This cross-sectional analysis used data collected as part of a cluster-randomized control trial evaluating the effectiveness of a bystander intervention to reduce SVP in 26 Kentucky high schools (two schools in each of thirteen geographic regions across the state) from 2010-2014 (U01CE001675). For this analysis only self-administered survey data from students attending the 26 participating high schools and completed between Jan and May 2011 (before the intervention) were used; 21,608 were present on the day of surveying across all schools (Coker, Bush, Brancato, Clear, & Recktenwald, 2018; Coker et al., 2017).

The survey was a 99-item, paper and pencil, anonymous questionnaire that took approximately 25-40 minutes to complete. The survey was conducted using the model familiar to schools in the Youth Risk Behavior Survey (YRBS) (Brener et al., 2013). Study personnel administered surveys during school hours. Resources were included, such as websites and toll-free numbers for national agencies available at all times to address domestic violence, sexual assault, depression, or suicidal ideation at the end of the survey booklet and on the pencils used to complete the survey.

All parents and guardians of enrolled students were mailed a letter describing the study at least two weeks prior to surveying. Parents and guardians who did not want their student to complete the survey were asked to call or email the study staff with the name and school of their child. Study staff worked with teachers to identify those students in individual classrooms.

Study staff read elements of the assent form to students who were directed to review the full assent in the survey booklet. Those opting not to complete the survey or those whose parent(s) did not grant permission for participation were asked to quietly remain in their seat. These students received a booklet with the same cover as their peers but no questionnaire was included. Study staff provided the same instructions and assent information to all students; students who did not wish to participate were asked to indicate so on the forms. Study protocol was approved by the Institutional Review Board.

The questions about sexual violence were prefaced with the following introduction: “The next questions are about sexual activities. Some of the questions might make you uncomfortable. Remember that this survey is anonymous. Your name will not be linked to your answers. You may also skip questions that make you uncomfortable.” Three items were used to measure SVP: 1) “In the last 12 months, how many times have you had sexual activities with a high school student because you either threatened to end your friendship or romantic relationship if they did not or because you pressured the other person by arguing or begging?” 2) “In the last 12 months, how many times have you had sexual activities with a high school student because she/he were drunk or on drugs?” 3) “In the last 12 months how many times have you had sexual activities with another high school student by threatening to use or used physical force (for example, twisting their arm or holding them down)?” Response options for all items ranged from never, 1-2 times, 3-5 times, 6-9 and 10 or more times, and yes, but not in the past 12 months. An indicator variable was created for each item to represent the proportion reporting each tactic occurring one or more times in the past 12 months (all dichotomous yes/no variables).

The same three items were rephrased to determine SVV: 1) “In the last 12 months, how many times have you had sexual activities even though you didn’t really want to because either they threatened to end your friendship or romantic relationship if you didn’t or you felt pressured by the other person’s constant arguments or begging” 2) “In the last 12 months, how many times have you had sexual activities when you didn’t want to because the other person threatened to use or used physical force (like twisting your arm, holding you down) if you did not agree?” 3) “In the last 12 months, how many times have you had sexual activities when you didn’t want to because you were drunk or on drugs.” The ordinal response options for these items were similar to the SVP questions, and we created dichotomous indicator variables for those who reported they had experienced each tactic one or more times in the past 12 months.

Mutually exclusive categories of SV experience were created as follows: 1) both perpetration and victimization (SVP & SVV), 2) perpetration alone (SVP-only), 3) victimization alone (SVV-only), and 4) the comparison group with no SVP nor SVV (no SV).

The following question, based on past YRBS surveys, was used to define self-reported symptoms of depression: “During the past 12 months, did you ever feel so sad or hopeless almost every day for 2 weeks or more in a row that you stopped doing some usual activities.” A yes response was used to define depressive symptoms in the past 12 months.

Symptoms of post-traumatic stress (PTSS) were defined using the following two questions with the introduction: “Have you personally ever had any experience that was so frightening, horrible or upsetting that in the past month you: 1) had nightmares about it [experience described] or thought about it when you did not want to? 2) Tried hard not to think about it or went out of your way to avoid situations that reminded you of it?” Post-traumatic symptoms were defined for this analysis as a yes answer to both of the above two questions.

Several demographic variables were included in the survey. Sex (male or female) and grade in school (9^th-12^th) were analyzed as asked. Race was asked using the question, “How would you describe yourself?” The response options were: “American Indian or Alaska Native,” “Asian,” “Black or African American,” “Hispanic or Latino/Latina,” “White,” and “Other.” Students were allowed to select only one option. For the analysis, responses were dichotomized to “White” and “Non White.”

Participants were asked about sexual minority status, dating/relationship status, sexual activity status, and witnessing parent-partner violence. Sexual minority status was operationalized with the following item addressing sexual attraction: “People are different in their sexual attraction to other people. Which best describes your feelings?” The response options were: 1) “only attracted to females,” 2), “mostly attracted to females,” 3) “equally attracted to females and males,” 4) “mostly attracted to males,” 5) “only attracted to males,” and 6) “not sure.” Students reporting exclusive heterosexual attraction were coded as such, an indication of sexual majority status. All others were coded as not exclusively attracted to the opposite sex, indicating sexual minority status. We assessed current dating or relationship status with this question: “During the past 12 months, have you been in a relationship with a boyfriend or girlfriend? By a relationship we mean either having a partner for a planned event like a school dance or going to the movies, having a sex partner, or hanging out in a group with a boyfriend or girlfriend.” Response options were collapsed to indicate whether the student had or had not been in a relationship in the past 12 months. Because sexual violence is highly correlated with sexual activity, students were grouped in to sexually active and not sexually active based on their response to the question, “how old were you when you had sexual intercourse for the first time?” Those responding that they had never had sexual intercourse were coded as not sexually active and all others were coded as sexually active. We measured parent partner violence with the item: “In your family, how often did you see or hear one of your parents or guardians being hit, slapped, punched, shoved, kicked or otherwise physically hurt by their spouse or partner?” Response options were dichotomized as no or yes.

Of the 21,608 students present on the day of the survey, 2,624 refused (12.15%; 11.65% student refusals; 0.5% parental refusals). Incomplete responses were defined as completing less than 30 of 99 items, and incomplete responses were excluded from analyses (n=1857; 8.59%).

Of the remaining 17,126 students, 38 were excluded because they were not in grades 9-12, 17 did not report a sex, and the remaining students were missing data on depression or PTSS (n =235) or current relationship status (n=34). The final analytic sample included 16,784 students.

There were no significant demographic differences between those who agreed and completed the survey and those who did not with the following exception: males were twice as likely (p<.0001) as females to refuse to complete the survey.

The frequency of SVP was estimated for all students and by specific items: coerced sex, drug or alcohol facilitated sex, and physically forced sex. SVP was investigated within and across the three items (Table 1) among all students and by sexual attraction (sexual minority and majority status). Demographic and violence risk factors potentially correlated with SVP (with and without SVV) were investigated (see Table 2). Multivariate analyses were used to determine the association between SVP and SVV and symptoms of depression (Table 3) and PTSS (Table 4). Sex (male, female) and sexual attraction were investigated as modifiers for the association between SVP and the two outcomes of depressive symptoms and PTSS in the past 12 months by including an interaction term in multivariable models. Because effect modification was identified analyses were conducted for all students and by sexual attraction.

Comparisons were made using generalize estimating equations, which allow for the relaxation of independence conditions (students attending the same school), adjustment for the study design by condition (intervention v control assignment at the school level) and for the control of potential confounding variables. Two models were used to estimate the association between experiencing SVP with and without co-occurring SVV and symptoms of depression and PTSS in the past 12 months. Models were used to estimate the prevalence rate ratios of having symptoms of depression and PTSS in the past 12 months given SVP and SVV experiences. These models included adjustment for potential confounders and the parent randomized control study design by incorporating the condition as a covariance and including the sampled high school (n=26) in generalized estimating equations (Coker et al., 2017). All analyses were conducted using Statistical Analysis Software, SAS, version 9.4; SAS Institute; Cary, North Carolina.

Although current evaluation protocols for pediatric health professionals who examine patients with concerns for PTSD or depression recommend asking about a child or adolescent’s history of SVV, but omit questions about SVP (Cohen, 2010; King & AACAP Work Group on Quality Issues, 1997). Most trauma history screeners include sexual victimization experiences and few, if any, dimensionalize sexual violence experiences to include perpetration, even though this behavior is well recognized as a response to maltreatment in childhood (Glasser et al., 2001; Seto, Babchishin, Pullman, & McPhail, 2015). For instance, the American Academy of Pediatrics (AAP) provides guidance to health care professionals regarding screening and assessment of sexualized behavior, and victimization in youth (Kellogg, 2009). The clinical report on the evaluation of sexual behavior suggests that developmental considerations, types and frequency of sexualized behavior, parental response to the behavior, situational factors, access to sexually explicit material, home environment and history of abuse or neglect are important domains of interest that influence sexual behaviors and perceptions of healthy sexual boundaries. Including perpetration experiences into a comprehensive assessment of sexual and mental health, including child traumatic stress symptoms facilitates the process of differential therapeutics, and the tailoring of interventions to meet the needs of a diverse population of adolescents.

Current SVP prevention initiatives (e.g. creating safe environments, social-emotional learning, teaching health and safe dating skills, promoting healthy sexuality, empowerment based training) target the social ecology of sexual violence, including the community and relationship factors that influence individual behavior, in order to reduce the rates of SVP (Basile et al., 2016; DeGue et al., 2014). Implementing changes in protocols for community and school reporting and responding to sexual violence, such as advocated by Shifting Boundaries, can reduce both sexual harassment and peer sexual violence (Taylor, Stein, Mumford, & Woods, 2013). Psychoeducation establishes the physical and psychological foundations for healthy intimacy by communicating to youth their rights and responsibilities for their own bodies and behaviors in order to facilitate safety in close relationships with others (Roberts, Rosario, Corliss, Koenen, & Austin, 2012). Teaching about healthy relationships and intimacy in programs such as Safe Dates can build skills needed to resolve conflicts and decrease attitudes that contribute to peer and sexual violence (Foshee et al., 1998; Foshee et al., 2000). Bystander interventions, such as Green Dot or Coaching Boys into Men, and other socio-environmental approaches can interrupt the pathways to and between victimization and perpetration (Coker et al., 2017; Miller et al., 2012; Wolfe et al., 2009). Including additional information about emotional health in prevention protocols, such as psychoeducation about depression and PTSD, may strengthen the ability to reach potentially vulnerable adolescents.

Further work is needed to ensure access to effective individual treatments for those adolescents perpetrating sexual violence (Borduin, Henggeler, Blaske, & Stein, 1990). Because trauma symptoms may be present in those with both victimization and perpetration experiences, evidence-based trauma interventions could address both PTSD and alterations in mood presenting as depressive symptoms. Recognizing, responding to, and intervening to address SVP can allow both early treatment of depression and PTSD and addressing the individual risks of SVP to reduce subsequent repeated sexual assaults.

Another unique contribution of this paper was our examination of SVP and both mental health outcomes within sexual minority and majority groups. Prior studies have found SVV rates to be higher among adolescents of sexual minority (e.g., LGBTQ+) compared with adolescent of sexual majority (i.e., exclusively heterosexual) (Halpern, Young, Waller, Martin, & Kupper, 2004; Martin-Storey & Fromme, 2016; Pathela & Schillinger, 2010). In this study, sexual minority youth reported higher rates of SVP and SVV than sexual majority youth, yet the association between SV and both mental health outcomes was attenuated compared to that among sexual majority students.

Rates of depression were higher for sexual minority youth, even among those who had not experienced SV (43.3%), compared with sexual majority youth. Although the rates were high, they are consistent with prior studies that have found high rates of depression and suicidality among sexual minority youth. Recent studies, for example, found 40% of sexual minority adolescents had contemplated suicide within the past year (Caputi, Smith, & Ayers, 2017; Remafedi, French, Story, Resnick, & Blum, 1998). Trauma symptoms in sexual minority youth who were not exposed to sexual violence within the past year were also higher than non-exposed sexual majority youth, which could represent other pathways to post-traumatic symptoms, such as peer victimization, family violence, or child maltreatment (Friedman et al., 2011; Ybarra, Mitchell, Palmer, & Reisner, 2015).

Adolescents who identify as sexual minorities may need additional support, given the higher rates of trauma and depression symptom expression noted in this study compared with sexual majority students. Sexual minorities seek treatment at higher rates than their counterparts, suggesting opportunities for screening and intervention (Grella, Cochran, Greenwell, & Mays, 2011). Future research is needed to further elucidate the interactions between SVP and SVV in sexual minorities in particular, so that existing prevention and treatment approaches can be tailored to capture teachable moments that influence sexual behavior. Given that primary and secondary prevention strategies have shown promise to reduce SVP, additional research is needed to develop and test programming specifically including sexual minority adolescents (DeGue et al., 2014; Miller et al., 2013).

As with all studies, the present study had limitations of note. Adolescent self-reports may be limited by recall of experiences or emotions. Sampling was limited to adolescents who are within the regular school environment; there may be differences in those who are not in traditional educational settings. Because our questionnaires used descriptive wording familiar to most high school students, some of the questions could have been understood differently by different students. For instance, because we were focused on peer perpetration experiences, our questions asked about perpetration specifically against another high school student, but could have been interpreted to exclude perpetration against those peers who were not enrolled in a school environment and other perpetration experiences. Youth trauma responses could have fallen outside those symptoms tapped by the trauma history screener used, if so posttraumatic stress responses may be underestimated. As this is a cross-sectional design, the associations do not establish a causal relationship between sexual violence and emotional distress (depression or post-traumatic symptoms) and may reflect pre-existing mental health concerns that place adolescents at risk for sexual violence and/or sexual violence associated with increased risk of mental health issues. However, given the dearth of information on SVP and mental health in the literature and the need to improve both screening and interventions, our study fills a gap that may have important implications for adolescent mental health. Finally, while the diversity of our sample was reflective of the surrounding community, only 17.7% of our sample identified as non-white and 13.9% as sexual minority. Given our sample’s limited diversity, we did not have sufficient power for subgroup analyses investigating associations between SVP and either mental health outcome.