The designation of violent and abusive behavior as a U.S. public health priority is evidenced by its inclusion in the Healthy People 2000 and 2010 objectives (U.S. Department of Health and Human Services, 2000a, 2000b). Intimate partner violence is an important subdomain of such behavior. Overall, studies indicate that the prevalence of non-sexual courtship violence ranges from 9% to 65%, depending on the definitions and research methods used (Centers for Disease Control, 2000). Moreover, it is now apparent that the reach of relationship violence extends well into the world of adolescence. In acknowledgment of this, a series of Healthy People health status objectives specifically targets physical partner abuse and assault leading to injury among females as young as 12 years of age.
Two probability samples (Silverman, Raj, Mucci, & Hathaway, 2001; Halpern, Oslak, Young, Martin, & Kupper, 2001) and several nonprobability samples (Foshee, Linder, Bauman, Langwick, Arriaga, Health, McMahon, & Bangdiwala, 1996; Jackson & Foshee, 1998; Gray & Foshee, 1997; Foshee, 1996) have provided the basis for what is currently known about the prevalence of adolescent dating violence and some of its psychosocial correlates. The Centers for Disease Control (2000), in summarizing the extent of dating violence in young adults, reported that the average prevalence for high school and college students is 22% and 32% respectively. Recently, Silverman et al. (2001), utilizing Massachusetts Youth Risk Behavior Survey (YRBS) data from 1997 and 1999, reported that about one in five adolescent high school girls has experienced dating violence. Examining data from the National Longitudinal Study of Adolescent Health, collected during the 1994-95 school year, Halpern et al. (2001) calculated the prevalence of psycholo gical and minor physical violence victimization among 7ththrough 12th-grade adolescents in romantic heterosexual relationships, and found that 32% reported experiencing some kind of violence in dating relationships in the 18 months prior to the interview. While the majority of these reports involved psychological violence, such as swearing, 12% reported physical victimization experiences (e.g., being shoved, having something thrown at them).
In an attempt to establish a theoretical framework for understanding and preventing adolescent dating violence, research has moved forward on several fronts, including a focus on individual and interpersonal influences, as well as ecological contexts such as the home, school, and community (Foshee et al., 1996). A lifespan perspective has led to the examination of developmental histories of youth who report such experiences (Jackson & Foshee, 1998). Behavioral correlates associated with dating violence victimization, among girls at least, include use of a variety of illicit substances, unhealthy weight control practices, sexual risk behaviors, and suicidality (Silverman et al., 2001). Conceptually, a problem behavior model as articulated by Jessor (Jessor, 1982, 1991), Dryfoos (1990), and others (DiClemente, Hansen, & Ponton, 1996; Resnick, Bearman, Blum, Bauman, Harris, Jones, Tabor, Beuhring, Sieving, Shew, Ireland, Bearinger, & Udry, 1997), to account for the apparent clustering of risk behaviors among cer tain adolescents, may be relevant for dating violence as well. While there is debate over whether health risk behavior among adolescents represents a single behavioral syndrome or multidimensional structure, evidence suggests that the co-occurrence of multiple problem behaviors such as substance use, delinquency, and sexual behavior among certain youth (Basen-Engquist, Edmundson, & Parcel, 1996; Farrell, Danish, & Howard, 1992) may have underlying personality and cognitive problem-solving referents (Siegel, Cousins, & Rubovits, 1993). The risk profile of adolescent girls who are victims of dating violence fits this framework; however, the evidence is far from definitive.
The present study examines the prevalence of adolescent dating violence and its relationship to demographic, psychological, and behavioral risk factors among a large, nationally representative sample of U.S. youth. While there are strong indications that violence in adolescent dating relationships involves the reciprocal use of violence by both partners (Gray & Foshee, 1997; Foshee, 1996; Gaertner & Foshee, 1999; O'Keefe, Brockopp, & Chew, 1986), the focus here is on adolescent females. Many have argued that girls are more likely than boys to suffer injury from partner violence (Gelles, 1981; Lane & Gwartney-Gibbs, 1985). In addition, this orientation is consistent with national health objectives which target women as victims of assaultive dating violence.
This study advances our present knowledge in two important ways. First, it provides current and comprehensive information on dating violence among adolescent girls in the U.S. Consequently, findings from this study have national significance. Prevalence data on adolescent dating violence should provide the kind of empirical evidence necessary to evaluate whether Healthy People 2010 objectives are on target to be achieved. Specifically it provides baseline data, enabling comparisons with subsequent prevalence rates. Second, it aims to broaden the profile of at-risk female adolescents. To this end we will examine the relationship of physical dating violence with other forms of violence, emotional health, sexual behavior, and substance use, and attempt to establish a more complete cluster of risk factors.
METHOD
Sample
The 1999 national school-based Youth Risk Behavior Survey (YRBS) data were used for this study. The YRBS is one component of the Youth Risk Behavior Surveillance System (YRBSS) that was established by the Centers for Disease Control and Prevention (CDC) to monitor the prevalence of youth behaviors that most influence health. Comprehensive design and sampling procedures are available from the CDC's Morbidity and Mortality Weekly Report series, and are presented here in condensed form (CDO, 2000). The YRBS used a three-stage cluster sample design to produce a nationally representative sample of high school students in grades 9-12. The first stage contained 1,270 primary sampling units (PSUs), which consisted of large counties or groups of smaller, adjacent counties. From these, 52 were selected from 16 strata, formed on the basis of the degree of urbanization and the relative percentage of African-American and Hispanic students in the PSU. Each PSU was selected with probability proportional to the size of the s chool enrollment. The second sampling stage selected 187 schools, also with probability proportional to school enrollment size. The third stage randomly selected one or two intact classes of a required subject (e.g., English or social studies) from the 9th-l2th grades at the chosen school. All students in the selected classes were eligible to participate. A total of 15,349 students in 144 schools completed survey questionnaires. The school response rate was 77% and the student response rate was 86%, resulting in an overall response rate of 66%. For the purpose of this study, only female participants (N = 7,824) were selected for the analysis.
A weighting factor was applied to each participant to adjust for non-response and for the varying probabilities of selection, including those resulting from the oversampling of African-American and Hispanic students. The weights were scaled so that (a) the weighted count of students was equal to the total sample size and (b) the weighted proportions of students in each grade matched national population proportions. The data were representative of students in grades 9-12 in public and private schools in the 50 states and the District of Columbia.
Survey Procedures and Measures
Surveys were administered anonymously in order to protect the participants' privacy. Participation was voluntary, and students completed the self-administered questionnaire in their classrooms during a regular class period. They recorded their responses on a scannable answer sheet. Parental permission was obtained before survey administration.
The survey focused on health risk behaviors established during adolescence that result in the most significant mortality, morbidity, disability, and social problems for youths and, later, adults. Monitoring progress in these areas was relevant in assessing the degree to which national health objectives for the year 2000 were achieved. Results were also to be used to help focus programs and policies for comprehensive school health education on those behaviors that contribute most to the leading causes of mortality and morbidity.
