PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Adolesc Health. Author manuscript; available in PMC 2014 March 1.
Published in final edited form as:
PMCID: PMC3579999
NIHMSID: NIHMS401500

The Impact of Family and Peer Protective Factors on Girls’ Violence Perpetration and Victimization

Rebecca J. Shlafer, PhD,1 Barbara J. McMorris, PhD,1,2 Renee E. Sieving, PhD, RN, FSAHM,1,2 and Amy L. Gower, PhD3

Abstract

Purpose

This study investigates whether family and peer connections and prosocial norms buffer adolescent girls’ violence involvement, and whether a youth development intervention augments the power of these protective factors in reducing girls’ risk for violence.

Methods

Data were obtained from 253 13–17 year-olds enrolled in a randomized controlled trial of Prime Time, a youth development intervention offered through urban clinic settings to girls at high risk for pregnancy. Participants completed an A-CASI survey at baseline, 6, 12, and 18 months following enrollment. Protective factors included scales assessing family and peer connections and prosocial norms. Outcome variables were violence victimization and perpetration scales measured at 18 months.

Results

Family connections and prosocial norms independently protected girls against violence involvement. Peer prosocial norms also served as a protective buffer against violence perpetration and victimization; however, girls with strong peer connections had higher levels of violence perpetration. Participation in Prime Time augmented the protective effects of family and peer connections on girls’ violence victimization but not perpetration. Prime Time participants who had high levels of family connections reported the lowest levels of violence victimization at 18 months. Prime Time participants with strong peer connections trended toward lower levels of violence victimization than other girls.

Conclusions

Results suggest that effects of the Prime Time intervention on violence victimization were optimized among high-risk adolescent girls with strong connections to family and peers. The intervention was most potent in preventing violence victimization among girls with strong prosocial connections to family and peers.

Keywords: youth development, intervention, violence perpetration, violence victimization, adolescent girls, family, peers, protective factors

Introduction

In the last decade, increasing rates of violence among adolescent girls have garnered attention from scholars, practitioners, and policymakers.1 Girls are frequently exposed to and involved in violence at school, home, and in their communities.2 Between 1985 and 2007, the number of delinquency cases involving female youth increased by 101% while the number of cases involving male youth increased by only 30%.3 According to the national Youth Risk Behavior Survey, nearly one in four female 9th–12th grade students were involved in a fight during the last year.4 In Minnesota, 13% of high school girls were victims and 15% were perpetrators of physical violence in the last year.5

Violence perpetration and victimization have important consequences for adolescents’ physical and mental health. In 2008, over 650,000 youth aged 10–24 years were treated in emergency departments for violence-related injuries.6 Violence involvement has been linked to depression, emotional distress, externalizing behaviors, pregnancy and childbearing during adolescence.7,8 Further, violence victimization has been related to post-traumatic stress disorder, depression, and suicide among teens.911

The prevalence of violence among adolescent girls and the serious negative consequences associated with violent behavior provide a compelling rationale for identifying factors that help to prevent both victimization and perpetration. Within their daily social contexts, particular aspects of family and peer relationships are known to protect girls from violence involvement. For example, using National Longitudinal Study of Adolescent Health (Add Health) data, Resnick and colleagues12 found that high levels of parent-family connectedness along with clear parental expectations for academic achievement were protective against violence involvement among 9–12th grade boys and girls. In subsequent gender-specific analyses, Resnick and colleagues13 confirmed that family connectedness buffered girls from violence involvement using longitudinal data. Research also suggests that having friends with prosocial norms reduces adolescents’ risk for violence involvement. In a sample of African American youth, Smith and colleagues14 found that adolescents who had close relationships with their parents were most likely to select prosocial friends; in turn, having prosocial friends decreased adolescents’ violence involvement.

A decade ago, Youth Violence: A Report of the Surgeon General15 called for the adoption of evidence-based approaches to preventing youth violence, including strategies that employ a dual approach of building protective factors that buffer young people from violence involvement, in addition to addressing risks. This call for a dual approach is reinforced by a body of research demonstrating compensatory effects of protective factors within individual, relational and environmental domains on cumulative risk for violent behavior.16 Despite the need for evidence-based approaches to youth violence prevention, there is a dearth of evidence regarding effective strategies for reducing violence involvement, particularly among girls.1

In the current study, we examine violence outcomes among participants in Prime Time, a youth development intervention targeting adolescent girls at high risk for early pregnancy and violence involvement. A randomized controlled trial of Prime Time involved girls recruited from urban primary care clinics. At the end of the 18-month intervention, the intervention group reported less relational aggression than controls.17 While no overall intervention effects were seen for physical violence,17 qualitative analyses of intervention summary reports18 led to a hypothesis that the intervention was helpful in reducing violence among participants who had existing prosocial supports.

We address two questions related to protective factors buffering high-risk adolescent girls from violence involvement in the present study. First, we examine whether known protective factors in general youth populations – including strong connections to family and peers, as well as prosocial family and peer norms – act as buffers against violent behaviors among adolescent girls for whom exposure to violence within their daily contexts is commonplace.18 Next, we investigate whether a youth development intervention designed for use by primary care clinics augments the power of these protective factors in reducing girls’ risks for violence involvement.

Methods

Participants

Adolescent girls at high risk for early pregnancy were the focus of this study. The sample consisted of sexually active 13–17 year old girls recruited from four clinics. Girls met one or more of the following risk criteria, assessed through a brief screening tool:19,20 clinic visits involving negative pregnancy test or treatment for sexually transmitted infection; young age (i.e., 13–14 years); aggressive and violent behaviors; sexual and contraceptive risk behaviors; and behaviors indicating school disconnection. Girls who did not understand consent materials, were married, pregnant or had given birth were not eligible. Of 1270 girls who completed study screening, 571 (45%) met eligibility criteria. Of these, 253 (44%) agreed to participate and provided written informed consent.17

After consent was obtained, participants provided baseline data via audio computer-assisted self-interview (A-CASI) and were randomized into intervention (n=126) and control (n=127) conditions. All participants completed an A-CASI survey at baseline, 6, 12, and 18 months following enrollment. All study protocols were approved by institutional review boards of the University of Minnesota and participating clinics.

Study attrition was minimal, with 94.5% (n=239) of participants completing an 18-month follow-up survey. An attrition analysis yielded very few significant differences in baseline characteristics between participants lost to follow-up and those completing the 18-month assessment.17 The study sample’s demographic characteristics are summarized in Table 1. There were no significant differences in baseline demographic indicators between intervention and control groups completing the 18-month survey,17 although group differences in violence perpetration were noted at baseline (see Table 2).

Table 1
Demographic Characteristics of Prime Time Participants by Intervention Condition.
Table 2
Descriptive Statistics for Violence Involvement at Baseline and 18 months, by Intervention Condition

Intervention

Guided by a resilience paradigm22 and social cognitive theory,24 Prime Time sought change in selected psychosocial attributes associated with pregnancy risk behaviors.22,24 Girls were involved in intervention programming for 18 months. All programs were led by case managers experienced in working with urban teens from diverse cultural backgrounds. Intervention components are summarized below; detailed information is provided elsewhere.20,21

Case Management

The overall goal of Prime Time case management was to establish a trusting relationship in which a teen and her case manager worked together to address attributes targeted by this intervention.25 One-on-one visits focused on core topics including social-emotional skills, healthy relationships, and positive family and peer involvement. As a client-centered approach, the capacities, interests and needs of individual participants determined topics employed during any particular visit. Monthly visits occurred for the duration of the 18-month intervention in locations convenient for individual participants.

Peer Leadership Programs

Designed to complement case management, youth leadership groups provided hands-on skill-building experiences.

Peer Educator Training

The goal of this component was to provide opportunities for positive social involvement by engaging participants as peer health educators. Peer educator training employed a standard 15-session curriculum addressing life skills, healthy relationships, and sexual decision-making. Staff leading the training focused on creating a group atmosphere that was safe, accepting, and valued sharing and listening. Through weekly homework, participants explored training topics with adult family members. Starting with their first training session, girls were charged to reach and teach others outside of their peer educator group. Girls received $5 for each documented contact, for up to 50 contacts. After completing the peer educator training curriculum, girls engaged in a 7-session group teaching practicum.

Service Learning

This component focused on expanding participants’ social-emotional skills and their real-world experience in youth. An initial unit focused on building group cohesion and identifying participants’ leadership skills. In a second unit, groups explored community needs, assets, and potential service projects. A third unit consisted of implementing groups’ service projects. Each implementation session included a group reflection highlighting the impact of service on recipients and on group members themselves.

Measures

Family and Peer Protective Factors

We examined four key protective factors: family prosocial norms, peer prosocial norms, family connections, and peer connections. Protective factors were assessed via A-CASI surveys at baseline, 6, 12, and 18 months.

Prosocial Norms

Family prosocial norms were measured using a 5-item index (α=0.40–0.73 across time points) assessing the degree to which participants felt their families would be upset if they were involved in various antisocial behaviors (e.g., “Would your family be upset with you if you beat someone up?”). At each time point, responses (no=0, yes=1) were summed to create an index (Range=0–5) with higher scores indicating stronger prosocial norms.

Peer prosocial norms were measured using a 5-item index (α=0.60–0.69) assessing the degree to which participants felt their friends would be upset if they were involved in antisocial behaviors (e.g., “Would your friends be upset with you if you used a weapon to hurt someone?”). Responses (no=0, yes=1) were summed to create an index (Range=0–5) with higher scores indicating stronger prosocial norms.

Connections

A 5-item scale (α=0.90–0.92), adapted from an Add Health measure,26 assessed the degree to which participants felt connected to their families (e.g., “My family understands me”). At each time point, responses (not at all=0, a little =1, some=2, a lot=3) were averaged across items; higher scores indicated greater connection to family (Range=0–3).

Similarly, a 6-item scale (α=0.89–0.90) assessed the degree to which participants felt connected to their peers (e.g., “My friends care about me”). At each time point, responses (not at all=0, a little=1, some=2, a lot=3) were averaged across items; higher scores indicated greater connection to peers (Range=0–3).

Outcome Measures: Violence Victimization and Perpetration

A 4-item scale (α=0.77 at 18 months), adapted from an Add Health measure,13,26 assessed violence victimization in the past 6 months (e.g., having had a weapon used or threatened to be used against you). Responses (never=0, once or twice=1, 3–5 times=2, 6 or more times=3) were summed across items; higher scores indicated greater violence victimization (Range=0–9).

Similarly, a 5-item scale (α=0.79 at 18 months) assessed perpetration of physical violence in the past 6 months (e.g., having hurt someone badly enough to need bandages or care from a doctor or nurse). Responses (never=0, once or twice=1, 3–5 times=2, 6 or more times=3) were summed across items; higher scores indicated greater violence perpetration (Range=0–15).

Analysis Methods

Because violence victimization and perpetration measures were skewed, we utilized natural log transformed versions of these measures.27 Descriptive statistics for these measures are presented in Table 2. For data reduction purposes, we averaged family and peer prosocial norms scores and family and peer connections scores across the four survey points to create composite measures characterizing the average family and peer contexts that girls experienced during 18 months (Table 3). Using a median split, girls were assigned to high (1) or low (0) groups for each composite measure (family norms, peer norms, family connections, peer connections). Interaction terms were created by multiplying the median split groups for connections and norms by intervention status (intervention (1), control (−1)).28

Table 3
Descriptive Statistics for Protective Factors

We utilized a stepwise regression approach to examine the influence of family and peer protective factors on violence outcomes at 18 months by entering family or peer predictors into a series of equations predicting each violence outcome, controlling for the baseline measure of each outcome. Our strategy was first to examine main effects of intervention status and either family or peer context variables on a given violence outcome, controlling for that outcome measured at baseline. We then stepped in interaction terms to determine whether relationships between family/peer context variables and the violence outcome were moderated by intervention participation. For example, violence perpetration at 18 months was first regressed on intervention status, family connections, and family prosocial norms, controlling for violence perpetration at baseline. Then, interaction terms (family connections X intervention, family prosocial norms X intervention) were added into the model.

Regression models were estimated using generalized estimating equations (GEEs) in SAS Version 9.2, to allow for adjustment of standard errors for inter-correlations between girls recruited from the same clinic.29,30 Because GEE models are “population averaged models”,30 model estimates can be interpreted as means applicable to an “average” participant in this population.

Results

Violence Perpetration

A main effects model estimated the impact of family connections and family norms on violence perpetration. Both connections (b=−0.25, p<0.01) and prosocial norms (b=−0.21, p<0.01) significantly predicted violence perpetration at 18 months. In a second model including interaction terms, neither interaction term was significant (Table 4). Girls who reported high family connections and high family prosocial norms were significantly less likely to perpetrate violence at 18 months; however, the intervention did not bolster the effects of either family protective factor on violence perpetration.

Table 4
Results of Regressing Prosocial Norms and Connections to Family and Peers on Violence Perpetration at 18-months.

In a main effects model estimating the impact of peer variables, peer prosocial norms was a significant protective factor (b=−0.44, p<0.01). In contrast, peer connections behaved as a risk factor for violence perpetration at 18 months (b=0.10, p<0.05). In a second model including interaction terms, neither interaction term was significant (Table 4). Girls who reported high peer prosocial norms were less likely to perpetrate violence, but girls who reported high peer connections were more likely to perpetrate violence at 18 months. The intervention did not modify the effects of either peer variable.

Violence Victimization

In a main effects model estimating the impact of family contextual factors on violence victimization, both connections (b=−0.17, p<0.05) and prosocial norms (b=−0.25, p<0.05) were significant protective factors against violence victimization. In a second model including interaction terms, the relationship between family connections and violence victimization was moderated by intervention status (Table 5). In particular, girls who participated in Prime Time and had high family connections reported the lowest levels of violence victimization (Figure 1). Girls who reported high family connections and high family prosocial norms were significantly less likely to be victims of violence at 18 months; furthermore, the intervention appeared to augment the effects of family connections in reducing risk for violence victimization.

Figure 1
Adjusted Means for Violence Victimization – Family Connections X Intervention Condition
Table 5
Results of Regressing Prosocial Norms and Connections to Family and Peers on Violence Victimization at 18-months.

Finally, the effects of peer variables on violence victimization were examined. A main effects model indicated that peer prosocial norms (b=−0.36, p<0.01) functioned as a protective factor against victimization at 18 months. No significant impact of peer connections was noted. In a second model including interaction terms, the relationship between peer connections and violence victimization was marginally moderated by intervention status (Table 5). Specifically, Prime Time participants with high peer connections reported marginally lower levels of violence victimization than other groups.

Discussion

Previous research suggests that strong connections to family and peers, as well as prosocial family and peer norms, buffer youth from violence involvement.1214 This study examined whether these family and peer influences operate as protective factors within a sample of adolescent girls with high exposure to violence in their daily contexts. We were also able to test whether a youth development intervention augments the power of these protective factors to reduce girls’ risks for violence involvement. We discuss each set of findings in turn.

Consistent with previous research,1214,16 we found that family connections and prosocial norms acted as protective buffers against violence within a high-risk sample of adolescent girls. Girls who reported strong connections and prosocial norms within their families had lower levels of both violence perpetration and victimization. Peer prosocial norms also served as a protective buffer against violence perpetration and victimization within this high-risk sample, similar to other studies.14 However, contrary to our hypothesis, girls with strong peer connections over the 18-month period reported higher levels of violence perpetration, after accounting for peer norms and baseline levels of perpetration.

Family and peers are important contexts for adolescent socialization and the development of aggressive and violent behaviors, although the processes by which they influence adolescents’ outcomes are likely different.31 High quality parent-child relationships, characterized by warmth, support, and expectations for prosocial behavior, have been linked with positive adjustment throughout the lifecourse.31 In contrast, peers play an increasingly important role in social and emotional development beginning in early adolescence. In the current study, we found that strong prosocial peer norms protected girls from violence perpetration and victimization. This complements previous research demonstrating harmful effects of deviant peers on adolescents’ antisocial behavior.31 We were surprised to find that strong connections to peers was associated with increased violence perpetration. One explanation for this finding is that youth who note strong connections to peers may place more value on the role of peers and as a result may be more susceptible to negative peer influences (e.g., violence involvement).

The current study also examined whether a youth development intervention bolstered the effects of family and peer protective factors. We found that participation in Prime Time augmented the protective effects of social connectedness in reducing girls’ violence victimization. Specifically, girls with strong family connections who participated in Prime Time reported the lowest levels of violence victimization after 18 months. In addition, Prime Time participants with strong peer connections trended towards lower levels of violence victimization than other girls. This study found no evidence that the intervention bolstered the effects of family or peer connections on violence perpetration. Further, we found no evidence that the intervention augmented the protective effects of family or peer prosocial norms.

Prime Time intervention activities explicitly sought to build girls’ prosocial relationships;20 the intervention was not designed to modify family or peer norms regarding antisocial behaviors. Previous studies have documented higher levels of family connectedness among the intervention group versus controls.17,21 By valuing supportive, prosocial relationships, the intervention may have increased the strength of social connectedness as a buffer against violence victimization.

The intervention bolstered the protective power of social connectedness to reduce risk for violence victimization but not for violence perpetration. This finding may be due to underlying differences in pathways to these two forms of violence. Prime Time aimed to build girls’ capacities for self-awareness and personal agency. Girls who receive consistent messages from others within their daily contexts that they are valued and worthy of respect and who participate in an intervention that reinforces these messages may be motivated to stand up for themselves and as a result, less likely to be victims of violence.13 Beyond building personal agency, Prime Time did not include explicit messages about alternatives to physical violence. In contrast to being a victim, instigating physically aggressive behaviors may be a means for self-protection, connection and friendship among girls living in contexts where violence is normative.32 Unless expectations for non-violence are clearly conveyed and consistently reinforced, girls within such contexts may use physically aggressive behaviors as known means to accomplish desired social goals.

The current study has limitations. First, our perpetration and victimization measures do not account for nuances in the context (e.g., school, home) of violence. It is possible that family and peer protective factors operate differently depending on the context of violence. For example, the association between family connections and violence victimization may vary depending on whether an adolescent is being victimized by a family member versus a peer. Future research should consider whether the strength of family and peer protective factors varies depending on the context and severity of girls’ violence.

Second, we averaged measures of connections and norms across four time points. Although this data reduction strategy has advantages, it does not account for changes in family or peer contexts over time. Consistent with previous research indicating that the quality of parent-adolescent relationships remains relatively stable over time,33 measures of family connections were relatively stable across survey points in this sample (r=0.64–0.74). In contrast to family connections, correlations among measures of peer connections (r=0.39–0.63) and peer prosocial norms were less stable over time (r=0.40–0.56). More than one-third of this study’s sample reported moving two or more times in the previous six months. Residential and school mobility have been associated with poor connections to peers and increased involvement with deviant peers which, in turn, increase adolescents’ risk for violence involvement.34 In the current study, we did not consider changes in adolescents’ peer contexts or how these changes may influence their violence involvement. These are valuable areas for future inquiry.

Implications and Contributions

Despite these limitations, our findings reaffirm previous research on peer and family protective factors1214 and add to the literature in several important ways. Our sample consisted of adolescent girls at very high risk for violence who accessed community and school-based clinic services. Thus, our work has important implications for adolescent health services, as clinics may be an ideal setting for identifying adolescents at risk for violence. By routinely screening girls for violence involvement as well as risk and protection at individual, family, peer, and community levels,13 health care providers can play an important role in preventing youth violence and related negative health outcomes.35

This study reinforces the potential power of parent- and family-based prevention efforts. Girls who reported strong connections and prosocial norms within their families had lower levels of both violence perpetration and victimization. Our findings complement an existing body of research demonstrating the effectiveness of programs that promote strong family bonds and prosocial norms for children’s behavior in reducing violence involvement during adolescence.36,37

The current study also has important implications for clinic-linked prevention efforts with vulnerable groups of youth. Prime Time was designed to reduce multiple risk behaviors associated with early pregnancy among girls from high-risk social contexts. In prior reports, the intervention has been associated with reductions in sexual risk behaviors21 and relational aggression.17 Extending these findings, results of our analyses demonstrate that Prime Time augmented the protective power of social connections in reducing physical violence victimization. Modifying the intervention to explicitly convey social expectations for non-violence14 and build girls’ repertoire of non-violent strategies to realize desired social goals38 may make Prime Time more effective in reducing instigation of physical violence among girls living in contexts where violence is normative.

Acknowledgments

This project was supported with funds from the National Institute of Nursing Research (5R01-NR008778), the Centers for Disease Control and Prevention (U48-DP001939) and the Bureau of Health Professions (T32HP22239). The views presented do not necessarily reflect those of the funders. The Prime Time study would not have been possible without the cooperation and contributions of the young women, clinics and research staff involved with this project.

Footnotes

This paper was presented on March 15, 2012 at the Society for Adolescent Health and Medicine Annual Meeting, as part of Research Papers Session I – New Investigators.

Implications and Contributions: This study has implications for health services and clinic-led prevention with vulnerable youth. Clinics may be an ideal setting for identifying adolescents at risk for violence. Through screening and identification of risk and protective factors, health care providers play an important role in preventing youth violence and related health outcomes.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

1. Zahn M, Hawkins S, Chiancone J, et al. Girls study group: Charting the way to delinquency prevention for girls. Washington, DC: Office of Juvenile Justice and Delinquency Prevention; 2008.
2. Substance Abuse and Mental Health Services Administration, Office of Applied Studies. The NSDUH Report: Violent Behaviors among Adolescent Females. Rockville, MD: 2009.
3. Puzzanchera C, Adams B, Sickmund M. Juvenile court statistics 2006–2007. Pittsburgh, PA: National Center for Juvenile Justice; 2010.
4. Centers for Disease Control and Prevention. Youth risk behavior surveillance – United States, 2009. MMWR Morb Mortal Wkly Rep. 2010:59.
5. Minnesota Department of Education. Minnesota Student Survey 2007. Roseville, MN: 2007.
6. Centers for Disease Control and Prevention. Youth violence: Facts at a glance. Centers for Disease Control and Prevention; Washington, DC: 2010.
7. Kirby D, Lepore G. A Matrix of Risk and Protective Factors Affecting Teen Sexual Behavior, Pregnancy, Childbearing, and Sexually Transmitted Disease. Scotts Valley, CA: ETR Associates; 2007.
8. Buka S, Stichick T, Birdthistle I, et al. Youth exposed to violence: Prevalence, risks, and consequences. Am J Orthopsychiatry. 2001;71:298–310. [PubMed]
9. Fitzpatrick KM, Boldizar JP. The prevalence and consequences of exposure to violence among African-American youth. J Am Acad Child Psychiatry. 1993;32:424–30. [PubMed]
10. Roscoe B, Callahan J. Adolescents’ self-report of violence in families and dating relationships. Adolescence. 1985;20:545–53. [PubMed]
11. Singer MI, Anglin TM, Song LY, et al. Adolescents’ exposure to violence and associated symptoms of psychological trauma. JAMA. 1995;273:477–82. [PubMed]
12. Resnick MD, Bearman PS, Blum RW, et al. Protecting adolescents from harm: Findings from the National Longitudinal Study on Adolescent Health. JAMA. 1997;278:823–32. [PubMed]
13. Resnick M, Ireland M, Borowsky I. Youth violence perpetration: What protects? What predicts? Findings from the National Longitudinal Study of Adolescent Health. J Adolesc Health. 2004;35:424. e1–10. [PubMed]
14. Smith P, Flay B, Bell C, et al. The protective influence of parents and peers in violence avoidance among African-American youth. Matern Child Health J. 2001;5:245–52. [PubMed]
15. Department of Health and Human Services. Youth violence: A report of the surgeon general. Washington, DC: 2001.
16. Fergus S, Zimmerman M. Adolescent resilience: A framework for understanding healthy development in the face of risk. Annu Rev Public Health. 2005;26:399–419. [PubMed]
17. Sieving R, McMorris B, Secor-Turner M, et al. Prime Time: 18-month violence outcomes of a clinic-based intervention. Prev Sci. under review. [PubMed]
18. Secor-Turner M, Garwick A, Seppelt A, et al. Characteristics of violence among adolescent girls at high risk for pregnancy. Journal of School Nursing. under review.
19. Duke N, Sieving R, Pettingell S, et al. Associations between health screening questions and sexual risk behaviors in adolescent female clinic patients: Identifying a brief question format to yield critical information. Clin Pediatr. 2008;47:564–72. [PubMed]
20. Sieving R, Resnick M, Garwick A, et al. A clinic-based, youth development approach to teen pregnancy prevention. Am J Health Behav. 2011;35:346–58. [PubMed]
21. Sieving R, McMorris B, Beckman K, et al. Prime Time: 12-month sexual health outcomes of a clinic-based intervention to prevent pregnancy risk behaviors. J Adolesc Health. 2011;49:172–79. [PMC free article] [PubMed]
22. Resnick MD. Protective factors, resiliency and healthy youth development. Adolesc Med. 2000;11:157–64. [PubMed]
23. Bandura A. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice–Hall; 1986.
24. Kirby D. Emerging Answers 2007: Research Findings on Programs to Reduce Teen Pregnancy and Sexually Transmitted Diseases. Washington, DC: National Campaign to Prevent Teen Pregnancy; 2007.
25. Tanner A, Secor-Turner M, Garwick A, et al. Engaging vulnerable adolescents in the Prime Time program to reduce unintended pregnancy: Perspectives of case managers. J Pediatr Health Care. in press.
26. Sieving R, Beuhring T, Resnick M, et al. Development of adolescent self-report measures from the National Longitudinal Study of Adolescent Health. J Adolesc Health. 2001;28:73–81. [PubMed]
27. Osgood DW, Finken L, McMorris BJ. Analyzing multiple-item measures of crime and deviance II: Tobit regression of transformed scores. J Quant Criminol. 2002;4:319–347.
28. Baron RB, Kenny DA. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51:1173–82. [PubMed]
29. Ballinger G. Using generalized estimating equations for longitudinal data analysis. Organizational Research Methods. 2004;7:127–50.
30. Liang K, Zeger S. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73:13–22.
31. Pettit GS. The developmental course of violence and aggression: Mechanisms of family and peer influence. Psychiatric Clinics of North America. 1997;20:283–99. [PubMed]
32. Adamshick PZ. The lived experience of girl-to-girl aggression in marginalized girls. Qual Health Res. 2010;20:541–55. [PubMed]
33. Allen JP. The Attachment System in Adolescence. In: Cassidy J, Shaver PR, editors. Handbook of Attachment. 2. New York: Guilford; 2008. pp. 419–35.
34. Haynie DL, South SJ. Residential mobility and adolescent violence. Social Forces. 2005;84:361–74.
35. Taliaferro LA, Borowsky IW. Beyond prevention: Promoting healthy youth development in primary care. Am J Public Health. 2012;102:S317–321. [PubMed]
36. Tolan PH, Guerra N, editors. An empirical review of the field. 3. Boulder, CO: Cener for the Study and Prevention of Violence; 2002. What works in reducing adolescent violence.
37. Thornton TN, Craft CA, Dahlberg LL, Lynch BS, Baer K. Best Practices of Youth Violence Prevention: A Sourcebook for Community Action (Rev.) Atlanta: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2002.
38. Wilson S, Lipsey M. School-based interventions for aggressive and disruptive behavior: Update of a meta-analysis. Am J Prev Med. 2007;33:S130–S143. [PMC free article] [PubMed]