys and Girls of Color: An Under-Recognized Factor in the Epidemic of Childhood Overweight. J Adolesc Health 2011;48:109–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Austin SB, Nelson LA, Birkett MA, et al. Eating disorder symptoms and obesity at the intersections of gender, ethnicity, and sexual orientation in US high school students. Am J Public Health 2013;103:e16–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Fournier ME, Austin SB, Samples CL, et al. A comparison of weight-related behaviors among high school students who are homeless and non-homeless. J Sch Health 2009;79:466–73. [DOI] [PubMed] [Google Scholar]
- [8].Wylie S, Corliss HL, Boulanger V, et al. Socially Assigned Gender Nonconformity: A brief measure for use in surveillance and investigation of health disparities. Sex Roles 2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Lampis J, Cataudella S, Busonera A, et al. The moderating effect of gender role on the relationships between gender and attitudes about body and eating in a sample of Italian adolescents. Eat Weight Disord EWD 2019;24:3–11. [DOI] [PubMed] [Google Scholar]
- [10].Calzo JP, Sonneville KR, Scherer EA, et al. Gender conformity and use of laxatives and muscle-building products in adolescents and young adults. Pediatrics 2016;138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Gordon AR, Meyer IH. Gender nonconformity as a target of prejudice, discrimination, and violence against LGB individuals. J LGBT Health Res 2007;3:55–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Roberts AL, Rosario M, Slopen N, et al. 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;52:143–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Kosciw JG, Greytak EA, Zongrone AD, et al. The 2017 National School Climate Survey: The Experiences of Lesbian, Gay, Bisexual and Transgender Youth in Our Nation’s Schools. New York: GLSEN; 2018. [Google Scholar]
- [14].Gordon AR, Conron KJ, Calzo JP, et al. Gender Expression, Violence, and Bullying Victimization: Findings From Probability Samples of High School Students in 4 US School Districts. J Sch Health 2018;88:306–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Meyer IH. Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: conceptual issues and research evidence. Psychol Bull 2003;129:674–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Rodgers RF, McLean SA, Paxton SJ. Longitudinal relationships among internalization of the media ideal, peer social comparison, and body dissatisfaction: implications for the tripartite influence model. Dev Psychol 2015;51:706–13. [DOI] [PubMed] [Google Scholar]
- [17].Griffiths S, Murray SB, Touyz S. Extending the masculinity hypothesis: An investigation of gender role conformity, body dissatisfaction, and disordered eating in young heterosexual men. Psychol Men Masculinity 2015;16:108–114. [Google Scholar]
- [18].Croll J, Neumark-Sztainer D, Story M, et al. Prevalence and risk and protective factors related to disordered eating behaviors among adolescents: relationship to gender and ethnicity. J Adolesc Health Off Publ Soc Adolesc Med 2002;31:166–75. [DOI] [PubMed] [Google Scholar]
- [19].Thompson JK. Exacting beauty : theory, assessment, and treatment of body image disturbance. 1st ed. Washington, DC: American Psychological Association; 1999. [Google Scholar]
- [20].Centers for Disease Control and Prevention. Methodology of the Youth Risk Behavior Surveillance System—2013. MMWR 2013;62:1–25. [Google Scholar]
- [21].Goffman E Stigma; notes on the management of spoiled identity. Englewood Cliffs, NJ, Prentice-Hall 1963; 1963. [Google Scholar]
- [22].Gordon AR, Krieger N, Okechukwu CA, et al. Decrements in health-related quality of life associated with gender nonconformity among U.S. adolescents and young adults. Qual Life Res Int J Qual Life Asp Treat Care Rehabil 2017;26:2129–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].SAS Institute Inc. SAS Software, Version 9.4 of the SAS System for Windows. Cary, NC: SAS Institute Inc.; 2013. [Google Scholar]
- [24].Kinasz K, Accurso EC, Kass AE, et al. Does Sex Matter in the Clinical Presentation of Eating Disorders in Youth? J Adolesc Health 2016;58:410–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Kann L, Kinchen S, Shanklin S, et al. Youth Risk Behavior Surveillance--United States, 2013. MMWR. Centers for Disease Control and Prevention; 2014. [PubMed] [Google Scholar]
- [26].California Department of Education. County Summary - Los Angeles (Census Day 2018-2019). Available at: http://www.ed-data.org/county/Los-Angeles. AccessedApril 18, 2020.
- [27].San Diego Unified School District. About San Diego Unified School District. Available at: https://www.sandiegounified.org/about-san-diego-unified. AccessedApril 18, 2020.
- [28].Chicago Public Schools. Chicago Public Schools Stats and Facts (2019-2020). Available at: https://cps.edu/About_CPS/At-a-glance/Pages/Stats_and_facts.aspx. AccessedApril 18, 2020.
- [29].Broward County Public Schools. Broward County Student Enrollment Counts. Available at: https://www.browardschools.com/Page/34033. AccessedApril 18, 2020.
- [30].U.S. Census Bureau. Characteristics of Teenagers 15-19 Years Old (2018 American Community Survey 1-year Estimates). Washington, D.C.: U.S. Census Bureau; n.d. [Google Scholar]
- [31].Copeland WE, Bulik CM, Zucker N, et al. Does childhood bullying predict eating disorder symptoms? A prospective, longitudinal analysis. Int J Eat Disord 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Pont SJ, Puhl R, Cook SR, et al. Stigma Experienced by Children and Adolescents With Obesity. Pediatrics 2017:e20173034. [DOI] [PubMed] [Google Scholar]
- [33].Calzo JP, Blashill AJ, Brown TA, et al. Eating Disorders and Disordered Weight and Shape Control Behaviors in Sexual Minority Populations. Curr Psychiatry Rep 2017;19:49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [34].Henrichs-Beck CL, Szymanski DM. Gender expression, body–gender identity incongruence, thin ideal internalization, and lesbian body dissatisfaction. Psychol Sex Orientat Gend Divers 2017;4:23–33. [Google Scholar]
- [35].Mason TB, Lewis RJ, Heron KE. Disordered eating and body image concerns among sexual minority women: A systematic review and testable model. Psychol Sex Orientat Gend Divers 2018;5:397–422. [Google Scholar]
- [36].Murray SB, Touyz SW. Masculinity, Femininity and Male Body Image: A Recipe for Future Research. Int J Mens Health 2012;11. [Google Scholar]
- [37].VanKim NA, Corliss HL, Jun HJ, et al. Gender Expression and Sexual Orientation Differences in Diet Quality and Eating Habits from Adolescence to Young Adulthood. J Acad Nutr Diet 2019;119:2028–2040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Lowry R, Johns MM, Gordon AR, et al. Nonconforming Gender Expression and Associated Mental Distress and Substance Use Among High School Students. JAMA Pediatr 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].Adams DF, Behrens E, Gann L, et al. Gender conformity, self-objectification, and body image for sorority and nonsorority women: A closer look. J Am Coll Health 2017;65:139–147. [DOI] [PubMed] [Google Scholar]
- [40].Rieger G, Linsenmeier JAW, Gygax L, et al. Sexual orientation and childhood gender nonconformity: Evidence from home videos. Dev Psychol 2008;44:46–58. [DOI] [PubMed] [Google Scholar]
- [41].American Academy of Pediatrics, Committee on Adolescence. Achieving Quality Health Services for Adolescents. Pediatrics 2016;138. [DOI] [PubMed] [Google Scholar]