See other articles in PMC that cite the published article. Abstract Prior research has examined parental and peer influences on teen dating violence TDV , but fewer studies have explored the role of broader social contexts. Using data from the Toledo Adolescent Relationships Study TARS , the present research examines the effect of variations in school context on teen dating violence perpetration, while taking into account parental, peer, and demographic factors.
School-level dating norms non-exclusivity in relationships also contribute indirectly to the odds of experiencing TDV.
However, a more general measure of school-level violence toward friends is not strongly related to variations in TDV, suggesting the need to focus on domain-specific influences. Implications for theories emphasizing social learning processes and for TDV prevention efforts are discussed.
Numerous studies demonstrate linkages between early exposure to violence within the family and the odds of experiencing violence within later adult romantic relationships e. This association is also demonstrated in studies of the phenomenon of teen dating violence TDV Simon and Furman ; Wolfe and Wekerle ; Wolfe et al.
Although adolescent romantic relationships generally do not involve the same constraints and levels of interdependence as adult relationships, recent research documents troubling rates of perpetration and victimization during this phase of the life course e. Research also demonstrates associations between early adolescent exposure and risk for later IPV e.
Recognizing areas of overlap in risk factors i. Specifically, the second decade of life is characterized by increased interest in friendship and peer relationships Brown and Bakken This suggests the utility of investigating further the potential impact of such extra-familial influences on the odds of experiencing teen dating violence. In prior examinations of peer effects, researchers have explored the role of social skills deficits Linder and Collins , antisocial friends Capaldi ; Foshee et al.
One arena that has not been researched extensively in prior studies is school influence, a context where adolescents are exposed to both close friends and a broader network of similarly situated peers.
Second, we consider that a comprehensive understanding of the role of peers requires assessment of the broader social climate within which both friendships and romantic relationships develop. We construct indicators of the normative climate of schools by aggregating reports about hitting friends and hitting romantic partners to create school-level indices.
Finally, we recognize that school climates may encompass a range of attitudes and behaviors concerning dating and sexuality, and not simply variations in the presence or absence of violence. Accordingly, we determine whether school environments characterized by relatively liberal attitudes about dating relationships indexed by non-exclusivity in relationships influence the odds of experiencing TDV perpetration.
Moreover, in a recent review, Foshee, Reyes and colleagues conclude that child abuse, as the more direct form of victimization, is more consistently linked to later violence perpetration within intimate relationships. Studies using both retrospective and prospective reports Cui et al.
However, some researchers suggest that it is inappropriate to conceptualize the intergenerational transmission of violence as an inevitable process Thornberry, Knight, and Lovegrove The Role of Adolescent Friendships Recognizing that peer relationships are central to child and adolescent development e.
Some of this research emphasizes attachment processes, which generally stress continuity between early family dynamics and the quality of ties formed later in the life course Bowlby ; Cook, Buehler and Fletcher ; Cui et al. This theoretical perspective leads to a social skills deficit hypothesis, including the idea that a lack of strong attachment to parents may be associated with a lack of attachment within peer and romantic contexts.
Some research has suggested that these skill deficits and the inability to form close ties are implicated in the experience of intimate partner violence perpetration and victimization Busby, Holman and Walker ; Dutton ; Dutton, Starzomski and Ryan ; Wekerle et al. Social learning theories adapt a more neutral stance on the intimacy of these ties, focusing greater attention on the content of attitudes and specific behavior patterns that are acquired and reinforced through recurrent interaction and communication with others Sutherland Findings from prior research do provide support for the current examination of extra-familial influences, although most of this research has been confined to assessments of the attitudes or behavior profiles of close friends.
In a recent exploration of the direct transmission of violence from peer to romantic relationship domains, longitudinal analyses indicate that adolescents with friends who perpetrate dating violence are significantly more likely to perpetrate dating violence themselves Foshee et al.
Conversely, research has shown that adolescent reports of hostile interactions and violence within their close friendships are associated with general hostility as well as both IPV perpetration and victimization within later romantic relationships Stocker and Richmond ; Williams et al. Longitudinal analyses have also shown that, for males, deviant peer associations and hostile talk about women among peers during adolescence are significantly and positively associated with aggression toward a romantic partner in young adulthood Capaldi et al.
Thus, it is generally recognized that friendships necessarily unfold against a broader socio- economic landscape that may directly and indirectly influence a range of behavioral outcomes, including violence Bronfenbrenner ; Cohen and Felson ; Morenoff, Sampson and Groves ; Sampson and Raudenbush ; Sampson, Raudenbush and Earls Although many studies have explored the role of neighborhood effects on crime and general violence, more recent work has begun to examine neighborhood contexts in relation to violence with intimate others among both adolescents and adults Browning ; Reed et al.
These studies of neighborhood effects suggest the general importance of the broader social context, yet scholars with an interest in youth development e. Within the confines of the school, unique status systems and norms emerge that draw from the broader neighborhood environments in which they are located but are never an exact replica Corsaro , ; Eder et al. In an examination of general patterns of violence, Felson et al. More recently, Klein, Cornell, and Konold analyze the relationship between school climate and a range of different risk behaviors, including carrying a weapon to school and engaging in physical fights.
The researchers find that students who feel that bullying and teasing are widespread at their school are more likely to engage in risk behaviors themselves. The studies described above examine how general exposure to violence may affect both generally violent and TDV-specific behavior, but most studies have not examined variations in school-level exposure to TDV itself.
In this study, school climate is measured by the percentage of students at each university who report being physically attacked or who injured a dating partner in the last year. In addition to the finding that child maltreatment neglect increases the likelihood of intimate partner violence, attending a university with a high level of dating violence is positively associated with IPV perpetration at the individual level. Further, the link between childhood maltreatment and engaging in violence against a partner is stronger at universities in which dating violence is more prevalent.
In the current analysis, we rely on responses of all individuals in the study who attend the same school to construct aggregated measures of TDV perpetration, as well as school-level reports about perpetrating violence against friends.
Yet it is potentially useful to broaden the scope of our inquiry and conceptualization of what constitutes the normative climate of a school to include nonviolent attitudes and behaviors that may also increase violence risk.
Specifically, in recognizing that dating violence may stem from dynamics that are unique to intimate relationships, we extend our assessment of school climate to include variation in norms about behavior within the dating realm. Adolescents are keenly interested in the world of dating and sexuality, but do not have an extensive backlog of experience about how best to conduct their romantic lives.
Thus, social learning and socialization processes are ongoing as individuals gradually develop understandings about appropriate ways to conduct this type of relationship in particular e.
While infidelity references dyadic behavior and a third party, a sense of what is considered desirable, acceptable, tolerable, or subject to derision is learned through processes of socialization Harris Related to this social learning process, definitions of what constitutes cheating, or non-exclusivity, in romantic relationships vary. As Wilson et al.
Thus, it is important to understand variations in the normative climates to which young people are exposed, as some contexts may be characterized by higher levels of non-exclusivity and related concerns that connect to negative emotions and conflicts within the romantic realm. In the current analysis, we explore whether the broader school climate with respect to norms about this aspect of dating behavior influences variations in TDV self-reports—in general and after taking into account other violence risk factors.
Social Learning beyond the Realm of Close Ties As noted briefly above, most research on the mechanisms underlying social learning theories have concentrated on the centrality of close intimate ties. Sutherland and other theorists e. The work of Eder et al. These researchers focus heavily on various forms of communication that are not limited to the small circle of close friends—such as gossip, storytelling, teasing and ridicule that serve to communicate and, in effect, create localized cultural worlds during the adolescent period Fine and Kleinman These forms of communication, combined with direct observation, create cultural knowledge and act as a socializing influence that transcends the attitudes and opinions of close friends.
Given its more intimate and private nature, compared to other forms of violence, TDV may be less likely to be observed by those outside the romantic partnership.
However, adolescents may observe some TDV at school or in other social settings. Yet, as with traditional forms of school violence, the communication that surrounds a given act of violence e. For example, in a series of focus groups, Johnson and colleagues found that some teens view partner violence as reflecting that a given relationship is a serious one, or a sign of being in love, rather than a destructive pattern that calls for immediately breaking up with the offending partner.
Acts that are seen as constituting various types of infidelity are subject to direct observation and gossip, as well as other forms of communication about how such violations of trust should be viewed and managed. Yet this investigation was limited to a relatively advantaged school environment, suggesting the need to explore variations in these normative climates around such dating norms i.
This investigation of extra-familial influences thus includes attention to school-level a violence, b reports of non-exclusivity, and c TDV, as well as traditional peer and family indicators. Subsequently, we explore whether these factors differentially influence male and female self-reports of TDV perpetration.
We assess whether three aspects of school normative climate prevalence of non-exclusivity, violence within the friendship context, and reports of TDV make a difference for understanding TDV, net of experiences within the more immediate circle of family and close friends. This analysis relies on data derived from a study with a strong relationship emphasis, the Toledo Adolescent Relationships Study TARS and includes respondents who attend a range of different schools.
We develop a third indicator focusing on dating norms; specifically, the prevalence of non-exclusivity across the various school contexts. Due to the young age of those in the sample and in line with more inclusive concepts of what constitutes cheating or non-exclusivity Feldman and Cauffman ; Wilson et al.
We examine whether school-level reports of non-exclusivity are significantly related to reports of intimate partner violence. Thus, we assess how specific domains of the school climate general violence, the prevalence of TDV and non-exclusivity are tied to TDV. The sampling frame of the TARS study encompasses 62 schools across seven school districts, and the initial sample was drawn from 7th, 9th, and 11th grade enrollments records, although school attendance was not a requirement for inclusion in the study.
The names and addresses of potential participants were obtained through a complete roster of all students enrolled in Lucas County schools, available under the Ohio Open Records Act.
Accordingly, all public school districts within the larger county participated in the present study. Devised by the National Opinion Research Center, the stratified random sample includes over-samples of Black and Hispanic adolescents.
Based on analysis of U. Census data, the TARS sample is similar to the national population in estimates of race and ethnicity, family status and income, and education. For example, among those 12—18 years old in , The TARS structured data were collected in the years , , , , and , and the quantitative analyses presented here rely on structured interviews conducted at wave 1, when respondents are, on average, 15 years of age.
The final analytic sample thus consists of respondents male and female respondents. Certainly victimization experiences figure into a comprehensive portrait of dynamics within each context; yet we also estimated models relying on alternative indices i. Due to the skewed distribution of responses to these items, TDV perpetration is dichotomized, where individuals reporting perpetrating any of these acts are coded as 1.
A victimization scale composed of the same measures is used in supplemental analyses in which victimization is the dependent variable. Key Independent Variables Individual-level social learning constructs Parental violence is based on an item that asked respondents how often their parents push, slap, or hit them during disagreements.
Due to the young age of the sample and in line with more inclusive concepts of what constitutes infidelity Feldman and Cauffman ; Wilson et al. Respondents reporting positively about at least one of these measures are coded as 1, whereas only those respondents reporting they have never engaged in any of the three behaviors are coded as 0. Contextual factors School-level friend violence is based on responses to four items: Each of these individual-level responses is dichotomized, where only those respondents reporting that they have never been violent toward their friends are coded as 0 and 1 otherwise.
Based on these percentages, we categorize schools into low contrast category , high, and midrange categories. Thus, the bottom third of schools with the lowest percentage of friend violence are categorized as low, the top third of schools with the highest percentage of friend violence are categorized as high, and the remaining schools are categorized as midrange. School-level partner violence is based on responses to the four CTS perpetration items used in the measurement of the dependent variable.
Thus, these items include how often the respondent has: As with school-level friend violence, this measure reflects the percentage of respondents within each school who report being violent with their partners.
These percentages are then used to classify schools into low, serving as the contrast category; high, and midrange levels of partner violence, as was done with the school-level friend violence measure.
The distinction between this measure and the school-level friend violence measure allows for the examination of whether and to what degree exposure to TDV-specific violence is significantly related to TDV perpetration, when general school-level violence exposure has been taken into account. Finally, school-level non-exclusivity is the aggregated response to the individual-level non-exclusivity measure.
The measure included in analyses is the percentage of respondents within each school who report non-exclusivity within the context of their current or most recent relationship. Controls Sociodemographic indicators gender, age, race, family structure, and socioeconomic status and relationship status are included in the analyses and represent traditional predictors of TDV. Socioeconomic status is based on two separate indicators.