Being the target of physical, verbal, or relational aggression is unfortunately a relatively common experience for children and adolescents. Prevalence rates indicate that between 40 and 80% of school-aged youth have been exposed to victimization by peers, and approximately 10 to 15% of these youth experience chronic victimization (Juvonen & Graham, 2001). A growing body of research documents the detrimental effects of peer victimization on children's psychosocial development. Peer victimization has been related to internalizing behaviors such as anxiety, depression, and low self-worth (e.g., Hawker & Boulton, 2000), externalizing behaviors such as aggression and delinquency (e.g., Khatri, Kupersmidt, & Patterson, 2000; Schwartz, Proctor, & Chien, 2001), and peer rejection (e.g., Hodges & Perry, 1999). The majority of studies in this area focus on physical (e.g., being hit) or more general verbal (e.g., being teased) forms of victimization. An important extension to this literature is the study of relational victimization (e.g., Crick & Bigbee, 1998; Crick & Grotpeter, 1996). Relational victimization represents a distinct construct in that it encompasses behaviors that are specifically designed to inflict harm by damaging or manipulating the victim's relationship with other peers (Crick et al., 2001; Crick, Casas, & Nelson, 2002). Examples include lying and spreading rumors to make others dislike a peer, excluding a peer from group activities, and threatening to withdraw friendship unless a peer acquiesces to specific demands (Crick et al., 2002).

The study of peer victimization and its association with psychosocial adjustment is particularly relevant during early adolescence as youth transition from adult- to peer-centered relationships (Nansel, Haynie, & Simons-Morton, 2003). As autonomy from parents increases, adolescents spend more time with peers and rely on peer support to a greater extent (Prinstein, Boergers, & Vernberg, 2001). Hierarchies develop within peer groups and concerns about social status and popularity become increasingly salient (Espelage, 2002). The formation of cliques that represent small groups of youth with similar backgrounds, interests, or reputations is common and adolescents may also become more possessive of their friendships, demanding exclusivity (Brown & Lohr, 1987). Within dyadic and peer group contexts, supportive peer relationships fill needs for belonging and acceptance by peers and provide opportunities to develop adaptive coping skills and improved social competence (Yoon, Barton, & Taiariol, 2004). Such relationships may also help youth better adapt to contextual changes in middle school that may create feelings of anonymity including a more bureaucratic administration, larger student body, and moving from class to class (Nansel et al., 2003).

Although physical victimization clearly places youth at risk for injury, physical and relational victimization are also harmful because they limit opportunities to develop supportive peer relationships (e.g., Crick et al., 2001; Coie, Dodge, & Kupersmidt, 1990). Both types of victimization are used to establish and maintain boundaries and status within cliques, and victims are frequently rejected from or marginalized within these peer groups (e.g., Rodkin & Hodges, 2003; Underwood, 2003). In particular, developmental changes in the quality and meaning of friendships underscore the hurtful and harmful nature of relational victimization. The quality of relationships changes from being unilateral or reciprocal in middle childhood to being more mutual and interdependent (Selman, 1981). Relationships are viewed as enduring and committed and involve more intimate disclosure of personal issues (Maccoby, 1988). Themes of conversation include avoidance of rejection, inclusion by peers, and concerns about self-presentation, with gossip and evaluation of others becoming more normative (Parker & Gottman, 1989). Relational victimization may be particularly threatening because this behavior betrays the trust placed in peer relationships and the high level of sharing and openness make youth vulnerable to having personal information used against them in this context (Prinstein et al., 2001). Overall, the combined developmental changes occurring in early adolescence (shifts in understanding of friendships and peer relations, underlying themes of conversations with peers, and the development of cliques) highlight the significance of investigating peer victimization.

The study of relational victimization provides important insight into a unique category of victimization experiences for girls and boys. In contrast with physical victimization that more frequently occurs among boys, several studies documented similar prevalence rates for relational victimization among adolescent boys and girls (e.g., Prinstein et al., 2001; Storch, Brassard, & Masia-Warner, 2003). This may be explained in part by peer networks moving from predominantly same-gender peer groups during middle childhood to mixed-gender peer groups and romantic relationships in adolescence, thus providing new contexts for relational victimization across gender (Crick et al., 2001). Although boys and girls may experience comparable rates of relational victimization, key differences exist in the way these events are perceived (e.g., Galen & Underwood, 1997; Paquette & Underwood, 1999). In a study including elementary, middle, and high school youth, girls rated social victimization (a construct that encompassed relational victimization) as significantly more hurtful than did boys. In addition, girls perceived social and physical victimization to be equally hurtful, while boys rated physical victimization as more hurtful than social victimization (Galen & Underwood, 1997). A subsequent study of early adolescents in seventh and eighth grade found that boys experienced similar levels of negative affect following social and physical victimization, but girls reported significantly more negative affect in response to social victimization compared to boys. Results also indicated that girls tended to ruminate more about experiences of social victimization than did boys (Paquette & Underwood, 1999). Overall, these findings suggest that social forms of victimization such as relational victimization are particularly distressing for girls, potentially reflecting the strong emphasis placed by girls on social bonds with peers (Crick & Zahn-Waxler, 2003).

Although a number of studies have examined the relation between peer victimization and internalizing behaviors (see Hawker & Boulton, 2000, for a recent meta-analysis), fewer have focused on peer victimization as a risk factor for externalizing behaviors such as aggression, delinquency, and drug use (e.g., Khatri et al., 2000; Schwartz et al., 2001; Weiner et al., 2003). The identification of potential risk factors for externalizing behaviors is particularly important given the increase in these behaviors during early adolescence (e.g., Farrell & Sullivan, 2004; Moffitt, 1997). Although these behaviors typically decline by adulthood, they exact significant negative short- and long-term consequences. They not only place adolescents at risk for adverse health consequences (e.g., physical injury) but may also interfere with normative developmental processes. As an example, adolescents who rely on drug use or aggressive behavior to cope with stressors may limit opportunities to learn more adaptive problem-solving strategies. Involvement in delinquent behavior may result in legal sanctions that displace youth from typical school environments and curtail involvement with nondelinquent peers. Finally, onset of drug use in early adolescence is related to continued use and abuse of substances at later stages of development (Everett et al., 1999; Grant & Dawson, 1997).

Several studies have examined peer victimization as a risk factor for increased aggressive and delinquent behavior among elementary school children (e.g., Hodges, Boivin, Vitaro, & Bukowski, 1999; Khatri et al., 2000). For example, Khatri et al. (2000) investigated whether verbal victimization (i.e., being teased or called names a lot) predicted increased delinquency and aggression among a sample of 471 fourth, fifth, and sixth graders. Verbal victimization resulted in subsequent increases in delinquency for girls but not for boys 1 year later, but it was not related to changes in aggression for either gender. In a sample of 393 French Canadian fourth and fifth graders, peer victimization based on a combined measure of verbal and physical victimization predicted increased externalizing behavior (including aggression and delinquency) over a 1-year period (Hodges et al., 1999).

Building on research conducted with elementary school students, Paul and Cillessen (2003) investigated longitudinal associations between a composite peer nomination measure of relational, physical, and verbal victimization and levels of disruptive behavior (including aggression and delinquency) among a sample of predominantly Caucasian sixth grade middle school students. Students nominated as victims reported higher levels of disruptive behavior than did nonvictims 1 year later. Among victims, girls reported higher levels of disruptive behavior than did boys. The authors speculated that the finding for girls may reflect the inclusion of relational victimization items, but the composite measure precluded establishing any unique contributions. Peer victimization (a composite measure of physical, relational, and verbal victimization) was also positively associated with teacher reported externalizing behavior among 785 African American and Latino youth in middle school (Graham, Bellmore, & Juvonen, 2003). In contrast, Prinstein and colleagues (2001) examined associations between relational and physical victimization and externalizing behavior (including both aggression and delinquency) among 566 high school students. Results indicated that externalizing behavior was positively associated with physical victimization for boys and girls, but not with relational victimization. One complication of much of the existing literature is the use of composite measures for externalizing behavior and victimization, making it difficult to isolate the contributions of the various forms of these behaviors. Previous studies have highlighted the importance of differentiating between relational and physical victimization (e.g., Crick & Bigbee, 1998; Crick & Grotpeter, 1996) and between aggression and delinquency (e.g., Farrell, Kung, White, & Valois, 2000).

Few studies have examined potential links between physical and relational victimization by peers and drug use. One exception is a study by Weiner et al. (2003) that explored associations between relational victimization, anger control, and drug use among a sample of 3,922 predominantly Caucasian 8th, 10th, and 12th graders. Results indicated that higher levels of relational victimization were associated with lower levels of anger control, which in turn, was related to higher levels of gateway drug use. However, these researchers did not measure the direct association between relational victimization and gateway drug use. Support for the relation between physical forms of peer victimization and drug use is provided by a related line of research that documents a positive association between violent victimization and drug use among adolescents (e.g., Berenson, Wiemann, & McCombs, 2001; Kilpatrick et al., 2000; Martin, Clark, Lynch, Kupper, & Cilenti, 1999; Simantov, Schoen, & Klein, 2000). For example, adolescent girls who reported experiencing violent acts including robbery, physical attack, rape, or threats against their life were 2–4 times more likely to use cigarettes, alcohol, and marijuana than those who had neither witnessed nor experienced these types of violence (Berenson et al., 2001). Moreover, among a sample of 4,023 adolescents ages 12–17, victimization experiences including physical and sexual assault were significant risk factors for alcohol, marijuana, and hard drug abuse or dependence (Kilpatrick et al., 2000). Additional research is needed to determine whether these relations also generalize to victimization by peers.

Stress and coping theories provide insight into why physical and relational victimization may be related to higher levels of externalizing behaviors. Peer victimization is a salient stressor for youth (e.g., Boulton & Underwood, 1992) that elicits coping responses that may be adaptive or maladaptive (e.g., behavioral, cognitive) in attempts to relieve or manage stress associated with these experiences (Lazarus & Folkman, 1984). The hurtful nature of physical and relational victimization has been documented in several studies (e.g., Crick, Bigbee, & Howes, 1996; Galen & Underwood, 1997; Paquette & Underwood, 1999). Maccoby (1988) highlighted several developmental changes in the nature of stressful events that suggest peer victimization may be particularly distressing for early adolescents. When peer relationships are damaged, adolescents compared to younger children are more upset by the loss of emotional support. In addition, self-concept during this time period is increasingly driven by social comparison with peers. Youth are invested in creating and projecting a certain image and stressors that humiliate or embarrass them in front of peers may be especially painful. Finally, youth are generally more distressed when they experience difficulty in areas that impact their self-concept where they wish to excel and have invested considerable energy (Maccoby, 1988). For early adolescents, substantial effort and importance is placed in developing and maintaining supportive peer relationships. Concomitant with the above changes are shifts in coping behavior. As youth move from childhood into adolescence their coping repertoires increase. Specifically, youth acquire and use more emotion-focused coping strategies (Compas, Worsham, & Ey, 1992), some of which are less adaptive than others (e.g., avoidant strategies such as drug use).

The stress-coping model for drug use developed by Wills and Filer (1996) provides a framework for hypothesizing a link between peer victimization and aggression and delinquent behaviors. At a general level, stressful events are characterized as risk factors for drug use and other problem behaviors (Wills, Vaccaro, & Benson, 1995). This premise is supported by empirical findings linking peer victimization to higher levels of drug use (e.g., Weiner et al., 2003) and to aggression and delinquent behaviors (e.g., Paul & Cillessen, 2003). More specifically, drug use is viewed as an avoidant coping strategy whereby drugs are used to reduce the negative emotional arousal generated by a stressful experience (Wills & Filer, 1996). Several studies have revealed positive associations between peer victimization and increased negative affect such as anxiety or depression (e.g., Hawker & Boulton, 2000), and one reason that adolescents report using drugs is to reduce stress (e.g., Brennan & Moos, 1991; Novacek, Raskin, & Hogan, 1991; Simantov et al., 2000).

Aggression and delinquent behaviors could also be characterized as maladaptive coping strategies to deal with the emotional arousal created by peer victimization experiences. In a recent study, Kochenderfer-Ladd (2004) found that elementary school children responded to peer victimization with negative affect including being afraid, embarrassed, upset, or angry. Being angry was the most common response to a victimization experience, and the most frequent way youth coped with this anger was to get back at the victim by using aggressive behavior. Crick and Bigbee (1998) theorized that some youth who experience physical or relational victimization respond by forming negative opinions of their aggressive peers, leading to difficulty with self-restraint and retaliatory behavior. For example, seventh and eighth graders in middle school perceived that the most frequent goal of an aggressor was to seek revenge (Paquette & Underwood, 1999). Youth who experience peer victimization may thus engage in aggression or delinquent behaviors (e.g., destroying another's property) in retaliation. Although the function of externalizing behaviors may differ with aggression and delinquency being used to retaliate more directly against an aggressor and drug use being used as an avoidant coping strategy to blunt negative affect, each of these responses represents a maladaptive coping response for dealing with peer victimization experiences.

The present study examined the association between physical and relational victimization and externalizing behaviors including drug use, aggression, and delinquent behaviors. This study differed from most prior studies in several important respects. First, unlike many previous studies of relational victimization, this study employed a sample of predominantly African American urban middle school students. This study thus provides information about the generalizability of prior research examining the association between relational victimization and psychosocial adjustment to urban African American adolescents. We examined the association between relational and physical forms of aggression and victimization. Crick and Zahn-Waxler (2003) argued that relational aggression should be included in the categorization of externalizing behaviors to provide a more comprehensive understanding of externalizing behaviors across gender. Unfortunately, most studies examining associations between peer victimization and aggression have not included separate constructs for physical and relational aggression. In addition, although studies have examined the association between relational victimization and externalizing behaviors such as aggression and delinquent behaviors (e.g., Khatri et al., 2000), few have focused specifically on drug use (e.g., Weiner et al., 2003). We also examined the potential moderating effect of gender in the relation between peer victimization and externalizing behaviors. For boys, compared to girls, physical victimization occurs more frequently (e.g., Crick & Bigbee, 1998; Prinstein et al., 2001), is salient in establishing status and hierarchies within peer groups (Rodkin & Hodges, 2003), and is rated as more hurtful than social forms of victimization (Paquette & Underwood, 1999). Thus, we anticipated that physical victimization would be more strongly related to externalizing behaviors for boys than for girls. In contrast, girls rated social forms of victimization as more hurtful and harmful than did boys, and these behaviors undermine the development of social bonds and intimate relationships with peers, which are especially important goals for girls during adolescence (Crick et al., 2001). Therefore, we anticipated that relational victimization would be more strongly related to externalizing behaviors for girls than for boys.

Results

Prevalence of physical and relational victimization and drug use

The prevalence of physical and relational victimization for eighth graders is reported by gender in Table 1. For physical victimization, approximately 40% of the students reported being hit, pushed, or shoved in the past 30 days, 20% reported being kicked or having their hair pulled, and 11% indicated being threatened with physical violence. Boys reported higher frequencies than did girls for all items except being kicked or having their hair pulled. Overall, 49% of students reported being the victim of at least one physically aggressive act in the past 30 days, and 33% experienced multiple acts of physical victimization. For relational victimization, the prevalence of individual items varied from 40% for another student told lies about you to make other kids not like you anymore to 13% for another student said they would not like you unless you complied with his or her demands. Boys and girls reported similar frequencies of relational victimization across all items with the exception that girls reported being excluded from a peer group by a student who was mad at them more often than did boys. Sixty-one percent of adolescents reported experiencing at least one act of relational victimization in the past 30 days, and 38% experienced multiple acts of relational victimization.

Prevalence of overt and relational victimization among eighth grade boys and girls in an urban school system during the past 30 days

For drug use, prevalence rates indicated that 31% of eighth graders reported cigarette use, 48% reported alcohol use, 24% reported advanced alcohol use, and 22% reported marijuana use in the past 30 days. Comparison to national surveys assessing drug use among adolescents revealed that these prevalence rates were somewhat higher than those found for eighth graders (e.g., Johnston, O'Malley, & Bachman, 2001) and more typical of rates found among high school students (e.g., Youth Risk Behavior Surveillance, 2004). However, studies indicate that youth living under conditions of poverty in disadvantaged neighborhoods may be at risk for higher levels of drug use irrespective of ethnic background (Oetting & Beauvais, 2002).

Descriptive statistics and associations among physical and relational victimization and aggression, delinquent behaviors, and drug use variables

The mean, standard deviation, and observed ranges for physical and relational victimization and drug use, aggression, and delinquent behaviors are reported by gender in Table 2. Analysis of variance (ANOVA) indicated that boys reported significantly higher mean frequencies for physical victimization, F (1, 271) = 5.36, p < .05, and delinquent behaviors, F (1, 271) = 8.59, p < .01, compared to girls. Boys also reported a higher frequency of marijuana use, F (1, 271) = 4.78, p < .05, than did girls.

Means, standard deviations, and observed ranges for victimization and externalizing behavior variables by gender for eighth graders

Correlations among peer victimization and externalizing behaviors are presented in Table 3. Physical and relational victimization were moderately correlated supporting the notion that they are related but distinct constructs. Correlations among physical victimization and physical and relational aggression were significant. Similarly, relational victimization was significantly correlated with these forms of aggression. Physical and relational victimization were both significantly correlated with delinquency. With regard to drug use, physical victimization was significantly correlated with alcohol use only. Correlations among relational victimization and drug use variables were significant for cigarette, alcohol, and advanced alcohol use.

Correlations among physical and relational victimization, delinquency, aggression, and drug use variables for eighth graders

Comparisons between independent correlations (Cohen & Cohen, 1988) were used to determine if correlations among peer victimization and externalizing behavior variables were different for boys and girls. Out of 36 correlations, only four significant differences were found at p < .05. For boys compared to girls, physical victimization was more strongly correlated with physical aggression (r = .56 for boys, r = .28 for girls; z = 2.78, p < .01), delinquent behaviors (r = .44 for boys, r = .16 for girls; z = 2.51, p < .05), and alcohol use (r = .40 for boys, r = .02 for girls; z = 3.26, p < .01). Correlations among relational victimization and relational aggression were also stronger for boys than for girls (r = .65 for boys, r = .35 for girls; z = 3.31, p < .01).

Combined effects of physical and relational victimization on delinquency, aggression, and drug use variables

Hierarchical regression analyses were used to examine the effects of physical victimization on externalizing behaviors and the unique contribution of relational victimization in the concurrent prediction of these behaviors after controlling for physical victimization (see Tables 4 and 5). Gender differences in these relations were also examined. Separate analyses were conducted for physical and relational aggression, delinquency, and each category of drug use (cigarettes, alcohol, advanced alcohol, and marijuana). Physical and relational victimization variables were converted to z scores. The scores for physical victimization ranged from 4 to 21, and the scores for relational victimization ranged from 6 to 32. Because many students scored at the bottom of the scale (i.e., 51% for overt victimization, 39% for relational victimization), these scores were represented at the lower end of the plotted scores. The cutoff for the upper end of the plotted scores was set at the 90th percentile. Gender (coded −1 for girls and 1 for boys) was entered at Step 1, physical victimization at Step 2, relational victimization at Step 3, the Gender × Physical Victimization interaction at Step 4, and the Gender × Relational Victimization interaction at the final step. Consistent with the recommendations of Aiken and West (1991), interaction terms were created by multiplying gender by each of the victimization variables. Drug use variables were log transformed to normalize their distribution and multiplied by 10 to facilitate reporting.

Summary of hierarchical regression analyses using physical and relational victimization to predict concurrent levels of delinquency and aggression among eighth grade boys and girls

Summary of hierarchical regression analyses using physical and relational victimization to predict concurrent levels of drug use among eighth grade boys and girls

For delinquency, a significant main effect for gender was found with boys reporting higher levels of this behavior than girls. With gender controlled, physical victimization accounted for an additional 8% of the variance. Relational victimization accounted for 6% of the variance beyond that explained by gender and physical victimization. The Gender × Physical Victimization interaction was significant, and Figure 1 represents the frequency of delinquent behaviors as a function of gender and physical victimization. Boys reported slightly higher frequencies of delinquent behaviors than girls at low levels of physical victimization. For boys, higher levels of physical victimization were associated with an increased frequency of delinquent behaviors. In contrast, there was little change in the frequency of delinquent behaviors for girls as a function of physical victimization levels. No gender differences were found in the association between relational victimization and delinquency.

Regression lines for reported frequency of delinquent behaviors as a function of gender and physical victimization.

No gender differences were found in the frequency of physical aggression. Physical victimization accounted for 16% of the variance in physical aggression after controlling for gender. Relational victimization accounted for a significant portion of the variance with the R² change = 3% when gender and physical victimization were controlled. Physical victimization was more strongly related to physical aggression for boys than for girls, and the pattern of this relation was similar to that found for delinquent behaviors. A significant Gender × Relational Victimization interaction was also found. Boys reported higher levels of physical aggression than girls at low levels of relational victimization. For girls, higher levels of relational victimization were associated with an increased frequency of physical aggression, but boys showed little change in physical aggression as a function of relational victimization levels.

Boys and girls reported similar frequencies of relational aggression. Main effects for physical and relational victimization were both significant. Physical victimization accounted for 14% of the variance in relational aggression after controlling for gender, and relational victimization, entered at Step 3, explained an additional 8% of the variance. The relation between physical victimization and relational aggression was stronger for boys than for girls. Girls reported higher frequencies of relational aggression than did boys at low levels of physical victimization. Higher levels of physical victimization were associated with increased relational aggression for boys, but girls showed little change in relational aggression as a function of physical victimization levels. Finally, a significant Gender × Relational Victimization interaction was found. At low levels of relational victimization, higher frequencies of relational aggression were found for girls compared to boys. Higher levels of relational victimization were associated with increased relational aggression for both boys and girls, but this relation was stronger for boys (see Figure 2).

Regression lines for reported frequency of relational aggression as a function of gender and relational victimization.

In terms of drug use, results revealed only one significant main effect for gender with boys reporting a higher frequency of marijuana use than girls in the past 30 days. Physical victimization, entered at Step 2, was related to cigarette and alcohol use, accounting for an additional 2% of the variance in cigarette use and 4% in alcohol use, but not to advanced alcohol or marijuana use. Relational victimization, entered at Step 3, accounted for a significant increase (2–5%) in the variance across all drug use categories. The Gender × Physical Victimization interaction was significant for both categories of alcohol use, but not for cigarette or marijuana use. Higher levels of physical victimization were associated with increased alcohol and advanced alcohol use for boys. However, for girls the frequency of use decreased as the level of physical victimization increased from low to high. Figure 3 represents the frequency of alcohol use as a function of gender and physical victimization; the pattern of findings for advanced alcohol use was similar. The Gender × Relational Victimization interaction was significant for marijuana use. For girls, higher levels of relational victimization were associated with an increased frequency of marijuana use. However, there was little change in the frequency of marijuana use for boys as a function of relational victimization levels (see Figure 4).

Regression lines for reported frequency of alcohol use as a function of gender and physical victimization.

Regression lines for reported frequency of marijuana use as a function of gender and relational victimization.