OBJECTIVES: To assess gender differences in the proportion of adolescents reporting teen dating violence (TDV) and the frequency of TDV at multiple age points across adolescence in a high-risk sample of youth with previous exposure to violence.
METHODS: A cross-sectional, high-risk sample of boys and girls (n = 1149) ages 11 to 17 years completed surveys assessing TDV and self-defense. Indices of TDV included perpetration and victimization scales of controlling behaviors, psychological TDV, physical TDV, sexual TDV, fear/intimidation, and injury.
RESULTS: More girls reported perpetrating psychological and physical TDV, whereas twice as many boys reported sexual TDV perpetration. More girls reported fear/intimidation victimization than boys. When comparing the frequency of TDV across adolescence, boys reported more sexual TDV victimization at younger ages, and girls demonstrated a trend toward more victimization at older ages. Likewise, younger boys reported more fear/intimidation and injury perpetration and injury victimization than younger girls. However, by age 17, girls reported more injury perpetration than boys, and reports of injury victimization and use of self-defense did not differ. Notably, despite potential parity in injury, girls consistently reported more fear/intimidation victimization associated with TDV.
CONCLUSIONS: Contrary to data suggesting that girls experience far more sexual TDV and injury, these data suggest that at specific times during adolescence, boys among high-risk populations may be equally at risk for victimization. However, the psychological consequences (fear) are greater for girls. These findings suggest a need to tailor strategies to prevent TDV based on both age- and gender-specific characteristics in high-risk populations.
- TDV —
- teen dating violence
What’s Known on This Issue:
Youth exposed to violence in the home or community are at heightened risk of dating violence victimization. Results of studies examining gender differences in the rates of dating violence victimization and perpetration differ depending on the age of the sample.
What This Study Adds:
Gender differences in rates of dating violence vary depending on the age of adolescents responding. In high-risk samples of adolescents such as this one, teens in later adolescence may be at equal risk of injury by a dating partner.
Research on gender differences in intimate partner violence has led to 2 different conclusions: (1) intimate partner violence is primarily a male-perpetrated phenomenon, versus (2) both genders perpetrate equally.1 Further complicating this picture, a number of studies indicate that among adolescents, girls may perpetrate teen dating violence (TDV) at rates commensurate to or greater than boys, whereas in adult populations, these differences are reversed.1–7 Some have suggested that the incongruous findings on gender symmetry/asymmetry may be explained by the severity of violence, degree of coercive/controlling behavior, and fear/intimidation involved in measuring violence.2,8 Studies reporting only proportions of perpetrators and victims or the frequency with which perpetration and victimization occur fail to capture the seriousness, contexts, or consequences of such acts (eg, injury, fear, and distress).6 Recently, Hamby and Turner2 reported rates of TDV victimization in a national probability sample. They found that when broad inclusion criteria for TDV were used (ie, any physical act), male rates of victimization were nearly double that of female rates. However, when TDV was defined more restrictively as a physical act that was fear-inducing or injurious, the victimization rate for girls was twice that of boys.2 These data suggest that discrepant findings in gender differences for TDV may be a consequence of measurement and lack of uniform definitions.1 Moreover, attitudes about the acceptance of violence seem to differ by gender and may vary across adolescence, which may influence rates of perpetration and victimization among genders at different stages of development.6,9–11
The current study adds to the literature by investigating gender differences for both perpetration and victimization of psychological, physical, and sexual TDV in a high-risk sample of youth (ie, youth exposed to violence in the home or community). High-risk populations have been underrepresented in TDV research and are an important population to study given the amplified likelihood of future violence victimization and perpetration.2,12 For a number of reasons (eg, economic disadvantage, dangerous family environments or neighborhoods, earlier initiation of dating and/or sex, cumulative risk), violence-exposed youth have developmental trajectories distinct from non–violence-exposed youth, including greater risk of TDV.12,13
We additionally examine gender differences for several TDV variables associated with severe forms of relationship violence that have rarely been assessed among adolescents, including physical self-defense, fear/intimidation, injury, and controlling/coercive behaviors.8,14,15 Finally, as adolescence reflects a developmental period of significant change over a relatively short period of time,16,17 and TDV rates, attitudes about violence, relationship characteristics, and risk behaviors change significantly during this time period,6,18 we examine how gender differences vary from age 11 to 17 years in this high-risk cross-sectional sample.
Participants and Procedures
Participants were 1141 sixth- to 12th-grade students (mean age 14.4 years, SD 1.6, range 11–17) from 35 schools in Texas. The sample comprised 61.9% self-identified females (n = 706) and 38.1% self-identified males (n = 435), with 53.4% self-identifying as Hispanic/Latino (n = 609), 16.4% African-American (n = 187), 12.7% non-Hispanic white (n = 145), 12.9% multiracial (n = 147), and 3.8% “other” (n = 43); 10 students did not respond. Data for the present research are derived from the baseline assessments of adolescents participating in an evaluation of the Expect Respect TDV prevention program.19 Fig 1 displays selection procedures for the final analytic sample.
Students were referred for screening by school counselors or social workers if they suspected, for any reason, the student had been exposed to violence in the home or community, as these youth are at significantly higher risk of TDV victimization, particularly those exposed to multiple forms of violence.13 During a brief intake assessment, semistructured interviews were conducted to assess if students had been the witness, victim, or perpetrator of (1) dating violence, (2) peer violence, (3) adult partner violence, (4) child abuse, or (5) some other form of violence in the home or community at any point in their life. Youth that verbally reported ≥1 type of violence exposure were eligible to participate in the study. Most students (73%) reported multiple forms of violence exposure. Students were informed that all information would be confidential except for disclosures of child abuse or homicidal or suicidal threat, which were reported to the appropriate agencies specified by law.
Data were collected between 2011 and 2013 via paper-and-pencil surveys. A waiver of parental consent was obtained, and an informational letter was mailed to parents or guardians, who were able to opt out by either mail or phone. During the initial interview, facilitators explained confidentiality and mandatory reporting requirements to students, who gave their written assent before participating. All procedures were approved by the Institutional Review Board at the Centers for Disease Control and Prevention and by participating school districts.
Students reported on 6 indices of TDV perpetration and victimization: controlling behaviors (5 items, α = 0.70 and 0.78), psychological TDV (8 items, α = 0.72 and 0.80), physical TDV (5 items, α = 0.76 and 0.82), sexual TDV (6 items, α = 0.69 and 0.76), fear/intimidation (2 or 3 items, α = 0.56 and 0.82), injury (3 items, α = 0.75 and 79), and 1 item pertaining to the use of self-defense. Questions were adapted from the Conflict in Adolescent Dating Relationships Inventory20 and the Safe Dates TDV scales.21 Additionally, items were developed where necessary. Students responded to items on a 4-point scale where 0 = never, 1 = rarely, 2 = sometimes, and 3 = often. Consistent with previous research on TDV, a 3-month reporting period was used owing to the short-lived nature of adolescent relationships as well as to minimize recall error common to retrospective reporting.6 Supplemental Information contains all survey items and instructions.
As the present data are nested within schools, we first tested the effect of clustering. Small intraclass correlations (<0.04), small average cluster size, and nonsignificant variances of parameter estimates between schools indicated the propriety of ordinary least squares regression and χ2 cross-tabulation to test gender differences in TDV. Pairwise deletion was used, as response rates for all TDV indices were high (see Table 1), indicating that missing data should not bias parameter estimates.
We examined the mean frequency of TDV behaviors for boys and girls as well as the proportion of boys and girls reporting presence of TDV. To assess the percentage of youth reporting TDV, 3 sets of binary response items (0 = no TDV; 1 = TDV) were created for the 12 TDV scales and self-defense. Previous research indicates that the degree of “horseplay” and less serious physical behavior is greater among adolescents. Merely measuring the presence of a single act may result in false-positives for TDV, particularly among girls.2,22 Thus, we created increasingly conservative binary outcomes wherein the most liberal outcome grouped students with a score of ≥1 on a scale as positive for the presence of TDV and score of 0 on a scale as no TDV present for that particular scale. The most conservative grouping criteria classified students with ≥3 as positive for the presence of TDV and a score of 0 to 2 as no TDV present.
Differences by age level were tested via moderation and simple slope analysis.23 Specifically, we conducted multiple linear regressions with age, gender, and their interaction term. A significant interaction indicates that gender differences in TDV varied by age. When age-by-gender interactions were significant, we conducted simple slope analysis by testing the coefficient for gender at 1 and 2 SDs above and below the mean age. This process provides a t test comparing TDV frequency between genders at ages 11.2, 12.8, 14.4, 16.0, and 17.6 years. Because of the exploratory nature of this research, we explicated all interactions at P < .10 with simple slopes analysis so that readers may inspect trends in the data. However, we interpret only interactions significant at P < .05.
Proportion of Adolescents Reporting TDV
Table 2 displays the percentage of youth reporting self-defense, TDV perpetration, and TDV victimization by gender. There were no differences in the proportion of boys and girls reporting the use of self-defense, controlling behavior, or injury, regardless of how conservative the classification was for the presence of TDV. There was a pattern wherein more girls than boys reported perpetrating psychological and physical TDV whereas more boys reported sexual TDV perpetration. The percentage of boys and girls reporting TDV victimization tended to be relatively equivalent across indices, with the exception of fear. More girls than boys reported experiencing fear/intimidation in relation to TDV by a partner. Likewise, more girls reported psychological victimization, but this difference dissipated as classification became more restrictive (ie, a pattern of behavior versus a single act). Similarly, more boys reported sexual and physical TDV victimization than did girls, but these differences became nonsignificant as the classification was restricted.
Frequency of TDV
Table 3 displays mean comparisons of the frequency of self-defense, TDV perpetration, and TDV victimization by gender aggregated across ages 11 to 17. For self-defense, controlling behavior, and injury, there were no differences between genders. Girls reported more psychological and physical perpetration, whereas boys reported more sexual TDV perpetration. Girls reported more psychological TDV victimization and fear/intimidation associated with TDV.
Gender Differences by Age
Four outcomes manifested significant interactions, indicating that gender differences changed with age at which they were compared: fear perpetration (β = 0.11, P < .05), injury perpetration (β = 0.15, P < .005), sexual victimization (β = 0.13, P = .01), and injury victimization (β = 0.12, P < .05). At younger ages, boys reported more fear and injury perpetration and sexual and injury victimization than girls, but these patterns dissipated or reversed (ie, injury perpetration) at later ages. Notably, in reference to injury victimization, analyzing only the most severe injury item (ie, “I went to a doctor or nurse because of an injury”) yielded results consistent with the full scale. However, we report the full scale because of increased base rates and stability in standard errors. The interaction terms for physical perpetration (β = 0.09, P = .09), physical victimization (β = 0.10, P = .06), and fear victimization (β = 0.09, P = .09) neared significance. Table 4 presents gender differences at the 5 five different age points.
We examined gender differences among youth exposed to violence in both the frequency of TDV and the proportion of boys and girls reporting different forms of TDV. Findings indicate that gender differences (and lack of gender differences) in the proportion of youth reporting TDV perpetration were mostly consistent across classification criteria from broad (ie, a single act) to restrictive (ie, a pattern of behavior), with more boys reporting sexual TDV perpetration and more girls reporting physical and psychological TDV perpetration. These data also suggest that among youth previously exposed to violence, the proportion of boys and girls that report TDV victimization may be comparable, apart from the number of girls reporting fear. Notably, the proportions of TDV dropped by approximately half for all indices when comparing the broadest to the most restrictive inclusion criteria (see Table 2). This is consistent with findings of Fernández-González et al,24 who reported that correcting for aggression in a joking context reduced the proportion of youth reporting TDV by roughly half.
We also found that age moderates the gender differences in the frequency of TDV in this high-risk sample. At younger ages, boys reported more sexual TDV victimization, but these differences dissipated as age increased and trended toward girls reporting more victimization. It seems surprising that boys would report more sexual TDV victimization than girls at any point in adolescence. However, this may be because our measure of sexual TDV is more representative of sexual coercion (eg, pressuring partner to have sex, grabbing or touching partner’s private parts without consent, spreading sexual rumors) than acts of physical force to penetrate or complete a sexual act, which may be more likely to involve male-to-female perpetration. Nevertheless, this sexual coercion should not be dismissed as minor or inconsequential. Young boys’ sexually coercive victimization could potentially contribute to the development of maladaptive beliefs about the appropriateness of such behaviors and increase their risk to perpetrate similar and more severe acts of sexual violence later in adolescence and adulthood. These findings suggest a need to engage both boys and girls with a history of violence exposure in sexual violence prevention efforts and to focus on clarifying boundaries and consent.
Referring to injury, boys reported significantly more perpetration and victimization at younger ages. However, by age 17, girls reported perpetrating more injury, and there were no differences in victimization. These findings remained unchanged when we looked at only the most severe injuries that necessitated medical attention. The parity in injury rates in this sample is hard to reconcile with nationally representative studies that suggest girls/women are at greater risk.2,25 However, our discrepant findings may be attributable to the use of a high-risk sample rather than national probability samples. For example, in another study of TDV in high-risk populations, girls that carried weapons were 5 times more likely to commit physical and sexual TDV and stalk their partners compared with non–weapon-carrying girls.12 Likewise, in the Baltimore Prevention Project, among high-risk adolescents, girls committed aggression commensurate to that of boys.26 Fernández-González et al27 found that injury rates among boys were commensurate with girls and peaked at age 17 before dropping significantly during ages 18 to 20. However, rates of injury for girls continued to increase through age 20. Thus, it is possible that, in the present sample, the parity in the frequency of injury would not persist beyond age 17.
It is important to note that despite the overall equivalence of injury across genders, the rates of fear associated with TDV for girls were double the rates for boys. Moreover, gender differences in the frequency of fear/intimidation held across age groups: girls consistently reported more victimization than did boys. In fact, this gender difference demonstrated a trend wherein it appeared to increase with age. This finding is consistent with previous research on adolescents and adults, indicating that the psychological consequences of relationship violence, including fear and loss of power or control, are usually greater for females than males.2,25,28 This may also be related to growing differences in physical strength and size as adolescents become older and girls’ perceptions of danger increase. Hence, even when the physical consequences are equal, the lasting psychological consequences of TDV may be greater for girls with a history of violence exposure, which can result in chronic physical and mental health issues.
Importantly, our data suggest that there are significant changes in patterns of dating behaviors over short periods of time (ie, ∼18 months) among high-risk youth. Treating an 11-year-old equally to a 13-year-old or treating a 15-year-old similar to a 17-year-old may be as erroneous as aggregating adolescents with young adults. For example, a single TDV prevention strategy is often administered to students across multiple grade levels.21,29,30 However, developmental differences over 2 or 3 years may be too great for a single program to be effective with youth of differing ages. Rather, it may be necessary to implement prevention strategies tailored to age level in these high-risk populations.12,31
In addition, these findings may inform anticipatory counseling for pediatricians who may be in a unique position to identify TDV; raise awareness and educate caregivers, school officials, and/or authorities about the victimization; and identify the most appropriate treatment strategies. The American Academy of Pediatrics recommends a comprehensive approach to violence prevention and lays out the roles of physicians in screening and assessing for abuse, reporting, educating parents and caregivers, and liaising with other treatment professionals and community agencies.32–35 The American Academy of Pediatrics Connected Kids: Safe, Strong, Secure program32 does provide some of this targeted information, such as “Staying Cool When Things Heat Up” in early adolescence. Another consideration is embedding violence risk screening questions in an electronic health record to prompt clinicians to screen and, when appropriate, refer for interventions among high-risk adolescents.36 For example, screening for the potential psychological consequences of dating violence and their apposite treatment strategies may be particularly important among adolescent girls who have previous violence exposure, and increasingly so as they age.
Our study suggests that direct assessment of sexual abuse at young ages should not be restricted to abuse by adult figures when treating violence-exposed youth. In particular, sexual abuse by a dating partner may start as early as 11 years old for boys; thus, screening questions should similarly begin early. Another promising strategy, brief clinical interventions and counseling for students in school health clinics, demonstrated reductions in relationship abuse.37 Pediatricians could potentially provide education about the appropriateness of sexually coercive behaviors and how to respond to young boys that have been victims, just as they could for girls in later adolescence. Likewise, pediatricians treating older adolescents for injuries should consider and screen for potential victimization by a dating partner for boys and girls alike among high-risk youth. Whether changing anticipatory counseling to include sexual violence and psychological trauma in earlier adolescence to conducting brief interventions or referrals, pediatricians have an important role in preventing TDV in these populations. It will be important for parents, pediatricians, and other practitioners to focus on healthy relationships in both boys and girls and give careful consideration to assessing sexual victimization and injury for boys and girls alike when treating youth exposed to violence. Additionally, giving special attention to assessing the experience of fear/trauma and associated mental health sequelae for girls is critical, as thes experiences can result in chronic health issues into adulthood.7,25 Moreover, considering age in relation to TDV could improve efficacy and precision of prevention strategies for such high-risk youth.
Finally, it is necessary to stress that the generalizability of these results is specific to youth who have been exposed to violence. Rates, causes, correlates, and prevention strategies for TDV have been underresearched in high-risk populations,2,12,13,31 and undertaking work with this population is crucial. The present research may ultimately help inform the development of tailored selective and indicated prevention strategies targeted toward youth most at risk for TDV and consequently result in the greatest reductions in such violence for boys and girls alike.
- Accepted November 16, 2015.
- Address correspondence to Dennis E. Reidy, Division of Violence Prevention, Centers for Disease Control and Prevention, Atlanta, GA 30341. Fax: (770) 488-1662; e-mail:
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by Centers for Disease Control and Prevention Contract #200-2010-34099.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Foshee VA
- Swahn MH,
- Simon TR,
- Arias I,
- Bossarte RM
- Diaz-Aguado MJ,
- Martinez R
- Messinger AM,
- Fry DA,
- Rickert VI,
- Catallozzi M,
- Davidson LL
- Schwartz SJ,
- Côté JE,
- Arnett JJ
- Ball B,
- Tharp AT,
- Noonan RK,
- Valle LA,
- Hamburger ME,
- Rosenbluth B
- Foshee VA,
- Bauman KE,
- Linder F,
- Rice J,
- Wilcher R
- Aiken LS,
- West SG
- Fernández-González L,
- O’Leary KD,
- Muñoz-Rivas MJ
- Breiding MJ,
- Chen J,
- Black MC
- Fernández-González L,
- O’Leary KD,
- Muñoz-Rivas MJ
- Houry D,
- Rhodes KV,
- Kemball RS, et al
- ↵Committee on Injury, Violence, and Poison Prevention (2009). Role of the pediatrician in youth violence prevention. Pediatrics. 2009;124(1). Available at: www.pediatrics.org/cgi/content/full/124/1/e393
- Thackeray JD,
- Hibbard R,
- Dowd MD; Committee on Child Abuse and Neglect; Committee on Injury, Violence, and Poison Prevention
- Sigel E,
- Harpin SB,
- Tung G
- Copyright © 2016 by the American Academy of Pediatrics