[Skip to Navigation]
Sign In
Oct 2011

School Bullying Perpetration and Other Childhood Risk Factors as Predictors of Adult Intimate Partner Violence Perpetration

Author Affiliations

Author Affiliations: Harvard School of Public Health, Boston, Massachusetts (Mss Falb and McCauley and Dr Silverman); Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (Dr Decker); Yale School of Public Health, New Haven, Connecticut (Dr Gupta); and Boston University School of Public Health, Boston (Dr Raj).

Arch Pediatr Adolesc Med. 2011;165(10):890-894. doi:10.1001/archpediatrics.2011.91

Objective To assess the relationship between bullying peers as a child and adult intimate partner violence perpetration in a clinic-based sample of adult men. School bullying perpetration and intimate partner violence perpetration are both thought to stem from desire for power and control over others.

Design A cross-sectional survey was conducted between January 2005 and December 2006.

Setting Three urban community health centers in Boston, Massachusetts.

Participants Men aged 18 to 35 years (n = 1491) seeking services at participating community health centers.

Main Exposure School bullying perpetration.

Outcome Measure Past-year physical or sexual violence perpetration against a female partner (intimate-partner violence [IPV]).

Results Two-fifths of men reported perpetrating school bullying as a child (n = 610; 40.9%). Men who rarely bullied in school were 1.53 times more likely to perpetrate past-year IPV than men who did not bully (95% confidence interval [CI], 1.02-2.29); this risk was elevated to 3.82 times more likely to perpetrate any past-year IPV for those men who bullied peers frequently (95% CI, 2.55-5.73).

Conclusions The present study indicates that bullying peers in school as a child, especially frequent bullying perpetration, is associated with increased risk for men's perpetration of IPV as an adult. The effect remains strong after controlling for common prior risk factors for both bullying and IPV perpetration. Future research is needed to discern the mechanisms and underlying root causes of abusive behavior, such as power and control, as a means to prevent violence perpetration across settings and life stages.

Approximately 1 in 4 women will experience intimate-partner violence (IPV) in her lifetime.1 Evidence indicates that IPV victimization is associated with major physical and mental health concerns, including increased risk for sexually transmitted infection, human immunodeficiency virus infection, depression, and substance abuse, making IPV a serious public health concern.2,3 Since the vast majority of IPV perpetrators are men and the health consequences of such victimization are more pronounced for women,4,5 a critical step in preventing and mitigating the effects of IPV is to identify and address risk factors for IPV perpetration among men.

Previous research drawing on samples from community-based health clinics has estimated that between 20% and 40% of adult men have ever perpetrated IPV,6-8 and a recent study estimated a 4% prevalence of past-year IPV perpetration in a national sample of men.9 There is clear and consistent evidence that adverse childhood experiences may increase risk for adult IPV perpetration; exposure to community violence, exposure to parental IPV, and physical or sexual child abuse during these crucial developmental years have been associated with later adult perpetration of IPV.9-13 Limited evidence suggests participation in youth delinquency is also related to adult IPV, although distinctions between nonviolent and violent youth delinquency are lacking in the literature.14,15

Within the literature concerning childhood experiences as factors in adult IPV perpetration, surprisingly little is known about the role of school bullying perpetration. School bullying has sparked recent public health and media attention and is defined as physical or psychological aggression repeatedly directed toward another student who is perceived by the aggressor as weaker and less powerful than himself or herself.16 Like IPV perpetration, it is fueled by desire for power and control over other individuals, leading to concern that it may be a precursor to IPV.17 In particular, a recent theoretical article proposed that “sexual bullying,” or unwanted sexual teasing, harassment, or threats, overlaps dating violence in young adolescents.17 Other work, guided by a social-ecological framework, extends this hypothesis and asserts that there is significant overlap between sexual harassment and bullying behaviors in students such that youths may participate in multiple forms of abuse and aggression concurrently.18 One previous study of bullies documented less positive and equitable attitudes toward dating partners and more reports of aggression.19 However, no study, to our knowledge, has examined the association of bullying with other forms of physical or sexual IPV in adulthood.18

Beyond the conceptual basis for such investigation, recent evidence strongly indicates that bullying peers in school may share common prior causes with IPV perpetration. Specifically, exposure to community violence,20 witnessing parental IPV,16,21,22 delinquent behavior,23,24 and childhood physical or sexual abuse25 have all been found to be relevant to both IPV perpetration and bullying perpetration in separate studies. Evidence also indicates that bullying others in school may be an important predictor of other types of adult criminal behavior.23,26,27

Although bullying and IPV share a common framework and common developmental factors, to our knowledge there has been no empirical quantitative investigation of the relationship of these concerns. To address this gap in knowledge, the current study seeks to assess the relationship between history of school bullying perpetration and recent adult IPV perpetration among men, and to determine whether exposure to other developmental risks affects this relationship.


The current study uses cross-sectional data from the Men's Ecological Systems, Development, and Abuse Study28 that took place between January 2005 and December 2006. Institutional review board approval was granted by the Harvard Human Subjects Committee. The study used a convenience sample of young men recruited from 3 urban community health centers (CHCs) in Boston, Massachusetts. Participants were aged 18 to 35 years and were fluent in English, Portuguese, or Spanish. Trained researchers recruited all men attending the clinic during designated recruitment hours by asking them if they would like to participate in a brief survey; men indicating interest were screened for eligibility in a semiprivate area. Staff members obtained verbal informed consent from eligible and willing participants.

After informed consent procedures, participants completed a 30-minute confidential health survey via an audio computer-assisted survey instrument (ACASI), which allowed them to hear questions and answer choices read aloud through headphones in the language of their choice. The ACASI has been shown to reduce literacy barriers and minimize the underreporting of sensitive behaviors, rendering it particularly useful for violence research.29-34 After completion of the interview, participants were given a $20 gift card to compensate them for their time and a list of local community resources for violence-prevention and health-promotion services.

Of the 3430 men approached for the study, 2229 (65%) agreed to participate; 75 of these participants were excluded owing to extensive missing data, yielding a final effective sample size of 2154. The sample was then further restricted to those who reported ever having sexual intercourse, and that provided complete data on the outcome (past year IPV perpetration) and predictors of interest, which yielded a final sample size of 1491.


Demographic assessments for age, race, ethnicity, and educational attainment were modified from the National Behavioral Risk Factor Surveillance System.35 Past-year IPV perpetration was assessed using modified physical assault, sexual assault, and injury subscales from the Conflict Tactics Scale 2 (CTS-2).36 A dichotomous summary variable of past-year IPV perpetration was then created. Past-year IPV perpetration was selected as the outcome of interest (rather than lifetime perpetration) to ensure that school bullying and other childhood predictors preceded IPV perpetration. School bullying perpetration data were obtained by a single survey item: “How often did you take part in bullying other students?” Results were coded as “never,” “rarely (once or twice),” or “frequently (once a month or more).” Childhood risk factors, including bullying victimization, exposure to parental IPV, exposure to community violence, experiencing physical or sexual child abuse, and participating in nonviolent or violent delinquency were also assessed. Adolescent nonviolent delinquency was defined as ever participating in graffiti, vandalism, or stealing property between ages 12 and 18 years. Adolescent violent delinquency was defined as ever participating in physical fighting, pulling a gun or knife on a nondating and/or sexual partner, stabbing or shooting a person they were not in a dating or sexual relationship with, or participation in gang fighting between ages 12 and 18 years. All delinquency items were drawn from the National Longitudinal Study of Adolescent Health.37 Both nonviolent and violent delinquency were coded as dichotomous summary variables such that any affirmative response to the items was coded as positive for nonviolent and violent delinquency, respectively.

Statistical analysis

Descriptive statistics for prevalence of past-year IPV perpetration, frequency of bullying in childhood, demographics, and exposure to childhood risk factors were generated. We used χ2 analyses to assess for differences in predictors based on past-year IPV perpetration (P < .05). We used a logistic regression model to assess the relationship between recent past-year IPV perpetration and school bullying in childhood, controlling for potential risk factors (eg, bullying victimization, exposure to parental IPV, exposure to community violence, experiencing physical or sexual child abuse, and participating in nonviolent or violent delinquency) and demographics.


Demographics and childhood predictors

Demographic characteristics are listed in Table 1. Almost half of participants were aged between 18 and 24 years (46.2%), with a mean age of 25.7 years; 48.6% of the sample was on-Hispanic black, while almost one-third self-identified as Hispanic (31.2%). A smaller portion of the sample self-identified as non-Hispanic white (8.4%); 11.8% of the sample were self-classified as “other.” Approximately 27% of participants had less than a high school education.

Table 1. Characteristics of the Study Sample by Past-Year Physical or Sexual IPV Perpetration
Table 1. Characteristics of the Study Sample by Past-Year Physical or Sexual IPV Perpetration
Table 1. Characteristics of the Study Sample by Past-Year Physical or Sexual IPV Perpetration

Frequent and rare school bullying perpetration was reported by 16.3% (n = 243) and 24.6% (n = 367) of participants, respectively. Over one-quarter of participants were exposed to parental IPV during childhood (27.9%), while 43.6% experienced physical child abuse, and 20.5% experienced sexual child abuse. Frequent exposure to community violence was reported by 14.1% of the sample; over half the sample reported occasional exposure to community violence (56.1%). Roughly one-quarter reported taking part in nonviolent youth delinquency (27.7%), with slightly more reporting violent youth delinquency (30.6%). One in 10 respondents reported frequent bullying victimization at school (10.4%).

Approximately 16% of men reported perpetrating physical or sexual IPV in the past year (n = 241). Of these men, almost 40% bullied other students frequently. Slightly less than 20% of men who reported perpetrating bullying rarely went on to perpetrate IPV. Less than 10% of men who did not report bullying other students perpetrated past-year IPV. The main predictor, school bullying, and each of the childhood risk factors was statistically significantly associated with IPV perpetration in the χ2 analyses.

Association between history of bullying students and recent ipv perpetration

Those who reported frequent bullying of other students were 5.63 times more likely to report perpetrating past-year IPV (95% confidence interval [CI], 4.00-7.93); those who rarely bullied peers were almost twice as likely to report perpetrating past-year IPV (odds ratio [OR], 1.92; 95% CI, 1.35-2.72) as those who never reported school bullying perpetration in the unadjusted model (Table 2). After the inclusion of childhood risk factors and demographic characteristics, frequent bullying remained strongly associated with perpetration of any past-year IPV (adjusted OR, 3.82; 95% CI, 2.55-5.73). The effect of rarely bullying other students was slightly attenuated, such that those men were 1.5 times more likely to perpetrate past-year IPV (adjusted OR, 1.53; 95% CI, 1.02-2.29) than those who never reported bullying others.

Table 2. Crude and Adjusted Regression Models of Bullying and IPV Perpetration
Table 2. Crude and Adjusted Regression Models of Bullying and IPV Perpetration
Table 2. Crude and Adjusted Regression Models of Bullying and IPV Perpetration


Men who reported bullying their childhood peers in school were found to be significantly more likely to physically or sexually abuse their female partners as adults. Over two-fifths of men reported perpetrating any bullying of students, indicating the importance of addressing bullying peers as a potential risk marker for future IPV perpetration.

Consistent with previous evidence, sexual and physical child abuse, exposure to parental IPV, and participation in nonviolent or violent delinquency were also strong independent predictors of adult IPV perpetration. However, even after accounting for these other childhood risk factors, frequently bullying peers was the strongest predictor of past-year IPV perpetration. Bullying victimization or exposure to community violence was not found to be associated with past-year IPV perpetration in the final adjusted model.

Critically, this analysis demonstrates that those reporting school bullying are significantly more likely to perpetrate physical or sexual IPV, even after adjusting for potential confounders. This relationship was stronger for those who reported frequent bullying than for those who reported rare school bullying. Bullying others at school and perpetration of IPV are both defined by concepts of power and control over others.15,18 Thus, this finding was not surprising and provides empirical data to support the previously hypothesized links between bullying and IPV perpetration such that those who perpetrate school bullying by asserting power and control over classmates may also attempt to assert power and control over intimate partners as adults. Further research is needed to assess specific mechanisms, including investigation into whether bullying peers serves as a risk marker, as suggested by the convergence of risk factors for both bullying and recent adult IPV perpetration.18

These findings must be interpreted in light of several limitations. Although our analysis uses a large sample size, participants were recruited using a convenience sampling method, which may limit the generalizability of results to the larger population. Furthermore, only 65% responded, and we were unable to collect demographic information about those who chose not to participate, limiting our ability to understand any potential biases. Bullying perpetration was assessed via a single item, limiting our ability to understand the extent to which various aspects of bullying may be most relevant to IPV perpetration. In addition, bullying has been previously defined as one student having power and control over another student.17 However, our measure was self-defined by the participant, which may lead to a discrepancy between our conceptualization of school bullying and the participant's own definition. Although the frequency of bullying others in school was captured, we do not have data on the timing during the participant's schooling, severity, or specific forms of covert or overt bullying, which may have implications for potential programs to effectively and efficiently reduce the bullying of peers in school. Due to the cross-sectional design of the survey, we cannot discern whether other childhood factors temporally preceded school bullying thereby confounding the relationship of school bullying and IPV perpetration. Childhood exposures were assessed retrospectively and are subject to potential errors and biases in recall; therefore, causality cannot be ensured.

The present study indicates that bullying peers in school as a child, especially frequent bullying perpetration, is associated with increased risk for men's perpetration of IPV as an adult. The effect remains strong after controlling for common prior risk factors for both bullying and IPV perpetration. These findings suggest that individuals who are likely to perpetrate abusive behaviors against others may do so across childhood and into adulthood. Furthermore, these abusive behaviors—bullying peers in school and perpetrating violence against an intimate partner—may co-occur within individuals. Future research is needed to discern the mechanisms and underlying root causes of abusive behavior, such as power and control, as well as specific forms and timing of bullying peers in school. Potential programs that may seek to reduce bullying peers during school may also be effective avenues to reduce future violence perpetration within intimate partner relationships by focusing on the reduction of abusive behaviors and the promotion of equitable attitudes across settings, life stages, and relationships.

Back to top
Article Information

Correspondence: Kathryn L. Falb, MHS, Department of Society, Human Development, and Health, Harvard School of Public Health, 677 Huntington Ave, Boston, MA 02115 (kfalb@hsph.harvard.edu).

Accepted for Publication: March 11, 2011.

Published Online: June 6, 2011. doi:10.1001/archpediatrics.2011.91

Author Contributions: Dr Silverman had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Falb, McCauley, and Silverman. Acquisition of data: Decker, Gupta, and Silverman. Analysis and interpretation of data: Falb, McCauley, Decker, Gupta, and Raj. Drafting of the manuscript: Falb and McCauley. Critical revision of the manuscript for important intellectual content: McCauley, Decker, Gupta, Raj, and Silverman. Statistical analysis: Silverman.

Financial Disclosure: None reported.

Black M, Breiding M. Adverse Health Conditions and Health Risk Behaviors Associated With Intimate Partner Violence—United States, 2005. Atlanta, GA: Centers for Disease Control and Prevention; 2008
Campbell JC, Lewandowski LA. Mental and physical health effects of intimate partner violence on women and children.  Psychiatr Clin North Am. 1997;20(2):353-3749196919PubMedGoogle ScholarCrossref
Hankin A, Smith LS, Daugherty J, Houry D. Correlation between intimate partner violence victimization and risk of substance abuse and depression among African-American women in an urban emergency department.  West J Emerg Med. 2010;11(3):252-25620882145PubMedGoogle Scholar
Reed E, Raj A, Miller E, Silverman JG. Losing the “gender” in gender-based violence: the missteps of research on dating and intimate partner violence.  Violence Against Women. 2010;16(3):348-35420133922PubMedGoogle ScholarCrossref
Tjaden P, Thoennes N. Extent, Nature, And Consequences of Intimate Partner Violence: Findings From the National Violence Against Women Survey. Washington, DC: Department of Justice; 2000
Raj A, Santana MC, La Marche A, Amaro H, Cranston K, Silverman JG. Perpetration of intimate partner violence associated with sexual risk behaviors among young adult men.  Am J Public Health. 2006;96(10):1873-187816670216PubMedGoogle ScholarCrossref
Decker MR, Seage GR III, Hemenway D, Gupta J, Raj A, Silverman JG. Intimate partner violence perpetration, standard and gendered STI/HIV risk behaviour, and STI/HIV diagnosis among a clinic-based sample of men.  Sex Transm Infect. 2009;85(7):555-56019625287PubMedGoogle ScholarCrossref
Reed E, Silverman JG, Welles SL, Santana MC, Missmer SA, Raj A. Associations between perceptions and involvement in neighborhood violence and intimate partner violence perpetration among urban, African American men.  J Community Health. 2009;34(4):328-33519343487PubMedGoogle ScholarCrossref
Roberts AL, Gilman SE, Fitzmaurice G, Decker MR, Koenen KC. Witness of intimate partner violence in childhood and perpetration of intimate partner violence in adulthood.  Epidemiology. 2010;21(6):809-81820811285PubMedGoogle ScholarCrossref
Fang X, Corso PS. Child maltreatment, youth violence, and intimate partner violence: developmental relationships.  Am J Prev Med. 2007;33(4):281-29017888854PubMedGoogle ScholarCrossref
Herrenkohl TI, Mason WA, Kosterman R, Lengua LJ, Hawkins JD, Abbott RD. Pathways from physical childhood abuse to partner violence in young adulthood.  Violence Vict. 2004;19(2):123-13615384450PubMedGoogle ScholarCrossref
Raghavan C, Mennerich A, Sexton E, James SE. Community violence and its direct, indirect, and mediating effects on intimate partner violence.  Violence Against Women. 2006;12(12):1132-114917090690PubMedGoogle ScholarCrossref
Reed E, Silverman JG, Raj A,  et al.  Social and environmental contexts of adolescent and young adult male perpetrators of intimate partner violence: a qualitative study.  Am J Mens Health. 2008;2(3):260-27119477789PubMedGoogle ScholarCrossref
Herrenkohl TI, Kosterman R, Mason WA, Hawkins JD. Youth violence trajectories and proximal characteristics of intimate partner violence.  Violence Vict. 2007;22(3):259-27417619633PubMedGoogle ScholarCrossref
Casey EA, Beadnell B. The structure of male adolescent peer networks and risk for intimate partner violence perpetration: findings from a national sample.  J Youth Adolesc. 2010;39(6):620-63320422351PubMedGoogle ScholarCrossref
Baldry AC. Bullying in schools and exposure to domestic violence.  Child Abuse Negl. 2003;27(7):713-73214627075PubMedGoogle ScholarCrossref
Fredland NM. Sexual bullying: addressing the gap between bullying and dating violence.  ANS Adv Nurs Sci. 2008;31(2):95-10518497586PubMedGoogle Scholar
Basile KC, Espelage DL, Rivers I, McMahon PM, Simon TR. The theoretical and empirical links between bullying behavior and sexual violence perpetration.  Aggress Violent Behav. 2009;14(5):336-347Google ScholarCrossref
Connolly J, Pepler D, Craig W, Taradash A. Dating experiences of bullies in early adolescence.  Child Maltreat. 2000;5(4):299-31011232258PubMedGoogle ScholarCrossref
Chaux E, Molano A, Podlesky P. Socio-economic, socio-political and socio-emotional variables explaining school bullying: a country-wide multilevel analysis.  Aggress Behav. 2009;35(6):520-52919739091PubMedGoogle ScholarCrossref
Bauer NS, Herrenkohl TI, Lozano P, Rivara FP, Hill KG, Hawkins JD. Childhood bullying involvement and exposure to intimate partner violence.  Pediatrics. 2006;118(2):e235-e24216882768PubMedGoogle ScholarCrossref
Lepistö S, Luukkaala T, Paavilainen E. Witnessing and experiencing domestic violence: a descriptive study of adolescents.  Scand J Caring Sci. 2011;25(1):70-8020384972PubMedGoogle ScholarCrossref
Sourander A, Brunstein Klomek A, Kumpulainen K,  et al.  Bullying at age eight and criminality in adulthood: findings from the Finnish Nationwide 1981 Birth Cohort Study [published online December 1, 2010].  Soc Psychiatry Psychiatr Epidemiol19305936PubMedGoogle Scholar
Sourander A, Jensen P, Rönning JA,  et al.  Childhood bullies and victims and their risk of criminality in late adolescence: the Finnish From a Boy to a Man study.  Arch Pediatr Adolesc Med. 2007;161(6):546-55217548758PubMedGoogle ScholarCrossref
Duke NN, Pettingell SL, McMorris BJ, Borowsky IW. Adolescent violence perpetration: associations with multiple types of adverse childhood experiences.  Pediatrics. 2010;125(4):e778-e78620231180PubMedGoogle ScholarCrossref
Klomek AB, Sourander A, Kumpulainen K,  et al.  Childhood bullying as a risk for later depression and suicidal ideation among Finnish males.  J Affect Disord. 2008;109(1-2):47-5518221788PubMedGoogle ScholarCrossref
Roeger L, Allison S, Korossy-Horwood R, Eckert KA, Goldney RD. Is a history of school bullying victimization associated with adult suicidal ideation? a South Australian population-based observational study.  J Nerv Ment Dis. 2010;198(10):728-73320921863PubMedGoogle ScholarCrossref
Decker MR, Raj A, Gupta J, Silverman JG. Sex purchasing and associations with HIV/STI among a clinic-based sample of US men.  J Acquir Immune Defic Syndr. 2008;48(3):355-359Google ScholarCrossref
Abbey A. Lessons learned and unanswered questions about sexual assault perpetration.  J Interpers Violence. 2005;20(1):39-4215618559PubMedGoogle ScholarCrossref
Edwards SL, Slattery ML, Murtaugh MA,  et al.  Development and use of touch-screen audio computer-assisted self-interviewing in a study of American Indians.  Am J Epidemiol. 2007;165(11):1336-134217379618PubMedGoogle ScholarCrossref
Ghanem KG, Hutton HE, Zenilman JM, Zimba R, Erbelding EJ. Audio computer assisted self interview and face to face interview modes in assessing response bias among STD clinic patients.  Sex Transm Infect. 2005;81(5):421-42516199744PubMedGoogle ScholarCrossref
Metzger DS, Koblin B, Turner C,  et al; HIVNET Vaccine Preparedness Study Protocol Team.  Randomized controlled trial of audio computer-assisted self-interviewing: utility and acceptability in longitudinal studies.  Am J Epidemiol. 2000;152(2):99-10610909945PubMedGoogle ScholarCrossref
Simoes AA, Bastos FI, Moreira RI, Lynch KG, Metzger DS. A randomized trial of audio computer and in-person interview to assess HIV risk among drug and alcohol users in Rio De Janeiro, Brazil.  J Subst Abuse Treat. 2006;30(3):237-24316616168PubMedGoogle ScholarCrossref
Simoes AA, Bastos FI, Moreira RI, Lynch KG, Metzger DS. Acceptability of audio computer-assisted self-interview (ACASI) among substance abusers seeking treatment in Rio de Janeiro, Brazil.  Drug Alcohol Depend. 2006;82:(Suppl 1)  S103-S10716769438PubMedGoogle ScholarCrossref
Centers for Disease Control and Prevention.  Behavioral risk factor surveillance system, state questionnaire. http://www.cdc.gov/brfss/questionnaires/pdf-ques/2004brfss.pdf. Accessed April 14, 2011
Strauss MA, Hamby SL, Boney-McCoy S, Sugarman DB. The Revised Conflict Tactic Scales (CTS-2).  J Fam Issues. 1996;17:283-316Google ScholarCrossref
University of North Carolina, Carolina Population Center.  The National Longitudinal Study of Adolescent Health: Wave III in-home codebook. http://www.cpc.unc.edu/projects/addhealth/codebooks/wave3. Accessed April 14, 2011