Sexual Minority Status, Bullying Exposure, Emotion Regulation Difficulties, and Delinquency Among Court-Involved Adolescent Girls

Abstract

Sexual minority adolescent girls are overrepresented in the justice system. This study used the minority stress model and psychological mediation framework to investigate a pathway for this disparity among court-involved girls ages 14–18 (N=226; mean age: 15.58; 48% sexual minority). The hypotheses were that sexual minority status would be associated with delinquency, bullying exposure would be associated with delinquency indirectly via emotion regulation difficulties, and the relationship between bullying exposure and emotion regulation difficulties would be stronger for sexual minority girls. Bullying exposure and emotion regulation difficulties were not related. Sexual minority status was related to delinquency, and emotion regulation difficulties mediated this relationship. The findings suggest interventions to build emotion regulation skills may reduce delinquency for sexual minority girls.

Introduction

Although the extant literature has established that sexual minority adolescent girls are overrepresented in the juvenile justice system (Jonnson, Bird, Li, & Viljoen, 2019), potential pathways explaining this disparity have not been adequately explored. Identifying the mechanisms through which sexual minority adolescent girls enter and remain in the system is necessary for developing effective programs to prevent both initial and continued involvement for this population. The current study contributes to this literature by investigating whether sexual minority status is associated with higher levels of delinquency among court-involved adolescent girls and examining a potential pathway accounting for such a relationship based on the minority stress model (Meyer, 2003) and psychological mediation framework (Hatzenbuehler, 2009) through which bullying exposure and emotion regulation difficulties contribute to disparities in delinquency based on sexual minority status.

Juvenile Justice System Involvement Among Sexual Minority Adolescent Girls

Sexual minority adolescent girls are disproportionately represented at multiple levels of the juvenile justice system. An investigation conducted with a population-based, national sample collected as part of the National Longitudinal Study of Adolescent Health found that young women who identified as sexual minorities were significantly more likely than heterosexual peers to report having been stopped by the police (19.5% vs. 9.5%), having been arrested before the age of 18 (4.1% vs. 1.2%), and having been convicted of a crime before the age of 18 (1.3% vs. 0.4%; Himmelstein & Brückner, 2011). Similarly, an epidemiological study found that 48% of court-involved, non-incarcerated adolescent girls who were first-time offenders were sexual minorities (Hirschtritt, Dauria, Marshall, & Tolou-Shams, 2018). An analysis using a sample of adolescents in detention centers – sites that hold adolescents between the time of arrest and going to court – conducted in seven different sites across the United States reported that 32% of girls identified as sexual or gender minorities (Irvine & Canfield, 2015). There is also evidence of disparities among incarcerated youth. In a national study of youth in custody, 39% of adolescent girls identified as sexual minorities (Wilson et al., 2017), and, in a study of youth confined in facilities for those with felony convictions in Ohio, 27% of adolescent girls reported sexual minority identities (Belknap, Holsinger, & Little, 2012).

These findings regarding the representation of sexual minority adolescent girls at different levels of juvenile justice system involvement stand in stark contrast to estimates of the representation of sexual minority girls among adolescents in general, with a national population-based study of adolescent girls in schools finding only 12% identified as such (Kann et al., 2016). Furthermore, there is evidence that system involvement is a problem for sexual minority adolescent girls in particular: findings from a meta-analysis indicated that there was no evidence of a similar disparity for sexual minority adolescent boys (Jonnson et al., 2019). The extant literature on sexual minority adolescent girls in the juvenile justice system provides clear evidence of disproportionality. However, none of these studies investigated the factors contributing to this disparity, and so their results do not shed light on potential mechanisms that lead to sexual minority adolescent girls entering the justice system at such high rates.

One critical step in the path for sexual minority adolescent girls’ involvement in the juvenile justice system is engagement in delinquent behavior, which puts them at risk for arrest, conviction, and incarceration, depending on severity. Few studies have examined differences in delinquency between sexual minority and heterosexual adolescents, and results from those studies have been mixed. In a population-based sample, sexual minority adolescents reported higher levels of minor and moderate delinquent, but not violent behaviors than heterosexual peers (Himmelstein & Brückner, 2011). These findings suggest that sexual minority adolescents may be disproportionately represented in the justice system because they engage in more of some types of delinquent behavior. An analysis conducted with court-involved first-time offending adolescents found no difference in reports of past delinquency between sexual minority and heterosexual participants (Hirschtritt et al., 2018). This finding suggests that a disparity in delinquency may contribute to sexual minority adolescents’ overrepresentation in the justice system by providing evidence that sexual minority adolescents who do engage in delinquent behavior are as likely to become court-involved as heterosexual peers engaging in similar levels of such behavior. Neither of these studies reported analyses by gender, and so the extent to which findings differ for adolescent boys and girls cannot be determined.

One study investigated delinquency among sexual minority adolescent girls using a sample of unstably and under-housed youth (Frederick, Ross, Bruno, & Erickson, 2011). The results indicated that sexual minority participants reported more delinquent behaviors than same-gender heterosexual peers; whether these findings generalize to girls with stable housing is unclear. Given the evidence that sexual minority adolescent boys and girls differ in terms of their risk for justice system involvement (Jonnson et al., 2019), further research is needed to investigate differences in delinquency based on sexual minority status among adolescent girls in particular and the underlying mechanisms that account for such differences.

The Minority Stress Model and Psychological Mediation Framework

The minority stress model (Meyer, 2003) provides a theoretical framework to account for higher rates of delinquency among sexual minority adolescent girls and thus a potential explanatory pathway for their disproportionate representation in the juvenile justice system. According to the model, sexual minority individuals experience unique stress processes not encountered by heterosexuals as a direct result of holding a sexual minority status in environments and systems that privilege heterosexuality, and the added burden of this minority stress leads to negative mental health outcomes for sexual minority individuals. The model differentiates distal and proximal minority stress processes. Distal minority stress processes refer to events and interactions in which the individual experiences discrimination based on their sexual minority status, ranging from severe events such as physical assault or more subtle ones such as microaggressions in a workplace. Proximal minority stress processes encompass cognitive and affective changes resulting from experienced or threatened discrimination, such as expecting to be victimized, developing internalized homophobia, or hiding one’s sexual minority status in certain settings. The minority stress model identifies pathways from both distal and proximal minority stress processes to negative mental health outcomes, which can include internalizing and externalizing disorders.

The psychological mediation framework extends the minority stress model by proposing that changes in universal psychological processes mediate the relationship between minority stress and specific mental health outcomes (Hatzenbuehler, 2009). According to this framework, one process through which minority stress affects mental health functioning is emotion regulation. The model hypothesizes that, when confronted with minority stressors, sexual minority individuals marshal coping resources to manage negative affect and maintain emotional control. The experience of a particularly severe minority stress process (e.g., losing one’s job) or a chronic one (e.g., concealing one’s sexual minority status on a daily basis at one’s job) depletes these faculties, leading to difficulties with emotion regulation for the individual, which in turn contribute to negative mental health outcomes.

As originally conceptualized, this model focuses on alcohol use as the primary externalizing outcome associated with emotion dysregulation. However, extending the model to include delinquent behaviors provides a potential mechanism to explain juvenile justice system involvement among sexual minority adolescent girls. For example, emotion regulation difficulties may lead sexual minority girls to engage in transgressive behaviors to manage negative affect (e.g., substance use to alter mood, theft to satisfy an urgent desire, destruction of property to “blow off steam”), as a result of emotional control deficits in situations requiring it (e.g., escalating an argument into a physical fight), or due to emotion dysregulation interfering with decision-making (e.g., making threats when speaking to an authority figure like a teacher or parent about a stressful topic). Thus, incorporating both the minority stress model and the psychological mediation framework leads to a theoretical causal chain in which minority stress predicts emotion regulation difficulties, and emotion regulation difficulties then predict delinquency for sexual minority adolescent girls.

Bullying Exposure, Mental Health Outcomes, and Delinquency

Experiencing peer bullying serves as one potential distal minority stress process to initiate this chain. Previous research has indicated that sexual minority adolescent girls experience high rates of such victimization. A national, population-based study found that sexual minority girls are more likely than heterosexual peers to report being bullied on school property, being electronically bullied, and being threatened or injured with a weapon on school property (Kann et al., 2016). Likewise, results from a meta-analysis indicated that sexual minority adolescent girls were more than three times as likely as heterosexual peers to report a peer threatening or assaulting them with a weapon at school (Friedman et al., 2011).

Of note, the minority stress model holds that bullying exposure and other peer victimization¹ constitutes a minority stress process only if it occurs due to the target’s sexual minority status; however, the significant disparity in victimization rates for sexual minority adolescent girls suggests that their sexual minority status likely played a role in peers targeting them. Studies measuring sexual minority status-based peer victimization have found that 21% of sexual minority high school students in an urban population-based sample (Barnett, Molock, Nieves-Lugo, & Zea, 2019) and 36% of sexual minority adolescents recruited through community organizations (Huebner, Thoma, & Neilands, 2015) report such experiences. There is also evidence that sexual minority adolescents report experiencing more homophobic name-calling (Tucker et al., 2016), homophobic teasing (Espelage, Aragon, & Birkett, 2008), and homophobic bullying, threats, or harassment (Poteat, Mereish, DiGiovanni, & Koenig, 2011) than heterosexual peers. Consistent with the minority stress model, research has found that peer victimization based on sexual minority status was associated with negative health outcomes for sexual minority adolescents, including anxiety symptoms (Tucker et al., 2016), depression or suicidal ideation (Espelage et al., 2008), and substance use (Huebner et al., 2015).

The body of knowledge concerning the relationship between bullying exposure and other peer victimization and delinquency is less developed. An investigation of the relationships between peer victimization, punishable infractions (consisting of cigarette, alcohol, and marijuana use as well as truancy and carrying a weapon at school), and school suspension and incarceration among adolescents found evidence that sexual minority status predicted peer victimization, which then predicted punishable infractions, which then in turn predicted suspension and incarceration (Poteat, Scheer, & Chong, 2016). Indirect effects were significant though small in size. These results provide some support for the conceptualization that minority stress contributes to delinquency and ultimately system involvement for sexual minority adolescents. This study also investigated an alternative model by testing whether the magnitude of the paths from victimization to infractions and from infractions to suspension and incarceration differed by sexual minority status. Findings indicated that the path from peer victimization to punishable infractions was significantly lesser in magnitude for sexual minority adolescents than heterosexual ones. This result runs counter to the minority stress model by suggesting that peer victimization had a less deleterious effect on sexual minority participants. However, the sample was 50% male, and so the extent to which findings differ for sexual minority adolescent girls is not clear. Furthermore, the punishable infractions measure was limited and did not include many other delinquent behaviors, such as fighting, theft, and vandalism.

An investigation of school victimization and discipline among a large cohort of sexual and gender minority youth (Palmer & Greytak, 2017) sheds additional light on the potential pathway from minority stress to juvenile justice system involvement. The study examined whether school victimization predicted multiple forms of school discipline as well as juvenile justice system involvement arising from problems at school. Results indicated there was a significant association between school victimization and system involvement, again providing evidence that this minority stress process contributes to juvenile justice system involvement for sexual minority adolescents. Once more, the analysis grouped adolescent boys and girls together. It also did not include a measure of delinquency, and so the extent to which higher rates of delinquency contributed to system involvement was not tested. Given the mixed results of these two studies, further research is warranted to understand the relationships among sexual minority status, bullying exposure, and delinquency. Moreover, the disparity of juvenile justice system involvement is specific to sexual minority adolescent girls (Jonnson et al., 2019), and so there is a critical need for studies on contributing factors that are conducted with samples consisting entirely of adolescent girls.

The underlying psychological processes through which bullying exposure contributes to negative mental health outcomes in general and delinquent behavior specifically are also not well understood. The psychological mediation framework identifies emotion regulation difficulties as one potential mechanism (Hatzenbuehler, 2009). The relationship between bullying exposure and emotion regulation processes among sexual minority adolescents in particular has not been sufficiently studied. Investigations with presumably predominantly heterosexual adolescent samples have conceptualized difficulties with emotion regulation as increasing the likelihood that an adolescent will be targeted for victimization. Findings have indicated that maladaptive emotion regulation is associated with electronic bullying cross-sectionally (Vranjes, Erreygers, Vandebosch, Baillien, & De Witte, 2018) as well as bullying (Chervonsky & Hunt, 2018) and electronic bullying victimization (Hemphill, Tollit, Kotevski, & Heerde, 2015) longitudinally. Further investigation is needed to determine whether bullying exposure contributes to subsequent difficulties with emotion regulation for adolescents and if that association differs between sexual minority and heterosexual populations, as would be expected when bullying exposure constitutes a minority stress process for sexual minorities.

Potential Confounding Constructs for Mediation Analysis.

The psychological mediation framework posits that the relationship between sexual minority stress processes and negative mental health outcomes is mediated by universal psychological processes. Rigorous evaluation of such a mediation relationship requires controlling for potential confounding of the exposure-mediator, exposure-outcome, and mediator-outcome relationships as well as the exposure variable influencing a third variable that confounds the mediator-outcome relationship (VanderWeele, 2016). The extant literature on bullying exposure, sexual minority status, emotion regulation, and delinquency points to several potentially confounding relationships. Neighborhood disorder has been associated with bullying exposure (Hong & Espelage, 2012), emotion regulation (Criss, Morris, Ponce-Garcia, Cui, & Silk, 2016), and delinquency (Byrnes, Chen, Miller, & Maguin, 2007). Socioeconomic status has been associated with bullying exposure (Hong & Espelage, 2012), emotion regulation (Herd, King-Casas, & Kim-Spoon, 2020), and antisocial behavior, which conceptually overlaps with delinquency (Piotrowska, Stride, Croft, & Rowe, 2015). Finally, childhood maltreatment has been associated with sexual minority status (McGeough & Sterzing, 2018), bullying exposure (Hong & Espelage, 2012), and emotion regulation (Kim & Cicchetti, 2010). Investigating whether emotion regulation difficulties mediate the relationship between bullying exposure and delinquency therefore necessitates inclusion of these constructs to control for potential confounding relationships.

Current Study

The extant literature provides strong evidence that sexual minority adolescents girls are disproportionately represented in the juvenile justice system, but further investigation is needed to identify mechanisms accounting for this disparity. The present study contributes to closing this gap by investigating a pathway based on the minority stress model and the psychological mediation framework involving sexual minority status, bullying exposure, emotion regulation, and delinquency. Specifically, the study investigated if sexual minority status is associated with higher levels of delinquency (Hypothesis 1) and if bullying exposure is associated with more difficulties with emotion regulation (Hypothesis 2). Based on the minority stress model and psychological mediation framework, the study examined whether the relationship between bullying exposure and emotion regulation difficulties is stronger for sexual minority participations (Hypothesis 3), and whether bullying exposure is associated with more delinquent behavior indirectly via more difficulties with emotion regulation (Hypothesis 4). Figure 1 depicts the conceptual model.

Figure 1.

Conceptual Model Depicting Hypothesized Relationships Among Bullying Exposure, Sexual Minority Status, Emotion Regulation Difficulties, and Delinquency

Methods

Recruitment

The present study was conducted using the baseline data for a randomized controlled trial of Project Date SMART (Rizzo et al., 2018), an intervention to prevent dating violence and sexual risk behavior among court-involved adolescent girls. Eligibility criteria for participation were as follows: (1) be an adolescent girl between the ages of 14 and 18; (2) have an open family court petition; (3) if under 18, have a parent or legal guardian who could provide informed consent; and (4) be English speaking. The sample for the present study was recruited through the truancy and formal court calendars as well as the juvenile intake department for a family court in the Northeast United States. Girls were directly referred by judges, magistrates and intake workers, as well as community agencies serving justice-involved youth (e.g., group homes). Others were approached by study staff prior to scheduled appointments at the courthouse. Legal guardians for referred girls under age 18 completed a consent to contact form which gave permission to study staff to contact the family and provide details about the program. At meetings with study staff, parental consent/adolescent assent were obtained for all minor participants and informed consent was obtained for all adult participants (age 18). All study procedures were approved by the affiliated hospital’s Institutional Review Board.

Procedure

Participants completed a baseline assessment prior to randomization. Baseline measures were administered via Audio Computer-Assisted Self Interview (ACASI) on tablet computers, with the exception of the Childhood Trauma Questionnaire, which was administered as part of a separate paper-and-pencil survey battery. Participants could elect to skip individual questions for all measures. Compensation of $40 was provided to participants who completed baseline assessments.

Measures

Study variables.

Sexual minority status.

Sexual minority status was assessed by participants self-reporting sexual behavior, sexual attraction, and sexual identity via three separate items. Participants who endorsed any attraction to females, any previous sexual behavior with females, or gay/lesbian/queer or bisexual sexual identities were coded as sexual minorities, and other participants were coded as heterosexual. Sexual minority status was operationalized in this manner because sexual orientation encompasses behavior, identity, and attraction, and an individual need not espouse a sexual minority identity in order to be bullied or otherwise victimized on the basis of their perceived sexual orientation, particularly during adolescence. Classifying sexual minority status among adolescents based on more than one dimension is consistent with previous studies with court-involved samples (Hirschtritt et al., 2018) and sexual minority adolescents broadly (e.g., analysis of Youth Risk Behavior Surveillance data in Kann et al., 2016). Furthermore, there is evidence that sexual identity labels are frequently unstable among sexual minority adolescents (see Savin-Williams (2011) for a review and discussion).

Bullying exposure.

Participants reported whether they had experienced specific types of bullying in the past year using eight items adapted from the Internet Experiences Questionnaire (Raskauskas & Stoltz, 2007). A brief definition of bullying as repeated acts perpetrated by “someone such as a friend or classmate” preceded the items, as recommended in the literature (Thomas, Connor, & Scott, 2015). Items assessed types of in-person bullying (e.g., being pushed, hit, or shoved at school) and electronic bullying (e.g., being bullied through text messaging). Measurement of in-person and electronic bullying experiences concurrently has also been advised in the literature (Thomas et al., 2015). Participants additionally responded to a ninth item assessing “other” bullying and then were asked to describe that type of bullying in a free-response item. The “other” bullying item was administered prior to the electronic bullying items, and free responses appeared to overlap substantially with other items (e.g., “bullied online,” “beaten up”). Since these responses did not constitute distinct separate types of bullying, this item was not included in the sum score. The sum of the types of bullying experiences endorsed was used as the measure of bullying exposure, which is consistent with treatment of this construct in previous studies (Vivolo-Kantor, Martell, Holland, & Westby, 2014).

Emotion regulation difficulties.

Difficulties with emotion regulation were measured with the Affect Dysregulation Scale (ADS; Brown et al., 2012). The scale consists of six items assessing frequency of experiences of dysregulation during the past three months (e.g., “people have suggested I calm down,” “I have felt overwhelmed by strong feelings”) with response options “not at all,” “a little”, “sometimes,” and “often.” The mean was calculated for the six items, with higher scores indicating more difficulties with emotion regulation. Cronbach’s alpha for the sample was 0.77.

Delinquency.

Delinquency was measured using the Delinquency Activities Scale (DAS; Reavy, Stein, Paiva, Quina, & Rossi, 2012), a self-report measure partially based on the Self-Reported Delinquency Scale (SRD; Elliott, Huizinga, & Ageton, 1985). The measure assesses participants’ engagement in 40 delinquent behaviors, ranging from relatively minor (e.g., stealing something worth less than five dollars; lying about one’s age to obtain an age-restricted product or entry into an age-restricted setting) to more severe (e.g., sexual assault; involvement in gang fights). For each behavior, participants were asked a series of questions, including whether they have ever engaged in it and if they have in the past three months, with yes/no response options. For the present analysis, the types of behaviors reported in the past three months were summed to create a count of recent delinquent behaviors.

Potential confounding variables.

Childhood maltreatment.

Childhood maltreatment was measured with the Childhood Trauma Questionnaire (CTQ; Bernstein & Fink, 1998), a self-report measure that assesses lifetime experiences of abuse and neglect, including physical, sexual, and emotional abuse and physical and emotional neglect ranging from less severe (e.g., “I felt loved” [reverse-scored]) to more severe (e.g., “I got hit or beaten so badly that it was noticed by someone like a teacher, neighbor, or doctor”). Response options ranged from “never true” to “very often true.” The 25 items comprising the emotional abuse, emotional neglect, physical abuse, physical neglect, and sexual abuse subscales were used. The mean was calculated for the 25 items for a measure of overall abuse and neglect. Cronbach’s alpha for the sample was 0.95.

Neighborhood environment.

Neighborhood environment was measured using the sum of six items from the Neighborhood Environment Index (NEI; Crum, Lillie-Blanton, & Anthony, 1996). Items consisted of statements about the presence of signs of disorder in the participant’s neighborhood (e.g., abandoned or boarded up buildings) with response options “true” and “false.” Cronbach’s alpha for the sample was 0.81.

Neighborhood cohesion.

The Neighborhood Cohesion Scale (NCS; Tolan, Gorman-Smith, & Henry, 2001) assessed participants’ social connection to neighbors (e.g., knowing the names of neighbors, feeling comfortable asking neighbors to watch home while away). Participants completed six items with response options “strongly agree,” “agree,” “neither agree nor disagree”, “disagree,” and “strongly disagree.” The sum of the items was calculated with higher scores indicating less neighborhood cohesion. Cronbach’s alpha for the sample was 0.92.

Socioeconomic status.

Socioeconomic status was measured with an item asking participants how much they would have to pay at school for a full lunch, with the following response options: “full price,” “a reduced price,” and “nothing (it would be free).”

Demographics.

Participants self-reported their race, ethnicity (Hispanic or Latinx or not), and age.

Court involvement history.

Participants self-reported history of arrest in the previous three months and reasons for their most recent arrest during this timeframe. Study staff recorded source of recruitment into the study at time of enrollment.

Analytic Plan

Analyses were conducted using SAS v9.4. The test of the association between sexual minority status and delinquency was conducted with negative binomial regression. Tests of the association between bullying exposure and emotion regulation difficulties and whether sexual minority status moderated that relationship were conducted with linear regression. Tests of indirect effects were conducted with the causal mediation framework (Imai, Keele, & Tingley, 2010). This framework provides clear definitions for direct and indirect effects using a counterfactual approach. Furthermore, causal mediation does not depend upon a specific statistical model and can be applied to a range of outcome types, including counts, as was used in the present analysis. PROC CAUSALMED was employed for testing direct and indirect effects as well as confounding for the mediator and outcome models. For each hypothesis test, childhood maltreatment, neighborhood environment, neighborhood cohesion, and socioeconomic status were included as potential covariates due to their relationship with at least two of the study constructs. Race, Latinx ethnicity, and age were also entered into the models as potential demographic covariates.

Missing data were treated as follows. For the emotion regulation difficulties and childhood maltreatment measures, means were calculated for participants who completed at least 60% of items. For the bullying exposure, delinquency, neighborhood environment, and neighborhood cohesion scales, sums were calculated for participants who completed at least 60% of the items, assuming zero scores for skipped items. Participants who were missing on constructs assessed with a single item or missing on an entire scale according to the criteria above were excluded from the analysis (N=18, 7%). The cases dropped due to missing data did not differ from the included cases on any of the study variables.

One participant reported “Undecided (Questioning)” as her sexual orientation, reported she had never had sexual contact, and did not answer the question regarding sexual attraction. This participant was excluded from analysis because her responses concerning sexual minority status were ambiguous. All other participants reporting “Undecided (Questioning)” as their sexual orientation reported either sexual contact with girls or attraction to girls and were classified as sexual minorities.

Results

Descriptive Statistics

The analytic sample consisted of 226 participants. Descriptive statistics are presented in Table 1. Consistent with previous literature, a high proportion of the sample (48%) was classified as sexual minority. Either Pearson’s chi-square test of independence or Fisher’s exact test, depending on cell size, was used to test differences based on sexual minority status for categorical variables. T-tests were used to test the differences between the two groups for continuous variables. More than twice the proportion of sexual minority participants as heterosexual participants reported being bullied in the past year, and a greater proportion of sexual minority participants reported experiencing each of the eight types of bullying victimization assessed, as compared to their heterosexual peers. At the bivariate level, sexual minority participants also reported higher levels of delinquent behavior in the previous three months and higher levels of emotion regulation difficulties and childhood maltreatment. There was no difference between the two groups in recent history of arrest.

Table 1.

Descriptive Statistics

	Total Sample	Sexual Minority	Heterosexual
Demographics
	M (SD)	M (SD)	M (SD)	t	p a
Age	15.58 (1.10)	15.58 (1.12)	15.57 (1.09)	−0.04	0.97
	n (%)	n (%)	n (%)	χ2 (1, 226)	p
Hispanic or Latinx ethnicity	109 (48%)	49 (45%)	60 (51%)	0.91	0.34
Race				--	0.02 b
American Indian or Alaskan Native	12 (5%)	7 (6%)	5 (4%)
Asian	3 (1%)	1 (1%)	2 (2%)
Black or African American	45 (20%)	20 (18%)	25 (21%)
Native Hawaiian or Pacific Islander	2 (1%)	0 (0%)	2 (2%)
White	69 (31%)	37 (34%)	32 (27%)
Multiracial	44 (19%)	28 (26%)	16 (14%)
Other	51 (23%)	16 (15%)	35 (30%)
				χ2 (2, 226)	p
School lunch payment status				2.32	0.31
Full price	50 (22%)	20 (18%)	30 (26%)
Reduced price	28 (12%)	16 (15%)	12 (10%)
Free	148 (65%)	73 (67%)	75 (64%)
Sexual attraction
Females only	12 (5%)	12 (11%)	0 (0%)
Females more than males	10 (4%)	10 (9%)	0 (0%)
Females and males equally	44 (19%)	44 (40%)	0 (0%)
Males more than females	41 (18%)	41 (38%)	0 (0%)
Males only	119 (53%)	2 (2%)	117 (100%)
Sexual behavior
Only with females	8 (4%)	8 (7%)	0 (0%)
Mostly females	3 (1%)	3 (3%)	0 (0%)
Equal males and females	10 (4%)	10 (9%)	0 (0%)
Mostly males	22 (10%)	22 (20%)	0 (0%)
Only males	86 (38%)	30 (28%)	56 (48%)
Never had sex	96 (42%)	36 (33%)	60 (51%)
Not reported	1 (<1%)	0 (0%)	1 (1%)
Sexual orientation
Heterosexual (straight)	128 (57%)	11 (10%)	117 (100%)
Homosexual (gay, lesbian, queer)	13 (6%)	13 (12%)	0 (0%)
Bisexual	70 (31%)	70 (64%)	0 (0%)
Undecided (questioning)	15 (7%)	15 (14%)	0 (0%)
Court involvement history
	n (%)	n (%)	n (%)		p
Recruitment source c					<0.01 b
Intake appointment	153 (68%)	63 (58%)	90 (77%)
Truancy court	19 (8%)	12 (11%)	7 (6%)
Drug court	2 (1%)	1 (1%)	1 (1%)
Court-appointed special advocate	8 (4%)	2 (2%)	6 (5%)
Group home	42 (19%)	29 (27%)	13 (11%)
Other	1 (<1%)	1 (1%)	0 (0%)
	n (%)	n (%)	n (%)	χ2 (1, 224)	p
Arrested in past three months d	57 (25%)	27 (25%)	30 (26%)	0.02	0.88
Reasons for most recent arrest in past three months e				χ2 (1, 56)	p
Assault (aggravated)	5 (9%)	3 (12%)	2 (7%)	--	0.65 b
Assault (simple)	25 (45%)	16 (62%)	9 (30%)	5.61	0.02
Curfew and loitering	2 (4%)	2 (8%)	0 (0%)	--	0.21 b
Disorderly conduct	18 (32%)	7 (27%)	11 (37%)	0.61	0.44
Larceny/theft	8 (14%)	1 (4%)	7 (23%)	--	0.06 b
Property crime	4 (7%)	3 (12%)	1 (3%)	--	0.33 b
Robbery	2 (4%)	1 (4%)	1 (3%)	--	1.00 b
Vandalism	5 (9%)	4 (15%)	1 (3%)	--	0.17 b
Other	10 (18%)	4 (15%)	6 (20%)	--	0.74 b
Study variables
	n (%)	n (%)	n (%)	χ2 (1, 226)	p
Ever bullied in past year	61 (27%)	45 (41%)	16 (14%)	21.83	<0.01
Bullying exposure type				χ2 (1, 226)	p
Hit, pushed, shoved at school	19 (8%)	15 (14%)	4 (3%)	7.84	0.01
Teased or called names	48 (21%)	38 (35%)	10 (9%)	23.36	<0.01
People started rumors about you	43 (19%)	33 (30%)	10 (9%)	17.29	<0.01
Left out on purpose	27 (12%)	25 (23%)	2 (2%)	24.17	<0.01
Bullied through text messages	43 (19%)	33 (30%)	10 (9%)	17.29	<0.01
Bullied on social media	35 (15%)	29 (27%)	6 (5%)	19.89	<0.01
Someone posted embarrassing photos or videos	19 (8%)	16 (15%)	3 (3%)	10.76	<0.01
Someone took pictures without permission	18 (8%)	16 (15%)	2 (2%)	12.95	<0.01
	M (SD)	M (SD)	M (SD)	t	p
Number of types of bullying reported	1.12 (2.18)	1.88 (2.71)	0.40 (1.12)	−5.28	<0.01
Childhood maltreatment	0.89 (0.86)	1.11 (0.92)	0.69 (0.75)	−3.71	<0.01
Delinquency in past three months	2.11 (3.20)	2.73 (3.86)	1.52 (2.28)	−2.85	<0.01
Emotion regulation difficulties	2.53 (0.78)	2.75 (0.70)	2.32 (0.80)	−4.23	<0.01
Neighborhood environment	1.67 (1.91)	1.85 (2.03)	1.50 (1.79)	−1.37	0.17
Neighborhood cohesion	22.91 (6.65)	22.56 (7.06)	23.24 (6.25)	0.77	0.44

Note. N =226, except where otherwise reported below.

^aDifferences in demographic and study variables between sexual minority and heterosexual subsamples assessed using Chi-Square Tests of Independence for categorical variables and t-tests for continuous variables.

^bDue to small cell sizes, Fisher’s exact test was conducted for these tests of independence.

^cN = 225 due to data missing for one participant.

^dN = 224 due to data missing for two participants.

^eParticipants could endorse more than one reason for most recent arrest. For these variables, Chi-Square tests of independence compared sexual minority and heterosexual samples among participants endorsing any arrest in previous 3 months (N = 56; data were missing for one participant).

Hypotheses Tests

Results of the test of association between sexual minority status and delinquency are reported in Table 2. Consistent with the first hypothesis, sexual minority status was associated with higher levels of delinquency when controlling for the potential covariates (Risk Ratio=1.52, 95% C.I.=1.02 – 2.24). Linear regression was used to test the association of bullying exposure and emotion regulation difficulties. Variables were entered into the model in two steps. In the first step, the independent variables (bullying exposure, sexual minority status) and covariates were entered. In the second step, the bullying exposure by sexual minority status interaction term was added. Final results for both models are presented in Table 3. Contrary to hypotheses, there was no evidence that bullying exposure was related to emotion regulation difficulties or that sexual minority status moderated this relationship. Thus, the component of the conceptual model based on the minority stress model was not supported. Without evidence of a direct relationship between bullying exposure and emotion regulation, there was no reason to test for an indirect relationship between bullying exposure and delinquency via emotion regulation. For this reason, Hypothesis 4 was not tested.

Table 2.

Association of Sexual Minority Status and Delinquency Among Court-Involved Adolescent Girls

	B	SE	95% CI		p
Effect			LL	UL
Independent variables
Sexual minority status a	0.42	0.20	0.02	0.81	0.04
Covariates
Age	−0.13	0.09	−0.32	0.05	0.15
Race b
American Indian / Alaskan Native	0.23	0.45	−0.65	1.10	0.61
Asian	0.57	0.78	−0.97	2.10	0.47
Black / African American	0.42	0.27	−0.11	0.94	0.12
Native Hawaiian / Pacific Islander	0.50	1.04	−1.55	2.55	0.63
Multiracial	0.26	0.27	−0.27	0.79	0.33
Other	0.01	0.32	−0.61	0.63	0.97
Ethnicity c	−0.40	0.22	−0.83	0.03	0.07
School lunch payment status d
Reduced price	0.40	0.33	−0.24	1.04	0.22
Free	0.15	0.24	−0.32	0.61	0.54
Childhood maltreatment	0.22	0.12	−0.01	0.45	0.06
Neighborhood environment	0.10	0.05	0.00	0.20	0.04
Neighborhood cohesion	0.00	0.01	−0.02	0.03	0.88
Intercept	1.96	1.50	−0.98	4.91	0.19
Dispersion	1.30	0.20	0.96	1.77	--

Note. N = 226.

LL = lower limit; UL = upper limit.

^a0 = heterosexual, 1 = sexual minority.

^bReference = White.

^c0 = non-Hispanic or Latinx, 1 = Hispanic or Latinx.

^dReference = full price.

Table 3.

Association of Bullying Exposure and Emotion Regulation Difficulties Among Court-Involved Adolescent Girls

	Model 1			Model 2
Effect	B	SE	p	B	SE	p
Independent variables
Bullying exposure	0.02	0.03	0.43	−0.04	0.06	0.52
Sexual minority status a	0.35	0.11	<0.01	0.30	0.12	0.01
Sexual minority status x bullying exposure	--	--	--	0.07	0.07	0.29
Covariates
Age	0.05	0.05	0.27	0.05	0.05	0.31
Race b
American Indian / Alaskan Native	0.12	0.24	0.61	0.11	0.24	0.63
Asian	0.50	0.45	0.27	0.50	0.45	0.26
Black / African American	0.16	0.15	0.27	0.14	0.15	0.33
Native Hawaiian / Pacific Islander	−0.31	0.55	0.58	−0.34	0.56	0.54
Multiracial	−0.05	0.15	0.75	−0.06	0.15	0.69
Other	−0.06	0.16	0.71	−0.07	0.16	0.65
Ethnicity c	−0.19	0.11	0.11	−0.18	0.11	0.11
School lunch payment status d
Reduced price	0.01	0.18	0.94	0.04	0.18	0.83
Free	−0.01	0.13	0.92	−0.02	0.13	0.89
Childhood maltreatment	0.06	0.07	0.35	0.06	0.07	0.37
Neighborhood environment	0.01	0.03	0.62	0.02	0.03	0.57
Neighborhood cohesion	−0.00	0.01	0.96	−0.00	0.01	>0.99
Intercept	1.51	0.79	0.06	1.60	0.79	0.04

Note. N = 226. LL = lower limit; UL = upper limit.

^a0 = heterosexual, 1 = sexual minority.

^bReference = White.

^c0 = non-Hispanic or Latinx, 1 = Hispanic or Latinx.

^dReference = full price.

Consistent with past literature, sexual minority status was significantly associated with difficulties with emotion regulation, suggesting potential partial support for the component of the model based on the psychological mediation framework – specifically, that emotion regulation difficulties accounted for the disparity in delinquency for sexual minority adolescent girls. The conceptual model held that the minority stress process of bullying exposure would contribute to delinquency via emotion regulation difficulties. To further explore the links among sexual minority status, delinquency and emotion regulation, an adjusted model evaluated whether sexual minority status was indirectly associated with delinquency via emotion regulation, as depicted in Figure 2. In this model, sexual minority status was used as a proxy for experiences of sexual minority stress processes not measured in the present study, including distal ones (e.g., discrimination based on sexual minority status and parental rejection) and proximal ones (e.g., internalized homophobia and expectations of victimization). Although this model is only a partial test of the psychological mediation framework, this approach has been used previously (e.g., Hatzenbuehler, McLaughlin, & Nolen-Hoeksema, 2008).

Figure 2.

Exploratory Conceptual Model Depicting Hypothesized Relationships Among Sexual Minority Status, Emotion Regulation Difficulties, and Delinquency

Based on the psychological mediation framework, sexual minority status was hypothesized to be associated with more delinquency indirectly via more emotion regulation difficulties. The parameter estimates for the causal mediation analysis are presented in Table 4. Sexual minority status was directly associated with delinquency (Natural Direct Effect: B=1.20, 95% C.I.=0 .75 – 1.64; SE=0.22; p<0.01), and indirectly associated with delinquency via emotion regulation difficulties (Natural Indirect Effect: B=1.32, 95% C.I.=1.09 – 1.56; SE=0.12; p<0.01). The estimate for percent mediated was 66.46% (95% C.I.=19.92 – 113.00; SE=23.74; p=0.01), and there was no evidence of interaction between sexual minority status and emotion regulation difficulties (B=11.36, 95% C.I.= −2.83 – 25.55; SE=7.24; p=0.12). In the outcome model in which emotion regulation difficulties were included as a mediator, the relationship between sexual minority status and delinquency was no longer significant.

Table 4.

Association of Sexual Minority Status and Delinquency via Emotion Regulation Difficulties Among Court-Involved Adolescent Girls

Variables	Mediator Model			Outcome Model
	B	SE	Wald χ2	B	SE	Wald χ2
Independent variable
Sexual minority status a	0.37	0.10	12.96**	0.18	0.19	0.90
Mediator
Emotion regulation difficulties	--	--	--	0.76	0.13	36.61**
Covariates
Age	0.05	0.05	1.07	−0.17	0.08	4.14*
Race b
American Indian / Alaskan Native	0.13	0.23	0.31	0.19	0.41	0.21
Asian	0.48	0.43	1.24	0.42	0.71	0.35
Black / African American	0.14	0.14	1.06	0.38	0.24	2.49
Native Hawaiian / Pacific Islander	−0.31	0.54	0.33	0.96	0.95	1.01
Multiracial	−0.06	0.14	0.17	0.39	0.25	2.46
Other	−0.07	0.15	0.19	0.17	0.29	0.34
Ethnicity c	−0.19	0.11	2.94	−0.24	0.20	1.39
School lunch payment status d
Reduced price	0.01	0.17	0.01	0.19	0.30	0.39
Free	−0.01	0.12	0.01	0.09	0.22	0.16
Childhood maltreatment	0.08	0.06	1.58	0.14	0.11	1.74
Neighborhood environment	0.01	0.03	0.21	0.10	0.04	5.51*
Neighborhood cohesion	−0.00	0.01	0.02	0.00	0.01	0.01
Intercept	1.64	0.74	4.84*	0.60	1.37	0.19
Scale	0.72	0.03	--	--	--	--
Dispersion	--	--	--	0.94	0.17	--

Note.

N = 226.

^a0 = heterosexual, 1 = sexual minority.

^bReference = White.

^c0 = non-Hispanic or Latinx, 1 = Hispanic or Latinx.

^dReference = full price.

^*p<0.05

^**p<0.01.

As a sensitivity analysis, the models were evaluated without the covariates, and there was no substantive change in the conclusions.

Discussion

The existing juvenile justice literature provides robust evidence that sexual minority adolescent girls are overrepresented in the juvenile justice system, with the results of a meta-analysis indicating that the percentage of adolescent girls in custody who are sexual minorities is nearly twice that found in community samples (Jonnson et al., 2019). Despite the consistent evidence of this disparity across multiple studies, mechanisms explaining why so many sexual minority adolescent girls are involved in the justice system have not been identified. Understanding the factors that precipitate these girls’ entry into the system and maintain their involvement is critical for prevention programs and tailored services for sexual minority adolescent girls currently in juvenile justice settings. Findings from this study indicate that higher levels of delinquency among sexual minority adolescent girls compared to their heterosexual peers may contribute to their disproportionate representation in the system and that difficulties with emotion regulation may account for their engagement in this behavior.

This study is the first to examine differences in delinquency based on sexual minority status among court-involved adolescent girls in particular. Similar to previous findings among a mixed-gender population-based sample (Himmelstein & Brückner, 2011) and a sample of unstably and under-housed young women (Frederick et al., 2011), sexual minority status was associated with more reported delinquency in this sample. This result contrast with findings from a previous study of court-involved adolescent boys and girls that found no differences in delinquency based on sexual minority status (Hirschtritt et al., 2018). This difference may be due to the court-involved sample for this study consisting entirely of adolescent girls. Just as the justice system overrepresentation based on sexual minority status is specific to adolescent girls (Jonnson et al., 2019), sexual minority adolescent girls may be at particular risk for delinquent behavior compared to their sexual minority male counterparts. This finding suggests that sexual minority adolescent girls may enter into the juvenile justice system at disproportionate rates because they engage in delinquent behavior at higher rates, thus placing them at risk for contact with the system. Research with community samples of adolescent girls is needed to shed further light on this risk factor.

This study also investigated a potential pathway to delinquency among sexual minority adolescent girls based on the minority stress model (Meyer, 2003) and the psychological mediation framework (Hatzenbuehler, 2009) with the hypotheses that bullying exposure was associated with emotion regulation difficulties and this relationship was stronger for sexual minority adolescent girls. Results from this analysis did not support either hypothesis. Given the importance of peer relationships during adolescence, bullying exposure has the potential to be a particularly salient form of minority stress for teenagers, and previous studies have found that sexual orientation-based peer victimization is associated with negative mental health outcomes for sexual minority adolescents such as substance use (Huebner et al., 2015) and depression or suicidal ideation (Espelage et al., 2008), but it is not the only deleterious form of minority stress (e.g., Mereish, Miranda, Liu, & Hawthorne, 2021).

The lack of an association in the present analysis may be due to limitations in the measure of bullying exposure, which did not specifically assess victimization on the basis of sexual minority status. Instead, the items used in this study measured types of bullying experiences regardless of the target’s perception of the perpetrator’s motivation. Conversely, the disparity in emotion regulation difficulties among sexual minority adolescent girls may result from the cumulative effect of multiple minority stress processes, and bullying victimization alone may not be sufficient to predict differences. Future studies assessing sexual minority status-based bullying are needed to clarify whether the null findings observed in this study were due to how the construct was measured or suggestive of a genuine lack of association.

Using the psychological mediation framework, an exploratory analysis of the relationships among sexual minority status, emotion regulation difficulties, and delinquency using a modified model evaluated whether sexual minority status was indirectly associated with delinquency via emotion regulation difficulties. This alternative model conceptualized sexual minority status as indicative of experiences of other, unmeasured minority stressors. This model included the assumption that the court-involved sexual minority adolescent girls in this sample had experienced distal and proximal minority stressors aside from bullying exposure. Previous findings that distal minority stress processes such as hearing homophobic remarks at school (Kosciw, Clark, Truong, & Zongrone, 2020) and family rejection based on sexual minority status (Ryan, Huebner, Diaz, & Sanchez, 2009) are common among sexual minority adolescents supported this premise. Furthermore, a recent daily diary study of sexual minority adolescents found that 82% of participants reported experiencing a sexual minority status-based stressor on at least one day during the 21 days assessed, and that, on average, participants reported experiencing such stressors on 4.52 days during that period (Mereish et al., 2021), providing evidence that these stressors are not only pervasive but frequent.

Findings from the exploratory analysis indicated that sexual minority status was associated with emotion regulation difficulties and delinquency and that emotion regulation difficulties mediated 66% of the relationship between sexual minority status and delinquency. This result suggests that court-involved sexual minority adolescent girls may engage in more delinquent behaviors due to difficulties with emotion regulation. Furthermore, sexual minority status was not directly associated with delinquency when emotion regulation was included in the model as a mediator, indicating that this construct plays a significant role in accounting for the observed disparity in delinquency.

The present study was not able to provide evidence for specific minority stress processes driving differences in emotion regulation. It is possible that such links will be found with more precise measures of minority stress. It is also possible that another mechanism accounts for the disparity in emotion regulation difficulties among sexual minority adolescent girls involved in the juvenile justice system. Although such a mechanism need not be specific to sexual minority adolescent girls, results from this analysis suggest emotion regulation is at the least associated with sexual minority status and delinquency. Therefore, results from this study point to building emotion regulation skills as a potential intervention target to prevent future delinquency among court-involved sexual minority adolescent girls.

The overrepresentation of sexual minority adolescent girls in the juvenile justice system (Jonnson et al., 2019) provided the impetus for the present analysis, which broadly sought to investigate a potential mechanism that may ultimately lead to these young women entering the system. The sample in this analysis was not recruited on the basis of sexual minority status, and so the finding that 48% of participants were sexual minorities provides further evidence of the disproportionality in the juvenile justice system. It warrants mention that adolescent girls who were already court-involved comprised the sample in this analysis, and, as such, findings may not generalize to all sexual minority adolescent girls in the community. Nonetheless, the mechanism identified in this analysis may indicate that emotion regulation difficulties increase court-involved sexual minority adolescent girls’ likelihood of continuing to engage in delinquent behaviors and thus place them at risk for continued involvement in the juvenile justice system.

This analysis focused on predictors of delinquency as a potential pathway to justice system involvement. Sexual minority adolescent girls reported higher levels of delinquency than their heterosexual peers, suggesting that engaging in more of this behavior may place them at greater risk for continued involvement. There is also evidence that sexual minority adolescents may face disproportionate punishment for the same types of behaviors as their heterosexual peers. One previous investigation found that the path from delinquent behaviors to incarceration was more than four times as strong for sexual minority adolescents than it was for heterosexuals (Poteat et al., 2016). Similarly, a population-based study of adolescents found that identifying as a sexual minority was associated with increased odds of reporting having been stopped by the police, being arrested before the age of 18, and being convicted before the age of 18 even when controlling for reported delinquent behavior (Himmelstein & Brückner, 2011). It is possible then that the disproportionality of sexual minority adolescent girls in the juvenile justice system stems from two overarching processes, the first of which drives greater rates of delinquent behavior among these girls, placing them at greater risk for contact with the system, and the second of which leads to them facing more severe consequences at those points of contact. This possibility further highlights the need for more investigations into the causes of justice system involvement for sexual minority adolescent girls as well as the development of prevention interventions.

Findings from the present analysis should be interpreted in the context of several limitations. First, the study used cross-sectional data. Measures assessed the constructs of interest in time periods ranging from more distal to more recent based on the hypothesized relationships (past year for bullying exposure; past three months for emotion regulation difficulties and delinquency) to approximate temporal order for the mediation relationship. Future investigations using longitudinal methods replicating these results are needed to strengthen evidence for the hypothesized mediation relationships. Second, the measure of bullying exposure was limited to types of bullying experienced and did not assess whether participants were targeted based on their sexual minority status, nor did it assess the frequency of bullying. Additional research with more robust measures of minority stress processes are needed to determine if they contribute to the differences observed in emotion regulation difficulties based on sexual minority status.

Third, as previously noted, the sample consisted of court-involved adolescent girls, and so the results of this analysis may not generalize to adolescent girls who have not had contact with the juvenile justice system. However, court-involved adolescent girls constitute a vulnerable population, and so the findings are nevertheless significant even if they do not generalize beyond this group. Furthermore, the sample was recruited through multiple sources including truancy court, drug court, and agencies serving adjudicated adolescents, which increased the variability in terms of participants’ previous and current level of court involvement. Fourth, this study was a secondary analysis of an existing data set that was not collected for this purpose, and so the sample size may have lacked sufficient statistical power to detect all of the relationships evaluated in it. Despite these limitations, these findings remain important for identifying a psychological process associated with delinquent behavior for court-involved sexual minority adolescent girls, who may be at risk for continued and more significant justice system involvement.

Conclusion

Although multiple investigations have found that sexual minority adolescent girls are disproportionately represented in the juvenile justice system, the research on processes that contribute to this disparity is relatively underdeveloped. This investigation found that court-involved sexual minority adolescent girls reported higher levels of delinquent behavior than their heterosexual counterparts, suggesting that one step in the causal path leading to system involvement may be that these girls engage in more behavior that puts that at risk for contact with the system. This study also evaluated whether bullying exposure, as a minority stress process, was associated with emotion regulation difficulties as a hypothesized mediator to delinquency. While there was no evidence of a relationship between bullying exposure and emotion regulation difficulties, results indicated that emotion regulation difficulties mediated the relationship between sexual minority status and delinquency, denoting that this psychological process may play an important role in the pathway to justice system involvement for sexual minority adolescent girls. Further research is needed to investigate drivers of emotion regulation difficulties for these girls, and the minority stress model identifies many other distal (e.g., family rejection) and proximal stress processes (e.g., concealment of sexual minority status) that warrant evaluation as potential precipitants in this causal chain. These lines of research inquiry are critically important to elucidate the mechanisms that account for sexual minority adolescent girls’ engagement in delinquent behavior and ultimately develop effective program and policy interventions to promote positive outcomes for this vulnerable group.