Perpetration of Physical Assault Against Dating Partners, Peers, and Siblings Among a Locally Representative Sample of High School Students in Boston, Massachusetts FREE

Dating violence (DV) is a common adolescent health problem with substantial public health consequences. As many as 1 in 10 US high school students reports having been “hit, slapped, or physically hurt on purpose by their boyfriend or girlfriend” in the past year.¹ Research on victims of DV has demonstrated that they are at risk for a range of negative consequences, including death, injury, suicidal thoughts, substance use, disordered eating, and psychiatric disorders.^2- 6 In contrast to the relatively rich information available about victims of DV, far less is known about DV perpetrators and how to prevent their aggression, which may be one of the reasons that only 2 DV prevention programs have been found effective to date.^7- 8

It has been suggested that at least a portion of DV perpetration may be attributable to general antisocial traits and that DV is one of numerous problem behaviors that tend to co-occur in adolescence, including substance use, delinquency, and peer aggression.^9- 10 However, no prior study of which we are aware has quantified the extent of the overlap among DV and sibling and peer violence. Moreover, previous studies that have found that DV and peer violence are associated or that DV and sibling violence are associated have described those associations by using odds ratios, correlation coefficients, or mean scale scores.^10- 14 Although these are useful metrics for researchers, practitioners still lack an easily interpretable quantification of the overlap among the different types of abuse, such as a proportion that represents the prevalence of the overlap. The size of the overlap could have important implications for prevention; if the overlap is substantial, rather than delivering tailored messages to what are presumed to be unique subgroups of perpetrators, educators could conjointly address shared risk factors for peer, community, sibling, and dating aggression.

The purpose of the present study was 2-fold. First, we sought to establish the prevalence of past-month physical assault of dating partners among 9th through 12th graders in public high schools in Boston, Massachusetts. We refer to this form of DV as physical DV to distinguish it from the full spectrum of dating abusive behaviors, which may include emotional, sexual, and physical abuse. Second, we explored the relationship between DV perpetration and other adolescent problem behaviors, including sibling and peer violence perpetration, substance use, knife carrying, and school truancy. We expected to find that physical DV was common, that it was associated with both peer and sibling assault perpetration, and that it co-occurred with other adolescent risk behaviors and experiences.

We collected self-report survey data from 9th through 12th graders in Boston public schools via the Boston Youth Survey (N = 1878).¹⁵ Participation in the survey was voluntary, and students' answers were confidential. The survey used a sampling method similar to that of the Boston Youth Risk Behavior Survey (YRBS)¹⁶ and is representative of Boston public high school students.

All public high schools in Boston were invited to participate, with the exception of 9 specialty schools that serve nontraditional students (eg, adults in night school). Twenty-two of the 32 schools invited participated. The remaining 10 schools were unable to participate because of scheduling conflicts and did not differ from participating schools in terms of their dropout rates, standardized test scores, proportion of students receiving free or reduced-price lunches, or student race. To generate our sample, we assembled a list of unique classrooms within each participating school, stratified by grade, and randomly selected classrooms for survey administration. Every student within selected classrooms was invited to participate. Classrooms were selected until the total number of enrolled students in each grade reached 100 to 125 per school. Of the 2725 students in classrooms selected for participation, 1878 completed the survey (American Association for Public Opinion Research response rate, 68.9%).¹⁷ Noncompleters were absent from school on the day of the survey (n = 724), declined to participate (n = 99), or were not permitted to participate by their parents (n = 24). Of the students selected for participation and not absent on the day of the survey (n = 2001), 93.9% completed surveys. The demographic composition of the total sample was comparable to the population of Boston public high school students in terms of sex, nativity, race, ethnicity, and age.

Trained staff administered the survey between January and April of 2008 during 50-minute class periods. Passive consent was sought from parents, and students were read a statement regarding assent/consent before survey administration. The Human Subjects Committee at the Harvard School of Public Health approved all procedures, and research protocols were approved by the Office of Research, Assessment, and Evaluation at the Boston Public Schools.

Past-month violence against dating partners was assessed via questions adapted from the Revised Conflict Tactics Scales.¹⁸ Respondents were instructed to “think about someone you were or are dating” in the past 30 days, with the specification that “by dating, we mean a girlfriend or boyfriend, or someone who you were romantically or sexually involved with.” Respondents were further instructed “when answering these questions, do not include times when you or someone else was playing or joking around.” They were then asked to indicate the number of times in the past month that they (a) had a yelling argument with him or her; (b) swore or cursed at him or her or called him or her fat, ugly, stupid, or some other insult; (c) threatened to hit, punch, kick, or hurt her or him; (d) pushed, shoved, or slapped him or her; and (e) hit, punched, kicked, or choked him or her. For the present analysis, respondents who indicated that they had perpetrated physical violence 1 or more times (ie, endorsed items d or e) were classified as having perpetrated physical DV.

Respondents were asked to indicate how many times in the past 30 days they had perpetrated various physically violent acts against “kids, including those in your school or neighborhood” but not in their family and not including someone whom they were dating. The violent acts were defined as (a) pushed, shoved, or slapped and (b) hit, punched, kicked, or choked. Respondents who reported perpetrating any of the violent acts at least 1 time were classified as perpetrating peer violence.

Respondents were asked to indicate the number of times in the past 30 days that they had perpetrated physically violent acts against “kids in your immediate family, meaning the kids who live in your home.” The acts were identical to those described under peer violence perpetration. Respondents with siblings who reported perpetrating any of the violent acts at least 1 time were classified as perpetrating sibling violence.

Substance use was assessed via 3 similarly worded items about past-month use of alcohol, tobacco, or marijuana (eg, “In the past 30 days, on how many days did you drink alcohol?”) adapted from the 2005 national YRBS.¹⁹ Respondents were classified as having been delinquent if they reported involvement with the juvenile justice system in the 12 months preceding survey administration or if they reported that they were involved in a gang. Respondents were classified as having been exposed to community violence if, within the 12 months preceding survey administration, they reported having seen someone (a) punched, kicked, or choked; (b) attacked or threatened with a weapon other than a gun; (c) threatened with a gun; (d) shot or shot at; or (e) killed. Academic performance was assessed through 2 items, the first asking respondents about their grades (ie, “In the past 12 months, how would you describe your grades in school?”) and the second asking about truancy (ie, “In the past 30 days, how many times did you skip school when you were not sick?”). The former item was adapted from the YRBS,¹⁹ and the latter was adapted from the National Longitudinal Study of Adolescent Health.²⁰ Students who reported having mostly Ds or mostly Fs were classified as having failing grades, and those who reported skipping school 1 or more times in the past month were classified as truant. A question adapted from the YRBS assessed whether students carried a knife in the past 30 days. Using questions that were based on physical violence items from the Conflict Tactics Scales,²¹ we also assessed whether students had been physically victimized by peers in the past year or physically assaulted by an adult who lived in their home in the past year.

Demographic characteristics included grade (9th-12th), sex, Hispanic/Latino ethnicity, nativity status (US or foreign born), and race.

We used SAS software, version 9.1.3, for statistical analysis.²² We tabulated descriptive information on DV perpetration by sociodemographic characteristics for the full sample and then stratified by sex. We used the Pearson χ² statistic to assess the statistical significance of differences in the prevalence of acts of DV by sex. We then examined differences in sociodemographic characteristics among perpetrators of physical DV compared with nonperpetrators, stratified by sex. Statistical significance was assessed using Pearson χ² statistics. Next, we calculated the prevalence of physical assault by relationship type (ie, toward dating partners, siblings, and peers), stratified by sex, and restricted the analyses to those with siblings. These results are presented in proportional Venn diagrams.²³ We examined bivariate associations between physical DV and other adolescent problem behaviors and experiences and conducted multivariable regression analyses to determine whether the associations persisted after statistical adjustment for potential confounders (eg, age and race). We controlled for the potential effect of school clustering by using a generalized estimating equations model with a compound symmetric error covariance structure matrix. Finally, we computed prevalence ratios using a binomial distribution and log-link function.²⁴ All models converged.

Of the 1878 students in the sample, 350 had not dated anyone in the past month, 126 did not answer the questions about dating partner assault, and 4 did not indicate their sex. Those respondents did not differ from the rest of the sample in terms of race, age, or grade level and were excluded, bringing the analytical sample to 1398. Most of the respondents were black (44.2%) or Hispanic/Latino (35.6%), were born in the United States (69.5%), and lived with siblings (86.1%). The mean age was 16.4 (SD, 1.3) years. For analyses of the overlap among prevalence of physical DV and peer and sibling violence, we further restricted the sample to respondents who had siblings and had perpetrated at least 1 form of violence (n = 1084).

Approximately one-fifth of the students reported past-month physical DV perpetration (18.7%; Table 1), whereas 41.2% reported peer violence and 31.2% reported sibling violence perpetration (not shown). A significantly larger proportion of girls than boys reported perpetrating physical DV (26.6% vs 9.6%) and each other form of DV (Table 1). Although there were no significant differences in physical DV perpetration by age for boys or girls, there were differences by nativity among girls (Table 2). Specifically, the prevalence of physical DV perpetration was lower among foreign-born girls compared with US-born girls (29.4% vs 19.0%; P = .004) (Table 2).

Of the 1084 respondents with siblings, 256 of the boys (50.8%) and 351 of the girls (60.5%) reported that they had physically assaulted a sibling, peer, and/or dating partner (Figure). Among those who reported perpetrating any form of past-month violence, just 7.9% perpetrated physical DV exclusively (not shown).

Among the 256 boys who reported perpetrating at least 1 form of violence, physical DV was the least commonly reported form of violence perpetrated (14.1%), whereas 84.4% reported perpetrating peer violence and 49.6% reported perpetrating sibling violence (Figure). There was a high degree of overlap between physical DV and peer violence by boys: 75.0% of DV perpetrators had also perpetrated peer violence (Figure). There was also a high degree of overlap between physical DV and sibling violence: 55.6% of male DV perpetrators had also perpetrated sibling violence (Figure). The fraction of boys who had perpetrated physical DV and no other form of violence was 2.3% (Figure).

Among the 351 girls who reported perpetrating at least 1 form of violence, 44.2% reported perpetrating physical DV, 65.2% reported perpetrating peer violence, and 59.8% reporting perpetrating sibling violence (Figure). Among the female perpetrators of DV, 59.4% had also perpetrated peer violence, and 50.3% of female DV perpetrators had also perpetrated sibling violence (Figure). Among girls, 12.0% reported perpetrating physical DV and no other form of violence (Figure).

The multivariable regression models demonstrated that physical assault of a dating partner was significantly associated with physical assault of a peer or of a sibling for both sexes, controlling for age and school clustering (Table 3). The association between physical assault of a sibling and assault of a dating partner was strong for boys (adjusted prevalence ratio [PR], 3.81; 95% confidence interval [CI], 2.07-6.99) and for girls (1.83; 1.44-2.31). Peer violence perpetration was strongly associated with DV perpetration for boys (adjusted PR, 5.13; 95% CI, 3.15-8.35) and for girls (2.57; 1.87-3.52).

Physical DV perpetration was associated with alcohol use, tobacco use, marijuana use, knife carrying, juvenile delinquency, school failure, school truancy, physical child abuse victimization, peer violence victimization, and exposure to community violence, controlling for respondents' age and school clustering (Table 3). Each of these associations was present for boys and girls when we stratified by sex, with the exception of school truancy, which was not significantly related to physical DV among girls (Table 3).

In this locally representative sample of public high school–attending youth, we found that past-month physical DV perpetration was common (18.7%), although less common than peer violence or sibling violence. We also found that the perpetration of physical DV only was relatively rare. Consistent with previous research, we found that physical DV perpetration was associated with peer violence perpetration^9- 10 and sibling violence perpetration.²⁵ Also consistent with previous research, we found that the relationship between physical DV and peer violence perpetration was stronger for boys than for girls.¹⁰ Our study adds to the existing literature by analyzing a locally representative sample of public high school students and by quantifying the extent of the overlap between physical DV and other forms of interpersonal violence perpetration. Two strengths of our study are that we assessed past-month violence (rather than past-year or lifetime history), which may have limited recall bias, and that our measure of DV perpetration specified that playful aggression was not included.

Consistent with at least 16 previous studies, girls in this sample were more likely to report physical DV perpetration than boys.^7,12,26- 39 There are several possible explanations for this finding. First, it has been suggested that assessments of DV that do not solicit information about sexual violence may yield underestimates of perpetration by boys.⁴⁰ Second, assessments that do not take into account the severity of the assault or injuries mischaracterize perpetration by obscuring sex-based severity differences.^41- 42 However, 2 studies have found nearly equivalent rates of injury reported by male and female victims of DV,^7,27 and at least 3 studies have found that a greater proportion of girls than boys report perpetrating “severe” DV.^7,29,43 Third, boys may be more likely to be dishonest about their DV perpetration on surveys, and girls may be more likely to overestimate their perpetration (eg, report playful hitting as abuse),⁴⁴ although research suggests that adult men and women alike tend to underreport their own partner violence perpetration.⁴⁵ This issue has yet to be investigated among adolescents. Fourth, female reports of violence against dating partners may reflect self-defensive actions, whereas male reports may not.⁴⁶ Finally, girls in our sample may have reported their DV perpetration accurately. Because this study was not designed to explore the full situational context or consequences of DV perpetration, we were unable to determine why we found a higher perpetration rate among girls than boys.

Our hypothesis that physical DV perpetration would be associated with a range of other adolescent problem behaviors was supported by our results and is consistent with research on adult male batterers, which suggests that perpetration of physical violence against partners often co-occurs with other forms of interpersonal violence, criminality, and substance use.^47- 49 According to the problem-behavior theory of Jessor,⁵⁰ certain adolescents are prone to engage in a variety of co-occurring behaviors such as underage alcohol use, cigarette smoking, marijuana use, precocious sexual activity, delinquency, and “acting out” against society because of underlying psychological and environmental factors. We propose that physical DV perpetration be added to the list of problem behaviors that co-occur among adolescents who may be characterized as being at high risk for other delinquent and unhealthful activities.

Longitudinal research could help to determine whether peer or sibling violence tends to precede physical DV perpetration. Consistent with social learning theory, which posits that children learn dysfunctional relationship behaviors from those close to them, children who use violence with siblings or peers may be on a trajectory toward partnership violence.⁵¹ Peer and sibling relationships may be training grounds for dating relationship behavior, which in turn may be a rehearsal for adult intimate partnership. If this hypothesis is supported, intervening with potentially violent adolescents at earlier stages in their development, when they first begin to show signs of aggression toward peers and siblings, may help to prevent partner violence perpetration later. Clinicians are encouraged to consider asking parents and pediatric patients about sibling and peer violence and DV and to discuss the importance of curbing abusive behavior.

Our results are limited by several factors. Our DV definition is limited to physical abuse. Had we included emotional abuse in our definition, more than half the sample would have been classified as perpetrators. Our definition permitted us to focus on those who had engaged in behavior with the potential to produce physical injury and to assess the overlap with comparable forms of physical aggression toward peers and siblings.

We did not assess the intent or motivation for perpetration or whether the violence was perpetrated in self-defense. These contextual variables would be critically important if our goal was to attempt to explain the disparity in male and female perpetration rates. That type of inquiry was beyond the scope of this investigation. Also, data were self-reported, and there is a concern that violence perpetration and other socially undesirable behaviors might have been underreported. However, a comparison of our results with data from the YRBS from Boston high schools showed that self-reported rates of underage alcohol use, marijuana use, and suicidal ideation were nearly equivalent on the 2 surveys.⁵² Therefore, although underreporting may have been a factor, it does not appear to have affected our survey more than a comparable locally representative survey. Finally, our sample was restricted to those in school who reported dating in the past month and, for the overlap analysis, had siblings. Therefore, results may not be generalizable to students who are chronically truant, live in other cities, date less frequently or not at all, and have no siblings.

In conclusion, investigating and addressing the overlap among DV and peer and sibling violence may help reduce all these problem behaviors. It is imperative that researchers and health care practitioners who focus on DV consider the co-occurrence of physical DV perpetration and other adolescent problem behaviors, including violence against nonintimates. The discovery of root causes of these overlapping problems and methods for addressing them should be a priority for research and practice.

Correspondence: Emily F. Rothman, ScD, Department of Community Health Sciences, Boston University School of Public Health, 801 Massachusetts Ave, Floor 3, Crosstown Center, Boston, MA 02118 (erothman@bu.edu).

Accepted for Publication: May 19, 2010.

Author Contributions: Dr Rothman 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: Rothman and Johnson. Acquisition of data: Johnson and Azrael. Analysis and interpretation of data: Rothman, Johnson, Azrael, Hall, and Weinberg. Drafting of the manuscript: Rothman, Johnson, Azrael, Hall, and Weinberg. Critical revision of the manuscript for important intellectual content: Rothman, Azrael, Hall, and Weinberg. Statistical analysis: Johnson, Azrael, and Weinberg. Obtained funding: Azrael.

Financial Disclosure: None reported.

Funding/Support: This study was supported in part by grant U49-CE00740 from the Centers for Disease Control and Prevention (CDC), grant 1K01AA017630-01A1 from the National Institute on Alcoholism and Alcohol Abuse (NIAAA), grant R03DA025823 from the National Institute on Drug Abuse (NIDA), and a grant from the Robert Wood Johnson Foundation (RWJF) New Connections program.

Role of the Sponsor: The NIAAA, NIDA, and RWJF played no role in the design and conduct of the study; in the collection, management, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. A CDC staff person (Dr Hall) participated in data interpretation and manuscript preparation.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the CDC, NIAAA, NIDA, RWJF, or the National Institutes of Health. The views expressed in this article are those of the authors and should not be interpreted as those of the US Congressional Budget Office.

Online-Only Material: This article is featured in the Archives Journal Club. Go to here to download teaching PowerPoint slides.

Additional Contributions: The Boston Youth Survey was conducted by the Harvard Youth Violence Prevention Center in collaboration with the Boston Public Health Commission (Barbara Ferrer, PhD, executive director), Boston's Office of Human Services (Larry Mayes, MA, chief), and Boston Public Schools. The survey would not have been possible without the participation of the faculty, staff, administrators, and students of Boston Public Schools.