Introduction

In the field of peer victimization, there are two relatively distinct perspectives on how social standing should relate to victimization. According to one prevailing perspective, victims are viewed as largely unpopular, marginalized youth who have little social power and are actively rejected by their peers (Hodges and Perry 1999; Storch and Ledley 2005). This view emphasizes the maintenance of social status norms within the peer group (Berger and Dijkstra 2013; Coleman 1990). Victims are viewed as relatively powerless and submissive in the face of aggression, except perhaps when victims lash out and engage in retaliatory or reactive aggression against their attackers (Camodeca et al. 2002; Fox and Boulton 2006). Though this perspective on victimization has been well-supported, another perspective offers the opposite view of victimized youth; here, it is thought that victims are targeted because of their social power, with popular and well-liked youth serving as targets of victimization (Adler and Adler 1995; Closson 2009; Jamal et al. 2015; Merten 1997). This perspective emphasizes the idea that victimization is targeted toward youth who are liked, influential, and socially dominant, and that it occurs within or across friendship groups (Dyches and Mayeux 2012; Mishna et al. 2008; Peets and Hodges 2014).

These two perspectives represent opposing points of view and, thus far, have been insufficiently integrated in the literature on victimization. Although support has emerged for both, there is little understanding of the robustness of each of these models and few prior studies have been designed to simultaneously evaluate the relative merits of both perspectives (for an exception, see Faris and Felmlee 2014). Moreover, there are inconsistencies across prior studies in the operationalization of victimization and the assessment of gender differences (e.g., Merten 1997; Zimmer-Gembeck et al. 2013). Specifically, many prior studies have operationalized victimization as a general behavior (i.e., not differentiated by form; Hodges and Perry 1999) or they have operationalized one form of victimization but not another (i.e., relational or physical victimization; Zimmer-Gembeck et al. 2013). Further, prior studies have neglected to test for gender differences or have assessed one gender but not the other (e.g., Merten 1997). Consequently, the field is limited by an incomplete and insufficient conceptual understanding of where the risk for victimization lies. In other words, is risk for victimization largely a concern for socially marginalized youth, for socially powerful youth, or are both groups of youth at risk? Further, are the risk patterns similar or different for boys and girls? Answering these questions is central to advancing the knowledge base and to supporting translational efforts to mitigate peer-directed aggression. In the present study, we tackle this problem by examining both linear and curvilinear associations between middle schoolers’ social standing and relational and physical forms of victimization and by testing for gender moderation. We used a social networks approach to conceptualize and measure social standing (measured as social network prestige), which considers an individual’s importance and prominence within the peer network.

A Social Dominance Approach

We considered why social standing or social position might relate to victimization in early adolescence by drawing on a social dominance approach. A social dominance approach to the study of aggression highlights the idea that youth use aggression to obtain social goals by manipulating their own or another’s position within the social status hierarchy (Hawley 2003; Pattiselanno et al. 2015). That is, it is thought that youth within a social group compete with one another to gain power, prestige, and access to resources (e.g., preferred seats, social options on the weekend), and aggression may be a strategy used to gain these social rewards (Hawley 2003; Pattiselanno et al. 2015; Pellegrini 2001).

Such dominance striving is thought to be particularly important during the early adolescent, middle school years. The transition from elementary to middle school represents a time of substantial social change (Monahan and Booth-LaForce 2015; Pellegrini and Bartini 2000; Pellegrini 2002). Young adolescents not only typically move to a new school environment, but also attend multiple classes with multiple peer groups and multiple teachers throughout the day. As such, young adolescents form new relationships with unfamiliar peers and renegotiate their social standing in relation to that of their peers (LaFontana and Cillessen 2010; Merten 1997). The social pressures that youth face during this developmental period may help explain the heightened aggression that is seen at this time (Merten 2004; Moffitt 1993).

Though a social dominance approach has typically been used to explain the link between social standing and aggression, we expanded upon this approach and considered how it could be used to explain why youth’s social standing might put them at risk for victimization. That is, if aggressors are using aggression to manipulate their place within the social dominance hierarchy, the success of this aggression—that is, social success—may be dependent upon the social standing of the victim. However, as noted above, there are two distinct perspectives on what the social standing of victims is expected to be (i.e., Adler and Adler 1995; Hodges and Perry 1999). We consider, in more detail below, the relative support for each position.

The High Prestige Victim Perspective

From a social dominance approach, youth who occupy prestigious and powerful positions in the social network might be highly victimized because the aggressor’s underlying goal is to gain power (Hawley 2003; Pellegrini 2001; Savin-Williams 1979). Targeting (and beating) a powerful victim may raise aggressors’ status above that of their victim, affording them a large gain in power. Along these lines, Adler and Adler (1995) described how aggressive behavior was frequent among members of peer groups that were dominant and powerful, with highly influential youth acting as both aggressor and victim. Similarly, Merten (1997) found that dominant and powerful females aggressed internally against other dominant peer group members. Other researchers have reported similar findings, showing that victims are at least moderately popular, socially prominent, and have high social network centrality within their peer group (Dyches and Mayeux 2012; Faris and Felmlee 2014; Peets and Hodges 2014; Rodkin and Berger 2008). Challenges among peers at the upper end of the social hierarchy can be used to establish status and position within the upper echelons of the social group (i.e., Adler and Adler 1995). Further, it is possible that envy toward prestigious adolescents may promote aggression toward them. That is, adolescents high in social network prestige likely possess other desirable attributes, such as athleticism, physical attractiveness, and leadership abilities (Dijkstra et al. 2009; Rodkin et al. 2000). Thus, the high prestige victim perspective would suggest that adolescents who are high in social network prestige may also be highly victimized.

The Low Prestige Victim Perspective

According to ideas of norm enforcement (Coleman 1990; Faris and Felmlee 2014), aggressors may specifically target youth who violate social norms in order to maintain and uphold the social dominance hierarchy. Thus, it may be that aggressors target victims with low power and prestige because aggressors want to enforce social norms and distinguish themselves from their low prestige victims. In fact, Berger and Dijkstra (2013) found that (perceived) popular youth disliked unpopular peers, and hypothesized that this was driven by popular youth’s desire to maintain a distinction between social positions. This dislike and desire to differentiate social positions might encourage those occupying more powerful and influential social positions to victimize those occupying less powerful positions. Alternatively, it may be that aggressors are particularly likely to choose victims low in social network prestige because these individuals may have limited peer support and relatively few friends; thus, choosing a low prestige victim minimizes potential retaliation from supporters of the victim (Hodges et al. 1999), making them easy targets. Empirical work supports these explanations, showing that victimized youth have peer relationship difficulties (i.e., are withdrawn, have low social competence, and are rejected by their peers; Fox and Boulton 2006; Hodges and Perry 1999; Storch and Ledley 2005), and also that peer rejection leads to increased victimization longitudinally (Hodges and Perry 1999). Thus, it may be that youth low in social network prestige are frequent targets of victimization (i.e., the low prestige victim perspective).

Robustness of the Link Between Social Standing and Victimization: Gender and Form of Victimization

There are other factors that may be involved in the relationship between social network prestige and victimization, including form of victimization (i.e., relational vs. physical) and gender. For instance, the high prestige victim perspective suggests that aggressors may be targeting prestigious victims to maximally increase their own prestige (Adler and Adler 1995; Peets and Hodges 2014). However, this association might be stronger for relational than physical victimization. Relational aggression directly involves the manipulation of peer relationships (e.g., Rose et al. 2004). Thus, relational aggression may be a powerful tool for aggressors to gain position within the network, and is likely to be used strategically against targeted victims to achieve one’s goals (Hawley 2003; Merten 1997). Given that physical aggression is less directly related to the manipulation of one’s peer relationships, physical aggression may not be used in such targeted manner. Thus, there may be a stronger positive association between social network prestige and relational than physical victimization.

Associations also may be qualified by the gender of the victim. Some studies suggest that low prestige individuals are being targeted because they have low peer support and fewer friends who might defend them (Hodges et al. 1999). Given gender differences in peer relationships (e.g., Sainio et al. 2011; Salmivalli et al. 1996), we might then expect this to occur more so for boys than girls. Girls more often defend their peers who are victimized (e.g., Sainio et al. 2011). As such, boys with low social network prestige may be particularly at risk of victimization, given that they have few friends who might support or defend them from aggression, and the few friends they have are likely boys (Kovacs et al. 1996); both of which make it less likely that they will be defended. Thus, the low prestige victim perspective may be more likely to occur for boys.

Little is known about the joint impact of gender and form of victimization, in conjunction with social standing. Rodkin and Berger (2008) found that social standing (perceived popularity) was positively associated with victimization for girls, but negatively associated for boys. However, this study only examined victimization perpetrated by boys and did not differentiate between relational and physical victimization. In another study, relational victimization was associated with high power within the peer group (measured as a composite including popularity, influence, and leadership) for girls, but was unrelated for boys (Zimmer-Gembeck et al. 2013). These findings might suggest that, for girls and for relational victimization, there may be positive associations between social network and victimization, whereas for physical victimization, there may be weaker associations. Given the mixed findings for boys, it may be that the relationship is better captured by a curvilinear association.

Conceptualizing Social Standing: Social Network Prestige

We conceptualized social standing as social network prestige, which represents youth’s position within the grade-level social network, and is characterized as one’s importance, prestige, and potential for influence (Reynolds and Crea 2015; Suthers and Desiato 2012). Peers nominate friends within their grade; these nominations are used to construct a social network. The structure of this network is then used to calculate each individual’s position within that network (i.e., social network prestige). Specifically, social network prestige comes from the extensiveness and closeness of the peer relationships that one has and one’s potential reach to peers. The more social ties that one has with youth who are themselves closely tied with peers, the more prestige and potential for impact one has. For instance, an adolescent with high social network prestige would be seen as a friend by many peers who themselves are closely linked to many others.

We utilized this conceptualization of social standing because it is both theoretically and methodologically suitable to address our questions of interest. Social network prestige is an indicator of the social network. Methodologically, social network measures have advantages over reputation-based measures (e.g., perceived popularity, peer acceptance; Caravita et al. 2011; de Bruyn et al. 2010; Meter and Card 2015; Sentse et al. 2015) because they are calculated based on real relationship ties. Reputation-based measures may not capture the intricacies of real relationships (e.g., one may have a reputation as being not well liked, but may still have relationships with peers within the social network, and this needs to be accounted for in order to arrive at accurate estimates of peer effects; McNeilly-Choque et al. 1996; Zhang et al. 2014). Thus, some argue that social network measures reflect greater social complexity than reputational measures (Zhang et al. 2014). Moreover, social network prestige is an ideal choice among social network measures in that it is directly applicable to our theoretical perspective in ways that other network-based measures are not. For example, in contrast to measures of the number of friends one has (i.e., only considering the closest relationships), social network prestige captures one’s position within the entire grade-level social network. Further, because social network prestige is based on incoming friendship ties (i.e., how many peers want to be friends with an individual), it represents one’s potential power and importance within the peer group, which differs from measures based on outgoing friendship ties (how many peers an individual chooses as a friend). From a social dominance approach, we expected victims to be targeted based on their importance or influence within the overall social network; one’s position within the network (i.e., social network prestige) is, therefore, an appropriate indicator of whether or not one has this type of social influence and power.

Despite the benefits of this network-based measure and the conceptual appropriateness of this measure to the study of victimization, to our knowledge, no one has examined social network prestige in association with victimization. As such, empirical evidence that social network prestige should operate differently from other measures in relation to victimization is virtually non-existent. Thus, taking a conservative approach, we expected associations between social network prestige and victimization to be similar to past studies that utilized other measures of social standing (e.g., perceived popularity, social network centrality).

The Current Study

Two perspectives on the link between youth’s social standing and victimization are both empirically supported, yet suggest opposing mechanisms leading to peer victimization. Some have found that socially powerful and prominent youth are highly victimized (e.g., Dyches and Mayeux 2012; Faris and Felmlee 2014), whereas others suggest that socially marginalized and rejected youth are highly victimized (e.g., Hodges and Perry 1999; Storch and Ledley 2005). The main goal of the current study was to integrate these seemingly opposite perspectives (and consider whether both could be true) by examining linear and curvilinear associations between social network prestige and victimization in a sample of early adolescents. To further explore nuance in these associations, we examined the prestige-victimization link for both relational and physical victimization, and tested moderation by gender. Based on the high prestige victim perspective (i.e., aggressors target prestigious peers in order to gain maximal social benefits from defeating these peers) and the effectiveness of relational aggression as a strategic tool (Hawley 2003; Merten 1997), we expected positive associations for relational victimization and weaker associations for physical victimization. Further, we hypothesized positive associations for girls, and negative or curvilinear associations for boys. Boys who are low in prestige may be particularly at risk for victimization because their friends are less likely to defend them against aggressors (i.e., the low prestige victim perspective) (e.g., Sainio et al. 2011), yet mixed findings suggest the potential for a curvilinear association (e.g., Rodkin and Berger 2008; Zimmer-Gembeck et al. 2013). We examined these associations after taking into consideration relevant demographic characteristics (i.e., grade, ethnicity), as well as behavioral factors that may impact both social standing and victimization (i.e., aggression). Aggression is likely associated with both social network prestige and victimization, as suggested by studies demonstrating positive relations between aggression and indices of influence and social position, such as perceived popularity (e.g., Schwartz 2000). It may be that adolescents who are high in social network prestige are more likely to be highly aggressive, and this aggression leads to retaliatory victimization (Hodges et al. 1999). In this way, aggression may act as a confounding variable, and thus was included as a covariate.

Method

Participants

Participants were sixth to eighth grade students from an ethnically diverse, southwestern middle school in the US who had agreed to participate in a three-wave longitudinal study. The study took place across the spring of Year 1, the fall of Year 2, and the spring of Year 2. There were variations in measures and participants across the three waves, and the data for the present study are drawn from the second wave (fall of Year 2), when measures of relevance to the present study were collected. Information letters and consent forms (in English and Spanish) were distributed to the families of all students in the middle school. The study used passive consent, meaning that if parents did not specifically opt their child out of the study, consent was assumed. We used passive consent (with active student assent) due to the nature of the social network data collected in this study. In social network research, missing data on friendship ties renders the data unusable because it creates holes in the network that we aim to examine (Burt 1987). Because of the benefits and importance of using a social network-based measure to assess social standing (i.e., social network prestige), an inaccurate depiction of network structure due to missing friendship ties would be particularly problematic. Unfortunately, such missing data on friendship ties cannot be addressed using typical methods for missing data, such as multiple imputation. Additionally, Ellickson and Hawes (1989) found that 96 % of middle school parents who did not return consent forms did not object to the research, but rather lacked the motivation to return the consent form. Finally, the principal of the participating middle school agreed to act as loca parentis for the study. These recruitment procedures were approved by the participating school and by the university Institutional Review Board. Of the total 1052 students enrolled in the school at the time of data collection, 61 parents requested that their child not participate, 4 students refused participation at the time of survey administration, 17 students had withdrawn from the school by the time of survey administration, and 18 students were absent from school during survey administration. This resulted in a final sample of N = 952 participants (49.9 % girls; ns = 339 sixth graders, M age = 11.12 years, SD = .51; 300 seventh graders, M age = 12.12 years, SD = .52; and 313 eighth graders, M age = 13.15 years, SD = .52).

Participants were from ethnically diverse backgrounds: 44 % self-identified as Latina/o, 20 % as White, 18 % as Black or African American, 9 % as American Indian or Alaska Native, 3 % as Asian, and 6 % self-identified as other. Forty-four percent of participants and their parents were US-born, 12 % were US-born with one parent foreign born, 30 % were US-born with both parents foreign born, and 14 % were foreign born with both parents foreign born. Forty-six percent of students reported speaking English at home, 23 % spoke Spanish, and 27 % spoke both English and Spanish. District-provided data showed that 79 % of participants were eligible for free lunches, and 9 % were eligible for reduced-price lunches, indicating that the majority of participants were from low socioeconomic status families. Finally, 48 % of participants came from two-parent married families, 29 % came from single-parent families, and 16 % came from two-parent, unmarried families; the remaining 7 % were unsure.

Procedure

Measures used in the current study were collected in the fall semester. Participants completed a questionnaire in their classrooms. Researchers read aloud each item, while additional research assistants provided individualized assistance and answered questions. The questionnaire packet was administered on two consecutive days and took approximately 2 h to complete. The items used in the present study were completed on the second day of administration. Students received a small gift for completing the survey.

To assess friendships (from which social network prestige was calculated) and victimization, peer nomination procedures were utilized. Data were collected 3 months into the school year, allowing incoming middle school students time to form relationships with their peers. During survey administration, students were given a list of peers in their grade and were instructed to think of peers in their grade that fit each description (i.e., friend, victim). They recorded the peers’ first name, last initial, and a unique identification number (created for the study). Participants were instructed not to nominate themselves, but were told that they could nominate the same person for more than one description. If they could not think of peers who fit a particular description, they were instructed to leave the space blank.

Measures

Social Network Prestige

Participants were asked to nominate up to ten of their closest friends in their grade; thus, the social network was created at the grade level. In middle school, the social network typically consists of the entire grade (unlike elementary school, wherein the social network is typically the classroom). As such, allowing for nominations at the grade level prevents the social network from being artificially truncated into smaller groups. This provides an accurate assessment of one’s position within the entire grade-level social network.

Within the social network, each individual is termed a node, and a friendship nomination is a directional arrow between two nodes. That is, if individual i nominates j as a friend, there is an arrow (i.e., a friendship tie) from i to j. Using social network analysis, social network prestige (operationalized as proximity prestige; Wasserman and Faust 1994) was calculated for each individual as a function of how many peers nominated the individual as a friend, as well as how close in the network peers are to the individual. For instance, an individual has high social network prestige (i.e., is seen as important and influential; Reynolds and Crea 2015; Wasserman and Faust 1994) if he or she is nominated as a friend by many peers, and can be reached through the network by many peers. To test for a curvilinear association, a quadratic term for social network prestige was created by multiplying social network prestige by itself.

Victimization

Victimization was assessed using a peer nomination measure (modified from Rodkin and Berger 2008). The measure was modified such that adolescents were asked to nominate aggressors and their victims separately for relational and physical aggression. Items were chosen to match those commonly used in peer nominations of aggression and victimization (e.g., Peeters et al. 2010). One item assessed relational aggression and victimization; “Someone who gossips about others or excludes others” and one item assessed physical aggression and victimization; “Someone who hits, kicks, or pushes others”. For each item, students were asked to nominate up to three peers as aggressors. For each aggressor nominated, students could nominate up to three victims (i.e., “Who does person 1 gossip about or exclude the most?”). The number of times each participant was nominated as a victim was summed to yield total relational victimization and physical victimization scores. Sum scores were used, rather than proportion scores, because nominations were obtained at the grade-level. This resulted in very large divisors and, as such, very small proportion scores. Given that variations in sample sizes across grades were very small (ns by grade ranged from 300 to 339), findings obtained with sum scores of victimization were virtually identical to those obtained using proportion scores. For ease of interpretation, sum scores were retained.

Covariates

Grade, ethnicity, and aggression were considered as covariates. Participants self-reported their grade and ethnicity. In addition, they nominated peers as relationally and physically aggressive (see above). These nominations were summed to yield relational and physical aggression scores.

Results

The main goal was to examine the linear and curvilinear associations between social network prestige and victimization, using a measure of social network position that assesses one’s influence and importance within the grade-level friendship network. These associations were examined based on both relational and physical victimization, and we assessed moderation by gender. To do this, multiple regression analyses were conducted to predict relational and physical victimization from social network prestige and the gender of the victim.

Descriptive Analyses

Victimization was operationalized as the number of times each participant was nominated by peers as relationally and physically victimized. Relational victimization scores ranged from 0 to 23 (M = 2.82, SD = 2.93), and physical victimization scores ranged from 0 to 12 (M = 1.50, SD = 1.95). All variables were normally distributed. Independent samples t-tests (see Table 1, Panel A) indicated that girls were more likely than boys to be rated by their peers as relationally victimized [t (968) = 5.05, p < .001], and boys were more likely than girls to be rated as physically victimized [t (968) = −4.57, p < .001]. No gender differences were found for social network prestige. Zero-order correlations were computed separately for boys and girls to assess the relations among variables (see Table 1, Panel B).

Table 1 Means, standard deviations, and correlations of study variables
Full size table

Linear and Curvilinear Associations Between Social Network Prestige and Victimization

To assess the study goals, two hierarchical multiple regression models were conducted, one predicting relational victimization and the other predicting physical victimization. Predictors were centered prior to calculating interaction terms and running regression analyses. Interaction terms were created by multiplying social network prestige variables by gender. When significant interactions were found, simple slopes for boys and girls were examined.

In each model, dummy-coded ethnicity was included as a covariate in the first step (with Latino as the reference group, given that Latino was the majority ethnic group in this sample). Dummy-coded grade was also entered as a covariate, with sixth grade as the reference group. Aggression was also included as a covariate, entered such that form of aggression corresponded to form of victimization (i.e., relational aggression predicted relational victimization; physical aggression predicted physical victimization), because aggression and victimization are thought to be more closely related within form (Crick et al. 1999). In the second step, gender and the linear social network prestige term were entered. In the third step, the interaction between gender and linear social network prestige was entered. In the fourth step, the quadratic social network prestige term was entered. In the final step, the interactions between gender and the quadratic term was entered.

Relational Victimization

The model indicated a marginally significant interaction between quadratic social network prestige and gender (see Table 2, Panel 1). Examining simple slopes indicated that, for boys, there was a significant and positive quadratic effect of social network prestige (b = 155.32, p < .001) (see Fig. 1a). For girls, there was no evidence of a curvilinear association (b = 59.71, p > .05) (see Fig. 1b). However, there was also a significant linear social network prestige by gender association. Here, simple slopes indicated positive linear effects for both boys (b = 10.03, p < .001) and girls (b = 17.01, p < .001). Thus, this indicates that, controlling for ethnicity, grade, and aggression, girls who are high in social network prestige are frequently nominated as relational victims, and boys who are both high and low in social network prestige are nominated as relational victims (whereas boys at mid-levels of social network prestige are infrequently nominated as relational victims).

Table 2 Hierarchical multiple regression model with gender and social network prestige predicting victimization
Full size table
Fig. 1
figure 1

Curvilinear and linear associations between social network prestige and relational victimization for a boys and b girls. Note Data points have been jittered for readability

Full size image

Physical Victimization

As with relational victimization, there was a significant interaction between quadratic social network prestige and gender (see Table 2, Panel 2). Simple slopes indicated a significant and positive curvilinear effect for boys (b = 59.58, p < .05) (see Fig. 2a) but no curvilinear association for girls (b = −21.23, p > .05) (see Fig. 2b). There was also no significant linear social network prestige by gender interaction. However, there was a linear main effect of social network prestige (b = 9.17, p < .001), indicating that both boys and girls higher in social network prestige were more highly physically victimized. Thus, as with relational victimization, girls high in social network prestige, and boys both high and low in prestige were frequently nominated as physical victims.

Fig. 2
figure 2

Curvilinear and linear associations between social network prestige and physical victimization for a boys and b girls. Note Data points have been jittered for readability

Full size image

Discussion

Early adolescence is a critical time in terms of youth’s involvement in aggressive behavior; for one, because rates of peer victimization are high (Merten 2004; Moffitt 1993). Further, the middle school setting can be a particularly challenging social environment because of the changing social structure that accompanies the transition into middle school (Monahan and Booth-LaForce 2015; Pellegrini and Bartini 2000; Pellegrini 2002). Youth’s social standing within this social structure may place certain individuals at risk for victimization. However, there are two largely disparate perspectives regarding the associations between social standing and victimization; some suggest that rejected and marginalized youth are highly victimized (Hodges and Perry 1999; Storch and Ledley 2005), while others find that victimized youth are dominant and powerful members of the peer group (Adler and Adler 1995; Merten 1997). Thus, we aimed to integrate these competing perspectives and allow for the possibility that both might be true by assessing linear and curvilinear associations between social standing and victimization. Using social network prestige as a measure of social standing, we found that girls high in social network prestige were frequent targets of victimization, and boys both high and low in social network prestige were highly victimized. These findings contribute to research on adolescent victimization by integrating the predominate (and largely separate) perspectives on the link between social standing and victimization, showing simultaneous support for both perspectives (for boys) and support for the high prestige victim perspective over the low prestige victim perspective for girls. Further, by introducing a new measure of social standing to the study of victimization (i.e., social network prestige), we provide a foundation for future work using this methodology. As such, this study advances our understanding of youth’s risk for victimization and offers practical ways to decrease this risk.

The High Prestige Victim Perspective

Robust support was found for the high prestige victim perspective; positive associations were found across forms of victimization (relational and physical) and gender, supporting the perspective that victimized youth are popular and prominent within the social group (e.g., Adler and Adler 1995; Merten 1997). Although past work within the high prestige victim perspective has not often examined both relational and physical victimization (e.g., Dyches and Mayeux 2012; Rodkin and Berger 2008), our findings suggest that the processes through which social network prestige relates to victimization appear to be similar across forms of victimization. From a social dominance approach, prestigious individuals may be targeted because they provide aggressors with the maximal gain of influence and power within the social network (Adler and Adler 1995; Peets and Hodges 2014). That is, targeting high prestige victims may be beneficial to aggressors attempting to achieve goals of increased social network prestige.

The Low Prestige Victim Perspective

Support for the low prestige victim perspective was found for boys. For boys, curvilinear effects were found, such that those at both high and low levels of social network prestige were likely to be victimized. Testing curvilinear associations was a strength of the current study; it allowed us to integrate the two prevailing perspectives on the social standing-victimization link and test both simultaneously. In fact, our results suggest that aggressive behavior may be directed towards different individuals, for different purposes. This is not a new idea—hostile and instrumental aggression (or similarly reactive and proactive aggression) are used for different purposes (e.g., Crick and Dodge 1996). When used for such varied purposes, aggression is likely targeted toward different individuals. Thus, as is reflected in our results, victimization is a multifaceted construct, which may not be adequately explained using linear models.

Though support was found for the low prestige victim perspective for boys, no such support was found for girls; low prestige girls were not the targets of victimization. Considering victimization from the aggressor’s perspective, it might be that targeting a victim low in social network prestige is attractive because that victim is less likely to have supportive friends who can come to the victim’s defense (Hodges et al. 1999). For girls, however, perhaps even girls with relatively low prestige have friends who would defend them against victimization, whereas low prestige boys’ friends are less likely to defend (Hodges et al. 1999). Of course, this is assuming same-gender friendship (e.g., assuming that girls’ friendships are protective because girls’ friends are also girls). Although same-gender friendships are more common, cross-gender friendships do exist, especially during adolescence (Kovacs et al. 1996). Thus, future work is needed to examine the role of friendships (both same- and cross-gender) in the association between social network prestige and victimization.

Alternatively, this gender difference may be explained by considering that aggressors might target certain individuals in order to enforce social norms (Coleman 1990; Faris and Felmlee 2014). For youth, one of the most central and salient social organizers is gender (e.g., Maccoby 1998), and those who violate gender norms are socially sanctioned (i.e., are the targets of aggression; Drury et al. 2012; Young and Sweeting 2004; Zosuls et al., accepted). However, social sanctions against gender non-conforming boys are stronger than they are for girls (e.g., Young and Sweeting 2004). Thus, it may be that low prestige boys are being targeted due to their behavioral characteristics (such as non-conforming gender expression), by aggressors attempting to enforce norms against such behavior. Conversely, girls who violate gender norms may not be punished as harshly. Of course, gender norms are only one type of social norm that may place youth at risk for victimization, and we did not assess norms (gender or otherwise) in the current study. Thus, further research is needed to directly assess low prestige youth’s behavioral characteristics to determine whether this might account for gender differences in victimization.

Integrating Perspectives: The Importance of Clarifying Competing Findings

By simultaneously testing linear and curvilinear associations between social network prestige and victimization, we integrated the two predominant perspectives surrounding peer victimization. Importantly, we found strong, robust support for the high prestige victim perspective. Across boys and girls, as well as for both relational and physical victimization, those who were considered important and prominent within the social network were also victimized. We found modest support for the low prestige victim perspective, in that this was only found for boys. Thus, by simultaneously testing the applicability of both perspectives, we can conclude with some certainty that both perspectives have merit, yet more robust results were found for the high prestige victim perspective. As such, this study moves the field of peer victimization forward, toward a consensus regarding who gets victimized within the social group. This is impactful for teachers and interventionists, in that it pinpoints the upper levels of the social network as problematic in terms of victimization, as well as highlighting the unique risk of low social standing for boys.

Strengths, Limitations, and Future Directions

Based on theoretical and empirical work that suggests victims are targeted due to their social position (Adler and Adler 1995; Peets and Hodges 2014), we hypothesized that prestige would predict victimization. However, due to the concurrent nature of the data, the directionality of effects could not be determined. To our knowledge, very few empirical studies have examined associations between social position and victimization longitudinally. In one study, Faris and Felmlee (2014) found that social standing (how connected one was to others in the peer group) positively predicted increases in victimization a year later, although not for those very high in social standing. However, the reverse may be true (i.e., victimization predicts social position). For instance, researchers have theorized that cross-gender victimization signifies that the victim is attractive to the opposite gender, thus enhancing his or her overall status within the peer group (Pellegrini 2001). Though our goal was not to differentiate cross-gender and same-gender victimization, a focus on the gendered patterns of victimization might lead to the consideration of a causal link in the other direction. Additional longitudinal work is required to clarify these patterns and more accurately test directionality.

Our results were generally consistent across forms of victimization; this is similar to past work reporting no differences in the link between social standing and physical versus non-physical victimization (i.e., verbal harassment or relational victimization; Faris and Felmlee 2014). This may suggest that the form of victimization does not matter. However, it may be that a more complex examination of victimization and social network prestige—in particular, one that examines both aggressor and victim in tandem—is required to identify differences between forms of victimization. Prior research has shown that relational aggressors are high in perceived popularity and social network centrality (e.g., Rose et al. 2004). Taken together with the findings of this study, it might be that relational aggression is relatively contained among the upper levels of the social dominance hierarchy, with youth high in social network prestige attacking other youth who are high in social network prestige (Adler and Adler 1995; Merten 1997). Yet, this pattern is not so apparent for physical aggression; not all physical aggressors occupy high status positions (Rose et al. 2004), and physical aggression is associated with being less socially skilled and less regulated than relational aggression (e.g., Schwartz 2000). As such, even when targeting victims who are high in social network prestige, physical aggressors may be relatively dysregulated. A more relational perspective on aggressive behavior is needed; perhaps directly linking the social network prestige of aggressor-victim pairs (and examining both relational and physical aggression) would be useful to understand whether and how aggression is directed from prestigious aggressors to prestigious victims. The results of the current study are an important starting point to further this line of research.

We operationalized social standing as social network prestige, given the methodological benefits that this measure has, as well as the applicability of this measure (conceptually) within our theoretical framework. However, given the lack of extant research considering links between social network prestige and victimization (and our available measures), we cannot explicitly compare social network prestige with other measures of social standing. We know that social network prestige is positively correlated with social preference (i.e., being nominated as liked most), social network centrality (i.e., how well-connected one is within the social network), and the number of friends one has (Reynolds and Crea 2015; Zhang et al. 2014). Yet, given the methodological and conceptual distinctions (see Wasserman and Faust 1994; Zhang et al. 2014), it may be that there are nuances across these measures. By demonstrating the role of social network prestige in victimization, using a measure that is both methodologically sophisticated and theoretically driven, this study can serve as a basis for future work that can build upon the theoretical model described here. For instance, future work can be done to explicitly compare victimization across various indices of social standing to further develop our understanding of the complex social nature of victimization and the precise role of the social hierarchy to one’s risk of victimization.

One strength of this study was that our measure of social network prestige allowed us to examine youth’s position within the grade-level social network, which is the relevant social context for middle school youth. However, there may be additional contextual factors that impact the link between prestige and victimization, for instance, broader level contextual factors such as grade-level norms of victimization or centralization of victimization (i.e., more victimization was focused towards specific victims). Although these contextual factors have been examined at the classroom level for elementary school youth (see Sentse et al. 2007; Serdiouk et al. 2015), grade-level factors are likely important for middle school youth. These broad level factors are difficult to examine, due to the prohibitive costs associated with obtaining data from enough schools to measure grade-level contextual effects; yet, it remains that grade-level contextual factors are potentially important. Next steps in this line of work might entail obtaining the larger samples necessary to examine such effects.

Finally, in assessing victimization, participants were limited to nominating three aggressors for each form of aggression, and three victims for each aggressor nominated (i.e., were able to nominate up to nine victims for each form of aggression). Although limiting nominations of aggressors and victims is common (Giang and Graham 2008; Graham et al. 2006; Pellegrini and Bartini 2000; Rodkin and Berger 2008), it remains that only the most well known aggressors and victims are likely to be nominated. That is, it is possible that the peer reports of victimization used failed to capture victimization that is less well known or less frequent (but still potentially harmful for the victim). Future research might consider alternative methods of assessing victimization, or allowing for unlimited nominations of aggressors and victims to ensure that all victimized youth are considered.

Conclusion

This study extended knowledge in the area of victimization by integrating two prevailing perspectives on the link between youth’s social standing and their risk for victimization by testing both linear and curvilinear associations. We also assessed nuances in this link by considering both relational and physical victimization and assessing moderation by gender. We found robust support for the perspective that popular, high prestige youth are highly victimized (e.g., Dyches and Mayeux 2012; Merten 1997), and limited (i.e., only for boys) support that marginalized, low prestige youth are also victimized (e.g., Hodges and Perry 1999). From a theoretical standpoint, our results support the notion that aggressors might be strategically targeting high prestige victims who provide aggressors with a maximal amount of social benefits (e.g., Adler and Adler 1995; Peets and Hodges 2014). Further, it may be that low prestige boys are targeted for victimization because they are less likely to be defended by the few friends they have, and/or because boys receive greater social sanctions for violating social norms. Our study also adds to the field of adolescent development broadly by utilizing a new methodological approach—that is, measuring social standing as social network prestige. By introducing social network prestige to the study of peer victimization, we lay a foundation for future work that can utilize this complex and nuanced measure to understand youth’s aggressive behavior.

The benefits of this study also include practical implications. Our results point to the potential importance of considering structural features of the social environment to decrease victimization. From a social dominance perspective and based on our results, it seems that aggressors are targeting prestigious peers. Thus, decreasing the social network hierarchy and striving for a more egalitarian social structure may reduce the risk of victimization. In fact, Garandeau et al. (2014) found that aggression was more prevalent in classrooms with more status hierarchy (based on the variability of perceived popularity within the classroom). Similarly, Ahn and Rodkin (2014) found that, in classrooms with more egalitarian social networks (based on the dispersion of friendship nominations), aggressors decreased in perceived popularity over time. Therefore, perhaps in egalitarian social groups, there are fewer prestigious victims to target for victimization, thus aggressors are less successful in gaining the social network position they seek. Further, low prestige boys were at high risk for victimization. If this can be explained by considering boys’ friendships and the potential for friends to defend against victimization, efforts could be made by teachers or interventionists to support and encourage boys’ friends (particularly low prestige boys’ friends) in their defending behavior. Overall, our results highlight the targeted nature of victimization, particularly based on social network prestige, and point to ways in which we can reduce youth’s risk for involvement in such victimization.