Abstract
Decades of inquiry on intimate partner violence show consistent results: violence is woefully common and psychologically and economically costly. Policy to prevent and effectively intervene upon such violence hinges upon comprehensive understanding of this phenomenon at a population level. The current study prospectively estimates the cumulative incidence of sexual and physical dating violence (DV) victimization/perpetration over a 12-year timeframe (2010–2021) using diverse participants assessed annually from age 15 to 26. Data are from Waves 1–13 of an ongoing longitudinal study. Since 2010 (except for 2018 and 2019), participants were assessed on past-year physical and sexual DV victimization and perpetration. Participants (n = 1,042; 56% female; Mage baseline = 15) were originally recruited from seven public high schools in southeast Texas. The sample consisted of Black/African American (30%), White (31%), Hispanic (31%), and Mixed/Other (8%) participants. Across 12 years of data collection, 27.3% experienced sexual DV victimization and 46.1% had experienced physical DV victimization by age 26. Further, 14.8% had perpetrated at least one act of sexual DV and 39.0% had perpetrated at least one act of physical DV against a partner by this age. A 12-year cumulative assessment of physical and sexual DV rendered prevalence estimates of both victimization and perpetration that exceeded commonly and consistently reported rates in the field, especially on studies that relied on lifetime or one-time specified retrospective reporting periods. These data suggest community youth are at continued and sustained risk for DV onset across the transition into emerging adulthood, necessitating early adolescent prevention and intervention efforts that endure through late adolescence, emerging adulthood, and beyond. From a research perspective, our findings point to the need for assessing DV on a repeated basis over multiple timepoints to better guage the full extent of this continued public health crisis.
Physical and sexual dating violence (DV) is a pressing public health and safety concern among adolescents and emerging adults (Smith et al., 2018). DV encompasses a host of abusive interpersonal behaviors and is often defined by physically and sexually violent behaviors such as hitting, shoving, and sexually assaulting a partner. Estimates from the National Intimate Partner and Sexual Violence Survey indicate that 36% and 34% of adult women and men, respectively, are victims of intimate partner sexual violence, physical violence, and/or stalking in their lifetime (Smith et al., 2018). Beyond potential physical harm and injury, physical, psychological, and sexual DV victimization and perpetration are consistently associated with an array of adverse physical and mental health outcomes, including psychopathology, substance misuse, risk for future victimization, and chronic health consequences across the lifespan (Exner-Cortens et al., 2013; Jouriles et al., 2017; Smith et al., 2018). Moreover, partner violence is costly to Americans, with one estimate putting the population economic burden at $3.6 trillion over victims’ lifetimes (Peterson et al., 2018).
Rates of DV Perpetration and Victimization
Given its impact, DV prevention among adolescents and emerging adults has become a national public health and policy priority (e.g., American Psychological Association Resolution, 2019; ongressional Research Services, 2023; Williams et al., 2020). In efforts to better define and monitor DV, the public health model of violence prevention (Mercy et al., 1993) attempts to accurately estimate the prevalence and incidence of DV to understand the development of such violent behaviors, including risk and protective factors. Adolescents and emerging adults are at the highest risk for first incidents of DV victimization. Over 71% of victimized women and 55% of victimized men first experienced sexual or physical IPV before age 25 (Smith et al., 2018). Meta-analytic results of racially and ethnically diverse samples suggest that approximately 20% of teens experience physical DV and 9% experience sexual DV (Wincentak et al., 2017). This rate is consistent with US trends over a 12-year timeframe from 1999 to 2011, in which approximately 9% of adolescents reported experiencing past-year physical DV at each timepoint (Rothman & Xuan, 2014). DV perpetration and victimization peak during the transition from adolescence to emerging adulthood (Johnson et al., 2015; Shorey et al., 2017; Smith et al., 2018) clustering with other risk behaviors across adolescence (McNaughton Reyes et al., 2020; Orpinas et al., 2017). While the risk for DV victimization and perpetration in both adolescence and emerging adulthood has been well studied, the research has primarily been siloed, looking only at adolescents or emerging adults, which limits our understanding of the continumum of development in these critical periods. With respect to the link between dating violence and race and ethnicity, research has been largely inconsistent; however, accumululating studies show hightened rates of victimization for Black and Hispanic youth (Boothe et al., 2014). To be clear, being Black or Hispanic is not a risk factor for DV. Instead, it is the systemic and everyday racism and discrimination that people of color experience that may increase the likelihood of DV victimization and perpetration (Caldwell et al., 2004; Clark et al., 2004).
Assessing DV Prevalence
It is essential that we establish accurate and comprehensive estimates of DV during the critical development transition from adolescence into adulthood. Over the last several decades, researchers have sought to refine the measurement of DV, addressing concerns about context, intent, and gender parity (Follingstad & Bush, 2014; Hamby, 2016a, 2016b; Rothman, Cuevas, Mumford et al., 2022). Despite this work, few innovations have occurred in understanding population-level incidence of these data, especially as it relates to approaches to robustly establishing prevalence and onset. Extant literature on DV prevalence and incidence largely relies on estimates of DV victimization and perpetration assessed using lifetime or one-time specified retrospective reporting periods (e.g., within current relationship; within past 1–12 months) (Bonomi et al., 2013; Smith et al., 2018; Vagi et al., 2015; Ybarra et al., 2016). These approaches risk underreporting of DV due to respondents forgetting or misremembering events (i.e., when asking about lifetime) or limiting their recall to short spans of time or limited partners. Alternatively, cumulative assessment approaches aggregate repeated measurements across (often shorter) reference periods rather than rely on a single-point retrospective report (Caiozzo et al., 2016; Jouriles et al., 2005; Krauss et al., 2020). Since it is likely that reference period influences prevalence rates, a cumulative approach can mitigate some measurement error by capturing violence “missed” by lifetime lookbacks. Past use of cumulative assessment strategies resulted in higher prevalence rates of DV among adolescents and, importantly, improved criterion validity with psychological correlates (Jouriles et al., 2005; Krauss et al., 2020). Thus, cumulative assessment functions as a valid and potentially more accurate assessment of DV involvement than previously documented, making the onset of certain behaviors more identifiable and thus refining prevention approaches
Current Study
Cumulative violence estimates have yet to be applied across the developmental transition from adolescence into emerging adulthood, or beyond aggregated periods of 3 months. The objective of this study is to prospectively estimate the cumulative incidence of sexual and physical DV victimization and perpetration over a 12-year timeframe using community participants from diverse ethnic/racial and socioeconomic backgrounds assessed annually from age 15 to age 26 (2010–2021), spanning the transition from adolescence to emerging adulthood.
Methods
Participants and Procedures
Participants (N = 1,042) were drawn from on ongoing longitudinal study, Dating it Safe (Blinded for Review). Participants were initially recruited from ninth and tenth grade high school classes across seven campuses in five school districts of southeast Texas. The sample (56% female) is 30% Black, 31% White, 31% Hispanic, and 8% Other. Average age at baseline assessment was 15. Recruitment concluded in Spring 2010, with the aim of enrolling students from urban, rural, and suburban schools (62% response rate). Schools were recruited based on their representative make-up of ethnically diverse and low-income students. Every school district/high school approached agreed to participate in the study. Recruitment occurred during school hours in courses with mandated attendance to ensure a representative sample of adolescents at each school. All students were eligible to participate. Active and written parental consent was obtained at baseline. Participant assent was obtained at each wave and consent obtained when participants reached the age of majority.
The cohort has been assessed annually (except for 2018 and 2019, in which the authors lacked funding) and is ongoing. All procedures were approved by the first author’s Institutional Review Board. At the most recent completed data collection wave (2021), 21% of the sample were married, 35% had children, and 18% identified as a sexual minority. All original cohort participants, provided they have not already declined to participate (n = 67) or passed away (n = 3), are eligible to participate at every wave and attempts are made to collect data regardless of prior pattern of missing data. Data were collected via paper/pencil while participants were in high school and via a web-based service through email and text contact after graduating or dropping out of high school. Attrition analyses comparing Wave 1 to Wave 10 revealed little to no differences in demographic and main study variables, including all forms of DV perpetration and victimization.
Measures
Physical and sexual DV perpetration and DV victimization were assessed using physical (8 items) and sexual (6 items) abuse subscales of the Conflict in Adolescent Dating Relationships Inventory (CADRI) (see Table 1; Wolfe et al., 2001). As we have previously shown with this sample, the CADRI is a suitable behaviorally specific approach for measuring multiple types of DV over time across sex and race/ethnicity (Shorey et al., 2019). Participants indicated whether each act occurred (yes/no) during a conflict or argument with a current or former dating/marital partner during the past 12 months. Questions are asked twice—first in relation to perpetration, and then in relation to victimization. At each wave, a participant was coded as “yes” to victimization/perpetration for each subscale (sexual, physical) if they endorsed any of the items. At the initial assessment (Mage = 15) students were asked to recall lifetime experiences. All subsequent assessments were based on a 12-month timeframe. Baseline prevalence was estimated using the initial participant assessment.
Conflict in Adolescent Dating Relationships Inventory (Wolfe et al., 2001).
Note. Recall period for measure is last 12 months. Response options for each item were Yes/No. DV = Dating Violence.
Data Analysis
A cumulative estimate of DV victimization/perpetration across subsequent follow-ups was estimated by selecting a subsample of participants for each of the four outcomes selected on those reporting no behavior at baseline. Each subsample was then followed prospectively to assess subsequent experiences with DV. Starting with the first 12-month incidence estimate, for each follow-up, if a participant indicated victimization/perpetration their value was recoded to “yes.” This was done for each subsequent wave, which provided a cumulative estimate of victimization/perpetration across the follow-up period. If a participant was missing a follow-up observation, we conservatively assumed no victimization/perpetration during that 12-month interval. Rates were estimated both overall and stratified by sex.
Results
As shown in Table 2, subsamples of DV were similar across sex and race/ethnicity; however, the physical DV perpetration subgroup was composed of fewer females and slightly more White participants than the other subgroups.
Sample Demographics.
Note. DV = Dating Violence.
The baseline lifetime prevalence and 95% confidence intervals (CIs) for DV victimization and perpetration are presented in Table 3. For sexual DV victimization, the prevalence was 9.1% with females having significantly higher sexual DV victimization than males (12.5% vs. 4.7%, p < .001). The prevalence of sexual DV perpetration was 3.1% of the sample and did not significantly differ by sex. Baseline prevalence of physical DV victimization was 21.8% with males reporting higher levels of victimization than females (24.3% vs. 19.8%, p = .05). Physical DV perpetration at baseline was 21.7% and was significantly higher for females than males (29.4% vs. 12.0%, p < .001, respectively).
Baseline Prevalence of Dating Violence Victimization and Perpetration (Age 15).
Note. Sample size denotes those with valid response to all scale items. Bold = Sample level.
The cumulative incidence rates of sexual and physical DV perpetration and victimization post-baseline are presented in Table 4 for each of the four subgroups, including sexual DV victimization (n = 832), sexual DV perpetration (n = 890), physical DV perpetration (n = 711) and physical DV victimization (n = 713). New incidence of sexual DV victimization over the first consecutively assessed 7 years of follow-up was 20.7% (Mage = 22) and increased to 27.3% at the most recent follow-up (Mage = 26). Incidence of sexual DV perpetration over that same period was 11.2% (Mage = 22) and 14.8% (Mage = 26), respectively. The 7-year cumulative incidence of physical DV victimization was 41.1% (Mage = 22) increasing to 46.1% at the most recent follow-up (Mage = 26). Physical DV victimization had a 7 year cumulative incidence of 33.7% (Mage = 22), which increased to 39.0% at the most recent follow-up (Mage = 26). As shown in Figure 1, new incidences of DV victimization and perpetration appear to peak at around age 20.
Cumulative Incidence of Dating Violence Victimization and Perpetration.
Note. M = male, F = female; Study paused in 2018 and 2019 due to lack of funding. DV = Dating Violence
Cumulative incidence of DV age 15 to 26 among those with no prior experience.
Cumulative incidence rates by sex are also provided in Table 4 and indicate that at the most recent year of data collection (2021), just under half of males (43.4%) and females (48.2%) have been a victim of physical DV by the time they were, on average, 26 years old. Further, 24.2% of males and 53.3% of females have perpetrated at least one act of physical DV against a partner by this age. With respect to sexual DV, 19.7% of males and 33.8% of females have been victims by age 26, while 16.4% of males and 13.9% of females reported perpetration. Notably, participants endorsed items within these scales equally; in other words, a single item was not driving these cumulative rates.
Discussion
Using data from one of the longest longitudinal studies to annually assess the cumulative incidence of physical and sexual DV victimization and perpetration, results reveal alarmingly high rates: by age 26, over a quarter of participants had been victims of sexual DV and nearly half had been victims of physical DV. Similarly, 15% and 39% had reported perpetrating sexual and physical DV against a dating partner, respectively. Notably, these numbers are underestimates as we did not capture revictimization rates. Given the plethora of research detailing the negative mental, physical, and social consequences of experiencing DV (Exner-Cortens et al., 2013; Haynie et al., 2013; Jouriles et al., 2017; Orpinas et al., 2017; Smith et al., 2018), including risk for future partner violence, these numbers are concerning. That the rates are higher than estimates gleaned from cross-sectional (Haynie et al., 2013; Jouriles et al., 2022; Vagi et al., 2015), retrospective (Bonomi et al., 2013; Smith et al., 2018), and longitudinal studies that only consider one (Ybarra et al., 2016) or a few time-points suggest that partner violence may be even more prevalent than previously reported. Further, in using a cumulative incidence approach, we found higher rates of DV than even outlier (cross-sectional) studies using high-risk samples (Alleyne-Green et al., 2012; Martin-Storey, 2015) and comprehensive measures (Niolon et al., 2015). This pattern of results align with estimates derived from similar cumulative assessments over shorter timespans (Jouriles et al., 2005; Krauss et al., 2020; Smith et al., 2003). Importantly, rates exceeded the field’s consensus on the prevalence for both sexual and physical DV in just three aggregated follow-ups, suggesting that future longitudinal research need not expend extensive resources to capture these numbers. The current study contributes to our understanding of assessing DV among adolescents and emerging adults, supporting the use of repeated assessments to accurately obtain prevalence of DV, in part by mitigating the risk of participants forgetting or misremembering events that happened in the distant past.
Our findings also indicate that DV impacts diverse populations and both sexes. While we found no significant impacts by race or ethnicity, there were some differences by sex. By age 26, women report slightly higher levels of physical dating violence victimization and significantly higher rates of physical DV perpetration, as well as sexual DV victimization. That women reported perpetrating more physical DV across adolescence and emerging adulthood is consistent with other research (Niolon et al., 2015; Ontiveros et al., 2020; Straus et al., 1996; Ybarra et al., 2016). This may partly be a result of measurement approaches that are overly simplistic or lack contextualization about self-defense and “horseplay” (Hamby, 2016a, 2016b). Further, as articulated first by Johnson (2006), couples use violence in different ways, not all of which involve power and control. Measurement approaches used for DV among adolescents and emerging adults may differ in their attention to criterion validity, including the ability to predict future DV and past behaviors. Previous research has indicated that males tend to underreport and females over report their behaviors (Ackerman, 2018). Established research also indicates that women sustain more severe and injurious forms of violence than men (Archer, 2000), including sexual assault (Hamby, 2017; Hamby & Turner, 2013; Vagi et al., 2015) and homicide (Messing et al., 2021) from an intimate partner. Sex-related measurement differences for DV merit further examination and testing approaches, especially periods of developmental transition. Nevertheless, these findings emphasize the importance of DV prevention programs addressing all individuals.
Practice and Policy Implications
DV prevention should continue to be a national priority. These results underscore the scope of DV as it impacts adolescents and emerging adults. Thus, governmental funding mechanisms must increase and continue to support research and practice efforts in DV prevention, detection, and intervention. Earmarking funds for clinical science trainees, program delivery grants, and well-designed intervention trials is a worthy funding priority to produce and sustain mental health providers and researchers working with DV. Prevention work is needed across developmental spans, with adaptations made for the unique context of adolescents, including an emphasis on social networks, as well as sexual and relationship development. Primary prevention efforts should start early. At baseline, 4.9% and 13.9% of participants had already experienced sexual and physical DV, respectively. These data, coupled with other studies (Johnson et al., 2015; Shorey et al., 2017), indicate the importance of implementing effective prevention efforts early (e.g., middle school). In a cluster randomized controlled trial, we found that students in 7th grade who received a healthy relationships curriculum were less likely to perpetrate DV a year later, relative to their peers who received a standard health curriculum (Temple et al., 2021). Similarly, middle school students participating in Dating Matters exhibited significantly less physical DV perpetration and victimization over time, relative to standard of care students (Niolon et al., 2019), and a coach-delivered intervention for middle school males reduced physical and sexual DV perpetration among dating youth over a 1-year follow-up, compared to a no-treatment control group (Miller et al., 2020). Further, 6th grade students who received a classroom-computer hybrid DV program reported less DV perpetration and some specific types of DV victimization at the 1-year follow-up (Peskin et al., 2019). Thus, guided by dissemination and implementation science, these and other efficacious programs should be standard middle school curricula nationwide.
Findings also indicate the need for routine DV screening, universal education, and referral for support during adolescence and emerging adulthood. Youth and emerging adults should be regularly screened and provided universal education for potential DV involvement across healthcare, employment, extracurricular, and academic settings. This recommendation is widely supported in the literature, yet practical barriers to implementing wide-spread, universal screening present challenges (Rothman, Cuevas, Mumford et al., 2022; Rothman, Campbell & Hoch, 2022). Notably, presence of an internalizing or externalizing disorder is associated with subsequent involvement in physical DV victimization and perpetration before the age of 21 (McCauley et al., 2015). Thus, screening for DV among youth seeking psychological treatment and youth with childhood exposure to DV is especially encouraged. Given the risk for DV in emerging adulthood, college campuses, along with campus climate assessments, should routinely assess for DV in campus health care and provide supportive services (Krause et al., 2019).
Finally, primary, secondary, and tertiary prevention must be ongoing. These data also demonstrate that prevention efforts must extend throughout adolescence and into emerging adulthood. Indeed, one contribution of our study is that we observed steady development of risk for DV victimization and perpetration into emerging adulthood, with many participants’ first violent relationship not materializing until they were 21 or older (n = 66 new cases of physical DV victimization and n = 71 new cases of physical DV perpetration). Still, and consistent with the limited longitudinal research (Johnson et al., 2015), rates of violence, including new incidences of DV, appeared to peak at age 20. Risk windows for physical and sexual DV appear during adolescence (as teenagers begin dating and experimenting with risky behaviors) and the transition to emerging adulthood (as individuals enter college or the workforce and their intimate relationships intensify). Current results suggest the importance of providing all forms of DV prevention throughout this timeframe. Further, given the rate of new instances of DV, comprehensive primary prevention should not be limited to middle or high school—our findings indicate new incidences of victimization and perpetration peaking at age 20, indicating prevention should extend post-high school and engage emerging adults. We must address the impact of victimization and perpetration to prevent revictimization of DV. Secondary and tertiary prevention approaches, including advocacy, counseling, and economic support, are critical to preventing DV from reoccurring and mitigating negative health impacts (Ogbe et al., 2020).
Limitations
While our study had several strengths, including annual assessments over many years, solid retention, comprehensive DV measures, and a large contemporary and ethnically diverse sample, there were also limitations. Our DV measure is intended for adolescents and emerging adults and did not include more severe forms of DV typically assessed on other DV measures such as the National Intimate Partner and Sexual Violence Survey (NISVS; Smith et al., 2018), Measure of Adolescent Relationship Harassment and Abuse (MARSHA; Rothman, Cuevas, Mumford et al., 2022), and the revised Conflict Tactics Scale (Straus et al., 1996) (e.g., strangulation, coercive control, use of weapons, burning) nor does it attend to context like the Partner Violence Scale or the Youth Risk Behavior Survey (Hamby, 2016a, 2016b). By not including these latter items, we missed an opportunity to determine if severity of violence (as opposed to relying on acts of violence) differs by sex, race/ethnicity, and socioeconomic status. Future research examining DV into emerging adulthood should capture a broader range of physical DV tactics, as well as coercive control, economic, and emotional violence, and the impact of DV tactics. Another limitation, inherent in much of the DV literature, is how participants defined “romantic partner” over the years. It is likely that some participants characterized this as someone they dated once or twice while others considered romantic partners to be a boyfriend, girlfriend, or spouse (Collins et al., 2009).
As can be expected, there was not follow-up data for every participant at every point in time. The cumulative incidence coding scheme started with the first follow-up observation and only updated that response to change from a “no victimization/perpetration” to a “yes” at subsequent measures when data was available. For any wave that a subject was missing follow-up observation their cumulative value was carried forward to the next timepoint. Therefore, a missing observation would be a missed opportunity to convert from “no victimization/perpetration” to “yes,” which likely resulted in a slight downward bias of cumulative incidence and an underestimate of the actual value.
Due to funding constraints, we did not have data for two successive years, which likely resulted in lower rates of DV. The use of self-reports may have introduced reporting biases such as social desirability, especially in the assessment of DV perpetration (Sugarman & Hotaling, 1997). However, previous studies examining problem behaviors in this age group, including DV, support the validity of self-report measures (Wolfe & Mash, 2006). Further, stringent efforts were made to encourage honest reporting with assurances of confidentiality, including a federal certificate of confidentiality, use of participant identification numbers, and a message that information participants provide is kept private.
Conclusions
Despite peaking in the early 20s, risk of victimization and perpetration onset does not appear to completely wane with age, as might be expected. Emerging adults do not developmentally “grow out” of risk with improving executive functioning or as they become more experienced with relationships and more risk averse. Given the sheer volume of potential victims and perpetrators across any given population (e.g., student, patient, employee), we must screen for DV early, often, and in multiple settings. These findings underscore the urgent need for comprehensive and accessible services to help address the impacts of DV among victims, and the crucial need to build a more robust network of approaches to interrupt DV perpetration behaviors. Further, we must do a better job of disseminating effective and universal primary, secondary, and tertiary interventions to prevent the onset and revictimization (and perpetration) of DV. Structural, economic, and practice-oriented changes are needed to address high rates of DV and associated impacts. From a research perspective, our findings point to the need for assessing DV on a repeated basis over multiple timepoints to better gauge the full extent of this continued public health crisis.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interests with respect to the authorship and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: All phases of this study were supported by Award Numbers K23HD059916 and R01HD099199 (PI: Temple) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), as well as Award Number 2012-WG-BX-0005 from the National Institute of Justice (NIJ). The content is solely the responsibility of the authors and does not necessarily represent the official views of NICHD or NIJ.
