Violence perpetration prevalence among Colorado (United States) high school students across gender, racial/ethnic, and sexual identities

Abstract

Adolescent violence, including sexual violence, homophobic name-calling, and teen dating violence, are public health problems that cause harm to many adolescents in the United States. Although research on the perpetration of these forms of adolescent violence has increased in recent years, little is known about perpetration rates across gender, race/ethnicity, and sexual orientation. To address this gap, the current study descriptively examined perpetration rates between and across different identities, including self-identified race/ethnicity, sexual identity, and gender identity. In Fall 2017, 9th – 11th grade students (N = 4782) at 20 high schools in Colorado (United States) completed a survey that assessed demographics (e.g., race/ethnicity, sexual identity, and gender identity) and various forms of violence perpetration. Compared to female adolescents, male adolescents reported significantly higher perpetration rates for: any sexual violence (27% vs. 17%); sexual harassment (26% vs. 15%); unwanted sexual contact (8% vs. 4%); and homophobic name-calling (61% vs. 38%). Differences in perpetration rates were also observed among various racial/ethnic, sexual, and gender minority students compared to non-minority students. This emphasizes a need for more research on how minority stress that results from the dynamics of intersecting identities and societal systems of power–including racism, sexism, homophobia, and transphobia–contributes to violence perpetration. Evidence-based violence prevention approaches, particularly strategies targeted at changing social norms about violence, gender, and sexuality, need to be tailored and evaluated for students with diverse cultural and social identities to ensure safe school climates for all students.

1. Introduction

Adolescent interpersonal violence, including various forms of sexual violence (SV), homophobic name-calling (HNC), and teen dating violence (TDV), are public health problems that harm many adolescents in the United States (U.S.; Basile et al., 2020). These forms of violence often co-occur among adolescents (Espelage et al., 2021a; Humphrey and Vaillancourt, 2020). SV varies from verbal sexual harassment (SH) to unwanted touching to rape and can occur in-person or online (Basile et al., 2014). HNC includes the use of verbal epithets (e.g., fag, gay) to disparage youth perceived to be gender non-conforming or gender/sexual minorities (Meyer, 2008). TDV includes physical, sexual, psychological, or stalking aggression against a dating partner (Centers for Disease Control and Prevention, 2020). Adolescent violence is preventable, but successful prevention requires an understanding of perpetrator characteristics. Although research on adolescent violence perpetration has increased in recent years, little is known about perpetration across gender, racial/ethnic, and sexual identities.

Gender identity and expression, race/ethnicity, and sexual identity are distinct yet overlapping cultural and social identities. A framework of intersectionality (Bowleg, 2012; Warner, 2008) is necessary for examining how multiple social identities, in a reflection of larger systems of privilege and oppression, contribute to differences in interpersonal violence perpetration. In a sample of high school students in Colorado, the present study examined the prevalence of SV, TDV, and HNC perpetration among adolescents based on gender identity, race/ethnicity, sexual identity, and intersecting social identities. Understanding peer violence perpetration within and as a product of systems of privilege and oppression can inform efforts to create safe, affirming schools for all students regardless of personal demographic characteristics.

2. Adolescent violence perpetration and social identities

Racism, sexism, homophobia, and trans-prejudice are societal systems that place youth at high risk for internalized oppression and interpersonal stressors; individuals with multiple oppressed social identities (e.g., sexual/gender minority person of color) are at particularly high risk (Crenshaw, 1991; Nadal et al., 2021; Garofalo et al., 2006; Sokoloff and Dupont, 2005). While not considered direct risk factors for adolescent violence perpetration, race and ethnicity are seen as “markers for a constellation of external and malleable social contexts that are differentially allocated by racial/ethnic status in American Society” (Sampson et al., 2005, p. 224). Many scholars point to socioeconomic inequality embedded in racially segregated neighborhoods where individuals are differentially exposed to key risk and protective factors that contribute to differences in violence perpetration rates across race/ethnicity (Kaufman and Cooper, 2001; Sampson et al., 2005).

Historically excluded groups (e.g., women, people of color, transgender individuals) are often targets of repeated disparaging messages about their identity groups as they interact with peers, educators, and family (Nadal, 2018). These encounters, combined with systemic oppression (e.g., discrimination) and stereotypes about the groups they belong to (David, 2014), can lead to adolescents from historically excluded groups internalizing negative feelings and thoughts about themselves and their own social identity (Nadal et al., 2021). For youth experiencing identity-related internalized oppression and interpersonal stressors, the enforcement of these norms—as modeled by peers, family, social and broader societal forces—can be associated with violence perpetration (Garofalo et al., 2006; Nadal et al., 2021; Sokoloff and Dupont, 2005).

2.1. Sexual violence

In a U.S. nationally representative sample of adolescents, 23% of boys and 17% of girls reported SH perpetration in the last year, and 4% of boys and 2% of girls reported penetrative rape perpetration (Ybarra and Thompson, 2018). Unwanted online sexual solicitation (i.e., cyber-SV perpetration) was reported by 3% of 10–15 year old youth in the year before the study (Ybarra et al., 2007), with perpetrators being mostly White boys. Among a large population-based study conducted between 2006 and 2012, 8% of White high school students and 12% of students of color reported perpetrating SH in the last year; perpetration rate was 8% among students attracted to the opposite sex only and 14% among those attracted to the same/both sexes or unsure of their attraction (Clear et al., 2014).

2.2. Homophobic name-calling

A recent longitudinal study among middle school students found that girls reported significantly lower rates of HNC perpetration than boys, and African American students reported significantly higher rates than White students (Valido et al., 2021a). A study among high school youth (N = 4822) found that transgender students reported significantly higher rates of HNC perpetration than cisgender students; no significant differences were found between sexual minority and heterosexual students (Valido et al., 2021b).

2.3. Teen dating violence

In a meta-analysis of TDV, Wincentak et al. (2017) found gender differences in physical TDV perpetration (boys 13% vs. girls 25%) and sexual TDV (boys 10% vs. girls 3%). Similarly, Fedina et al. (2016) found in a predominantly Black and Latinx adolescent sample that girls reported higher rates of physical TDV perpetration than boys, but boys reported more sexual TDV. Race and ethnicity were not significant predictors of TDV perpetration when adjusting for gender, age, and sexually risky behaviors (Fedina et al., 2016). The only available data on TDV perpetration of lesbian, gay, bisexual, transgender, and/or queer (LGBTQ) persons was at the college level; nearly a third of LGBTQ students reported perpetration within the past 12 months, with 19.9% reporting physical violence, 12.5% reporting psychological violence, and 10.5% reporting SV perpetration (Edwards and Sylaska, 2013).

3. Gaps in literature and current study

Although the field has documented the overlap between adolescent SV, HNC, and TDV perpetration and identified shared and unique risk and protective factors in general population youth studies (Espelage et al., 2015; Humphrey and Vaillancourt, 2020; Ybarra and Thompson, 2018; Wincentak et al., 2017), the field knows little about variations in perpetration by gender identity, race/ethnicity, and sexual identity given challenges with identifying large enough samples or lack of demographic data. To address this gap, the current study examined perpetration rates between and across gender, race/ethnic, and sexual identities. This study represents an important starting place to better understand perpetrators of specific forms of violence so that prevention efforts can be adapted to improve their effectiveness.

4. Method

4.1. Participants

Data are from the baseline (i.e., pre-intervention) survey of a trial evaluating Sources of Strength, a universal peer-led school-based mental and behavioral health program targeting risk and protective factors associated with multiple types of violence (Espelage et al., 2021b). In Fall 2017, 20 high schools (11 urban, 8 rural, 1 frontier) agreed to school-level random assignment to immediate intervention condition or a waitlist condition and agreed to waiver of active parental consent. Eligible participants included all 9th–11th grade students.

4.2. Measures

Additional psychometric evidence for each measure is included in supplemental materials.

4.2.1. Demographics

Students self-reported gender, racial/ethnic, and sexual identities. For gender identity, students were asked “What is your gender? (Check all that apply)”: male, female, transgender, and other (open-ended). Responses for “other” included “non-binary,” “genderqueer,” and “gender fluid” and excluded fictitious responses (e.g., “alien”). Responses of transgender were coded into a separate variable (1 = transgender, 0 = cisgender). This allowed students to identify as transgender and male, female, or other. For sexual identity, students were asked “What is your sexuality? (Check all that apply)”: straight/heterosexual, gay/lesbian, bisexual, questioning/unsure, and some other sexuality (open-ended). Open-ended responses included “asexual”, “pansexual”, and “demisexual”; fictitious responses were excluded. Students who checked multiple identities were coded as other. For racial/ethnic identity, students were asked, “What is your race/ethnicity? (Check all that apply)”: African American/Black, American Indian/Alaskan Native (AI/AN), Asian, Hispanic/Latino, Multiracial, Native Hawaiian/Pacific Islander (NHPI), and White. Students who checked multiple identities were coded as multiracial.

4.2.2. SV perpetration

Students indicated how often in the past six months they perpetrated any of 13 unwanted behaviors toward other students. These items, from the American Association of University Women SH Survey – Perpetration Scale (Espelage et al., 2015), assessed: (1) SH (4 items, e.g., made sexual comments, jokes or looks), (2) unwanted sexual contact (5 items, e.g., pulled at clothing in a sexual way), (3) forced sexual contact (4 items, e.g., forced sexual acts). Response options were 0 (Never), 1 (1 or 2 times), 2 (3 or 4 times), 3 (5 or 6 times), and 4 (7 or more times). Dichotomous variables were created for each scale, plus the overall SV scale (13 items); students were coded as a perpetrator if they responded 1 or higher to any of the items. Scale reliability in the present sample was α = 0.99 and ω = 0.96 for overall SV, α = 0.93 and ω = 0.74 for SH, α = 0.98 and ω = 0.97 for unwanted sexual contact, and α = 1.00 and ω = 0.97 for forced sexual contact. We assessed technology-facilitated SV perpetration (cyber-SV perpetration) with a three-item scale (Ybarra et al., 2007). Students were asked how often, in the last six months, they used mobile apps, social networks, texts, or other digital communication to: (1) try to get someone to talk about sex when they did not want to, (2) ask someone to do something sexual that they did not want to do, or (3) post or publicly share a nude or semi-nude picture of someone. Students who responded 1 or more times to any items were coded as perpetrators. Scale reliability was α = 0.93 and ω = 0.78.

4.2.3. HNC perpetration

HNC perpetration was assessed using the 5-item Homophobic Content Agent Scale (Poteat and Espelage, 2005), which asked students “How many times in the last 30 days did YOU say homo, gay, lesbo, or fag to the following individuals?”: (1) a friend, (2) someone you did not know well, (3) someone you did not like, (4) someone you thought was gay or lesbian, and (5) someone you did not think was gay or lesbian. Students who responded 1 or more times to any items were coded as perpetrators. Scale reliability was α = 0.92 and ω = 0.82.

4.2.4. TDV perpetration

Students who reported having ever dated (n = 3012) responded to three items after the prompt, “During the past 6 months…How many times did you” do the following things to someone you were dating or going out with: (1) verbally hurt on purpose, (2) physically hurt on purpose, (3) force to do sexual things that they did not want to do? Students who responded 1 or more times to any items were coded as perpetrators. Scale reliability was α = 0.87 and ω = 0.61.

4.3. Procedure

Four IRBs (three universities and one public health department) approved the study. The study met the institution’s guidelines for the protection of human subjects concerning their safety and privacy. Parents could opt-out their child from the study; non-consented youth were removed from the room during survey administration. Students provided assent to participate by signing the front page of the survey. Data collection occurred during regular class times supervised by two researchers in each classroom; teachers were required by IRB to leave the room. Most students completed the survey online in English (98.5%); Spanish- and braille-translated surveys were available. Prior to survey administration, researchers informed students about the project’s purpose and participation rights. Students were informed they could skip any questions and/or stop responding at any point. Following survey completion, researchers provided students with a list of health and wellness support resources. Additional study administration procedural details are included in supplemental materials.

4.4. Analytic plan

All analyses were conducted in R (R Core Team, 2020). Descriptive statistics and statistical comparisons were computed for each binary outcome across gender, race/ethnicity, sexual identity, transgender status, and each two-way intersectionality. Counts and percentages were computed for each demographic and intersectionality group to indicate frequency of perpetration using the gtsummary package (Sjoberg et al., 2021). Each demographic and intersectionality variable was regressed on each binary outcome using logistic regression, and post-hoc statistical comparisons across groups were conducted with the Tukey test for multiple comparisons using the emmeans package (Lenth, 2020). To further control for increased Type I error rates due to multiple comparisons, only contrasts significant at p < .01 are reported. Statistical comparisons were only computed between groups with 10 or more participants.

4.5. Results

Of 6032 9th–11th grade students, 4817 (80%) completed a survey. Students missing data for demographic characteristics or perpetration were excluded from analysis. Missing data ranged from 0 to 2% for demographic characteristics and 1–4% for perpetration outcomes. In the final sample (n = 4782), 22% (n = 1058) reported some kind of SV perpetration; 21% (n = 986) perpetrated SH, 6% (n = 291) perpetrated unwanted sexual contact, 1% (n = 65) perpetrated forced sexual contact, and 6% (n = 284) reported cyber-SV perpetration. Half (50%, n = 2378) of the students reported HNC perpetration (totals not shown in table). Of the 3012 students who had ever dated, 14% (n = 408) reported TDV perpetration. Tables 1–4 present counts and percentages for each outcome by gender, race/ethnicity, sexual identity, and transgender status, respectively. Tables 5–10 present counts and percentages for intersectionalities formed by combinations of demographic characteristics.

Table 1.

Violence perpetration outcomes by gender.

Outcome, n (%)	aFemale, N = 2298	bMale, N = 2426	cOther, N = 55
Sexual violence	380 (17%)	640 (27%)a	15 (28%)
Sexual harassment	348 (15%)	602 (26%)a	14 (26%)
Unwanted sexual contact	91 (4%)	180 (8%)a	7 (13%)
Forced sexual contact	16 (1%)	41 (2%)	‡
Homophobic name calling	855 (38%)	1437 (61%)a	23 (43%)
Cyber sexual violence	110 (5%)	156 (7%)	8 (16%)a
Teen dating violence†	240 (16%)b	153 (11%)	‡

Note. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the column for the group with higher perpetration rate than the group signified by the superscript.

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for teen dating violence rate was the number of adolescents in each group reporting having ever dated (Female = 1483; Male = 1469; Other = 33).

Table 4.

Violence perpetration outcomes by transgender status.

Outcome, n (%)	aCisgender, N = 4796	bTransgender, N = 58
Sexual violence	1042 (22%)	16 (29%)
Sexual harassment	972 (21%)	14 (25%)
Unwanted sexual contact	284 (6%)	7 (12%)
Forced sexual contact	62 (1%)	‡
Homophobic name calling	2321 (49%)	34 (62%)
Cyber sexual violence	275 (6%)	9 (16%)a

Note. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are noted in the column for the group with the higher perpetration rate than the group signified by the superscript.

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for teen dating violence rate was the number of adolescents in each group reporting having ever dated (Cisgender = 2983; Transgender = 37).

Table 5.

Violence perpetration outcomes by intersectionality of gender identity and race/ethnicity.

Gender – Race/ethnicity, N, N of daters	Sexual violence	Sexual harassment	Unwanted sexual contact	Forced sexual contact	Homophobic name calling	Cyber sexual violence	Teen dating violence†
aFemale - ANHPI, N = 61, ND = 20	7 (11%)	6 (10%)	‡	‡	20 (33%)	‡	‡
bFemale - African American, N = 21, ND = 8	‡	‡	‡	‡	‡	‡	‡
cFemale - Hispanic, N = 1024, ND = 672	171 (17%)	159 (16%)	45 (5%)	11 (1%)	446 (44%)f	61 (6%)	120 (18%)l
dFemale - multiracial, N = 326, ND = 224	52 (16%)	50 (16%)	8 (3%)	‡	126 (39%)	9 (3%)	37 (17%)
eFemale - AI/AN, N = 16, ND = 11	‡	‡	‡	‡	6 (38%)	‡	‡
fFemale - white, N = 840, ND = 543	144 (17%)	128 (15%)	36 (4%)	‡	251 (30%)	36 (4%)	77 (14%)
gMale - ANHPI, N = 61, ND = 27	22 (38%)f	18 (31%)	10 (17%)d	‡	42 (72%)adf	5 (9%)	‡
hMale - African American, N = 42, ND = 29	12 (31%)	11 (28%)	‡	‡	24 (60%)	‡	‡
iMale - Hispanic, N = 1065, ND = 665	261 (25%)cf	248 (24%)cf	71 (7%)	19 (2%)	674 (65%)acdf	61 (6%)	70 (11%)
jMale - multiracial, N = 366, ND = 253	109 (30%)cdf	103 (29%)cdf	37 (10%)	8 (2%)	215 (60%)cdf	31 (9%)	40 (16%)
kMale - AI/AN, N = 30, ND = 12	8 (28%)	8 (28%)	‡	‡	15 (50%)	‡	‡
lMale - white, N = 846, ND = 480	225 (27%)cf	212 (26%)cf	55 (7%)	10 (1%)	461 (56%)cdf	54 (7%)	34 (7%)
mOther - Hispanic, N = 12, ND = 5	‡	‡	‡	‡	6 (50%)	‡	‡
nOther - multiracial, N = 11, ND = 3	5 (45%)	5 (45%)	‡	‡	5 (45%)	‡	‡
oOther - white, N = 23, ND = 18	5 (23%)	5 (23%)	‡	‡	6 (27%)	‡	‡

Note. ANHPI = Asian American, Native Hawaiian, or Pacific Islander; AI/AN = American Indian or Alaskan Native; N_D = Number of adolescents reporting having ever dated. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the row for the group with the higher perpetration rate than the group signified by the superscript. Intersectionalities with <10 individuals were not included to protect confidentiality (Other - AI/AN, Other - ANHPI, Other - African American).

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for Teen Dating Violence rate was the N_D in each group.

Table 10.

Violence perpetration outcomes by intersectionality of race/ethnicity and sexual orientation.

Race/ethnicity - sexual orientation, N, N of daters	Sexual violence	Sexual harassment	Unwanted sexual contact	Forced sexual contact	Homophobic name calling	Cyber sexual violence	Teen dating violence†
aANHPI - bisexual, N = 10, ND = 7	‡	‡	‡	‡	7 (70%)	‡	‡
bANHPI - heterosexual, N = 107, ND = 41	27 (26%)	22 (21%)	10 (10%)	‡	53 (51%)	7 (7%)	‡
cAfrican American - heterosexual, N = 60, ND = 39	18 (32%)	17 (30%)	‡	‡	28 (47%)	‡	7 (18%)
dHispanic - bisexual, N = 136, ND = 103	37 (28%)	35 (26%)	13 (10%)	‡	62 (46%)	15 (11%)	29 (28%)s
eHispanic – Gay/lesbian, N = 41, ND = 27	5 (13%)	5 (14%)	‡	‡	24 (60%)	‡	6 (22%)
fHispanic - other, N = 20, ND = 12	5 (25%)	5 (25%)	‡	‡	9 (45%)	‡	‡
gHispanic - questioning, N = 43, ND = 21	13 (30%)	12 (28%)	‡	‡	19 (44%)	‡	‡
hHispanic - heterosexual, N = 1842, ND = 1178	367 (20%)	346 (19%)	95 (5%)	23 (1%)	1005 (55%)γ	103 (6%)	152 (13%)
iMultiracial - bisexual, N = 87, ND = 69	18 (21%)	17 (20%)	5 (6%)	‡	32 (37%)	‡	15 (22%)
jMultiracial – Gay/lesbian, N = 25, ND = 14	8 (33%)	8 (33%)	‡	‡	11 (46%)	‡	‡
kMultiracial - other, N = 18, ND = 9	‡	‡	‡	‡	5 (28%)	‡	‡
lMultiracial - questioning, N = 31, ND = 18	7 (23%)	6 (19%)	‡	‡	15 (48%)	‡	‡
mMultiracial - heterosexual, N = 547, ND = 372	133 (25%)	127 (24%)	38 (7%)	7 (1%)	285 (53%)	34 (7%)	58 (16%)
nAI/AN - heterosexual, N = 37, ND = 19	8 (22%)	8 (22%)	‡	‡	18 (49%)	‡	‡
oWhite - bisexual, N = 131, ND = 94	34 (26%)	32 (25%)	11 (9%)	‡	55 (42%)	16 (12%)	21 (22%)
pWhite – Gay/lesbian, N = 31, ND = 22	7 (24%)	7 (24%)	‡	‡	15 (52%)	‡	‡
qWhite - other, N = 34, ND = 24	6 (18%)	5 (15%)	‡	‡	10 (29%)	‡	‡
rWhite - questioning, N = 67, ND = 28	9 (13%)	9 (14%)	‡	‡	23 (34%)	6 (9%)	‡
sWhite - heterosexual, N = 1441, ND = 874	317 (22%)	291 (21%)	79 (6%)	10 (1%)	619 (44%)	70 (5%)	81 (9%)

Note. ANHPI = Asian American, Native Hawaiian, or Pacific Islander; AI/AN = American Indian or Alaskan Native; N_D = Number of adolescents reporting having ever dated. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the row for the group with the higher perpetration rate than the group signified by the superscript. Intersectionalities with <10 individuals were not included to protect confidentiality (ANHPI - Gay/Lesbian, ANHPI - Other, ANHPI - Questioning, African American - Bisexual, African American - Gay/Lesbian, African American - Other, African American - Questioning, AI/AN - Bisexual, AI/AN - Gay/Lesbian, AI/AN - Other, AI/AN - Questioning,).

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for Teen Dating Violence rate was the N_D in each group.

4.5.1. Significant contrasts between identities

Compared to girls, boys reported significantly higher perpetration rates for: overall SV (27% vs. 17%); SH (26% vs. 15%); unwanted sexual contact (8% vs. 4%); and HNC (61% vs. 38%; Table 1). Girls reported significantly higher TDV perpetration than boys (16% vs. 11%). Students who identified as a gender other than male/female reported higher cyber-SV perpetration than girls (16% vs. 5%).

By race/ethnicity, compared to White students, African American students reported higher forced sexual contact perpetration (7% vs. 1%; Table 2) and Hispanic students reported higher HNC perpetration (55% vs. 43%).

Table 2.

Violence perpetration outcomes by race and ethnicity.

Outcome, n (%)	aANHPI, N = 131	bAfrican American, N = 76	cHispanic, N = 2114	dMultiracial, N = 712	eAI/AN, N = 48	fWhite, N = 1726
Sexual violence	34 (27%)	22 (31%)	438 (21%)	168 (24%)	11 (23%)	378 (22%)
Sexual harassment	29 (23%)	21 (30%)	412 (20%)	160 (23%)	11 (23%)	348 (21%)
Unwanted sexual contact	13 (10%)	7 (10%)	120 (6%)	48 (7%)	‡	95 (6%)
Forced sexual contact	‡	5 (7%)f	31 (2%)	10 (2%)	‡	13 (1%)
Homophobic name calling	68 (53%)	37 (51%)	1135 (55%)f	350 (50%)	23 (48%)	730 (43%)
Cyber sexual violence	10 (8%)	6 (9%)	125 (6%)	44 (6%)	‡	95 (6%)
Teen dating violence†	7 (13%)	10 (23%)	192 (14%)	79 (16%)	‡	114 (11%)

Note. ANHPI = Asian American, Native Hawaiian, or Pacific Islander; AI/AN = American Indian or Alaskan Native. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are noted in the column for the group with the higher perpetration rate than the group signified by the superscript.

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for teen dating violence rate was the number of adolescents in each group reporting having ever dated (ANHPI = 54; African American = 45; Hispanic = 1352; Multiracial = 484; AI/AN = 23; White = 1049).

Students who identified as other sexual identity reported higher rates of forced sexual contact perpetration than heterosexual students (6% vs. 1%; Table 3). Compared to heterosexual students, bisexual students reported higher cyber-SV (10% vs. 6%) and TDV perpetration (24% vs. 12%). Transgender students reported higher cyber SV perpetration than cisgender students (16% vs. 6%; Table 4).

Table 3.

Violence perpetration outcomes by sexual orientation.

Outcome, n (%)	aHeterosexual, N = 4051	bBisexual, N = 375	cGay/Lesbian, N = 102	dQuestioning, N = 151	eOther, N = 87
Sexual violence	874 (22%)	92 (25%)	21 (22%)	33 (22%)	18 (21%)
Sexual harassment	813 (21%)	87 (24%)	21 (22%)	31 (21%)	17 (20%)
Unwanted sexual contact	230 (6%)	31 (8%)	5 (5%)	9 (6%)	6 (7%)
Forced sexual contact	44 (1%)	6 (2%)	‡	‡	5 (6%)a
Homophobic name calling	2016 (51%)	161 (43%)	51 (52%)	63 (42%)	32 (37%)
Cyber sexual violence	219 (6%)	38 (10%)a	5 (5%)	13 (9%)	6 (7%)
Teen dating violence†	305 (12%)	68 (24%)a	11 (17%)	12 (17%)	8 (16%)

Note. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are noted in the column for the group with the higher perpetration rate than the group signified by the superscript.

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for teen dating violence rate was the number of adolescents in each group reporting having ever dated (Heterosexual = 2529; Bisexual = 278; Gay/Lesbian = 65; Questioning = 72; Other = 50).

4.5.2. Significant contrasts between intersectionalities

Twenty-eight of the 30 significant contrasts pertained to boys reporting higher perpetration rates than girls (Table 5). Multiracial boys reported higher overall SV, SH, and HNC perpetration than Hispanic, Multiracial, and White girls. Similar contrasts were found for White, Hispanic, and Asian/NHPI boys but not African American or AI/AN boys. Conversely, Hispanic girls reported higher TDV perpetration than White boys (18% to 7%). The only significant difference within gender by race/ethnicity was that Hispanic girls reported higher HNC perpetration than White girls (44% vs. 30%). No differences were found in forced sexual contact or cyber-SV.

Most significant contrasts between intersections of gender and sexual identity were between gender rather than sexual identities within gender (Table 6). Compared to heterosexual girls, heterosexual boys reported higher SH (25% vs. 14%), unwanted sexual contact (8% vs. 4%) and HNC (62% vs. 37%) perpetration. Bisexual girls reported higher TDV perpetration than heterosexual boys (26% vs. 10%). The only within-gender difference was bisexual girls reporting higher cyber-SV (10% vs. 4%) and TDV (26% vs. 14%) perpetration than heterosexual girls.

Table 6.

Violence perpetration outcomes by intersectionality of gender identity and sexual orientation.

Gender – Sexual orientation, N, N of daters	Sexual violence	Sexual harassment	Unwanted sexual contact	Forced sexual contact	Homophobic name calling	Cyber sexual violence	Teen dating violence†
aFemale - bisexual, N = 293, ND = 225	64 (22%)	60 (21%)	20 (7%)	‡	113 (39%)	30 (10%)e	58 (26%)ej
bFemale – Gay/lesbian, N = 49, ND = 39	6 (12%)	6 (12%)	‡	‡	26 (53%)	‡	7 (18%)
cFemale - other, N = 41, ND = 25	7 (17%)	7 (17%)	‡	‡	17 (41%)	‡	5 (20%)
dFemale - questioning, N = 92, ND = 41	14 (15%)	13 (14%)	‡	‡	25 (27%)	‡	7 (17%)
eFemale - heterosexual, N = 1793, ND = 1141	282 (16%)	256 (14%)	62 (4%)	11 (1%)	663 (37%)	74 (4%)	163 (14%)
fMale - bisexual, N = 65, ND = 44	23 (35%)e	22 (34%)e	9 (14%)	‡	37 (58%)	‡	8 (18%)
gMale – Gay/lesbian, N = 38, ND = 20	12 (32%)	12 (35%)	‡	‡	16 (43%)	‡	‡
hMale - other, N = 16, ND = 7	‡	‡	‡	‡	‡	‡	‡
iMale - questioning, N = 52, ND = 26	16 (31%)	15 (29%)	5 (10%)	‡	33 (63%)d	10 (19%)e	‡
jMale - heterosexual, N = 2223, ND = 1364	580 (27%)e	545 (25%)e	162 (8%)e	33 (2%)	1333 (62%)ade	139 (7%)	137 (10%)
kOther - other, N = 17, ND = 13	‡	‡	‡	‡	‡	‡	‡
lOther - straight, N = 19, ND = 12	5 (26%)	5 (26%)	‡	‡	10 (53%)	‡	‡

Note. N_D = Number of adolescents reporting having ever dated. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the row for the group with the higher perpetration rate than the group signified by the superscript. Intersectionalities with <10 individuals were not included to protect confidentiality (Other - Questioning, Other - Gay/Lesbian, Other - Bisexual).

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for Teen Dating Violence rate was the N_D in each group.

All statistically significant contrasts between intersections of gender and transgender status were among cisgender students (Table 7). Compared to cisgender girls, cisgender boys reported higher overall SV (27% vs. 17%), SH (26% vs. 15%), unwanted sexual contact (8% vs. 4.0%), and HNC (61% vs. 38%) perpetration. Cisgender-other gender adolescents also perpetrated unwanted sexual contact at higher rates than cisgender girls (15% vs. 4%). Cisgender girls reported higher rates of TDV perpetration than cisgender boys (16% vs. 11%).

Table 7.

Violence perpetration outcomes by intersectionality of gender and transgender status.

Gender – Transgender status, N, N of daters	Sexual violence	Sexual harassment	Unwanted sexual contact	Forced sexual contact	Homophobic name calling	Cyber sexual violence	Teen dating violence†
aFemale - cisgender, N = 2291, ND =1479	379 (17%)	348 (15%)	91 (4%)	15 (1%)	853 (38%)	110 (5%)	240 (16%)b
bMale - cisgender, N = 2411, ND = 1457	636 (27%)a	598 (26%)a	178 (8%)a	41 (2%)	1428 (61%)a	153 (7%)	152 (11%)
cOther - cisgender, N = 50, ND = 30	13 (27%)	12 (24%)	7 (15%)a	‡	21 (43%)	6 (13%)	‡
dMale - transgender, N = 15, ND = 12	‡	‡	‡	‡	9 (69%)	‡	‡

Note. N_D = Number of adolescents reporting having ever dated. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the row for the group with the higher perpetration rate than the group signified by the superscript. Intersectionalities with <10 individuals were not included to protect confidentiality (Other - Transgender, Female - Transgender).

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for Teen Dating Violence rate was the N_D in each group.

Between transgender status and race/ethnicity, Hispanic-cisgender students perpetrated HNC at higher rates than White-cisgender students (55% vs. 43%; Table 8). At intersections of transgender status and sexual identity, bisexual-cisgender students reported significantly higher TDV perpetration rates than heterosexual-cisgender students (24% vs. 12%; Table 9). Between race/ethnicity and sexual identity (Table 10), heterosexual White students reported lower HNC perpetration than heterosexual Hispanic students (44% vs. 55%) and lower TDV than bisexual Hispanic students (9% vs. 28%).

Table 8.

Violence perpetration outcomes by intersectionality of race/ethnicity and transgender status.

Race/ethnicity – Transgender status, N, N of daters	Sexual violence	Sexual harassment	Unwanted sexual contact	Forced sexual contact	Homophobic name calling	Cyber sexual violence	Teen dating violence†
aANHPI - cisgender, N = 128, ND = 53	32 (26%)	27 (22%)	13 (10%)	‡	66 (53%)	10 (8%)	7 (13%)
bAfrican American - cisgender, N = 70, ND = 41	18 (27%)	17 (26%)	5 (8%)	‡	33 (49%)	‡	7 (17%)
cHispanic - cisgender, N = 2098, ND = 1339	435 (21%)	409 (20%)	117 (6%)	31 (2%)	1126 (55%)h	121 (6%)	192 (14%)
dHispanic - transgender, N = 16, ND = 13	‡	‡	‡	‡	9 (60%)	‡	‡
eMultiracial - cisgender, N = 698, ND = 476	164 (24%)	157 (23%)	47 (7%)	9 (1%)	342 (50%)	43 (6%)	79 (17%)
fMultiracial - transgender, N = 14, ND = 8	‡	‡	‡	‡	8 (62%)	‡	‡
gAI/AN - cisgender, N = 47, ND = 23	11 (24%)	11 (24%)	‡	‡	22 (47%)	‡	‡
hWhite - cisgender, N = 1708, ND = 1038	375 (22%)	346 (21%)	94 (6%)	13 (1%)	720 (43%)	93 (6%)	111 (11%)
iWhite - transgender, N = 18, ND = 11	‡	‡	‡	‡	10 (56%)	‡	‡

Note. ANHPI = Asian American, Native Hawaiian, or Pacific Islander; AI/AN = American Indian or Alaskan Native; N_D = Number of adolescents reporting having ever dated. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the row for the group with the higher perpetration rate than the group signified by the superscript. Intersectionalities with <10 individuals were not included to protect confidentiality (ANHPI - Transgender, African American - Transgender, AI/AN - Transgender).

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for Teen Dating Violence rate was the N_D in each group.

Table 9.

Violence perpetration outcomes by intersectionality of sexual orientation and transgender status.

Sexual orientation – Transgender status, N, N of daters	Sexual violence	Sexual harassment	Unwanted sexual contact	Forced sexual contact	Homophobic name calling	Cyber sexual violence	Teen dating violence†
aBisexual - cisgender, N = 364, ND = 274	90 (25%)	85 (24%)	29 (8%)	5 (1%)	155 (43%)	36 (10%)	67 (24%)g
bBisexual - transgender, N = 11, ND = 4	‡	‡	‡	‡	6 (55%)	‡	‡
cGay/lesbian - cisgender, N = 95, ND = 62	19 (21%)	19 (22%)	‡	‡	46 (50%)	5 (6%)	11 (18%)
dOther - cisgender, N = 74, ND = 43	14 (19%)	14 (19%)	‡	‡	25 (34%)	‡	7 (16%)
eOther - transgender, N = 13, ND = 7	‡	‡	‡	‡	7 (58%)	‡	‡
Questioning - cisgender, N = 146, ND = 68	31 (21%)	29 (20%)	9 (6%)	‡	60 (41%)	12 (8%)	10 (15%)
gHeterosexual - cisgender, N = 4029, ND = 2510	868 (22%)	808 (21%)	228 (6%)	43 (1%)	2003 (51%)	215 (6%)	303 (12%)
hHeterosexual - transgender, N = 22, ND = 19	6 (27%)	5 (24%)	‡	‡	13 (62%)	‡	‡

Note. N_D = Number of adolescents reporting having ever dated. Superscript indicates significant contrast from a Tukey test for multiple comparisons with p < .01. Superscripts are only listed in the row for the group with the higher perpetration rate than the group signified by the superscript. Intersectionalities with <10 individuals were not included to protect confidentiality (Gay/Lesbian - Transgender, Questioning - Transgender).

^‡Cell counts <5 are suppressed to avoid presentation of unstable results.

^†Denominator for Teen Dating Violence rate was the N_D in each group.

5. Discussion

This study examined multiple forms of adolescent interpersonal violence perpetration across cultural and social identities. While varying perpetration rates were noted across race/ethnicity and sexual identity, boys in most intersectional identities reported greater perpetration than girls, except for TDV, where the trend was reversed. The finding that African American students had higher rates of forced sexual contact perpetration than White students may be a valid finding, or it may be a case of Simpson’s paradox - the association between two variables in the population is ameliorated or reversed in all or most subpopulations (Goltz and Smith, 2010). Possible causes are unequal distribution of a confounding variable within subpopulations and low frequency events. Both apply to this study sample. Boys had higher forced sexual contact perpetration rates, albeit non-significantly, than girls (2% vs. 1%); the sample of African American students was disproportionately composed of boys (67%), unlike the gender distribution of White students (50%). Paired with the low frequency of occurrences (n = 5 for African American students), the unequal gender distribution could contribute to Simpson’s paradox when comparing perpetration rates aggregated by identity to rates disaggregated by intersectionality.

Another possible explanation for the findings for African American students is found in minority stress theory. Extant literature suggests that minority stress — chronic stress resulting from actual, perceived, or anticipated interpersonal prejudice and discrimination –increases the likelihood of violence perpetration (Edwards and Sylaska, 2013). Minority stress experiences are associated with psychological symptoms such as depression, anxiety, and low self-esteem (Brooks, 1981; Meyer, 2003), all previously identified risk factors of violence perpetration. For instance, Edwards and Sylaska (2013) found that, among LGBTQ college students, TDV perpetration was associated with identity concealment, internalized homophobia, and sexual orientation-focused victimization. Also, the prevalence findings for this study are generally consistent with previous literature (Clear et al., 2014; Fedina et al., 2016; Valido et al., 2021a).

This study builds on previous studies by assessing perpetration rates across intersections of gender identity, race/ethnicity, and sexual identity (Birkett and Espelage, 2015; Clear et al., 2014; Edwards and Sylaska, 2013; Fedina et al., 2016; Wincentak et al., 2017). Given findings that boys reported higher perpetration rates than girls across violence types, a continued emphasis on prevention efforts that focus on boys and address hypermasculinity, patriarchy, misogyny, and homophobia is needed. Prevention strategies that seek to change social norms around violence and gender, such as bystander approaches or approaches focused on engaging men and boys as allies in violence prevention (Basile et al., 2016; Niolon et al., 2017), coupled with antiracist practices to create intersectional approaches (Garnett et al., 2014), are necessary to comprehensively address violence against and by adolescents with historically excluded intersectional identities.

However, there is a lack of evidence-based prevention approaches specifically tailored for racial/ethnic, sexual, and gender minority adolescents. Future studies are needed to examine the effectiveness of prevention efforts for youth with historically excluded intersectional identities and to evaluate adaptations of existing evidence-based approaches for their effectiveness among students with diverse cultural and social identities. Prevention strategies must center intersectional experiences and the structural systems of both racism and sexism (Bowleg, 2012) to address the larger structural conditions and systems of power that contribute to and perpetuate violence against and by students with historically excluded intersectional identities, such as school policies or social norms that discriminate against some students. Future research with larger, representative samples that allow an examination of violence perpetration by youth with historically excluded intersectional identities, including socioeconomic status, in greater detail would be beneficial for informing the content of prevention efforts. Additional perpetration-specific data on gender, sexual, and racial/ethnic minority youth would be useful for developing prevention strategies, shifting away from the existing practice of relying on victimization data to understand these groups and inform prevention strategies.

6. Limitations

Although the current study fills essential gaps in the literature, several limitations exist. First, results are descriptive and employ cross-sectional data from a non-representative convenience sample of Colorado high schools, limiting the generalizability of findings. Compared to state demographics, the study sampled a higher percentage of female, Hispanic, multiracial, and AI/ANstudents and a lower percentage of male, African American, White, and Asian American/NHPI (National Center for Education Statistics, 2022). Additionally, the sample size for some intersections was not large enough to be included in the contrasts analysis. Second, victim characteristics were not assessed. It is possible that the high rates observed among some of the groups were due to other participant characteristics not examined in this study. Future longitudinal studies that measure gender, racial/ethnic, and sexual identity and assess victim characteristics would improve understanding of risk factors and potential intervention points for perpetration. Third, violence perpetration was self-reported, which might introduce bias given social desirability; however, other informants (e.g., teacher, official records) would also introduce bias. Our intentional assessment of violence perpetration through numerous behaviorally-specific questions was employed to increase chances of disclosure. Sexual orientation was also captured through self-reported sexual identity alone; adolescents who engage in same-sex sexual behavior or attraction but do not identify as LGBTQ were not captured in this sample. Further, the gender variable was intended to capture gender identity rather than sex assigned at birth, but it cannot be confirmed whether all respondents interpreted the question and answered as such. Fourth, the HNC variable did not ask students about their intent behind the use of homophobic epithets; some students who identify as LGBTQ may use homophobic slurs as a form of reclamation (e.g., change in the meaning of queer; Hess, 2020; Rand, 2014) rather than bullying. Future studies could include more detailed measures that capture students’ intent for using homophobic language. Finally, the TDV measure was limited to three items representing verbal, physical, and sexual TDV; items were combined into one overall measure, so findings may mask differences across different TDV types. Future studies could examine different types of TDV separately utilizing multiple items.

7. Conclusions

Adolescent interpersonal violence perpetration is a critical public health problem. The present study adds to the dearth of literature on perpetration by students of diverse social/cultural identities. Boys reported perpetration most commonly for most types of violence, which held true across sexual identity and race/ethnicity. These findings contribute to the field’s understanding of how students from historically excluded social/cultural communities, whom literature has often identified as frequent victims of interpersonal violence, are also sometimes perpetrators. It is important to understand these findings in the context of the structural and social power hierarchies that exist in society. More data and research are needed in this area to advance prevention efforts, such as adapting proven prevention approaches that have not yet been tested in diverse samples, to get closer to societal goals of making school a safe haven for all students regardless of their social/cultural identities.