ETV in adolescence is associated with multiple forms of violent victimization throughout the life course including peer victimization, dating violence (i.e., an adolescent form of IPV), and IPV in adulthood (Ehrensaft et al., 2003; Gómez, 2010; Reijntjes, Kamphuis, Prinzie, & Telch, 2010; Tillyer, 2013; Turanovic & Pratt, 2015; Voith, Topitzes, & Reynolds, 2016; Widom, Czaja, & Dutton, 2014). ETV can be divided into direct and indirect exposures. Direct exposure is defined as experiencing violent victimization like physical, sexual, and psychological abuse, whereas indirect exposure to violence is defined as witnessing forms of violence like abuse between family members or violent crime in the community (e.g., physical fights, shootings) (Evans, Davies, & DiLillo, 2008; Finkelhor, Turner, Shattuck, & Hamby, 2015; Sternberg, Baradaran, Abbott, Lamb, & Guterman 2006). ETV is a prevalent risk factor for youth, with more than two thirds of US youth 17 years and younger experiencing at least one form of direct or indirect exposure to violence annually (Finkelhor et al., 2015). Of youth reporting ETV, 60.8% reported violent victimization (e.g., physical assault) and 24.5% reported witnessing violence in the forms of family conflict and community-level violent crime in the past year (Finkelhor et al., 2015). Yet, individuals face the greatest ETV risk in adolescence and emerging adulthood (approximate ages of 12 and 24 years), as youth in this age range experience more ETV than any other age group (Truman & Rand, 2009). This increased risk for exposure may be developmentally-related as adolescents begin to engage in more person-context interactions, develop interpersonal skills, and with increased autonomy spend more time outside of the house in community (including school) settings (Shanahan, 2000). African American youth are at particular risk for exposure to community-based violence due to their disproportionate overrepresentation in urban areas, characterized by significantly higher levels of violence than rural and suburban areas (Browning et al., 2017; Cooley-Strickland et al., 2009; Voisin, 2007). Not only are African American adolescents more likely than White and Hispanic youth to experience direct violent victimization, but also more likely to have witnessed a violent crime or know someone who has been in a violent altercation (Paxton, Robinson, Shah, & Schoeny, 2004; Perron, Gotham, & Cho, 2008). Adolescent exposure to forms of interpersonal violence may shape how youth understand and engage with others, which in turn can have lasting influences on how youth engage in early adult relationships (Heinze, Hsieh, Aiyer, Buu, & Zimmerman, 2020). As such, it is important to study the relationship between adolescent ETV exposure and adult IPV outcomes in populations most susceptible for experiencing violent events.

Researchers differentiate between isolated exposure to victimization, consistent or repeated experiences of one type of victimization, and polyvictimization (i.e., exposure to multiple and repeated forms of violence) and the sequelae of negative outcomes over the life course (Finkelhor, Baradaran, Abbott, Lamb, & Guterman, 2015; Heyman & Slep, 2002; Hughes, Parkinson, & Vargo, 1989; Sternberg et al., 2006; Turner, Shattuck, Finkelhor, & Hamby, 2017). As data from the 2015 National Survey of Children’s Exposure to Violence indicated, 38.7% of children surveyed reported more than one form of direct victimization in the past year, and of those children who reported any direct victimization, 64.5% reported multiple co-occurring forms of violence (Finkelhor et al., 2015). The tendency of violent events to co-occur underscores the need to assess violence exposure across socio-ecological levels (i.e., direct violence, indirect family violence, indirect neighborhood violence) as risk factors for later violent experiences. Moreover, developmental theory provides some explanation for how multiple types of violence exposure in the home and community may lead to later IPV risk directly by strengthening the expectation that violence is a method through which to resolve conflict, as well as indirectly through stress and coping processes (Reese-Weber & Kahn, 2005; Semenza, 2019; Wenzel, Glanz, & Lerman, 2002).

We approach this work from a developmental psychopathology perspective, as we are interested in understanding the mechanisms through which adolescent ETV may lead to adult IPV victimization, and our longitudinal design allows us to examine these processes as they emerge across sensitive developmental periods. The positive association between adolescent ETV and adult IPV victimization may be explained in part by Social Learning Theory. Social Learning Theory posits that children and adolescents develop behavioral norms, attitudes, and beliefs through observation and behavior-modeling and has been applied to IPV contexts (Bandura, 1971; Jouriles, Norwood, McDonald, & Peters, 2001). Adolescence is a particularly salient period of developing behavioral norms and learning to make autonomous choices, as well as a time when dating begins and conflict-solving strategies are necessary in developing friendships and relationships. Youth exposed to violence within multiple socio-ecological domains may expect and accept the use of violence as a means of communicating and solving interpersonal conflicts (Cascardi, 2016; Gómez, 2010; Huesmann & Guerra, 1997). Moreover, chronic stressors like ETV can compromise adolescents’ everyday decision-making and processing (Whitfield et al., 2003). We suggest this normalization of violent communication in adolescence may increase the risk of violence in intimate partnerships, specifically among vulnerable populations experiencing an undue burden of violence-related stressors (Johnson, 2006).

Social learning operates primarily at the individual and interpersonal-levels and is consistent with previous frameworks to understand the relationship between ETV and IPV which tended to focus on individual risk factors such as child abuse (i.e., direct ETV) or exposure to interparental IPV (i.e., indirect ETV) (Capaldi et al., 2012; Heyman & Slep, 2002; Voith et al., 2016). Yet, researchers have also noted that empirical studies need to include other levels of the social ecology to better understand the influence of factors beyond the individual-and interpersonal-levels (Capaldi et al., 2012; Heise, 1998). Ecological theory postulates that personal development is influenced by individual, family, and community influences and can be applied to understand how ETV at multiple levels may have a cumulative effect on experiences of IPV (Bronfenbrenner, 1986). Community-level factors, such as witnessing community violence, frequently co-occur with factors of an adolescent’s interpersonal environment and should be accounted for in order to understand the full ramifications of violence exposure. Researchers, for example, found predictive relationships between negative neighborhood factors and IPV victimization, including that concurrent exposure to community violence was positively associated with IPV victimization in adult populations (Beyer, Wallis, & Hamberger, 2013; Thulin, Heinze, Kusunoki et al., 2020). Combined with individual-level exposures, neighborhood violence compounds social learning of violence behavior and adds to risks associated with ETV, contributing both directly and indirectly to longer term violence outcomes, including IPV. Despite the increased risk, not all youth exposed to violence experience later IPV victimization. Given the myriad negative sequelae of ETV, it is necessary to consider whether intervening mechanisms can explain some of the heightened risk for IPV experienced by youth exposed to violence. We explore two documented consequences of adolescent ETV-- mental distress and substance use--that are also correlates of IPV victimization and are candidates for mediating mechanisms.

Adolescent ETV is associated with increased risk for internalizing problems (e.g., post-traumatic stress disorder, depression, anxiety) and high-risk coping behaviors (e.g., substance use, sexual risk-taking) (Aiyer, Heinze, Miller, Stoddard, & Zimmerman, 2014; Eisman, Stoddard, Heinze, Caldwell, & Zimmerman, 2015; Evans et al., 2008; Golding, 1999; Gorman-Smith & Tolan, 1998; Heleniak, King, Monahan, & McLaughlin, 2018; Heyman & Slep, 2002; Holt, Buckley, & Whelan, 2008; Howell, 2011; Jouriles, Rosenfield, McDonald, & Mueller, 2014; Moylan et al., 2010; Sternberg et al., 2006; Vu, Jouriles, McDonald, & Rosenfield, 2016). At the community level, exposure to community violence predicted depressive and anxiety symptoms in female African American youth while indirect community-level violence exposure can also lead to internalizing symptoms and emotional dysregulation (Chen, 2010; Heleniak et al., 2018). Moreover, poor mental health is a negative outcome of trauma, including ETV, and is often associated with increased substance use (Griffin, Botvin, Scheier, Epstein, & Doyle, 2002; Repetto, Zimmerman, & Caldwell, 2004). In addition to their own effect on well-being, negative outcomes of exposure to violence in childhood and adolescence, including both internalizing behaviors and substance use and other maladaptive coping strategies, are known to increase risk for IPV in adulthood (Capaldi et al., 2012; Walton-Moss, Manganello, Frye, & Campbell, 2005).

The Transactional Model of Stress and Coping is useful for evaluating mental distress and substance use as processes of maladaptive coping with stressors such as direct and indirect exposure to violence. The model suggests that stressful experiences are interpreted as transactions between an individual and their environments while encountering stressful events (Wenzel et al., 2002). Stressors have a negative effect on psychological well-being, which in turn is associated with negative coping mechanisms to protect against the deleterious effects of stressors (Lazarus & Folkman, 1984). Stress does not affect all people equally because the effects of an external stressor may be mediated by the person’s assessment of the stressor and the psychological, social, and cultural resources available to individuals (Lazarus & Folkman, 1984; Wenzel et al., 2002). ETV can be viewed as a significant chronic stressor that has long-lasting detrimental effects on an individual’s health and developmental adjustments, and the psycho-physiological effect of such stressors is particularly salient among African American youth (Aiyer et al., 2014; Heinze, Stoddard, Aiyer, Eisman, & Zimmerman, 2017). Consequently, adolescents with high levels of ETV may develop depression, anxiety, and substance use behaviors in response to excessive stress, each of which further link to antisocial behavior, social isolation, and heightened risks of IPV. From a developmental psychopathology perspective, ETV exposure for some adolescents may set off a cascade of maladaptive coping behaviors across emerging adulthood that in turn increase later IPV risk. Because mental distress and substance use frequently co-occur in youth, it is important to examine both as mediators to understand the pathways between ETV and IPV (Griffin et al., 2002; Repetto et al., 2004). We therefore examine mental distress and substance use as potential mediators of the pathway between ETV in adolescence and IPV in adulthood.

We examine processes by which trauma and stress resulting from exposure to violence may solidify and ultimately characterize the coping mechanisms and violence expressed in intimate partner relationships in a predominantly African American sample residing in a high risk, urban environment. We aim to fill several gaps in the literature focused on adolescent exposure to violence and its effect on IPV: we use a prospective longitudinal design; utilize a cumulative measure of exposure to violence capturing direct and indirect forms of exposure; and examine mediation effects of mental distress, measured as depression and anxiety, and substance use, measured as alcohol and marijuana use, on the relationship between adolescent ETV and adult IPV victimization (Figure 1). Specifically, we test the direct association between adolescent ETV (approximately ages 15-18) and later reported IPV (approximately age 32) and a hypothesized parallel mediation model where emerging adult (ages 20-23) substance use and mental distress explain some portion of the total effect between adolescent ETV and IPV experienced as an adult. ETV is operationalized as the cumulative effect of three types of violence during adolescence - direct violent victimization by any person, indirect observed family violence, and indirect exposure of observing violence in the community - across time. We expect that adolescent ETV will be positively associated with subsequent experience of IPV in adulthood. We hypothesize mental distress and substance use in emerging adulthood will mediate the associations between cumulative adolescent ETV and adult IPV victimization, such that ETV will be associated with both higher levels of mental distress and substance use, which in turn predict more IPV experiences. In addition, researchers have noted systematic sex, race/ethnicity, level of education, and socioeconomic status differences in both rates of violence exposure and IPV victimization (Fleming, McCleary-Sills, Morton, Levtov, Heilman, & Barker, 2015; Rennison & Planty, 2003; Turner et al., 2017). Age at baseline may also have a potential influence on violence exposure as older participants may accrue a higher number of violent experiences (Heinze et al., 2017). We therefore included these variables as covariates in the analyses to account for potential confounding due to demographic influences.

Bivariate correlations showed small to moderate positive associations between ETV and each IPV outcome (Table 1). ETV was moderately correlated with alcohol/marijuana use at both baseline and in emerging adulthood. Similarly, ETV was correlated positively with psychological distress at baseline and in emerging adulthood. Age was correlated with ETV experience with older students reporting more exposure. In contrast, ETV was negatively associated with the highest education achieved. No other demographic variables were correlated with ETV. Descriptive statistics for demographic covariates included in the analyses are reported in Table 2.

Our results demonstrate that individuals who are exposed to higher levels of violence in adolescence are at a higher risk for IPV exposure later in life. These findings are particularly important given that exposure to violence in adolescence predicted IPV over 15 years later. We found that ETV, a notable adolescent stressor, was consistently associated with mental distress in emerging adulthood, which aligns with predictions from developmental psychopathology theory as well as the Transactional Model of Stress (Lazarus & Folkman, 1984). According to psychopathology theory, stressful and traumatic experiences precipitate a sequelae of maladaptive coping behaviors and increase risk for negative health outcomes (Cox et al., 2010). Similarly, the Transactional Model of Stress suggests persistent stress environments are interpreted by individuals as threats and prompt an appraisal of coping resources. (Lazarus & Folkman, 1984; Wenzel et al., 2002). We build on previous studies which found associations between ETV and higher levels of anxiety and/or depression, but demonstrate how the effect of ETV on mental distress in turn increased the likelihood of reporting both physical and psychological IPV victimization (Gorman–Smith & Tolan, 1998; Heinze et al., 2017; Howell, 2011; Vu et al., 2016). Young adults with a history of ETV, who are managing depression and anxiety, may have access to fewer resources or perceived alternatives to violence as a means of communication than young adults without such a history, thus increasing the likelihood of future incidences of IPV victimization. Yet, the developmental timing of the increases in mental distress also points to a period for intervention given that most emerging adults are still forming schemas related to intimate relationships in emerging adulthood (Desmarais et al., 2012). Our hypothesis that substance use would also manifest as a contributing factor of ETV-IPV mechanisms was only partially supported. Researchers have suggested that youth use alcohol and marijuana as a way of self-medication to cope with emotional distress (Repetto et al., 2004; Repetto, Zimmerman, & Caldwell, 2008). Substance use, however, is one of many possible coping mechanisms which may explain why substance use does not operate as consistently as mental distress in our study as a mediator that predicts subsequent IPV experiences. Further, a study among young adults found that African American women reporting marijuana use had decreased odds of IPV victimization but increased odds through binge drinking, suggesting problematic substance use may function differently in predicting IPV outcomes depending on use behavior (Nowotny & Graves, 2013). The mixed evidence suggests further research into understanding substance use as a coping behavior in relation to ETV in adolescence and adult violence outcomes may be warranted.

Incorporating nine measurement occasions spanning 17 years, our study addresses a limitation of cross-sectional studies examining connections between ETV and later IPV experience (Voith et al., 2016). Our ETV measure captures four years of data collection which preceded the measurement of both the intermediate (substance use and mental distress mediators) and distal (IPV) outcomes addressing concerns of recall bias commonly introduced by retrospective reports of ETV in childhood and adolescence. Further, including adolescent ETV experiences that were temporally prior to the mediating influences during emerging adulthood, in addition to controlling for baseline levels of both substance use and mental distress, adds confidence in the identified indirect effects of ETV. An implication of our findings is that adolescent ETV can influence later IPV risk through multiple pathways. Full or partial mediation of ETV was present in both single mediation models, although the indirect effects for substance use attenuated in the parallel mediation model. Yet our test of two theoretically driven mechanisms that are associated with both ETV and IPV yielded generally expected results. Although across outcomes these mediators accounted for between 35-55% of the total effect, ETV is known to influence myriad emerging- and young-adulthood outcomes with emotional, social, and physiological ramifications. This suggests that more longitudinal testing of mediating mechanisms, such as negative peer influence or sexual risk behavior, can further inform how ETV contributes to IPV victimization (Hope et al., 2019; Voisin, Jenkins, & Takahashi, 2011).

The findings add to research on ETV by including a multi-faceted measure of exposure, combining family violence, victimization, and observed violence experiences. Researchers note the deleterious effects of such cumulative exposure, as well as the disproportionate burden of cumulative ETV on African American populations (Finkelhor et al., 2015; Voith et al., 2016). Critically, our multifaceted, cumulative measure of adolescent victimization was predictive of reported IPV, despite not including measures of dating or other relationship violence in adolescence. The effects of violence exposure on future violent experience, therefore, may not be specific to a particular domain of violence exposure (e.g., peer violence leading to more peer violence). Victims of peer or observed violence in adolescence, for example, may still be at higher risk for IPV later in life even though such experiences are not related to intimate or dating relationships. Consistent with Social Learning Theory, this may be the result of the normalization of violence as an acceptable method for handling family/partner discord (Patterson, 2016). Researchers have found that attitudes regarding violence as a way to solve problems mediates exposure to violence and violent behavior (Stoddard, Heinze, Choe, & Zimmerman, 2015). Considering participant conflict resolution strategies or endorsement of violence as a way to solve problems may be an additional mechanism through which ETV affects later IPV victimization.

The findings of the present study support a developmentally-informed conceptualization of risk and suggest that it may be important to assess violence exposures outside of the household during adolescence, as these may be particularly salient at this age and shape behavior (Desmarais et al., 2012; Johnson, Giordano, Manning, & Longmore, 2015). Other researchers have examined trajectories of risk across childhood and adolescence, finding relationships between early childhood exposures such as child abuse to be predictive of dating violence during youth and IPV in adulthood (Capaldi et al., 2012). Still others suggested that risk related to adverse childhood experiences fluctuates across time relative to developmental periods during childhood, preadolescence, and adolescence (Thulin, Heinze, & Zimmerman, 2020). Risk during childhood may be more dependent upon family experiences and relationships, but the increased autonomy across preadolescence and adolescence may confer developmentally-specific risks, with later implications for relationship functioning in adulthood. For example, youth often have their first romantic relationship during preadolescence or adolescence. As such, exposure to violence in different contexts may inform youth perceptions of interpersonal interactions, including respect for partners or the use of coercion or force.

The finding that being male was predictive of physical IPV victimization was not expected. In tradition with feminist theory, male victimization is often described as reactionary perpetration by a woman to a male partner (Carney, Buttell, & Dutton, 2007). While this may explain some of the male victimization in the present study, another potential explanation may be situational couples violence (Johnson, 1995; 2006). Researchers have found that situational couples violence is the predominant form of IPV in population-level studies (Graham-Kevan & Archer, 2003; Kelly & Johnson, 2008). In situational couples violence, male victimization is not abnormal, nor is it conceptualized as being the result of reactionary violence. Situational couples violence occurs due to poor interaction norms within a couple, which might be aggravated by stressful situations. ETV is an example of a persistent stressor disproportionately experienced by African-Americans, which can lead to differential experience of mental distress for African American men as compared with African American women (Paradies et al., 2015; Williams & Williams-Morris, 2000). This study gives support that persistent stress as a result of cumulative ETV increases both negative mental health outcomes and heightens the impact of contextual risk factors for IPV, when conceptualized as situational couples violence.

Our results suggest that future research that continues to expand the conceptualization of ETV experiences across ecological levels and over vital developmental periods would be useful for informing preventive interventions. In addressing adolescent populations more broadly, these exposures may include cultural norms around the use of violence which is portrayed in media, with a potential focus on the difference between general portrayal of interpersonal violence in non-romantic relationships and the portrayal of violence or coercion in intimate or sexual relationships. Additionally, future research on the relationship between adolescent ETV and subsequent experience of IPV may also benefit by including protective effects of factors like community engagement and social support, which have been identified as particularly relevant measures of resilience among vulnerable populations (Dumont & Provost, 1999; Eisman et al., 2015; Howell, 2011). In a systematic review of risk factors for IPV, Capaldi and colleagues (2012) identified mixed evidence of an association between community-level measures like collective efficacy, social cohesion, and social control and IPV. Thulin, Heinze, Kusunoki, and colleagues (2020) examined various types of community-level risk factors and found that the measurement of community-level risk factors is often constrained to factors derived from social disorganization theories like the Broken Windows Theory, which focuses on the association between the decay of the built environment and social risks like crime and violence (Sampson, Raudenbush, & Earls, 1997). Expanding our understanding of ETV from a developmental perspective, expanding Social Learning Theory to account for exposures across ecological levels, and examining both risk and protective factors are important future directions which would add nuance to our understanding of the relationship between ETV, mental health and coping behaviors, and IPV.

Our study provides a new understanding of the progression of victimization experienced by youth and young adults in a disproportionately vulnerable population. Our results indicate that exposure to violence is a persistent reality with multiple negative sequelae that extend well beyond adolescence. Practically, mental health professionals should consider complete histories of victimization, including community contexts and events resulting in persistent stress exposure that may contribute to future anxiety or depression. Prevention of adult IPV may begin with early detection of violence exposure among youth and monitoring for associated consequences such as depression, anxiety, and relationship conflict.