The impact of timing of exposure to violence on violent behavior in a high poverty sample of inner city African American youth.

Introduction

Victimization and exposure to violence (ETV) are a common set of experiences facing youth living in economically disadvantaged, urban neighborhoods (Margolin and Gordis, 2000). Growing concern about the effects of exposure to violence on child and adolescent development has resulted in a rapidly developing literature that has linked exposure to violence to increased risk for a variety of problematic outcomes (i.e., PTSD, school failure, depression, and risky sexual behavior) (Margolin and Gordis, 2000). The emergence of violence as a significant public health threat, especially among minority, inner city adolescents, has resulted in the examination of the relationship between ETV and violent behavior (Cooley et al., 1995; Farrell and Bruce, 1997; Miller et al., 1999). The "cycle of violence" has been used as a framework to link ETV and violent behavior (Widom, 1989). Early research by Dodge et al. (1990) and Widom (1989) found that children exposed to violence (in the form of abuse and neglect) are more likely to engage in violent behavior. Although not all abused and neglected children perpetuate this cycle (Cicchetti and Toth, 1995; Duncan and Raudenbush, 2001; Kruttschnitt et al., 1987), a growing body of evidence suggests that today's victims of violence (children who were abused and neglected) are at increased risk of becoming tomorrow's perpetrators of violent behavior (Widom, 2000).

While the cycle of violence has focused on the role of domestic violence (e.g., children as victims of abuse and neglect) on subsequent violent behavior, a related body of research has focused on the impact of ETV in the community on adolescent development. Shahnifar et al. (2000) define ETV in the community as violent victimization (e.g., direct or threatened harm), witnessing violence in the community, and/or violent victimization experiences of family and friends. The literature on ETV in the community has clearly documented that a substantial number of children living in violent neighborhoods are either directly the victims of violence or have been innocent bystanders who have witnessed violent acts committed against others in these communities (Margolin and Gordis, 2000; Osofsky, 1995). For example, in a sample of fifth and sixth grade children from Washington, D.C., Richters and Martinez (1993) found that almost three out of four had witnessed a homicide and two out of three had witnessed a serious assault in the community. On a similar note, Osofsky et al. (1993) found that 26% of a sample of nine to twelve year old children from New Orleans witnessed a shooting and over 90% had witnessed at least one act of violence in the community. In another sample of predominantly African American 14 to 23 year olds, over one third had seen someone shooting a gun and almost 20% had witnessed a fight involving knives (Farrell and Bruce, 1997).

Dodge et al. (1990) and Guerra et al. (2003) argue that youth who are exposed to violence in the community are at increased risk for engaging in violent behavior because they develop cognitive schemas that view the world as a hostile place. Over time, children and adolescents begin to view violence as normative, lose inhibitions surroundings its use, and increase the frequency of their violent behavior (Bandura, 1986). Consistent with the cycle of violence perspective, a growing body of research utilizing both cross sectional and longitudinal data has established ETV in the community as a strong and robust predictor of violent behavior among children and adolescents (e.g., Attar et al., 1994; Scarpa, 2001; Schwab-Stone et al., 1995). For example, DuRant et al. (1994) found that ETV was the strongest predictor of violent behavior for African American youth from 11 to 19 years old living in a community with high levels of violent crime. Similarly, Miller et al.'s (1999) longitudinal study of an urban sample of six to 10 year old boys found a significant relationship between lifetime prevalence of exposure to community violence and antisocial behavior 15 months later.

One unresolved issue is whether ETV has lasting effects on violent behavior. Lynch and Cicchetti's (1998) longitudinal study of 322 ethnically diverse children from disadvantaged backgrounds found that the relationship between ETV and violent behavior became nonsignificant after controlling for prior behavioral functioning (i.e., violent behavior). However, other longitudinal studies have found that ETV has an enduring effect on violent behavior of urban children and young adolescents after controlling for previous antisocial behavior (e.g., Farrell and Bruce, 1997; Gorman-Smith and Tolan, 1998; Miller et al., 1999). On a related note, there is a lack of research on the short term versus long-term effects of ETV on violent behavior. Although no studies have examined the issue of timing of exposure to community violence, two studies utilizing data from the Rochester Youth Development Study (or RYDS) examined the importance of timing of maltreatment on delinquency and other problematic outcomes (Ireland et al., 2002; Thornberry et al., 2001). These RYDS studies found that more recent or proximal maltreatment and more persistent maltreatment had the largest impact on delinquency and drug use during adolescence.

The purpose of this paper is to assess the impact of the timing of ETV on violent behavior in an urban sample of African American adolescents living in extreme poverty. Our objective is to determine whether ETV has a long term versus short-term impact on violent behavior. The following section contrasts three competing theoretical perspectives on the relevance of the timing of ETV to its impact on violent behavior. One theme in the literature from developmental psychology and psychopathology (DPP) stresses the importance of distal ETV given its focus on cumulative consequences, while Agnew's general strain theory (GST) emphasizes the short term behavioral consequences of proximal ETV. The third perspective argues that the relationship between ETV and violent behavior could be spurious due the interrelationship between other factors, ETV, and violent behavior. Next, these competing theoretical perspectives will guide the multivariate analyses conducted on 5 waves of longitudinal data from the Mobile Youth Survey (MYS), a large, community-based study of adolescent development in a high poverty, urban setting. Finally, the theoretical and practical implications of the findings for violence intervention programs will be discussed.

Theoretical framework

A variety of perspectives have examined the impact of developmental processes on violent behavior (Margolin, 2005). However, few studies have studied the developmental consequences of the timing of youth ETV on violent behavior. Three competing theoretical perspectives give differing accounts of the relevance of timing of ETV. Ireland et al. (2002) contrasted theoretical insights from developmental psychology and psychopathology (DPP) versus criminology (Agnew's general strain theory or GST) in their discussion of how maltreatment can affect adolescent development. Both perspectives recognize the importance of timing of ETV. However, both also diverge on the issue of whether ETV will have long term versus short term developmental consequences. One body of research from DPP argues that ETV will have long term, cumulative consequences and that earlier or more distal ETV will have the largest impact on violent behavior. Thus, negative life events like ETV will accumulate, worsen, and intensify over time and lead to problematic outcomes years or even decades later (Caspi, 1987; Caspi and Elder, 1988; Heusmann et al., 1984).

In contrast, Agnew's (1997) GST argues that negative life events like ETV have a short term impact on behavior. GST is concerned with relationships that either threaten to or actually (1) block the attainment of positive stimuli, (2) remove positively valued stimuli, or (3) present negatively valued stimuli (Agnew, 1997:104). These negative relationships, or strain, increase the chance the individual will experience negative affect, including possibly anger. Crime and violence are possible corrective actions the individual may take to reduce the strain or take revenge on those perceived as causing the strain (Agnew, 1992, 1997). Thus, according to GST, higher levels of strain are likely to lead to higher levels of crime.

Related specifically to the current discussion, Agnew (2001, 2002) has discussed criminal victimization, a primary type of exposure to violence, as one of the types of strain that is most likely to lead to delinquency. Agnew (2002:606) notes that "[p]hysical victimization, then, is one of the key types of strain in GST." Further, Agnew (2002) broadens this type of strain by arguing that not only victimization that is directly experienced by the individual, but also vicarious and anticipated victimization, should lead to increased delinquency. He empirically supported his argument when, using data from a national sample of adolescent boys, he found that that all three types of victimization were indeed related to delinquency (Agnew, 2002).

In addition to discussing the types of strain, such as victimization, which are related to crime, Agnew (1992, 2001) has also discussed characteristics of strains that will strengthen their relationship to crime. One of these characteristics is timing--that is, more recent strains are likely to have a stronger relationship to crime (Agnew, 1992, 2001). This is due to the fact that the impact of strains is likely to fade over time (Agnew, 1992), and thus "recent strains should have a …