Gender differences in caregiver–child relationship mediation of the association between violence exposure severity and adolescent behavior problems

Abstract

The main objectives of this study were to investigate (1) the relationship between mild, moderate, and severe violence exposure in the home and behavior problems in adolescents; (2) the caregiver–child relationship as a potential mediator in this relationship; and (3) gender differences. A series of path analyses were conducted using a sample drawn from the National Survey of Child and Adolescent Well-Being (NCSAW-I) of 848 adolescents (ages 11–15) who had been reported to Child Protective Services for maltreatment and who remained in their homes. Exposure to violence and the caregiver–child relationship were reported by adolescents. Both caregiver ratings and adolescent self-reports were used to assess adolescents’ behavior problems. Path analysis indicated that exposure to mild and severe violence was directly associated with higher levels of child-reported behavior problems. However, exposure to violence was not directly associated with caregiver ratings of adolescent behavior problems. The caregiver–child relationship mediated the relationship between mild and moderate violence on both caregiver and child-reported adolescent behavior problems. Gender differences also emerged; for girls, the caregiver–child relationship mediated the effects of mild and moderate violence, whereas for boys, it mediated the effects of severe violence on behavior problems. Study findings suggest caregiver–child relationships as a critical underlying mechanism in the association between violence exposure and adolescent behavior problems, highlighting the importance of adding the caregiver–child relationship factor to intervention efforts.

Introduction

Research regarding the effects of adolescent exposure to violence has grown significantly within the past three decades. Existing research reveals links between various types of childhood victimization and child outcomes (Evans et al., 2008, Finkelhor et al., 2009). Generally, outcomes associated with violence exposure include an array of internalizing and externalizing behaviors that can manifest from infancy into young adulthood (Evans et al., 2008), and those manifestations can have cascading effects on subsequent developmental domains (Begle et al., 2011, Holmes et al., 2014). However, few studies have examined mediating pathways, namely, the role of caregiver–child relationships, explaining the relation between children's violence exposure and behavior problems (e.g., Lieberman et al., 2005, Toth et al., 2002). Furthermore, little is known about possible gender differences that may exist in these relations. To address these limitations, the current study examined the relationship between violence exposure severity levels (i.e., mild, moderate, severe) and behavior problems in adolescents, the mediating role of caregiver–child relationships, and gender differences in these relations.

Adolescents’ exposure to violence in the home, defined as adolescents being a witness to or victim of acts which threaten or actually inflict physical harm in the home he or she lives in, is a widespread problem. According to McDonald, Jouriles, Ramisetty-Mikler, Caetano, and Green (2006), approximately 15.5 million (29.4%) children in the United States reside in homes in which intimate partner violence (IPV) occurred at least once in the previous year, and of those children, nearly half (13.3%) had been exposed to severe IPV. Likewise, according to the National Survey of Children's Exposure to Violence (NatSCEV), approximately 20% of children and adolescents in the United States witness family assault such as intimate partner violence, and 9% experience physical abuse over their lifetime (Finkelhor et al., 2009). Although direct victimization (e.g., physical abuse) and witnessing family violence (e.g., exposure to IPV) represent distinct types of violence exposure in the home, research has revealed that these types of violence exposures frequently co-occur (Finkelhor et al., 2009).

Negative outcomes commonly associated with violence exposure include aggressive or antisocial behavior, conduct problems, depression, anxiety, low self-esteem, posttraumatic stress, substance abuse, academic challenges, reduced social development, and poor physical health maintenance and somatization (Tailor et al., 2015, Wolfe et al., 2003). In particular, many previous studies have found a significant link between exposure to violence in the home and behavior problems in adolescents (Evans et al., 2008, Moylan et al., 2010). In these studies, exposure to violence was significantly associated with higher levels of general behavior problems, which include externalizing behavior problems such as aggression and delinquency and internalizing behavior problems such as anxiety and depression. Studies that examined gender differences in the effects of violence exposure on behavior problems have yielded mixed findings, with some studies suggesting no significant gender difference (Fagan and Wright, 2011, Moylan et al., 2010, Sternberg et al., 2006) and others suggesting stronger effects of violence exposure on girls’ behavior problems (i.e., exposure to violence at home better predicted girls’ rather than boys’ behavior problems; Cummings et al., 1999, O’Keefe, 1994, Spano et al., 2009).

Although numerous studies report a significant link between violence exposure and children's behavior problems, many of these studies were limited by relying on one source of information, usually the child's mother, to assess the child's behavior problems. The elevated levels of behavior problems in violence exposed children should be understood with caution as these may reflect caregivers’ biased perceptions in reporting their children's behavior problems. Many researchers argue that relying on one source of information about children's behavior problems may be problematic (Kerr, Lunkenheimer, & Olson, 2007). Therefore, this study addresses this issue by utilizing multiple informants: caregiver-report and child self-report. Some researchers further discuss the issue of the validity of perpetrating caregivers’ reports of their children's behavior problems. For example, Sternberg et al. (1993) indicate that the use of a single-informant is of a particular concern in studies focusing on the effects of violence exposure and child maltreatment because when caregivers are the perpetrators of violence, their ability to accurately measure and report their children's behavior problems may be compromised. In fact, previous studies have found that abusive caregivers are likely to over-report externalizing behavior problems in their children compared to non-abusive caregivers (Bradley and Peters, 1991, Lau et al., 2006).

In the literature examining the effects of violence exposure on behavior problems, some researchers have examined the role of caregiver–child relationships. Research on caregiver–child relationships in violent families indicates that caregiver–child relationships may be affected by the perpetrators’ identity. Physically abusive mothers were more likely to minimize their contribution in negative parent–child interactions as well as their children's contribution in positive ones (Bradley & Peters, 1991). Similarly, physically abused children tended to have negative perceptions of their abusive parents (Sternberg et al., 1994). However, positive caregiver–child relationships have been found to weaken the effects of violence exposure on behavior problems in low-income adolescents (Hardaway, McLoyd, & Wood, 2012). Similarly, findings have suggested that a supportive caregiver can have ameliorating effects on violence-exposed children's behavior problems (Margolin and Gordis, 2000, Sousa et al., 2011). For example, researchers found that IPV-exposed children with mothers rated high in maternal warmth exhibited lower levels of externalizing problem behaviors, compared with children with mothers reporting lower levels of warmth (Skopp, McDonald, Jouriles, & Rosenfield, 2007). In a study employing structural equation modeling, a significant pathway emerged between children's IPV exposure, parental stress, and child internalizing problems in elementary-aged children (Renner & Boel-Studt, 2013). However, this pathway did not remain significant for adolescent-aged children (Renner & Boel-Studt, 2013), suggesting that the protective effects of parenting (i.e., caregiver–child relationships) may diminish with the increasing autonomy of adolescence. Parental factors may be particularly salient during early adolescence, however, when children are just beginning to seek independence but still need a supportive caregiver relationship (Galambos, Barker, & Almeida, 2003).

Although empirical studies suggest that caregiver–child relationships may play an important role in linking children's traumatic life events and psychopathological outcomes (e.g., Kliewer et al., 2004, Toth et al., 2002), few studies to date have investigated the mediating effects of the caregiver–child relationship on behavior problems in violence-exposed children. Examining the effect of mother–child relationship quality on the relation between preschooler's IPV exposure and behavior problems, Lieberman et al. (2005) found evidence of a mediating effect. However, the mediating effects of maternal PTSD were stronger, suggesting the importance of maternal psychological response for younger children. For school-age children, Toth et al. (2002) found that the mother-child relationship mediated the effect of child maltreatment on child-reported internalizing and externalizing behavior symptoms. In this study, maltreated children had less-positive perceptions of their mothers, which in turn led to greater behavior problems. Results from a national random household survey of adolescents in South America indicated a mediating effect of family cohesion, including parental support, on violence exposure and behavioral problems; notably, this relationship was stronger for girls (Kliewer, Murrelle, Mejia, Torres, & Angold, 2001). These findings suggest that positive caregiver–child relationships may have ameliorative mediating effects on the behavioral outcomes of violence-exposed children and adolescents, and gender differences may exist. However, no strong conclusions can be drawn without additional research.

This study is guided by developmental psychopathology (Sroufe & Rutter, 1984) and attachment theory (Bowlby, 1969). The developmental psychopathology perspective posits that adverse life experiences such as exposure to violence may affect children's adaptation and functioning (Sroufe & Rutter, 1984). This perspective emphasizes that a child's ultimate adaptation or maladaptation occurs as a result of interplay between the evolving individual and changing contextual factors (e.g., family environment), underlining the importance of mediating variables that may contribute to outcomes. Attachment theory suggests that parents in IPV relationships or abusive caregivers may fail to provide warm, reliable, and responsive care to their children, resulting in patterns of insecure attachment in the child, which may lead to poor relationships with caregivers (Ainsworth, Blehar, Waters, & Wall, 1978). Children with poor caregiver–child relationships are likely to develop negative representational models (Bowlby, 1982) of self and others, which in turn, may lead to maladaptive emotional and behavioral functioning.

Using the developmental psychopathology and attachment framework, the current study aims to enhance knowledge of behavioral maladjustment among adolescents exposed to violence at home by examining (1) if exposure to mild, moderate, or severe violence is related to behavior problems in adolescents; (2) if the caregiver–child relationship mediates the relations between the three levels of violence exposure and behavior problems; and (3) if those relations differ by gender. Because previous research has identified a bias in which reliance on single report may lead to the overestimation of relationships because of problems of common method variance (Galambos et al., 2003), this study tests caregiver and adolescent reported behavioral problems. Based on the results of previous empirical research, it was hypothesized that exposure to mild, moderate, and severe violence would be associated with higher levels of behavior problems in adolescents. Additionally, it was hypothesized that higher violence exposure would be associated with poorer caregiver–child relationships and poorer caregiver–child relationships would be associated with higher adolescent behavioral problems. Finally, the models investigated gender differences across these relations; however, because of the relatively inconsistent examination of this moderator, specific hypotheses were not formulated.