Introduction

In this article, we present a short-term longitudinal study focusing on social standing and electronic forms of aggression among adolescents. Traditionally, research on social standing during childhood and adolescence has emphasized aggressive behaviors as predictors of rejection by peers (for a review, see Crick et al. 2009). More recent evidence has underscored a shift in the implications of aggression. Aggression becomes associated more closely with positive social outcomes throughout development (Cillessen and Mayeux 2004). By the middle years of adolescence, aggressive youths may even achieve a degree of social prominence (Cillessen and Borch 2006). From a risk perspective, this developmental shift is important to recognize because antisocial behaviors are linked to maladaptive outcomes in a number of domains (Nansel et al. 2001, 2004). Therefore, the potential for aggression to be reinforced by peers could have problematic consequences for some adolescents.

Social Standing and Aggression During Adolescence

As we seek to understand changes in the correlates of aggressive behavior across development, it is essential to adopt a multidimensional perspective on social standing during adolescence. There is a growing body of work distinguishing social acceptance and popularity (Cillessen and Rose 2005). Social acceptance is conceptualized as an index of likeability or positive peer regard, whereas popularity is an indicator of reputational status, prestige, and visibility among peers. Although social acceptance is linked primarily with prosocial behavioral attributes, popularity can present a more mixed behavioral pattern of social skills and aggressive attributes (Sandstrom and Cillessen 2006). Notably, popular youths often use aggression or other aversive strategies to reach and maintain their position in their peer hierarchy (Bowker et al. 2010). Interestingly, by adolescence, the overlap between social acceptance and popularity is modest, and many popular youths are not particularly liked by peers (Parkhurst and Hopmeyer 1998). Considering a developmental psychopathology framework, there are also differences in the implications of social acceptance and popularity for longer-term adjustment. Whereas social acceptance has either positive or neutral implications for development, popularity can be predictive of a number of problematic outcomes including academic disengagement, truancy, early sexual experimentation, and substance abuse (for a review, see Schwartz and Gorman 2011).

To understand the importance of aggression in the emergence and maintenance of popularity, a careful consideration of the multifaceted nature of aggressive behavior is also needed. Investigators repeatedly have validated the distinction between overt and relational forms of aggression (e.g., Crick and Grotpeter 1995). Whereas overt aggression involves direct attempts to injure others through physical acts, intimidation, or verbal insults, relationally aggressive behavior causes harm by damaging social relationships through the use of rumors, exclusion, and social manipulation (Crick et al. 2009). The distinction between these forms of aggression is particularly salient in adolescence. During this stage of development, cognitive maturation can enhance the sophisticated social skills involved in manipulation, while socialization yields a normative decline in the frequency of direct peer harassment (Cillessen and Mayeux 2004). Importantly, relationally aggressive behavior is associated more strongly with popularity than overt aggression in adolescence, particularly among females (Cillessen and Mayeux 2004).

The Emergence of Electronic Modalities of Aggression

As a more refined understanding of the role of aggression in the lives of popular youths is developed, it is important for investigators to recognize the emergence of new, electronic modalities for such behavior. Nearly all of today’s adolescents have access to a wide variety of digital communication tools, including cell phones, chat boards, and social networking websites. One research group found that 80 % of adolescents between the ages of 14 and 17 have a cell phone and 77 % of teens text message, while 95 % of 14- to 17-year-olds go online and 88 % of adolescents use social networking websites (Lenhart et al. 2010a, b). Such digital communication provides opportunities for social interaction, learning, and exploration of ideas. However, it also can serve as a potent outlet for aggressive acts, such as posting unflattering pictures on social networking websites and sending threats in text messages or e-mails (Tokunaga 2010).

Much remains to be learned about the processes underlying electronic aggression. However, one hypothesis emerging from the growing literature on this topic is that electronic aggression may be an iteration of the intentionally harmful behavior seen in overt and relational aggression in schools (Dooley et al. 2009). Although electronic aggression can be conceptualized as a phenomenon that occurs largely outside school, these behaviors also may occur as a manifestation of existing dynamics among peers. As evidence in this regard, the different forms of aggression are moderately to largely inter-correlated and similarly related to many intrapsychic, interpersonal, and adjustment constructs (Pornari and Wood 2010; Raskaukas and Stoltz 2007; Wang et al. 2009; Williams and Guerra 2007).

Even if there is overlap across diverse forms of aggressive behavior, electronic modalities of aggression may still bring unique features. Adolescents can engage in these behaviors wherever and whenever they are connected to the digital world, and their actions have the potential to reach large audiences of peers (Slonje and Smith 2008). Under some conditions, the perpetrators and recipients of electronic aggression also may remain anonymous to each other (Dooley et al. 2009). In contrast, conventionally aggressive acts are often highly salient behaviors, generally witnessed by a small number of peers and involving familiar individuals (Slonje and Smith 2008). Such qualitative differences suggest that a comprehensive portrait of aggression requires the assessment of electronic interactions (Sontag et al. 2011). Multifaceted evaluations of aggression also tend to yield a better understanding of youths’ psychosocial outcomes (Turner et al. 2011).

Social Standing and the Perpetration of Electronic Aggression

The primary objective of our study was to examine the unique role of electronic aggression in the lives of popular adolescents. Given that diverse forms of aggression can have distinct implications for social standing (Rose et al. 2004b), we investigated whether electronically aggressive acts relate to increases in popularity, after accounting for other forms of aggression. Although there may be topographical differences across aggression subtypes, there are likely to be shared commonalities (Dooley et al. 2009; Williams and Guerra 2007). Analyses with a multidimensional perspective on aggression are needed to support inferences about the specific role of behaviors enacted through electronic modalities.

There is preliminary evidence that electronic aggression is associated concurrently with lower levels of popularity and social acceptance among elementary school children (Schoffstall and Cohen 2011). The nature of these associations in adolescence, however, remains unclear. We therefore extended on work with younger samples and analyzed the prospective relationships between electronic aggression and social standing among adolescents. Previous researchers have demonstrated that, although overt and relational aggression are associated with lower levels of social standing during middle childhood, a shift occurs in the early middle school years (LaFontana and Cillessen 2002). By early adolescence, aggressive behaviors in school are more and more rewarded by the peer group with status, prestige, and prominence (Parkhurst and Hopmeyer 1998). Consistent with the literature on aggression in school, we hypothesized that electronically aggressive behaviors would be related to higher levels of subsequent popularity.

Throughout adolescence, aggressive and antisocial behaviors may help not only increase or gain status but also maintain or defend one’s elevated prominence (Crick et al. 2009). Longitudinal analyses indicate that popularity is associated with increasing trajectories of overt and relational aggression in adolescence (Prinstein and Cillessen 2003). Hence, we examined the unique prospective relationships of social standing with later electronic aggression, and we expected prominence among peers to be related to increases in electronically aggressive behaviors over time. To our knowledge, no previous studies have addressed this issue. Still, cross-sectional research has considered electronic aggression in adolescence as a concurrent predictor of social acceptance and related constructs. Indices of peer acceptance and competence are unrelated, for the most part, to electronic expressions of aggression (Calvete et al. 2010; Vandebosch and van Cleemput 2009; Wang et al. 2009).

Social Standing and the Receipt of Electronic Aggression

A secondary aim of our project focused on victimization via electronic means, a phenomenon associated with emotional distress and academic maladjustment among adolescents (Beran and Li 2007; Hay and Meldrum 2010). The targets of aggression often are rejected by their peers and maintain little prominence and influence (e.g., de Bruyn et al. 2010). Nevertheless, the most popular adolescents are at risk for victimization (Prinstein and Cillessen 2003). The aggressive and manipulative strategies used to gain and maintain status often place adolescents in danger of reciprocated rivalry (Cillessen 2009). Consequently, we examined the reciprocal relationships between social standing and electronic victimization. We expected a positive association between popularity and subsequent electronic victimization, adjusting for other forms of school-based victimization. Although no known work has considered popular adolescents, socially skilled and integrated youths have emerged as the targets of electronic aggression in cross-sectional studies (Katzer et al. 2009; Vandebosch and van Cleemput 2009), whereas those who are well liked by their peers are no more likely than others to be the targets of electronic victimization (Gradinger et al. 2012).

Gender Differences in Social Standing and Involvement in Electronic Aggression

As an exploratory goal, we also examined the potential moderating role of gender on the links between social standing and electronic forms of aggression and victimization. Despite inconsistencies in the past literature, available findings indicate that girls may be more involved in electronic aggression than boys, both as aggressors and as victims (Tokunaga 2010). A similar pattern of results has been described for relational aggression. One explanation for this pattern is that girls experience greater social benefits from the use of relational aggression and thus encounter greater retribution (Crick et al. 2009). Likewise, the use of electronic aggression might be rewarded more often with social prominence for girls than boys; in turn, girls with elevated status may be targeted more often in the digital domain than boys.

The Developmental Context of Adolescence

As we investigated the reciprocal associations between social standing and electronic forms of aggression and victimization, we focused on adolescence. Standing among peers increases in salience during this developmental period (LaFontana and Cillessen 2010) and the peer environment increasingly reinforces aggressive behaviors (Cillessen and Mayeux 2004). This developmental stage also brings greater access to electronic communication tools (Lenhart et al. 2010b), while involvement in electronic aggression peaks during the early high school years (Sumter et al. 2012). As a result, understanding the interplay between social standing and involvement in electronically aggressive interactions may be particularly important during the adolescent years.

Assessing Involvement in Electronic Aggression

One issue requiring special attention in our investigation relates to measurement. In prior research, the perpetration and receipt of electronic aggression primarily has been assessed with self-report tools (Tokunaga 2010). Given that a subset of electronically aggressive interactions involves anonymity (Kowalski and Limber 2007; Ybarra and Mitchell 2004), self-report measures are potentially advantageous, allowing for the capture of individual knowledge of acts hidden in the digital world. From a less positive perspective, estimates from such methods have potential limitations insofar as adolescents may respond in a socially desirable manner (Sontag et al. 2011). Perhaps, as a partial reflection of this measurement issue, interrater agreement of school-based aggression and victimization is consistently greater between outside sources (e.g., peers, teachers) than between youths and collateral informants (Branson and Cornell 2009; Ledingham et al. 1982; Pakaslahti and Keltikangas-Järvinen 2001).

Based on the recommendations of prior researchers (e.g., Schoffstall and Cohen 2011), we adopted an alternative approach to measuring electronic forms of aggression and victimization. Specifically, we relied on a peer nomination inventory, a methodology often viewed as optimal for identifying aggressors and victims in schools (e.g., Schwartz 2000). Such an approach offers strong psychometric properties, capitalizing on multiple informants (i.e., all those who complete the measure) rather than relying on a single judgment, and it avoids self-reporting biases (Roseth and Pellegrini 2010). We also suggest that a reputational assessment of this nature can yield a useful snapshot of involvement in electronic aggression. Despite the potential for digital interactions to remain anonymous, those involved in electronic aggression are often known not only to each other but also among their peers (Lenhart et al. 2010a). Reflecting this limited anonymity, one research group found that the majority of adolescents are able to report on and differentiate acts of electronic aggression within their peer environment (Huang and Chou 2010). Thus, peer nominations should provide a valid, reliable assessment of electronic aggression and victimization.

Current Study

Social standing can play a critical role in the perpetration and receipt of aggression during adolescence. In school settings, researchers have found that aggressive behaviors may serve to elicit power and visibility among peers and, in turn, elevated status may be maintained through the use of aggression (e.g., Cillessen and Borch 2006; Vaillancourt and Hymel 2006). Recently, investigators have emphasized the emergence of new, electronic modalities for aggressive acts (Tokunaga 2010). Yet, no known work has considered the interplay between status and involvement in electronic aggression.

The primary aim of the current project was to investigate the reciprocal longitudinal associations between social standing and electronic forms of aggression, taking into account overtly and relationally aggressive behaviors. We anticipated that elevated levels of popularity would be related to increases in electronic aggression over time and that, in turn, higher levels of electronic aggression would be linked to greater popularity over time. No significant associations were predicted between social acceptance and electronically aggressive behaviors. The second objective of our study was to examine the reciprocal links between social standing and electronic victimization, and we expected high levels of popularity but not social acceptance to be associated with being the target of aggression. Lastly, our study analyzed whether gender moderated these reciprocal relationships; we had no a priori hypotheses regarding gender.

Method

Participants

We recruited 415 adolescents (193 males, 222 females; M = 14.68 years, SD = 0.56) from an urban high school in Southern California for a two-wave longitudinal investigation. In the first year of the study, all ninth grade students at the recruited high school were invited to participate. Of the 827 eligible adolescents, 78 % returned positive parental permission and, of those, 73 % assented in writing to the project. Similar consent and assent rates have been obtained in prior investigations within Southern California (e.g., Fuligni et al. 2005). Twenty-eight students were either absent for the data collection or elected to withdraw their assent on the day that they were scheduled to participate, resulting in a sample of 443 adolescents for the first year of the study. In the second year of the project, 6 % of the sample at the first time point did not participate, either because they were no longer enrolled at the school or because they withdrew their assent. A final sample of 415 students took part in both the first (T1) and second (T2) years of the project. There were no significant differences between those who were retained and attrited at T2 for age (F(1, 441) = 0.001, ns), ethnic-racial background (χ2 = 8, df = 5, ns), or neighborhood of residence (χ2 = 29, df = 23, ns). However, males were less likely to have participated at T2 (χ2 = 4, df = 1, p < .05). Retained adolescents additionally did not differ from attrited students on any of the T1 measures, and none of these analyses approached significance (all F(1, 441) < 1.00, ns).

Seventy percent of the final sample was Hispanic American, 6 % was European American, 4 % was Asian American, 3 % was African American, 13 % was of other ethnic-racial descent, and 4 % did not have a classified background. This predominantly Hispanic American sample was consistent with the ethnic-racial composition of the surrounding community (US Census Bureau 2010). The families served by the school have been described as working poor, and 17 % of the youths in the surrounding area lived below the poverty line (California Budget Project 2007). Fifty-two percent of students at the school qualified for free or reduced-cost lunches (RAND California 2008).

The majority of today’s adolescents live with cell phones and online access, with the use of digital technologies cutting across demographic groups. Survey research has failed to detect ethnic-racial differences in cell phone ownership or internet use among adolescents (Lenhart et al. 2010a; Pew Research Center 2011). Although socio-economic status is one area where digital media rates vary, even in families with an annual income below $30,000, more than half of teens have a cell phone and text message, and more than 80 % go online and use social networking websites (Lenhart et al. 2010a, b; Pew Research Center 2011). Thus, relatively broad access to digital media was expected in our sample of adolescents, largely from working poor families of Hispanic origin. Although we were unable to measure electronic technology use due to constrained administration time, the recruited school had designated policies against text messaging and smart phone use in class, suggesting that a large portion of students had access to digital media similar to teens across the nation.

Measures

Overview

The project was approved by the university’s Institutional Review Board and by the involved high school. In the spring of two consecutive years, students completed a group-administered peer-nomination inventory. Consistent with prior sociometric research using adolescent samples (e.g., Cillessen and Mayeux 2004), adolescents were provided with a roster of all participating students, so that each youth was evaluated by 442 peers at Time 1 and by 414 peers at Time 2. Participants were asked to identify those who fit a series of descriptors. A trained research assistant read aloud standardized instructions and questionnaire items. Several additional research assistants were on-hand to provide one-on-one assistance to participants that had questions or were experiencing difficulty with the survey.

Social Standing, Aggression, and Victimization

Each dimension of social standing was assessed with one peer-nomination item. We evaluated popularity by having students nominate grademates that are “popular,” and measured social acceptance by asking participants to identify peers that they “really like.” Participants completed additional items assessing overt aggression (“students that hit or push others”; “students that start fights with others”), relational aggression (“students that gossip about others”; “students that are mean to others by ignoring or excluding them”), and electronic aggression (“students that use the internet or text messages to insult or be mean”). Items also evaluated overt victimization (“students that get hit, pushed, or bullied”; “students who get beat up”), relational victimization (“students who get mean things said about them”; “students that get left out of activities, excluded, or ignored”), and electronic victimization (“students who others insult or are mean to using the internet or text messages”).

For each peer-nomination item, students could nominate up to nine grademates. Although unlimited nominations are widely used in research with adolescent samples (e.g., Cillessen and Mayeux 2004), we opted for a limited choice approach (albeit with a large number of maximum nominations) in order to simplify task demands and reduce administration time and participant fatigue. Of note, only 8 % of adolescents on average made the maximum allowable nominations, and diverse assessment techniques typically yield similar findings across the peer relations literature (Cillessen 2009).

To generate social standing, aggression, and victimization scores, we tallied the number of nominations each adolescent received for the relevant construct. In order to compare the peer nomination variables across the two waves of our project, we divided raw tallies by the total number of informants at each time point (i.e., 442 peers at Time 1 and 414 peers at Time 2). We relied on single-item scales to measure several of our constructs because of limited administration time. Peer-nomination inventories, however, yield highly reliable indices even when one-item assessments are used, as a function of the high number of raters for each item (e.g., Coie et al. 1995). Furthermore, our single-item scales of electronic aggression and victimization were broadly worded in order to capture the vast array of behaviors that can characterize these phenomena (Smith and Slonje 2010). These scales were also validated against multi-item self-report measures in prior work (for a study description, see Duong and Schwartz 2011). Analyses of the convergent and construct validity of the peer nominations mirrored analyses of peer reports for school-based aggression and victimization (e.g., Perry et al. 1988).

Results

Descriptive Analyses

First, we examined the distribution of each of our variables and the bivariate relations among the study constructs. The descriptive statistics for our social standing, aggression, and victimization variables are depicted for boys and girls in Table 1. Gender differences for the social standing dimensions were not significant. However, boys were more overtly aggressive and victimized, and less relationally aggressive than girls across time points. Girls were more electronically aggressive than boys at the second time point.

Table 1 Descriptive statistics for social standing, aggression, and victimization by gender
Full size table

Correlational Analyses

The zero-order correlations between the study variables are summarized in Table 2. We evaluated the bivariate relations using a relatively conservative Type I error rate of .005 to maintain overall error rates. The social standing, aggression, and victimization constructs were moderately to highly stable over our two-wave study. In addition, overt, relational, and electronic forms of involvement in aggression were only partially overlapping, suggesting that assessments of aggression and victimization in varied contexts can provide meaningful insight into adolescents’ lives. At each wave, the three forms of aggression were intercorrelated positively, with medium to large effect sizes. The bivariate relations between school-based and electronic victimization were positive with small to medium effect sizes, whereas the positive links among school-based forms of victimization were large in magnitude. Moreover, popularity and social acceptance were correlated positively with the aggression constructs. At both time points, popularity was associated positively with relational and electronic victimization, and social acceptance only was related positively to electronic victimization.

Table 2 Bivariate correlations for social standing, aggression, and victimization
Full size table

Given moderately high correlations between social acceptance and popularity (T1: r = .61, p < .001; T2: r = .60, p < .001), we conducted follow-up analyses to examine the unique cross-sectional associations between each dimension of social standing and involvement in electronic aggression as either an initiator or a recipient. Table 3 summarizes the analyses involving popularity with social acceptance partialled out and social acceptance with popularity partialled out. As shown, popularity was correlated positively with electronic forms of aggression and victimization at both time points, even after taking into account social acceptance. However, the positive, unique association between popularity and electronic victimization was only significant for girls at the first wave and for boys at the second wave. Social acceptance was correlated uniquely and positively with electronic forms of aggression and victimization only at Time 1.

Table 3 Partial correlations between social standing, electronic aggression, and electronic victimization
Full size table

Cross-Lagged Panel Analyses

To examine the reciprocal associations between social standing and electronic forms of aggression and victimization, we ran cross-lagged panel models using structural equation modeling (SEM). Analyses using structural equation modeling incorporate the assumptions that relationships between the predictor and outcome variables are linear and that residuals are normally distributed with equal variance across the range of the outcome. Analyses using structural equation modeling also necessitate an evaluation of the assumptions of multicollinearity and singularity and a review of outliers and influential points. Our exploratory analyses checking these statistical requirements revealed deviations from the assumptions of linearity, homoscedasticity, and normality of the residuals for our models, and several potentially troublesome observations. We applied log transformations to address these deviations (Tabachnick and Fidell 2007). Although the transformations reduced deviations, they did not alter the pattern of results. Consequently, we present analyses with untransformed scores.

To assess the fit of our cross-lagged panel models, we considered a number of different indices, including the Chi-square (χ2) statistic. The χ2 statistic indexes the closeness of fit between the unrestricted and model covariance matrices. Small, nonsignificant χ2 statistics indicate good fit between the observed data and the hypothesized model. Because χ2 values are inflated by large sample sizes, we also relied on other fit indices that are less sensitive to sample size, including the Bentler Comparative Fit Index (CFI; Bentler 1990), the Jöreskog-Sörbom Goodness of Fit Index (GFI; Jöreskog and Sörbom 1982), the Steiger-Lind Root Mean Square Error of Approximation (RMSEA; Steiger 1990), and the Standardized Root Mean Square Residual (SRMR). The CFI, an incremental fit index, indicates the proportion in the improvement of the overall fit of the hypothesized model to the null model. The GFI is an absolute fit index estimating how much better the hypothesized model fits compared to the null model. The RMSEA provides a measure of the model fit relative to the data, taking into account the complexity of the model. The SRMR measures the overall difference between the observed and predicted correlations. Acceptable fit generally is indicated by CFI values equal to or >0.95, GFI values close to 1.0, RMSEA values of 0.05 or less, and SRMR values equal to or <.08 (Browne and Cudeck 1993; Kline 2010).

We first specified a cross-lagged panel model examining the reciprocal associations of social standing with aggression, as detailed in Fig. 1. The model included the cross-paths for each aspect of social standing (controlling for the other) with each form of aggression, and for each form of aggression (controlling for the two others) with each aspect of social standing. This approach allowed us to consider the unique associations of popularity and social acceptance with electronic aggression. The model also took into account the stability over time of each construct. Although not depicted in Fig. 1, we allowed variables at Time 1 and error terms at Time 2 to correlate. We specified these intercorrelations at each wave to account partially for the shared method used to assess social standing and aggression; a similar approach has been adopted in prior research (Cillessen and Mayeux 2004; Puckett et al. 2008). Next, we ran a similar model examining the reciprocal links of social standing with victimization.

Fig. 1
figure 1

Path model for the reciprocal associations between social standing and aggression over time, controlling for construct stabilities. Paths of interest (i.e., between social standing and electronic aggression) are depicted with solid lines and correspond to the standardized estimates presented in Table 4. Intercorrelations between variables at Time 1 and between error terms at Time 2 were specified in the model but are not illustrated. d = disturbance. A similar model was specified for the reciprocal associations between social standing and victimization over time

Full size image

As illustrated in Table 4, the fit for our models was adequate. Adolescents’ initial levels of popularity were related positively and uniquely to increases in their electronically aggressive behavior. However, initial levels of social acceptance were not associated significantly with increases in electronic aggression. In turn, there were no significant associations between electronic aggression and later social standing. The reciprocal links between social standing and electronic victimization were not significant.

Table 4 Model fit and standardized paths between social standing, electronic aggression, and electronic victimization
Full size table

To investigate whether the reciprocal links of social standing with electronic forms of aggression and victimization vary by gender, we specified our cross-lagged panel models with paths free to differ across boys and girls. Our models fit the data adequately for males and for females. Model fit indices and parameter estimates for boys and girls are included in Table 4. Next, we conducted tests of moderation using a multi-group modeling approach. Multiple group path models allowed us to test whether specific parameter estimates varied across groups by imposing cross-group equality constraints. Such constraints force the statistical program to derive equal estimates of those parameters across groups. In our analyses, we imposed cross-group equality constraints on the paths from social standing at Time 1 to electronic aggression and victimization at Time 2 and the paths from electronic aggression and victimization at Time 1 to social standing at Time 2. The modified models with the paths constrained to be equal across males and females resulted in a significant decrement in fit (ps < .001). We concluded that the reciprocal links between social standing and electronic aggression and victimization differed for boys and girls.

To parse out which paths differed between males and females, we imposed cross-group equality constraints one-by-one on each of the reciprocal paths between the different dimensions of social standing and electronic forms of aggression and victimization. The added constraints significantly decremented the model fit for T1 Popularity → T2 Electronic Aggression (Δχ2 = 249, df = 1, p < .001), for T1 Electronic Aggression → T2 Popularity (Δχ2 = 8, df = 1, p < .01), for T1 Popularity → T2 Electronic Victimization (Δχ2 = 42, df = 1, p < .001), and for T1 Social Acceptance → T2 Electronic Victimization (Δχ2 = 99, df = 1, p < .001). Notably, elevated levels of popularity were related more strongly to increases in electronic aggression for boys than for girls. In turn, initial electronic aggression was associated significantly with small decreases in popularity for boys, but small increases for girls. Additionally, greater acceptance from peers was related significantly to greater electronic victimization, but only for adolescent males.

Discussion

During the adolescent years, aggression can play a critical role in the dynamics of social standing (Cillessen and Mayeux 2004). Aversive and domineering strategies often are used by popular adolescents to reach and maintain their position in their social hierarchy (Bowker et al. 2010). To provide a better understanding of the link between aggression and popularity, the current study focused on the emergence of new, electronic modalities for aggressive acts. Given that different forms of aggressive behavior can have independent implications for social standing across development (Rose et al. 2004a), it is important to consider novel expressions of aggression. Therefore, we examined the unique longitudinal relationships of electronic aggression with social standing during adolescence. Our results highlight the distinct and growing significance of aggression in the digital domain for the structure of the peer hierarchy.

Social Standing and the Perpetration of Electronic Aggression

By the middle years of adolescence, aggressive behaviors in school are associated with higher levels of prestige and visibility among peers (LaFontana and Cillessen 2002). Similarly, our correlational analyses indicated that aggression enacted with electronic communication tools is related positively to popularity in adolescence. Cross-lagged panel models revealed that adolescent girls with elevated levels of electronic aggression experienced increases in popularity over time, even after adjusting for individual differences in school-based aggression. Popular adolescents also became more electronically aggressive than their peers over the 1 year of our project. The pattern of results suggests that some youths not only may rely on hostile and manipulative acts in the digital world to establish social dominance and status, but also may use electronic aggression to uphold their elevated position with peers (Prinstein and Cillessen 2003). In addition, our findings underscore the reinforcement of electronic aggression in the adolescent peer system. These social benefits unfortunately may translate into maladaptive outcomes across different areas of adjustment (Rose and Swenson 2009; Schwartz et al. 2006).

Although aggressive adolescents are often disliked by the victims of their harassment and rejected by a substantial subset of their peers, they still may possess the social skills required to be liked by a portion of peers (Crick et al. 2009). As a result, aggression often is unrelated or only modestly related to peer liking (Card et al. 2008). In line with past research on aggression in schools and in the digital world (e.g., Card et al. 2008; Calvete et al. 2010), electronic aggression was related positively to concurrent social acceptance with a small effect size among our participants in ninth grade, and not significantly associated with peer liking in tenth grade. The prospective links between social acceptance and electronic aggression were not significant, after controlling for school-based aggression.

Social Standing and the Receipt of Electronic Aggression

A second focus of our project concerned the reciprocal relationships between electronic forms of victimization and social standing in adolescence. In school settings, those who experience the highest levels of victimization tend to maintain little prominence and influence among their peers (de Bruyn et al. 2010). Still, the most popular adolescents are often at risk for victimization, as the aggressive strategies that they use to gain and maintain status can place them in danger of reciprocated rivalry (Cillessen 2009). Therefore, we hypothesized a positive association between popularity and subsequent electronic victimization. Consistent with our expectation, there were small, positive correlations between popularity and electronic victimization at each time point of our study. However, the prospective links between popularity and electronic victimization were not significant after accounting for overt and relational victimization. The pattern of results suggests that popular youths are at an increased risk for victimization, regardless of the setting in which aggression may occur (i.e., in school vs. in the digital domain).

Typically, victimization in adolescent peer groups has a modest negative association with social acceptance (de Bruyn et al. 2010). However, we found a concurrent positive link between being accepted by peers and electronic victimization. Our cross-lagged panel analyses indicated that adolescent boys with elevated levels of social acceptance experienced greater electronic victimization over time than their less accepted peers. Given that offline social interactions often extend into the digital domain (Gross 2004), more accepted youths may have a more extensive social network with access to electronic media than others. Their increased potential for digital exchanges may place them at heightened risk for electronic victimization (Vandebosch and van Cleemput 2009). It will be important for future investigations to explore this possibility, assessing for individual differences in access and use of electronic technologies.

Gender Differences in Social Standing and Involvement in Electronic Aggression

A final aim of our project was to explore whether gender moderated the relationships between social standing and electronic forms of aggression and victimization. Consistent with prior work (e.g., Pornari and Wood 2010), adolescent girls were more electronically aggressive than boys. Moreover, electronically aggressive behaviors were related to increases in popularity for girls and decreases for boys, suggesting that perhaps only girls rely on electronically aggressive acts to establish social prominence. In contrast, our longitudinal analyses indicated that, having established prominence, boys may be more likely than girls to use aversive strategies in the digital world to maintain their position among peers. Qualitative research similarly has suggested that, whereas adolescent females may display aggression towards select peers while jockeying for social status, males may use aggressive behaviors in person and through digital means to maintain their position among a broad group of peers (Pronk and Zimmer-Gembeck 2010).

In line with previous investigations (e.g., Pornari and Wood 2010), adolescent males and females faced similar levels of electronic victimization. However, being accepted by peers was related positively to electronic victimization concurrently and prospectively only for boys. As previously suggested, more accepted youths may have a more extensive social network with access to electronic media than others and consequently be at greater risk for electronic victimization. Given that boys typically have larger peer groups than their female counterparts (Benenson 1990), accepted males may be particularly vulnerable to electronic victimization. Interestingly, prior research suggests that social acceptance may be unrelated to media access (Schoffstall and Cohen 2011), and that boys and girls do not differ in their access or use of electronic communication tools (Lenhart et al. 2010b; Underwood et al. 2011). Such work, however, focuses on youths’ individual characteristics, and no known studies have focused on adolescents’ peer networks with regards to the use of and access to digital technology.

Strengths, Limitations, and Future Directions

Before we move on to our concluding comments, some potential strengths and limitations of the current project should be identified. In particular, the focus on change within adolescence is an important contribution of this study. Prior work indicates that the social correlates of aggression and victimization shift during this stage of development (Cillessen and Mayeux 2004). Yet, the available findings regarding social standing and electronic forms of aggression and victimization largely have been circumscribed to young children (Schoffstall and Cohen 2011). Our study offers insight into the functions of electronic aggression and victimization in the lives of high standing youths beyond research with younger samples. Still, it will be important for future research to consider the developmental trajectories of social standing and electronic forms of aggression and victimization across a broad span of ages. For instance, the links between popularity and electronic aggression may be stronger in late adolescence, as involvement in aggression may peak at later ages within the digital world than the school context (Sumter et al. 2012).

A related strength of our project is the longitudinal aspect of our findings, which not only allows for the descriptions of developmental change but also provides more efficient estimates of effects than cross-sectional designs. Although our approach represents an important advance beyond existing cross-sectional research (e.g., Calvete et al. 2010), the conclusions that can be drawn from our results are limited by the short-term nature of our design. Meaningful changes in the explored social constructs may be easier to detect with longer term approaches (Cillessen and Borch 2006). Indeed, we found high stability in social standing. Perhaps, as a result, our results were often characterized by small effect sizes.

A notable methodological complexity of our study relates to our use of a peer-nomination approach to assess electronic aggression and victimization. The adoption of a peer-nomination tool extends beyond the typical utilization of single- or multiple-item self-report inventories of electronic forms of aggression and victimization (Menesini and Nocentini 2009). Importantly, the peer-nomination instrument often has been the method of choice for identifying aggressors and victims within the school context (e.g., Schwartz 2000), and may offer important insight into how peer dynamics play out in the digital world. Still, it may be worthwhile to replicate our analyses with multi-informant assessments of electronic aggression and victimization. Self-report tools or logs from cell phones, social networking sites, and e-mails may provide additional data on the frequency of individual adolescents’ electronic aggression and victimization. They also may capture incidents hidden to peers under the veil of anonymous communication tools. Despite adolescents’ preference for electronic mediums that provide identifying information (such as text messages, Lenhart et al. 2010b), a notable portion of the victims of electronic aggression report not knowing their aggressor (Kowalski and Limber 2007).

Another methodological limitation of our study was our focus on a single high school. Although the recruited institution was typical of public schools in the region in terms of demographics, scholastic performance and disciplinary outcomes (California Department of Education 2008), our findings may not generalize to all high school populations. Future research may benefit from replicating our results in diverse school settings across the nation (e.g., varying in urban development, economic resources).

Furthermore, future work may benefit from exploring the moderators and mediators of the associations between social standing and electronic forms of aggression and victimization. For instance, it may be fruitful to investigate how the structure of peer networks is related to the social standing of electronic aggressors and victims (as in, Ahn et al. 2010). It also may be beneficial to examine norms for hostile and manipulative behaviors within the peer context, as these have been shown to moderate the relationships between social standing and school-based aggression (Dijkstra et al. 2008). Finally, future investigations may benefit from considering individual attributes and cognitions. For example, prior work has noted that the link between popularity and school-based aggression is moderated by differences in social and emotional characteristics (e.g., sociability), peer-valued traits (e.g., athleticism), and representations of the peer environment (e.g., perceived social standing) (Mayeux and Cillessen 2008; Rosen and Underwood 2010; Vaillancourt and Hymel 2006).

Another fruitful avenue for future research may be to examine the consequences of electronic aggression and victimization for youths with elevated social standing. Prior research on aggression in schools notably has considered the interplay of aggression and victimization with social standing on psychological, academic, and interpersonal adjustment. Investigators have found that popularity buffers adolescents who hurt others through relationally aggressive acts from internalizing symptoms but does not protect overt aggressors (Rose and Swenson 2009). Additionally, for highly aggressive adolescents, increases in popularity are associated with increases in unexplained absences and decreases in grades (Schwartz et al. 2006). Finally, relational aggression is associated with having conflictual friendships for disliked youths but not for youths who are popular among their peers (Rose et al. 2004a). The adjustment implications of involvement in electronic aggression, similarly, may not be identical for youths with high levels of popularity and social acceptance as for lower standing adolescents.

Conclusion

Our longitudinal study considered electronic forms of aggression and victimization as correlates of social standing during adolescence. Our analyses suggested that adolescents rely on electronically aggressive acts to establish status and dominance with peers and use electronic aggression to maintain their privileged position in the social hierarchy. In addition, our results indicated that popular and accepted youths, who likely possess a large, digitally-connected social network, may be at increased risk for electronic victimization. Further research with multiple informants and in diverse school settings seems warranted. Future studies may benefit from considering the processes that account for, modify, and result from the links between social standing and electronic forms of aggression and victimization.