Evaluation of the expect respect support group program: A violence prevention strategy for youth exposed to violence

Article history: Received 26 October 2016 Received in revised form 4 May 2017 Accepted 7 May 2017 Available online 11 May 2017 In the present study, we assess the effects of the Expect Respect Support Groups (ERSG) on frequency of teen dating violence (TDV) andgeneral youth violence. ERSG is a school-based violence prevention program for youthwhohave been exposed to violence in their home, school, or community. Boys and girls (N= 1,678, Mage = 14.3, S.D. = 1.7, Range = 11–17) from 36 schools in Texas participated in this accelerated longitudinal (7-year trajectory) study beginning in 2011. Latent growth curve analyses were conducted using three waves of data from three crosssectional cohorts of adolescents. Among boys, the number of ERSG sessions attended related to incremental declines in psychological TDV perpetration and victimization, physical TDV victimization, sexual TDV perpetration and victimization, reactive aggression, and proactive aggression. Girls attending ERSG demonstrated reductions in reactive and proactive aggression. The present findings suggest ERSG may be an effective cross-cutting strategy to reduce TDV and other forms of violence among high-risk boys and possibly girls. This information provides valuable understanding of TDV and youth violence in high-risk populations andmay be useful in tailoring future prevention efforts to different groups of teens. © 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Teen dating violence (TDV) is a serious public health problem with chronic deleterious consequences including psychiatric symptoms, substance use/abuse, poor educational attainment, etc. (Kann et al., 2014;Banyard and Cross, 2008;Ackard et al., 2007;Smith et al., 2003;White and Smith, 2004). Currently, few programs have shown success in reducing/preventing TDV. Taylor and colleagues (2015) examined the effectiveness of the universal 'Shifting Boundaries,' program which includes a classroom-based curriculum and a building-level intervention (i.e., school-based restraining orders, increased security in violence "hot spots," and posters to increase awareness of sexual TDV). Findings indicated the building-level intervention reduced sexual TDV perpetration and victimization for boys and girls alike at six months postintervention, but the curriculum component of the intervention had no effect. This distinction between environmental (i.e., building-level) and individual (i.e., curriculum) components is important because it suggests that the program reduced the opportunity, rather than the propensity, to perpetrate sexual violence. Thus, the prevention effects of this program may be less likely to generalize across settings and be sustained over time compared to strategies that are also effective at changing attributes of the individual.
Conversely, two existing curriculum-based programs have shown promise in preventing TDV. The universal school-based 'Fourth-R' curriculum demonstrated efficacy in reducing physical TDV perpetration among boys over 30-month follow-up (Wolfe et al., 2009). Foshee and colleagues (Foshee et al., 2005) confirmed the effectiveness of the universal school-based 'Safe Dates' curriculum which reduced psychological TDV, moderate physical TDV, sexual TDV perpetration and moderate physical TDV victimization among both boys and girls over a four year follow-up period.
To date, the extant programs demonstrating promise for TDV prevention have been universal programs devised for the general population, while relatively little TDV research has been conducted among high-risk populations (Niolon et al., 2015). Yet, there is growing recognition that youth exposed to violence in the home and/or community are at particularly heightened risk not only for TDV, but multiple forms of subsequent violent delinquency (Crooks et al., 2011;Foshee et al., 2015;Hamby et al., 2012;Baskin and Sommer, 2014;Turner et al., 2016). For example, exposure to intimate partner violence (IPV) during childhood has been linked not only to TDV perpetration (Lichter and McCloskey, 2004)

Contents lists available at ScienceDirect
Preventive Medicine j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / y p m e d IPV perpetration as adults (Ehrensaft et al., 2003), but also to general aggression and violence outside of intimate relationships (Brown et al., 2002;Fishbein and Ajzen, 1975;Litrownik et al., 2003). Likewise, exposure to community violence has been linked to perpetration of both violence in the community and TDV (Farrell and Sullivan, 2004;Gorman-Smith et al., 2004;Malik et al., 1997). Given the considerable overlap of risk for disparate types of violence (Wilkins et al., 2014;Klevens et al., 2012;Tolan et al., 1995), it is not surprising that exposure to one form of violence would generalize to the perpetration of multiple forms of violence.
There are likely numerous mechanisms through which exposure to violence confers risk for subsequent perpetration of violence. For instance, witnessing aggression may provide a model for the learning and acceptance of aggressive behaviors by children (Abbey and Jacques-Tiura, 2009). Regardless of the type of exposurewhether witnessing community violence, IPV between parents, or being a victim of physical or sexual abusethe consequences are similar: exposed youth are more likely to be diagnosed with a psychological disorder and show difficulties with early attachment, emotion regulation, peer relationships, school adjustment, pro-social behaviors, and endorse attitudes condoning violence (Farrell and Sullivan, 2004;Cyr et al., 2010;Guerra et al., 2003;Wekerle and Wolfe, 2003). And, each of these risk factors increases the propensity for perpetrating violence against strangers, peers, and dating partners (Lichter and McCloskey, 2004;Ehrensaft et al., 2003;Herrenkohl et al., 1997;Lansford et al., 2007;Moynihan et al., 2011;Smith and Thornberry, 1995). Thus, there is reason to suspect youth exposed to violence may represent a unique population demonstrating rates of multiple forms of violence, including TDV, that differ from the general population. This heightened risk for multiple forms of violence is an important consideration as these high-risk youth may require the intensity of a selective intervention in place of, or in combination with, universal strategies. As such, identification of targeted cross-cutting violence prevention programs for youth exposed to violence is warranted and necessary (Wilkins et al., 2014).
In its efforts to build the evidence base for programs that prevent males' sexual violence perpetration, CDC identified the Expect Respect Support Group (ERSG) intervention (Ball et al., 2009(Ball et al., , 2012) as a targeted program already in implementation and ready to undergo evaluation. ERSG was initially developed as an empirically-informed TDV prevention program targeting high-risk adolescents with a history of violence exposure. However, the ERSG program addresses a number of shared risk factors for multiple forms of violence (e.g., attitudes condoning violence, emotion regulation, dynamics of health peer relationships) (Ball et al., 2009(Ball et al., , 2012, and there is therefore reason to suspect it may impact multiple forms of violence among its high-risk participants. 1 Notably, an initial pilot study using a single-group design indicated youth demonstrated increases in positive dating behaviors and decreases in emotional and physical TDV from pretest to posttest (Ball et al., 2012). However, this study did not look at other forms of violence. Given these preliminary findings, CDC's focus on identifying cross-cutting violence prevention strategies (Wilkins et al., 2014), and the current state of implementation of this program, a more rigorous evaluation of the ERSG program's effects (on multiple forms of violence) is warranted. In the present study, we evaluate the effects of ERSG on TDV, reactive aggression, and proactive aggression. We hypothesized that participation in the ERSG intervention would be associated with significant reductions in all three forms of violence.

Participants & procedures
Students from 36 schools in Texas were referred by school counselors or social workers for screening if they suspected, for any reason, the student had been exposed to violence in the home, school, and/or community. During brief intake assessments, semi-structured interviews were conducted to assess if students had ever been the witness, victim, or perpetrator of 1) TDV, 2) IPV between parents, 3) peer violence, 4) child abuse, and/or 5) some other form of violence in the home or community. Youth that verbally endorsed at least one type of violence exposure and were 11-17 years old were eligible to participate in the study. Most students (73%) reported multiple forms of violence exposure. However, this likely an underestimate of violence exposure as we did not exhaustively query each type of violence exposure. 2 Students were considered eligible for participation if they 1) endorsed history of violence exposure, 2) had not previously participated in ERSG support groups, 3) were not deemed to need a more significant care, and 4) were between the ages of 11 and 17. Of the 2380 students referred for screening, 1678 (M age = 14.3, S.D. = 1.7, Range = 11-17) met inclusion/exclusion criteria, assented, and completed baseline surveys; 1278 students (76.2%) completed surveys at time 2; and 906 students (54.0%) at time 3. Attrition rates were comparable across the treatment conditions (χ(1) = 0.020, η = 0.003; p = 0.89) and are consistent with longitudinal school-based research (Foshee et al., 2005;Horton and Lipsitz, 2001). There were no significant differences between study completers and non-completers on baseline variables. All students with reactive and proactive aggression data were included in analyses of these outcomes. Only students reporting a history of dating in the preceding three months at one or more waves of data collection were included in the analysis of TDV indices. The analytic dating sample included 1330 students in grades 6-12 (M age = 14.3, SD = 1.6, Range = 11-17). Fig. 1 displays selection procedures for the analytic samples. Overall, the sample was generally representative of the participating school districts in terms of race primarily comprising minority populations. Table 1 provides the breakdown of students by ethnicity and grade level.
ERSG is the core component of a larger comprehensive, multi-level program including school-wide prevention through policy, staff training and education, and youth leadership development provided by the SAFE Alliance: a nonprofit child abuse, domestic violence, and sexual assault victim service provider. This evaluation was focused only on the support group (ERSG) component of the program. The treatment condition included 24 schools already receiving the ERSG intervention. Students in ERSG schools attended up to 25 weekly structured group support sessions focused on developing healthy relationship skills and modifying maladaptive norms about dating behavior.
The curriculum includes five multi-session units: 1) Developing Group Skills (creating group guidelines and practicing communication skills); 2) Choosing Equality and Respect (learning about qualities of healthy relationships, defining abuse and respect, questioning gender stereotypes, and recognizing the use and abuse of power); 3) Recognizing Abusive Relationships (naming the violence, recognizing the impact of violence in your life, and understanding the dynamics and warning signs of dating abuse); 4) Learning Skills for Healthy Relationships (handling jealousy and anger, setting boundaries, asking for consent, resolving conflicts,