OBJECTIVES

To apply an intersectional lens to disparities in emotional distress among youth, including multiple social positions and experiences with bias-based bullying.

METHODS

Data are from the 2019 Minnesota Student Survey (n = 80 456). Social positions (race and ethnicity, sexual orientation, gender) and 2 forms of bias-based bullying (racist, homophobic or transphobic) were entered into decision tree models for depression, anxiety, self-injury, suicidal ideation, and suicide attempts. Groups with the highest prevalence are described. Rates of emotional distress among youth with matching social positions but no bias-based bullying are described for comparison.

RESULTS

LGBQ identities (90%) and transgender, gender diverse, and questioning identities (54%) were common among the highest-prevalence groups for emotional distress, often concurrently; racial and ethnic identities rarely emerged. Bias-based bullying characterized 82% of the highest-prevalence groups. In comparable groups without bias-based bullying, emotional distress rates were 20% to 60% lower (average 38.8%).

CONCLUSIONS

Findings highlight bias-based bullying as an important point for the intervention and mitigation of mental health disparities, particularly among lesbian, gay, bisexual, transgender, gender-diverse, queer, and questioning adolescents. Results point to the importance of addressing bias-based bullying in schools and supporting lesbian, gay, bisexual, transgender, gender-diverse, queer, and questioning students at the systemic level as a way of preventing emotional distress.

What’s Known on This Subject:

Several studies have identified disparities in emotional distress among youth across social positions, but are limited by inclusion of only 2 social positions (eg, sexual orientation and race), use of regression models, and sample sizes that necessitated combining groups.

What This Study Adds:

LGBQ identities (90%) and transgender/gender-diverse/questioning identities (54%) were common among the highest-prevalence groups for emotional distress, often concurrently; bias-based bullying characterized 82% of highest-prevalence groups. In comparable groups without bias-based bullying, emotional distress rates were 20%–60% lower.

The United States is facing a mental health crisis, particularly among youth.1  Rates of depression and anxiety are high among adolescents, and suicide is a leading cause of death in this age group.2  According to the Centers for Disease Control and Prevention’s most recent Youth Risk Behavior Surveillance System, ∼9% of high school students have attempted suicide, with notable disparities across marginalized social positions. This rate is ∼12% among non-Latina/x/o (NL) Black students and 8% among NL white students;3  23% among those who identify as lesbian, gay, or bisexual and 6% among heterosexual students3 ; and 35% among transgender students and 6% to 9% among cisgender students.4 

Minority stressors, including experiences of structural oppression, compound general stressors, contributing to the disproportionately high prevalence of mental health disorders seen among people of color and lesbian, gay, bisexual, transgender, gender-diverse, queer, and questioning (LGBTQ+) people.5 7  Those facing multiple types of stigma simultaneously, such as racism, heterosexism, and cisgenderism, may face the greatest burden.8,9  For example, Mereish and colleagues explored numerous measures of racism and homophobia or transphobia and found distinct effects of intersectional minority stressors on depressive symptoms, above and beyond separate experiences of racism and homophobia or transphobia.8 

A robust body of evidence reveals that short- and long-term outcomes of bullying include mental health problems.10,11  Bullying targeting actual or perceived personal characteristics, such as racial identities or sexual orientation (ie, bias-based bullying), has received increased attention in recent years.12  This is due, in part, to an emerging understanding that bias-based bullying is as harmful or more harmful to health than bullying that does not focus on marginalized identities.13 

Youth with marginalized racial identities, sexual identities, and gender identities report disproportionate rates of general and bias-based bullying compared with their peers in more privileged social positions,13 18  and this disparity is even more pronounced for those with multiple intersecting marginalized identities.19  In previous research using large statewide datasets, bias-based bullying was disproportionately common among youth who identified as transgender or gender diverse (TGD) or were questioning their gender and held additional marginalized social positions (with regards to racial or ethnic identities and sexual identities). For example, 13% of California youth reported experiencing race-based bullying in the past year, but the prevalence was 2 to 3 times higher in Asian and Pacific Islander and Latina/x/o youth who were also TGD or questioning their gender.19  Bias-based bullying is a direct manifestation of structural power inequities20  and a reflection of the historical injustice experienced by those who share their identities.21  Ample empirical evidence indicates that victimization rooted in bias is a key contributor to the health disparities affecting minoritized groups.18,22,23 

Although the authors of numerous studies have examined the prevalence of bias-based bullying experiences among adolescents with intersecting marginalized social positions, relatively few have explored their associations with mental health disparities.24 26  Authors who have done so have focused on only 2 social positions (eg, sexual orientation and racial or ethnic identity).25,26  The authors of these preliminary studies found significant differences in suicidality across intersecting social positions, but they are limited by the use of regression models (eg, models could not accommodate numerous interaction terms). More nuanced research that delves into the experiences of youth facing multiple types of oppression simultaneously with greater attention to different identities is warranted to enrich our understanding of intersectional disparities.

Improving mental health among youth with multiple marginalized identities requires identifying and addressing the role of stigma and bias-based bullying in creating noted disparities in mental health. The current study, therefore, uses a large, population-based sample to explore which intersecting social positions (with regards to racial or ethnic identities, sexual identity, and gender) are most affected by emotional distress and how the experience of bias-based bullying affects these disparities.

Data come from the 2019 Minnesota Student Survey (MSS).27  The MSS is a large, triennial statewide survey of adolescents in grades 5, 8, 9, and 11. All school districts in the state are invited to participate in each wave; in 2019, 81% of all districts had at least 1 eligible grade participate in online data collection. The sample for this analysis was restricted to ninth- and 11th-grade students, who received all relevant questions, as detailed below (n = 80 456). This sample includes ∼66% of all ninth-grade students and 54% of all 11th-grade students enrolled in public or charter schools in the state. Likely mischievous responders were removed by the MSS managers during cleaning (∼2%) on the basis of highly improbable or impossible combinations of responses (such as daily use of multiple illegal drugs). The University of Minnesota’s Institutional Review Board determined that the present analysis was exempt from review, due to the use of existing, anonymous data.

Three social positions were included in this analysis. One survey item assessed racial or ethnic identities, and participants were grouped into 6 categories: NL American Indian/Alaskan Native, NL Asian or Pacific Islander, NL Black/African American, NL white, NL multiracial, and Latina/x/o. Those who did not respond were included in the “missing” category (0.7%, n = 545).

Six sexual identity categories were created from 1 item, “How do you describe yourself?” with response options of “heterosexual (straight),” “bisexual,” “gay or lesbian,” “questioning/not sure,” “pansexual,” “queer,” “I don’t describe myself in any of these ways,” and “I am not sure what this question means.” Based on previous findings of similarities between pansexual and queer youth,28  these groups were combined to maximize statistical power and comparability with previous work.29  Those who did not understand the question (1.5%, n = 1192) were included in a “missing/other” category, along with those who did not respond to this item (0.8%, n = 663). A total of 8.4% of participants (n = 6671) did not describe themselves using any of the listed terms. Based on previous findings,28  these youth were also included in the “missing/other” category for this analysis.

Two survey items were combined to create a gender variable with 4 categories. First, participants were asked to indicate their “biological” sex as male or female. Second, students were asked if they were transgender, genderqueer, or genderfluid (ie, gender modality).30  Those responding “no” to the second item were categorized as a cisgender boy or cisgender girl based on the response to the sex question. Those who responded “yes” were categorized as TGD, and those who responded “I am not sure about my gender identity” were categorized as “questioning gender,” regardless of sex assigned at birth. Participants who responded “I am not sure what this question means” (3.2%, n = 2609) or who did not answer the gender identity question (0.4%, n = 307) were set to missing for this analysis. The decision to combine sex and gender modality was made to maximize cell sizes in the analytic models, allowing for many interactions before cell sizes became smaller than the minimum node size for additional splits.

This analysis modeled 5 indicators of emotional distress. Elevated depression and anxiety symptoms were assessed by using the Patient Health Questionnaire-2 and Generalized Anxiety Disorder-2, respectively, both reflecting the last 2 weeks, with a cutoff score of ≥3 indicating a positive screen result, as recommended by the developers.31  Non-suicidal self-injury (NSSI) was assessed with 1 item asking “During the last 12 months, how many times did you do something to purposely hurt or injure yourself without wanting to die, such as cutting, burning, or bruising yourself on purpose?” with 6 response options ranging from 0 to 20 or more times. The responses were dichotomized as none or any events, given the skewness in the distribution. Suicidal ideation was based on the question “Have you ever seriously considered attempting suicide” and suicide attempt was based on the question “Have you ever actually attempted suicide,” both with responses options of “No,” “Yes, during the last year,” and “Yes, more than a year ago.” Suicide items were dichotomized as yes (ever) or no.

Several types of bias-based bullying were assessed32  and combined to create 2 variables by using the following item: “During the last 30 days, how often have other students harassed or bullied you for any of the following reasons? 1) race, ethnicity, or national origin, 2) gender (being male, female, transgender, etc.), 3) gender expression (your style, dress, or the way you walk or talk), 4) Because you are gay, lesbian, or bisexual or because someone thought you were.” Five response options ranging from “never” to “every day” were offered for each reason. Those who responded “never” for bullying regarding race, ethnicity, or national origin were compared with those who reported any bullying for this reason in the past 30 days. Those who indicated any bullying related to perceived sexual orientation, gender identity, and gender expression (SOGIE) were compared with those reporting no bullying for these reasons (∼3% of participants were included in the missing category because of missing responses on 1 or more of the bullying items). SOGIE-related items were combined to simplify analysis and recognize the frequent overlap of these experiences. The use of the never/any cut point is based on previous work indicating that even infrequent bullying victimization is associated with detriments to wellbeing.33 

Exhaustive χ-square automatic interaction detection (CHAID) is a recommended analytic approach for quantitative studies of intersecting identities34,35  and was used to examine our research questions. It is advantageous in cases in which all predictor variables are categorical and have many values (in which regression models may not be powered to run numerous interaction tests simultaneously). Exhaustive CHAID is a data-driven decision tree algorithm that allows for complex testing of how youth with specific combinations of social positions vary on emotional distress variables. The program cycles through all categorical predictors (eg, identities, bullying) and creates splits between categories with significantly different prevalences on the dependent variable. For example, if the first split is gender, the tree then forms “branches” within each gender grouping. Among cisgender girls, all predictors are examined to locate the most significant split. A similar process proceeds for the TGD and questioning branches, with the predictors examined separately within each branch, allowing for unique variation to emerge in the social identities and experiences that are important for each branch. Additional categorical splits are tested until no further significant differences are detected within groups, given the specified P value (Bonferroni-adjusted <.05) and minimum group size (specified as 40 for this analysis to avoid overfitting).36  The remaining groups that are not significantly different with regard to the dependent variable are “terminal nodes.” Youth with missing data on social positions were retained in the exhaustive CHAID models as a “missing” category. In addition, exhaustive CHAID models were rerun excluding missing categories; the overall pattern of results and conclusions was the same as those described below. We have, therefore, elected to retain participants with missing data on social positions to use all available data whenever possible. Ten-fold cross-validation was used,37  and the risk of misclassification ranged from 8% (suicide attempts) to 25% (elevated anxiety symptoms). Analyses were conducted in IBM SPSS(version 28).

Separate exhaustive CHAID models were run for each emotional distress variable; indicators included the 3 social positions and 2 types of bias-based bullying. For each model, the 10 terminal nodes with the highest prevalence are described with regard to patterns of identities with significantly elevated rates. To explicitly examine the role of bias-based bullying, we calculated the prevalence of each emotional distress variable among youth with the same intersecting social positions (ie, racial or ethnic identity, sexual identity, gender) but who reported “no” on both types of bias-based bullying as the 10 highest-prevalence nodes. These prevalences were tested against the matching high prevalence terminal nodes by using χ-square tests.

Most participants (69.8%) identified as NL white, with similar proportions of NL Asian/Pacific Islander, NL Black/African American, and Latina/x/o students (6.7%–8.5%). More than three-quarters of students were heterosexual, approximately one-half were cisgender boys, and approximately one-half were cisgender girls, with 1.4% identifying as TGD and 1.5% questioning their gender identity. Additional details of the sample are shown in Table 1.

TABLE 1

Characteristics of the 2019 MSS sample (N = 80 456)

n%
Demographic characteristics 
 Racial/ethnic identity   
  NL American Indian/Alaska Native 941 1.2 
  NL Asian/Pacific Islander 5390 6.7 
  NL Black/African American 5966 7.4 
  Latina/x/o 6826 8.5 
  NL White 56 163 69.8 
  NL multiracial 4625 5.7 
  Missing 545 0.7 
 Sexual identity   
  Straight 62 799 78.1 
  Gay or lesbian 1253 1.6 
  Bisexual 4515 5.6 
  Questioning sexual identity 1662 2.1 
  Pansexual/queer 1701 2.1 
  Missing/other 8526 10.6 
 Gender   
  Cisgender girl 37 661 46.8 
  Cisgender boy 37 486 46.6 
  TGD 1141 1.4 
  Questioning gender 1179 1.5 
  Missing/other 2916 3.6 
Emotional distress indicators 
 Elevated depression symptoms 16 641 23.2 
 Elevated anxiety symptoms 19 814 27.6 
 NSSI (past y) 12 253 16.9 
 Suicidal ideation (ever) 16 213 22.7 
 Suicide attempt (ever) 5999 8.4 
Bias-based bullying 
 About race/ethnicity   
  Yes 9082 11.6 
  No 69 030 88.4 
  Missing 2344 2.9 
 About SOGIE   
  Yes 15 756 20.3 
  No 62 029 79.7 
  Missing 2671 3.3 
n%
Demographic characteristics 
 Racial/ethnic identity   
  NL American Indian/Alaska Native 941 1.2 
  NL Asian/Pacific Islander 5390 6.7 
  NL Black/African American 5966 7.4 
  Latina/x/o 6826 8.5 
  NL White 56 163 69.8 
  NL multiracial 4625 5.7 
  Missing 545 0.7 
 Sexual identity   
  Straight 62 799 78.1 
  Gay or lesbian 1253 1.6 
  Bisexual 4515 5.6 
  Questioning sexual identity 1662 2.1 
  Pansexual/queer 1701 2.1 
  Missing/other 8526 10.6 
 Gender   
  Cisgender girl 37 661 46.8 
  Cisgender boy 37 486 46.6 
  TGD 1141 1.4 
  Questioning gender 1179 1.5 
  Missing/other 2916 3.6 
Emotional distress indicators 
 Elevated depression symptoms 16 641 23.2 
 Elevated anxiety symptoms 19 814 27.6 
 NSSI (past y) 12 253 16.9 
 Suicidal ideation (ever) 16 213 22.7 
 Suicide attempt (ever) 5999 8.4 
Bias-based bullying 
 About race/ethnicity   
  Yes 9082 11.6 
  No 69 030 88.4 
  Missing 2344 2.9 
 About SOGIE   
  Yes 15 756 20.3 
  No 62 029 79.7 
  Missing 2671 3.3 

Emotional distress was common in this sample (Table 1). For example, approximately one-quarter of students had elevated depression (23.2%) or anxiety symptoms (27.6%) or reported suicidal ideation (22.7%). A total of 11.6% reported bias-based bullying about race, ethnicity, or national origin, and 20.3% reported bias-based bullying about SOGIE.

Characteristics of the highest-prevalence terminal nodes are shown in Tables 26. For example, the rate of NSSI was 71.6% (>4 times higher than the overall sample rate of 16.9%) among bisexual TGD youth who had experienced bias-based bullying about SOGIE (or were missing on this item), regardless of their racial or ethnic identity and regardless of their experience with bias-based bullying about race (Fig 1; additional decision trees are shown in Supplemental Figs 25). LGBQ identities and TGD or questioning identities were common among the 10 highest-prevalence groups for each measure of emotional distress, emerging in 90% and 54% of the highest prevalence nodes, respectively, and often concurrently (additional details of common identities in high prevalence nodes are shown in Tables 26). For example, pansexual or queer youth who identified as TGD or were questioning their gender were in the 10 highest-prevalence groups for most emotional distress indicators, even without experiencing bias-based bullying (eg, 59.1% had elevated depression symptoms, 67.1% had elevated anxiety symptoms, 63.5% had thought about suicide). Racial or ethnic identities only rarely emerged as a characteristic of high-prevalence groups.

TABLE 2

Ten Terminal Nodes With the Highest Prevalence of Elevated Depression Symptoms; 2019 MSS*

Elevated Depression Symptoms (23.2%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
344 NL white, Latina/x/o, AIAN GL Cis girl, TGD, quest, miss All Yes, miss 68.9 43.7** 
847 All PQ All All Yes, miss 68.7 47.9** 
43 All Quest All Yes No 62.8 35.2** 
1548 All Bi Cis girl, TGD, quest, miss All Yes, miss 62.7 46.1** 
225 All PQ TGD, quest All No 59.1 58.7 
534 All Quest. All All Yes, miss 55.2 35.2** 
179 All Bi TGD, quest All No 53.6 54.5 
97 All Straight TGD, quest All Yes 53.6 22.1** 
316 All Bi Cis boy All Yes, miss 51.6 38.0** 
97 All Miss TGD, quest All Yes 48.5 29.8** 
Elevated Depression Symptoms (23.2%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
344 NL white, Latina/x/o, AIAN GL Cis girl, TGD, quest, miss All Yes, miss 68.9 43.7** 
847 All PQ All All Yes, miss 68.7 47.9** 
43 All Quest All Yes No 62.8 35.2** 
1548 All Bi Cis girl, TGD, quest, miss All Yes, miss 62.7 46.1** 
225 All PQ TGD, quest All No 59.1 58.7 
534 All Quest. All All Yes, miss 55.2 35.2** 
179 All Bi TGD, quest All No 53.6 54.5 
97 All Straight TGD, quest All Yes 53.6 22.1** 
316 All Bi Cis boy All Yes, miss 51.6 38.0** 
97 All Miss TGD, quest All Yes 48.5 29.8** 

AIAN, NL American Indian/Alaskan Native; BBB, bias-based bullying; Bi, bisexual; Cis, cisgender; GL, gay/lesbian; Miss, missing; PQ, pansexual/queer; Prev, prevalence; Quest, questioning.

*

For cells that include multiple (or “all”) categories, no significant differences in emotional distress were detected between groups in that cell.

**

P < .05 compared to the prevalence (%) in that row.

TABLE 3

Ten Terminal Nodes With the Highest Prevalence of Elevated Anxiety Symptoms; 2019 MSS*

Elevated Anxiety Symptoms (27.6%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
709 NL White, Latina/x/o, AIAN, multi, miss PQ Cis girl, TGD, quest. All Yes, miss 79.4 61.4** 
1554 All Bi Cis girl, TGD, quest, miss. All Yes, miss 70.1 54.5** 
398 All GL Cis girl, TGD, quest All Yes, miss 67.3 54.3** 
225 All PQ TGD, quest. All No 67.1 67.0 
82 All Bi Cis boy Yes, miss Yes, miss 64.6 39.0** 
535 All Quest. All All Yes, miss 64.5 40.5** 
909 NL White, Latina/x/o, multi Straight Cis girl Yes, miss Yes 57.5 27.7** 
80 All PQ Cis boy, miss All Yes, miss 57.5 35.6** 
388 All PQ Cis girl All No 55.9 55.8 
1945 All Bi Cis girl, TGD, quest, miss All No 55.0 54.5 
Elevated Anxiety Symptoms (27.6%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
709 NL White, Latina/x/o, AIAN, multi, miss PQ Cis girl, TGD, quest. All Yes, miss 79.4 61.4** 
1554 All Bi Cis girl, TGD, quest, miss. All Yes, miss 70.1 54.5** 
398 All GL Cis girl, TGD, quest All Yes, miss 67.3 54.3** 
225 All PQ TGD, quest. All No 67.1 67.0 
82 All Bi Cis boy Yes, miss Yes, miss 64.6 39.0** 
535 All Quest. All All Yes, miss 64.5 40.5** 
909 NL White, Latina/x/o, multi Straight Cis girl Yes, miss Yes 57.5 27.7** 
80 All PQ Cis boy, miss All Yes, miss 57.5 35.6** 
388 All PQ Cis girl All No 55.9 55.8 
1945 All Bi Cis girl, TGD, quest, miss All No 55.0 54.5 

AIAN, NL American Indian/Alaskan Native; BBB, bias-based bullying; Bi, bisexual; Cis, cisgender; GL, gay/lesbian; Miss, missing; Multi, NL multiracial; PQ, pansexual/queer; Prev, prevalence; Quest, questioning.

*

For cells that include multiple (or “all”) categories, no significant differences in emotional distress were detected between groups in that cell.

**

P < .05 compared to the prevalence (%) in that row.

TABLE 4

Ten Terminal Nodes With the Highest Prevalence of NSSI; 2019 MSS*

NSSI: Any, Past 12 mo (16.9%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
141 All Bi TGD All Yes, miss 71.6 53.2** 
407 All PQ TGD, quest All Yes, miss 71.5 53.9** 
100 All GL TGD All Yes 67.0 38.3** 
106 All Bi Cis boy, miss Yes, miss Yes, miss 63.2 28.3** 
376 All PQ Cis girl, miss All Yes, miss 62.8 42.5** 
135 All Quest. All Yes, miss Yes, miss 60.7 24.9** 
1348 All Bi Cis girl, quest. All Yes, miss 59.9 40.7** 
306 All GL Cis girl, quest. All Yes 55.9 38.5** 
224 All PQ TGD, quest All No 54.0 53.9 
95 All Straight TGD, quest All Yes 50.5 24.2** 
NSSI: Any, Past 12 mo (16.9%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
141 All Bi TGD All Yes, miss 71.6 53.2** 
407 All PQ TGD, quest All Yes, miss 71.5 53.9** 
100 All GL TGD All Yes 67.0 38.3** 
106 All Bi Cis boy, miss Yes, miss Yes, miss 63.2 28.3** 
376 All PQ Cis girl, miss All Yes, miss 62.8 42.5** 
135 All Quest. All Yes, miss Yes, miss 60.7 24.9** 
1348 All Bi Cis girl, quest. All Yes, miss 59.9 40.7** 
306 All GL Cis girl, quest. All Yes 55.9 38.5** 
224 All PQ TGD, quest All No 54.0 53.9 
95 All Straight TGD, quest All Yes 50.5 24.2** 

BBB, bias-based bullying; Bi, bisexual; Cis, cisgender; GL, gay/lesbian; Miss, missing; PQ, pansexual/queer; Prev, prevalence; Quest, questioning.

*

For cells that include multiple (or “all”) categories, no significant differences in emotional distress were detected between groups in that cell.

**

P < .05 compared to the prevalence (%) in that row.

TABLE 5

Ten Terminal Nodes With the Highest Prevalence of Suicidal Ideation; 2019 MSS*

Suicidal Ideation: Ever (22.7%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
139 All Bi TGD All Yes, miss 81.3 56.0** 
825 All PQ All All Yes 75.6 55.3** 
1377 All Bi Cis girl, quest, missing All Yes, miss 70.7 52.5** 
390 All GL Cis girl, quest, missing All Yes, miss 69.0 47.6** 
108 All Bi Cis girl, TGD, missing Yes No 66.7 52.1** 
222 All PQ TGD, quest All No, miss 63.5 62.6 
93 All Bi Quest All No 63.4 63.2 
309 All Bi Cis boy All Yes, miss 61.8 42.7** 
55 Multi, AIAN, missing Missing Cis girl, quest All Yes 60.0 24.5** 
46 All Quest All Yes, miss No, miss 58.7 33.8** 
Suicidal Ideation: Ever (22.7%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
139 All Bi TGD All Yes, miss 81.3 56.0** 
825 All PQ All All Yes 75.6 55.3** 
1377 All Bi Cis girl, quest, missing All Yes, miss 70.7 52.5** 
390 All GL Cis girl, quest, missing All Yes, miss 69.0 47.6** 
108 All Bi Cis girl, TGD, missing Yes No 66.7 52.1** 
222 All PQ TGD, quest All No, miss 63.5 62.6 
93 All Bi Quest All No 63.4 63.2 
309 All Bi Cis boy All Yes, miss 61.8 42.7** 
55 Multi, AIAN, missing Missing Cis girl, quest All Yes 60.0 24.5** 
46 All Quest All Yes, miss No, miss 58.7 33.8** 

AIAN, NL American Indian/Alaskan Native; BBB, bias-based bullying; Bi, bisexual; Cis, cisgender; GL, gay/lesbian; Miss, missing; Multi, NL multiracial; PQ, pansexual/queer; Prev, prevalence; Quest, questioning.

*

For cells that include multiple (or “all”) categories, no significant differences in emotional distress were detected between groups in that cell.

**

P < .05 compared to the prevalence (%) in that row.

TABLE 6

Ten Terminal Nodes With the Highest Prevalence of Suicide Attempts; 2019 MSS*

Suicide Attempt: Ever (8.4%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
198 AIAN, Latina/x/o, Multi Bi All Yes, miss Yes, miss 56.6 24.6** 
188 All PQ All Yes Yes 50.5 24.1** 
115 All Bi TGD No Yes, miss 45.2 24.0** 
119 All GL All Yes, miss Yes, miss 43.7 13.9** 
296 NL White, B/AA, API, miss Bi All Yes, miss Yes, miss 41.2 18.9** 
192 Multi, Latina/x/o, B/AA, AIAN, miss PQ All All No, miss 37.5 37.7 
629 All PQ All No, miss Yes 36.2 24.1** 
132 All Bi All Yes No 34.8 20.2** 
1218 All Bi Cis boy, cis girl, quest, miss No Yes, miss 30.1 20.0** 
503 All GL All No Yes, miss 28.8 13.9** 
Suicide Attempt: Ever (8.4%)
nRacial/ethnic identitySexual identityGenderBBB-raceBBB-SOGIEPrev, %Same with no BBB (%)
198 AIAN, Latina/x/o, Multi Bi All Yes, miss Yes, miss 56.6 24.6** 
188 All PQ All Yes Yes 50.5 24.1** 
115 All Bi TGD No Yes, miss 45.2 24.0** 
119 All GL All Yes, miss Yes, miss 43.7 13.9** 
296 NL White, B/AA, API, miss Bi All Yes, miss Yes, miss 41.2 18.9** 
192 Multi, Latina/x/o, B/AA, AIAN, miss PQ All All No, miss 37.5 37.7 
629 All PQ All No, miss Yes 36.2 24.1** 
132 All Bi All Yes No 34.8 20.2** 
1218 All Bi Cis boy, cis girl, quest, miss No Yes, miss 30.1 20.0** 
503 All GL All No Yes, miss 28.8 13.9** 

AIAN, NL American Indian/Alaskan Native; API, NL Asian or Pacific Islander; B/AA, NL Black/African American; BBB, bias-based bullying; Bi, bisexual; Cis, cisgender; GL, gay/lesbian; Miss, missing; Multi, NL multiracial; PQ, pansexual/queer; Prev, prevalence; Quest, questioning.

*

For cells that include multiple (or “all”) categories, no significant differences in emotional distress were detected between groups in that cell.

**

P < .05 compared to the prevalence (%) in that row.

FIGURE 1

CHAID tree for NSSI.

FIGURE 1

CHAID tree for NSSI.

Close modal

In almost all (82%) of the highest prevalence nodes, participants reported at least 1 of the 2 types of bias-based bullying analyzed here. For example, straight TGD or gender-questioning youth who reported experiencing bias-based bullying about SOGIE had high rates of elevated depression symptoms (53.6% vs 23.2% in the full sample) and NSSI (50.5% vs 16.9% in the full sample).

In comparison groups with the same intersecting social positions but without either type of bias-based bullying experience, emotional distress rates were substantially lower. For example, whereas 68.7% of pansexual or queer youth who reported bias-based bullying about SOGIE had elevated depression symptoms, the prevalence was 47.9% among pansexual or queer youth who reported no bias-based bullying. Similarly, more than one-half (56.6%) of bisexual youth who identified as Native American/Alaska Native, Latina/x/o, or multiracial and reported bias-based bullying related to both race, ethnicity, and national origin and SOGIE had attempted suicide. This prevalence was 24.6% in the comparison group that reported no bias-based bullying. Overall, the prevalence of emotional distress was ∼20% to 60% higher in the highest-risk nodes than in comparison groups without bias-based bullying experience, with an average difference of 38.8% (among the high-prevalence nodes that included bias-based bullying experience). For all high-prevalence groups with bias-based bullying experience, these comparisons were statistically significant (P < .05).

Our findings reveal high rates of emotional distress across 5 separate indicators in a large and diverse population-based sample of adolescents, with significant disparities. Rates of emotional distress were greatly elevated in LGBTQ+ youth who had experienced bullying related to their race, ethnicity, or national origin, or SOGIE (or did not respond to the measures of bias-based bullying), especially when compared with youth in the same marginalized social positions who did not experience these types of bullying. These findings clearly suggest that bias-based bullying is a contributing factor to emotional distress among youth with multiple marginalized social positions.

The results seen here are both consistent with and extend previous research in several ways. Disparities in bias-based bullying13 19  and emotional distress3,8,9  have been observed before in large samples of diverse youth by using models that have included intersecting identities.8,9,19  However, in the current study, we used exhaustive CHAID, which is specifically recommended for studies of this type because of its ability to simultaneously contrast numerous different indicators of social positions. In addition, the large, population-based sample means the present findings may have greater generalizability than previous research using an online sample of youth recruited through organizations and connections to LGBTQ+ social media. The specific inclusion of pansexual and queer sexual identities also uncovered findings not previously described in the published literature: that this group of youth was highly likely to experience distress even in the absence of the 2 types of bias-based bullying included here.

Interestingly, marginalized racial or ethnic identities only rarely appeared among the groups with the highest prevalence of emotional distress, and this contrasts with previous findings revealing that LGBTQ+ youth who also identified as Latina/x/o, Native American, Pacific Islander, or bi/multiracial had greater emotional distress than NL white LGBTQ+ youth.24,25  The differences in these findings may be due to the inclusion of bias-based bullying regarding race and ethnicity in the current study. Explicitly modeling even 1 aspect of the social experience of race (ie, racist bullying) may account for differences that were attributed to race in previous research. However, it is also possible that the setting of this study (ie, Minnesota, in which almost 70% of participants identified as NL white) limited the statistical power to detect differences in some racial and ethnic groups, particularly in their intersections with numerous LGBTQ+ identities.

It is important to note that even among the comparison groups that had not experienced bias-based bullying related to race, ethnicity, national origin or SOGIE (most of which were characterized by an LGBQ identity, TGD or questioning identity, or both), in most cases, the rates of emotional distress were still high compared with the overall sample. Bias-based bullying is just 1 example of interpersonal stigma that generates minority stress; in addition to other types of bias-based bullying (about religion, weight, or ability, for example), interpersonal and structural stigma are evident in education, housing, health care, and all other facets of society,38 40  and these experiences could not be accounted for here. If eliminating bias in these 2 discrete forms accounted for a 20% to 60% difference in emotional distress, eliminating all forms and expressions of stigma may achieve the goal of both reductions in emotional distress and health disparities.

Despite the strengths noted above, this study is also subject to several limitations. First, although school-based samples have better generalizability than those recruited through clinics, support organizations, or social media, they are not able to capture the experiences of youth who were absent on the days of survey collection. Because youth experiencing emotional distress or bullying victimization are more likely to be out of school on any given day, our findings may underestimate prevalence or disparities. Second, the data used here are cross-sectional and student-level only, making it impossible to determine causality and the role of larger structural and contextual factors of relevance to adolescents’ experiences of emotional distress. The authors of future research should consider ways in which location type (eg, urban, rural), community resources, public policy, and other contextual factors affect the disparities seen here. Third, all survey data were self-reported and may, therefore, be subject to bias, including recall and social desirability (eg, community norms around mental health and victimization may inhibit youth from some backgrounds from disclosing these concerns even on an anonymous survey), and previous research reveals that youth with marginalized racial and ethnic identities may underreport bullying victimization in particular.41  Fourth, even with a large sample, there were some small intersecting groups. Decisions made to maximize cell size limit the specificity of findings. Fifth, recoding participants into 6 mutually exclusive racial and ethnic categories does not account for diverse lived experiences of intersectionality (eg, for youth who identify as Latina/x/o and Black). Any attempt to reduce the whole group to a single code (eg, Latina/x/o in this case) is inherently reductionist. Finally, survey items did not assess additional characteristics, such as religion, immigration status, or speaking with an accent (which may have also affected experiences with bias-based bullying), intersectional bullying of racially marginalized LGBTQ+ youth, or broader intersectional racism, homophobia, and transphobia.

Data for this study were collected before the coronavirus disease 2019 pandemic, and the crisis in youth mental health and health disparities has come into high relief in its aftermath.1  Prevention efforts are critical. Findings from the current study highlight bias-based bullying as an important point for intervention and mitigation of mental health disparities among LGBTQ+ adolescents, with rates of emotional distress as much as 60% lower among adolescents with the same social positions but no bias-based bullying experience. Pediatricians are encouraged to inquire about experiences of bias-based bullying and other stigma as part of assessments of emotional wellbeing. This work also points to the importance of addressing bias-based bullying in schools and supporting LGBTQ+ students at the systemic level as a way of preventing emotional distress.

Future research is needed to develop, refine, and evaluate efforts to prevent bias-based bullying in schools and other settings, as well as to identify and address other sources of stigma and systemic oppression in the lives of marginalized adolescents. Understanding and preventing these adverse experiences using an intersectionality lens will promote mental health and mental health equity among young people.

MSS data were provided by public school students in Minnesota via local public-school districts and managed by the MSS Interagency Team. PIQTOC (Protection at the Intersections for Queer Teens of Color) coinvestigators, including Drs Lisa Bowleg, Ana-Maria del Río González, Ryan Watson, and Stephen T. Russell contributed to the overall study from which this manuscript is derived.

Dr Eisenberg conceptualized and designed the study and drafted the initial manuscript; Dr Gower contributed to the study design and conducted statistical analysis; Dr Rider contributed to the study design and aided in the interpretation of findings; Drs Lawrence, Eadeh, and Suresh aided in the interpretation of findings; and all authors reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.

FUNDING: Funded by the National Institutes of Health (NIH). The research reported in this publication was supported by the National Institute on Minority Health and Health Disparities under award number R01MD015722. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The sponsors had no role in the study design, collection, analysis, and interpretation of data, writing of the report, or decision to submit the manuscript for publication.

CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no potential conflicts of interest relevant to this article to report.

CHAID

χ-square automatic interaction detection

LGBTQ+

lesbian, gay, bisexual, transgender, gender-diverse, queer, and questioning

MSS

Minnesota Student Survey

NL

non-Latina/x/o

NSSI

non-suicidal self-injury

SOGIE

sexual orientation, gender identity, and expression

TGD

transgender and gender-diverse

1
U.S. Department of Health and Human Services
.
The U.S. Surgeon General’s Advisory: protecting youth mental health. Available at: https://www.hhs.gov/sites/default/files/surgeon-general-youth-mental-health-advisory.pdf. Accessed November 21, 2023
2
Centers for Disease Control and Prevention
.
10 leading causes of death, United States: 2020, both sexes, all ages, all races. Available at: https://wisqars.cdc.gov/data/lcd/home. Accessed November 8, 2022
3
Ivey-Stephenson
AZ
,
Demissie
Z
,
Crosby
AE
, et al
.
Suicidal ideation and behaviors among high school students - youth risk behavior survey, United States, 2019
.
MMWR Suppl
.
2020
;
69
(
1
):
47
55
4
Johns
MM
,
Lowry
R
,
Andrzejewski
J
, et al
.
Transgender identity and experiences of violence victimization, substance use, suicide risk, and sexual risk behaviors among high school students — 19 states and large urban school districts, 2017
.
MMWR Morb Mortal Wkly Rep
.
2019
;
68
(
3
):
67
71
5
Brooks
V
.
Minority Stress and Lesbian Women
.
Lexington Books, Lexington, MA
;
1981
:
71
90
6
Meyer
IH
.
Minority stress and mental health in gay men
.
J Health Soc Behav
.
1995
;
36
(
1
):
38
56
7
Mongelli
F
,
Perrone
D
,
Balducci
J
, et al
.
Minority stress and mental health among LGBT populations: an update on the evidence
.
Minerva Psichiatr
.
2019
;
60
(
1
):
27
50
8
Mereish
EH
,
Parra
LA
,
Watson
RJ
,
Fish
JN
.
Subtle and intersectional minority stress and depressive symptoms among sexual and gender minority adolescents of color: mediating role of self-esteem and sense of mastery
.
Prev Sci
.
2022
;
23
(
1
):
142
153
9
Eisenberg
ME
,
Gower
AL
,
Nic Rider
G
, et al
.
At the intersection of sexual orientation and gender identity: variations in emotional distress and bullying experience in a large population-based sample of U.S. adolescents
.
J LGBT Youth
.
2019
;
16
(
3
):
235
254
10
Reijntjes
A
,
Kamphuis
JH
,
Prinzie
P
,
Telch
MJ
.
Peer victimization and internalizing problems in children: a meta-analysis of longitudinal studies
.
Child Abuse Negl
.
2010
;
34
(
4
):
244
252
11
Flannery
DJ
,
Todres
J
,
Bradshaw
CP
, et al
.
Bullying prevention: a summary of the Report of the National Academies of Sciences, Engineering, and Medicine : Committee on the Biological and Psychosocial Effects of Peer Victimization: lessons for bullying prevention
.
Prev Sci
.
2016
;
17
(
8
):
1044
1053
12
Earnshaw
VA
,
Reisner
SL
,
Menino
D
, et al
.
Stigma-based bullying interventions: a systematic review
.
Dev Rev
.
2018
;
48
(
February
):
178
200
13
Salmon
S
,
Turner
S
,
Taillieu
T
, et al
.
Bullying victimization experiences among middle and high school adolescents: traditional bullying, discriminatory harassment, and cybervictimization
.
J Adolesc
.
2018
;
63
:
29
40
14
Bucchianeri
MM
,
Gower
AL
,
McMorris
BJ
,
Eisenberg
ME
.
Youth experiences with multiple types of prejudice-based harassment
.
J Adolesc
.
2016
;
51
:
68
75
15
Blake
JJ
,
Lund
EM
,
Zhou
Q
, et al
.
National prevalence rates of bully victimization among students with disabilities in the United States
.
Sch Psychol Q
.
2012
;
27
(
4
):
210
222
16
Iyanda
AE
.
Bullying victimization of children with mental, emotional, and developmental or behavioral (MEDB) disorders in the United States
.
J Child Adolesc Trauma
.
2021
;
15
(
2
):
221
233
17
Earnshaw
VA
,
Bogart
LM
,
Poteat
VP
, et al
.
Bullying among lesbian, gay, bisexual, and transgender youth
.
Pediatr Clin North Am
.
2016
;
63
(
6
):
999
1010
18
Bucchianeri
MM
,
Eisenberg
ME
,
Neumark-Sztainer
D
.
Weightism, racism, classism, and sexism: shared forms of harassment in adolescents
.
J Adolesc Health
.
2013
;
53
(
1
):
47
53
19
Gower
AL
,
Rider
G
,
del Rio Gonzalez
A
, et al
.
Application of an intersectional lens to bias-based bullying among LGBTQ+ youth of color in the United States
.
Stigma and Health
.
2023
;
8
(
3
):
363
371
20
Hatzenbuehler
ML
,
Shen
Y
,
Vandewater
EA
,
Russell
ST
.
Proposition 8 and homophobic bullying in California
.
Pediatrics
.
2019
;
143
(
6
):
e20182116
21
Farley
J
,
Gallagher
J
,
Richardson Bruna
K
.
Disrupting narrow conceptions of justice: exploring and expanding ‘bullying’ and ‘upstanding’ in a university honors course
.
Educ Citizsh Soc Justice
.
2020
;
15
(
3
):
258
273
22
Rosenthal
L
,
Earnshaw
VA
,
Carroll-Scott
A
, et al
.
Weight- and race-based bullying: health associations among urban adolescents
.
J Health Psychol
.
2015
;
20
(
4
):
401
412
23
Bucchianeri
MM
,
Eisenberg
ME
,
Wall
MM
, et al
.
Multiple types of harassment: associations with emotional well-being and unhealthy behaviors in adolescents
.
J Adolesc Health
.
2014
;
54
(
6
):
724
729
24
Park
IY
,
Speer
R
,
Whitfield
DL
, et al
.
Predictors of bullying, depression, and suicide attempts among youth: the intersection of race/ethnicity by gender identity
.
Child Youth Serv Rev
.
2022
;
139
(
May
):
106536
25
Bostwick
WB
,
Meyer
I
,
Aranda
F
, et al
.
Mental health and suicidality among racially/ethnically diverse sexual minority youths
.
Am J Public Health
.
2014
;
104
(
6
):
1129
1136
26
Lytle
MC
,
De Luca
SM
,
Blosnich
JR
,
Brownson
C
.
Associations of racial/ethnic identities and religious affiliation with suicidal ideation among lesbian, gay, bisexual, and questioning individuals
.
J Affect Disord
.
2015
;
178
:
39
45
27
Minnesota Center for Health Statistics
.
Minnesota student survey. Available at: www.health.state.mn.us/divs/chs/mss/. Accessed October 11, 2016
28
Gower
AL
,
Rider
GN
,
Brown
C
,
Eisenberg
ME
.
Diverse sexual and gender identity, bullying, and depression among adolescents
.
Pediatrics
.
2022
;
149
(
4
):
e2021053000
29
Eisenberg
ME
,
Gower
AL
,
Watson
RJ
, et al
.
Substance use behaviors among LGBTQ+ youth of color: identification of the populations bearing the greatest burden in three large samples
.
J Adolesc Health
.
2022
;
71
(
3
):
317
323
30
Ashley
F
. “Trans” is my gender modality: a modest terminological proposal. In:
Trans Bodies, Trans Selves: A Resource by and for Transgender Communities
. 2nd ed.
Oxford University Press
;
2022
31
Kroenke
K
,
Spitzer
RL
,
Williams
JB
,
Löwe
B
.
An ultra-brief screening scale for anxiety and depression: the PHQ-4
.
Psychosomatics
.
2009
;
50
(
6
):
613
621
32
Hamburger
M
,
Basile
K
,
Vivolo
A
;
Centers for Disease Control and Prevention
.
Measuring bullying victimization, perpetration, and bystander experiences: a compendium of assessment tools. Available at: https://www.cdc.gov/violenceprevention/pdf/bullycompendium-a.pdf. Accessed November 21, 2023
33
Gower
AL
,
Borowsky
IW
.
Associations between frequency of bullying involvement and adjustment in adolescence
.
Acad Pediatr
.
2013
;
13
(
3
):
214
221
34
Bauer
G
,
Churchill
S
,
Mahendran
M
, et al
.
Intersectionality in quantitative research: a systematic review of its emergence and applications of theory and methods
.
Soc Sci Med -Population Heal
.
2021
;
14
:
100798
35
Shaw
L
,
Chan
F
,
McMahon
B
.
Intersectionality and disability harassment: the interactive effects of disability, race, age, and gender
.
Rehabil Couns Bull
.
2012
;
55
(
2
):
82
91
36
Kass
G
.
An exploratory technique for investigating large quantities of categorical data
.
J R Stat Soc Ser C Appl Stat
.
1980
;
29
(
2
):
119
127
37
El-Muslih
S
,
Vionanda
D
,
Amalita
N
,
Salma
A
.
Comparison of error rate prediction methods in classification modeling with the CHAID method for imbalanced data
.
J Stat Data Sci
.
2023
;
1
(
4
):
321
328
38
Alvarez
K
,
Cervantes
PE
,
Nelson
KL
, et al
.
Review: structural racism, children’s mental health service systems, and recommendations for policy and practice change
.
J Am Acad Child Adolesc Psychiatry
.
2022
;
61
(
9
):
1087
1105
39
Hardeman
RR
,
Medina
EM
,
Kozhimannil
KB
.
Structural racism and supporting Black lives - the role of health professionals
.
N Engl J Med
.
2016
;
375
(
22
):
2113
2115
40
Hatzenbuehler
ML
.
Advancing research on structural stigma and sexual orientation disparities in mental health among youth
.
J Clin Child Adolesc Psychol
.
2017
;
46
(
3
):
463
475
41
Lai
T
,
Kao
G
.
Hit, robbed, and put down (but not bullied): underreporting of bullying by minority and male students
.
J Youth Adolesc
.
2018
;
47
(
3
):
619
635

Supplementary data