Introduction
Family violence, and especially child abuse, is viewed as a major risk factor for delinquency, and especially for violent crime Farrington 1991, Rivera and Widom 1990 and Smith and Thornberry 1995. Violent delinquency in particular has been associated with documented child maltreatment in light of a “cycle of violence” theory, although different studies present contrasting evidence as to the strength of this association (c.f., Smith and Thornberry 1995 and Widom and Maxfield 1996).
Much of the research linking child abuse to crime suffers from limitations in methodology, including over-reliance on cross-sectional or retrospective designs, uncontrolled confounding variables, and lack of control groups (Widom, 1989a). Another limitation in some of the recent research is the classification of abuse derives from past government records (cf., Smith and Thornberry 1995, Widom 1989b and Zingraff et al 1994). This sampling choice means that other risk factors, including those within the family that are known to co-occur with child abuse, remain unmeasured. In addition, undetected abuse practices in the control group are missed, rendering some between-group differences weaker than they might otherwise have been. The present research offers some strengths in that it is prospective, collects interview-based data on abuse in the families, and partials out the unique effects of marital violence and child abuse on children’s arrest records.
Linking exposure to child abuse and marital violence to crime
Abused children are more likely to express problems in a wide array of developmental domains, including social development and peer relations (Salzinger, Feldman, Hammer, & Rosario, 1993), mental health (Cicchetti, Rogosch, & Toth, 1994), school achievement (Eckenrode, Laird, & Doris 1993) and later, crime and antisocial behavior (Smith & Thornberry, 1995).
Although there are fewer research findings isolating the effects on children of witnessing marital violence, many of these youngsters show symptoms of distress congruent with those who are abused (Jaffe, Wolfe, Wilson, & Zak, 1986). These childhood symptoms range from generally elevated overall symptoms Fantuzzo et al 1991 and McCloskey et al 1995 to specific problems, as in the case of posttraumatic stress Graham-Bermann and Levendosky 1998 and McCloskey and Walker 2000 or difficulties with peers (McCloskey & Stuewig, 2001). Marital violence and physical child abuse are known to overlap (Appel & Holden, 1998); but whether either differentially predicts the development of antisocial behavior is unknown. Until now, there has been no longitudinal research examining whether exposure to marital violence in childhood uniquely contributes to later behavior problems in adolescence, especially delinquency.
Gender and delinquency
Many of the studies tracing the parenting origins of delinquency focus exclusively on boys (e.g., Farrington, Barnes, & Lambert, 1996). There is reason to believe that the origins of delinquency vary for girls and boys, especially vis-a-vis their family relationships. Some investigators, for instance, tie female crime directly to women’s victimization Browne et al 1999, Chesney-Lind 1997 and Whitbeck and Simons 1990. Chesney-Lind and Shelden (1998) contend that girls growing up in abusive households develop unique tactics of self-preservation, including running away, that ultimately subject them to criminal exploitation. Their heightened delinquency, therefore, is an indirect result of acting out against or escaping an abusive home-life. Indeed, a recent examination of national arrest statistics for 1998 found that girls accounted for a higher proportion of arrests for running away from home than boys (Snyder, 1999).
It is well known that boys are more delinquent than girls, but whether differences in family dynamics influence this pattern is unclear, and probably unlikely. National statistics indicate that more boys than girls are family targets of neglect and physical abuse (Sickmund, Snyder, & Poe-Yamagata, 1997). Whether victimization, however, is tied more closely to crime in males than in females is unknown (Lake, 1994). To date little is known about how exposure to family violence in childhood effects males and females differently with respect to subsequent delinquency. One aim of the present research is to describe gender differences in the response to family violence to inform juvenile justice services. If there are gender differences in the etiology of crime, these services need to be recast to take into consideration the unique needs of female as well as male offenders.
Results from prospective research indicate a specific relationship between child maltreatment and female delinquency. In one prospective study comparing abused and nonabused children’s juvenile and adult records, girls who were abused and neglected were significantly more likely than their nonmaltreated counterparts to have both a formal juvenile delinquency record and a formal adult criminal record (Widom, 1989b). In another analysis of the same data set, the researchers found that abused and neglected girls showed a tendency for increased risk of arrest for violent juvenile offending; moreover, this trend was not found for abused and neglected boys Rivera and Widom 1990 and Widom and White 1997. Because females tend to engage in violent offending at a much lower rates than males, it has been suggested that those who do engage in violence are responding to their own victimization (Peters & Peters, 1998), whereas boys may be engaging in aggressive acts for a wider variety of reasons (e.g., during the commission of a crime, self defense against other aggressive males, peer pressure, to defend one’s ‘reputation,’ etc.).
The purpose of the current study is: (1) to identify gender differences in the official delinquency status of the adolescents in our sample; (2) to examine the effects of witnessing marital violence and being a victim of child abuse on subsequent adolescent delinquency; and lastly, (3) to examine both forms of family violence as differential predictors of male and female juvenile delinquency. Family members were interviewed when children were between 6 and 12 years of age about conflict, marital violence, and corporal punishment of the children during 1990 through 1991. County juvenile court records then were collected approximately 5 years later (1995) to identify which children in the original sample had arrest records.
This study differs from previous studies in that rather than relying on retrospective data about family violence, mothers and children were interviewed directly at Time 1 about abuse. We examined the effects of witnessing spousal abuse and being a victim of physical abuse, defined as extreme corporal punishment from either parent in childhood. This approach improves on previous studies that are limited only to official reports of direct victimization (neglect or abuse), or retrospective reports of childhood victimization by adults. Interviews provide a richer portrait of family life than State records. Marital violence and physical child abuse were measured in terms of (1) mother’s report of violence by the partner towards the mothers; and (2) mother’s and child’s report of violence by either parent towards the child at Time 1 (1990) of the study.
Delinquency was analyzed in terms of official reports of juvenile arrests in 1995, 5 years after the original interview. Based on the examination of prior research on the trends of female offending (e.g., Snyder, 1999), it was hypothesized that boys would be more likely to be referred to court than girls in most offense categories except status offenses, running away, and petty theft. In addition, it was hypothesized that the main effects of witnessing abuse and being a victim of child abuse would be significant predictors of ever being referred to juvenile court and being referred for violent offenses. It was also anticipated based on some of the research reviewed that girls who commit violent offenses would be more likely to have abusive families than boys.
Results
We describe below the degree of exposure to abuse derived from the interviews (the independent variable), and the rates of court referral and criminal offending of boys and girls (the dependent variables). We then performed logistic regressions to test whether exposure to different forms of abuse in the home predicted criminal profiles, and in particular, whether gender differences emerged. A logistic regression was used because our dependent variables were created as dichotomous variables.
Descriptive statistics
Exposure to violence
Of the entire sample of children (n = 299), 43% witnessed marital violence and a total of 29% directly experienced escalated physical abuse. There was, however, substantial overlap in the forms of abuse to which children were exposed, with 17% exposed to both forms of violence (n = 50). A total of 132 children experienced either form of abuse (see Table 1).
Table 1.
Number of children having been exposed to spousal and/or child abuse
Type of abuse Number of children Males Females
Marital violence (total) 129 66 63
Marital violence only 79 40 39
Child Abuse (total) 88 49 39
Child abuse only 38 23 15
Both forms of abuse 50 26 24
Neither (no abuse) 132 63 69
Table options
Analyses by form of abuse (child abuse, marital violence) indicated that 31% of the children receiving escalated forms of abuse and 33% who witnessed marital violence were referred to juvenile court at least once. Seventeen percent of abused children and 17% of children who witnessed abuse were referred for a violent offense. Of those children without abuse in their childhood, 18% (n = 25) were referred to juvenile court and 5% (n = 7) were referred for a violent offense.
Gender differences in delinquency
Seventy-five adolescents out of the entire sample were referred to juvenile court by the fall of 1995. Chi-square analyses revealed that there were no gender differences in ever being referred to juvenile court, with 44% (n = 33) of those referred being girls. In contrast, there were differences in the profile of offenses. Table 2depicts the results of χ2 analyses run on each offense category by sex. Boys committed more violent crimes [23 boys vs. 9 girls; χ2(1) = 5.71; p = .05], property offenses, [χ2(1) = 4.10; p < .05], and felonies [χ2(1) = 4.32; p < .05] than girls. Girls were as likely as boys to be referred for status offenses, running away, and petty theft.
Table 2.
Proportion of boys and girls who were referred for specific offense categories
N Drug Felony Order Petty theft Property Run-away Status Violent
Males 42 21% 41%∗ 36% 36% 71%∗ 36% 43% 55%∗
Females 33 6% 18% 24% 39% 49% 49% 55% 27%
∗
p < .05.
Table options
One type of violent offense, domestic violence, emerged as a distinct class of arrests. Among the youths ever referred to juvenile court, 23.8% of the boys and 24.2.% of the girls were arrested for a domestic violence charge. These charges were usually parent-child altercations, with youth charged with assault against a parent. On first glance it appears that boys and girls were equally likely to commit domestic violence offenses if they have an arrest record; but domestic violence weighed more heavily in girls’ violent offending than in boys. Among the girls ever arrested for a violent offense, 89% were arrested exclusively for domestic violence. Boys’ violent arrests were more heterogeneous, and their violent crimes were often perpetrated outside of the home. This finding indicates that the context for violence differs for boys and girls, with the domestic sphere an especially central venue for girls’ violence.