Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Sep 21.
Published in final edited form as: Soc Dev. 2009 Aug 1;18(3):577–596. doi: 10.1111/j.1467-9507.2008.00497.x

Who Dislikes Whom, and For Whom Does It Matter

Predicting Aggression in Middle Childhood

Stephen A Erath 1, Gregory S Pettit 2, Kenneth A Dodge 3, John E Bates 4
PMCID: PMC2747797  NIHMSID: NIHMS139460  PMID: 19774109

Abstract

This study investigated the role of mutual dislike dyads (MDDs) in the development of aggressive behavior across the middle childhood years. Of particular interest was whether involvement in MDDs predicted later aggression, and whether the magnitude of the association between MDDs and later aggression varied based on characteristics of target children and ‘others’ involved in their MDDs. Data were collected on a community sample of 453 children participating in an ongoing longitudinal study. Classroom peer nomination and rating-scale measures were collected in kindergarten through third grade; aggressive behavior problems were assessed via teacher ratings in the early elementary years (kindergarten and first grade) and late elementary years (fourth and fifth grade). MDD involvement in the middle elementary years (second and third grade) was associated with higher levels of aggression in the late elementary years among boys (but not girls), and these predictions held after controlling for group-level peer disliking in the middle elementary years, aggression in the early elementary years, and demographic variables. The association between MDD involvement and subsequent aggression was also qualified by the aggressiveness of others in children’s MDDs: Having more MDDs predicted later aggression only among boys whose MDDs involved mostly non-aggressive others.

Keywords: mutual dislike, aggression, middle childhood, peer relationships

Introduction

Peer relationships play an important role in children’s development (Ladd, 2005). Group-level indicators of peer relationships, such as the degree of disliking and the level of peer rejection (Coie, Dodge, & Coppotelli, 1982), have received extensive study, as have dyadic-level indicators, such as reciprocated liking, called friendships (Newcomb & Bagwell, 1995). Research has sought to identify the extent to which group-based and dyadic-based measures of peer relationships and reputations make independent contributions to children’s behavioral adjustment (Bukowski & Adams, 2005; Masten, 2005). Perhaps the most common approach to this issue has been to contrast mutual liking (construed as an index of friendship) and group-based liking or disliking (construed as an index of peer acceptance) (e.g., Ladd, Kochenderfer, & Coleman, 1997). A less common approach has been to consider mutual dislike dyads (MDDs), a potentially important dimension of peer relationships (Abecassis, Hartup, Haselager, Scholte, & Van Lieshout, 2002; Hodges & Card, 2003).

In contrast to an index of dislike by the peer group, which can reflect unilateral forms of antipathy, mutual dislike is a dyadic peer experience in which a child is disliked by another individual, whom the child also reports disliking. These dyads may, at times, be marked by high levels of conflict (Card, 2007; Hartup, 2005). MDDs appear to be a relatively common experience among elementary and middle school children, with reports of between 29 percent and 67 percent of children and preadolescents having at least one MDD, or ‘enemy’, depending on the type of assessment (Card & Hodges, 2007; Parker & Gamm, 2003; Pope, 2003; Schwartz, Hopmeyer-Gorman, Toblin, & Abou-ezzeddine, 2003). It is likely that a subset of MDDs can be characterized as bully-victim dyads (Abecassis, 2003; Card & Hodges, 2007). However, bully-victim dyads are less common, with estimates in the 15-30 percent range (Nansel, Overbeck, Pilla, Ruan, Simon-Morton, & Scheidt, 2001; Veenstra et al., 2005), and must involve intent to harm, repeated aggression over time, and an imbalance of power to be characterized as such (Nansel et al., 2001; Olweus, 1993). Furthermore, sociometric ratings or nominations must be reciprocated to constitute an MDD, whereas unilateral reports are generally used to characterize bullying or victimization. Thus, although some overlap exists, MDDs are not necessarily bully-victim dyads, and bully-victim dyads are not necessarily MDDs.

Studies provide mixed support for the contention that mutual disliking is associated with maladjustment (e.g., aggression) above and beyond group-level peer disliking. Several studies spanning middle childhood through early adolescence have failed to find a significant association between MDDs and aggressive behavior when group-level acceptance or rejection was controlled (Parker & Gamm, 2003; Rodkin, Pearl, Farmer, & Van Acker, 2003; Schwartz et al., 2003), whereas several other studies spanning late childhood through adolescence have reported associations between MDDs and aggression independent of peer rejection (Abecassis et al., 2002; Pope, 2003; Witkow, Bellmore, Nishina, Juvonen, & Graham, 2005). The present study provides an important addition to this research. It used MDDs in second and third grade to predict teacher-reported aggression in fourth and fifth grade, controlling for aggression in kindergarten and first grade and group-level peer dislike in second and third grade.

The present study also addressed several issues that have not been examined in prior research, and which might account for inconsistent findings in the literature. Prior findings are primarily drawn from cross-sectional investigations in which antecedent adjustment was not controlled, and in which MDDs were measured concurrently with proposed outcomes. The design of these studies precludes conclusions about directionality and may fail to distinguish deviant from normative levels of MDD involvement. In addition, few studies have examined MDD involvement during middle childhood, despite the developmental significance of establishing positive peer relations and learning to inhibit aggressive behavior during this period (Parker, Rubin, Erath, Wojslawowicz, & Buskirk, 2006). Furthermore, only one study to date has examined characteristics of disliked ‘others’ (i.e., the other child in an MDD) as a potential moderator (see Card & Hodges, 2007), and few studies have examined characteristics of target children themselves as moderators. In the current study, we examined MDD involvement during the middle elementary years as an independent predictor of aggression during the late elementary years, and tested whether this relation is conditional upon the gender and aggression of target children and the aggression of others in their MDDs.

Disliking at the Group vs. Dyadic Level as Predictors of Aggression

Peer social preference (Coie et al., 1982) indexes relative liking over disliking and has been used to designate children as rejected in status. Disliking by the peer group is a unilateral construct, representing the view of the group toward an individual (Asher, Parker, & Walker, 1996; Bukowski & Hoza, 1989). In contrast, mutual disliking is defined as two peers who have reciprocated negative feelings toward one another (Abecassis, 2003; Hartup, 2003). There is evidence that children who are highly disliked by the peer group and those involved in numerous MDDs display similar behavioral characteristics (e.g., aggression and jealousy) in their interactions with the peer group (Abecassis, 2003; Hartup, 2003). It is not surprising, then, that MDDs occur more frequently among peer-rejected children than among popular or average status children (Pope, 2003).

In spite of this overlap, there are reasons to believe that peer disliking and mutual disliking may be associated in non-redundant ways with child aggression (Card & Hodges, 2007). Group-level peer disliking may contribute to aggressive behavior, above and beyond aggressive tendencies that may lead to the initial disliking, because the broad social exclusion associated with rejection limits opportunities for learning prosocial alternatives to aggression, and because it signals to the peer group that disliked individuals are acceptable targets for maltreatment (Bierman, 2004; Dodge et al., 2003; Ladd, 2006). A key point is that a child receiving group-level disliking does not necessarily have to experience mutual dislike relationships, although a child with a large number of MDDs would, of course, experience group-level disliking.

In contrast, mutual dislike is potentially an intensive, emotionally charged, personalized type of rejection, with reciprocated negative feelings and behaviors experienced in a dyadic context (Abecassis et al., 2002; Card, 2007). Excess involvement in MDDs might fuel aggression if repeated negative exchanges provide numerous opportunities for reinforcement of aggressive behaviors and social information-processing biases (Dodge & Pettit, 2003; Egan, Monson, & Perry, 1998; Reid, Patterson, & Snyder, 2002). Indeed, much of children’s attributional bias and aggressive behavior occurs in the context of dyadic relationships, or between specific pairs of children (Coie et al., 1999; Hubbard, Dodge, Cillessen, Coie, & Schwartz, 2001). Over time, however, aggressive interactions might disproportionately shape the general social-cognitive schemas and behavioral inclinations with which children approach new interactions or relationships. Such generalized aggressive behavior would be reflected in teacher reports of aggressive behavior during late elementary school, the outcome examined in the present study.

Characteristics of Target Children and ‘Others’ in Their MDDs

The mechanism through which MMD involvement promotes aggressive behavior may depend, at least in part, on characteristics of target children and other children in their MDDs. Indeed, to understand fully the impact of a dyadic relationship on an individual, one must understand characteristics of both individuals in the relationship (Adams, Bukowski, & Bagwell, 2005; Vitaro, Brendgen, & Wanner, 2005). To our knowledge, only one prior study has investigated behavioral characteristics of individuals involved in MDDs as moderators. Card and Hodges (2007) reported that sixth- and seventh-grade children were more victimized in MDDs when the other child was aggressive, physically strong, non-victimized, and had low levels of internalizing problems. In the present study, we considered whether the association between MDD involvement and subsequent aggression might be conditional upon involvement with more or less aggressive others in MDDs.

It is possible that involvement in MDDs with aggressive others would foster aggression due to the potentially high number of conflictual interactions. Such relationships would be expected to provide ample opportunities for observing and provoking aggression, and might foster hostile attributional biases especially (Coie et al., 1999; Hubbard et al., 2001). However, in mutually aggressive MDDs, aggression is likely ‘rewarded’ through submission at times, but ‘punished’ through reciprocated aggression at other times, which might be expected to deter further aggression. Thus, a competing possibility is that MDDs with non-aggressive others would amplify aggression to a greater degree, perhaps as a result of positive outcome expectancies for aggression (Coie et al., 1999) and more consistent reinforcement for aggressive behavior in such relationships (Reid et al., 2002; Schwartz, Dodge, & Coie, 1993). That is, aggression directed toward a non-aggressive disliked other might be met with submission or emotional reactivity more often, as well as reinforcement from peers (e.g., laughing, elevated status), rather than reciprocated aggression. In the present study, we tested the hypothesis that the aggression of others in MDDs moderates the relation between number of MDDs and subsequent aggression, such that MDDs with non-aggressive others would promote aggression to a greater degree than MDDs with aggressive others. Although both possibilities seemed feasible, our hypothesis was based on compelling evidence derived from social learning theory that reinforcing contingencies are a primary mechanism by which aggressive behaviors are maintained and increased (Reid et al., 2002; Tapper & Boulton, 2005). It is important to note that we assessed the general aggression levels of others in target children’s MDDs, via peer nominations of children who start fights, say mean things, and get mad easily. We did not assess aggressive behavior exhibited specifically within MDDs.

We also considered whether target children’s gender or earlier aggression moderated the association between MDD involvement and later aggression. These were viewed as exploratory analyses due to the absence or inconsistency of evidence in the literature. Inconsistent evidence has emerged for gender as a moderator of the independent association between MDD involvement and aggression, with studies finding significant associations for boys and girls (Abecassis et al., 2002), boys only (Rodkin et al., 2003), and neither boys nor girls (Pope, 2003; Witkow et al., 2005). Another potential moderator is target child aggression. It is possible that MDDs contribute to increased aggression only among children who were highly aggressive in the first place, because these children may provoke the type of conflictual exchanges that can foster further aggression. Alternatively, the greatest growth in aggression may be observed among children with low aggression initially, for whom aggression is drawn out and reinforced through negative exchanges with disliked others. Finally, we tested three-way interactions in which the hypothesized moderating role of others’ aggression would differ by target child gender or aggression.

In addition, we considered the ‘main effects’ of demographic variables. Several studies have found no gender differences in MDD involvement (Card & Hodges, 2003; Parker & Gamm, 2003; Pope, 2003; Witkow et al., 2005), whereas other studies have shown higher MDD involvement among boys than girls (Abecassis et al., 2002; Rodkin et al., 2003; Schwartz et al., 2003). Few ethnic and socioeconomic differences have been reported in the existing literature on MDDs, although one study reported that European American children were more likely to have enemies than were Hispanic children (Pope, 2003). Thus, as an exploratory aim, we examined whether these demographic variables (i.e., gender, ethnicity, socioeconomic satus (SES)) were correlated with MDDs and whether gender or ethnicity moderated the association between MDDs and later aggression. Research has generally shown that boys, lower income children, and racial/ethnic minorities exhibit higher levels of aggressive behavior than their demographic counterparts (Aber, Brown, & Jones, 2003). Given these demographic differences, we also controlled for gender, ethnicity, and SES in the regression analysis predicting late elementary aggression.

Given the limited literature on MDDs in the elementary years, we provide descriptive data on MDDs from kindergarten through third grade. However, in our main analyses we used a composite of MDDs across second and third grades. We used only these two years because we wished to insure that all major constructs were assessed across comparable periods of time (i.e., two years). Thus, a measure of the main covariate, aggression, was formed using kindergarten and first grade data; the independent variables (group-level dislike, MDDs, and MDDs with aggressive others) were constructed using second and third grade data; and the dependent variable, aggression, was assessed across fourth and fifth grades.

Method

Participants

This study included participants in the Child Development Project (CDP). CDP participants were originally recruited during kindergarten preregistration, in two cohorts (1987 and 1988), from three sites: Nashville and Knoxville, Tennessee and Bloomington, Indiana. Approximately 75 percent of families agreed to participate in the project. Data collection began the summer before participants entered kindergarten, and follow-up assessments were conducted annually. The initial sample of 585 participants was diverse in terms of gender (52 percent male), ethnicity (81 percent European American, 17 percent African American, 2 percent other ethnic groups), family composition (26 percent lived with single mothers), and socioeconomic status (Hollingshead index, M = 40, SD = 14). A more comprehensive description of sample characteristics and data collection procedures can be found in Pettit, Bates, and Dodge (1997).

Participants were excluded from the present study if they did not have kindergarten or first grade teacher-report data, peer nomination/rating data in second or third grade, and teacher-report data in fourth or fifth grade. There were 453 CDP participants with complete data according to these criteria. Comparisons of participants with and without complete data revealed no significant differences on gender, ethnicity, aggression, MDDs, or group-level disliking at kindergarten, and no differences on teacher-reported aggression during the late elementary years (all ts < 1.5, all ps > .10).

Procedures

Sociometric interviews were administered in each child’s classroom in kindergarten through third grade to all peers whose parents consented to their participation (about 75 percent of each class participated). Teachers completed the child behavior checklist-teacher report form (CBCL-TRF; Achenbach, 1991) during each year of elementary school.

Measures

Sociometric Procedure

Sociometric interviews were conducted in the winter of each school year. Sociometric interviews were conducted individually in kindergarten and first grade and in small groups in second and third grade classrooms. Each child was shown photographs (kindergarten), nametags (first grade), or a roster with names (second and third grade) of all classmates, and asked to nominate up to three peers they liked and three peers they disliked, as well as peers with prosocial and aggressive characteristics. Each child was also asked to rate how much he or she liked each classmate on either a three-point or five-point scale, with higher ratings indicating higher liking and a rating of ‘l’ as the lowest possible rating. Consistent with past research, we relied on the classroom as the unit of analysis for classifying children’s peer relationships (e.g., Hodges, Malone, & Perry, 1997; Ladd et al., 1997). In the participating schools, children remained with their classmates for the entire day; therefore, most of their social interactions occurred within the classroom peer context.

Mutual Dislike Dyads (MDDs)

An MDD was operationalized as reciprocated sociometric ratings of low-liking (i.e., ‘ls’ on either a three-point or five-point scale) between two same-sex peers, consistent with prior investigations (e.g., Parker & Gamm, 2003; Pope, 2003). Ratings of ‘l’ (i.e., lowest rating) on the sociometric scale were accompanied by a sad face, and described as appropriate for ‘someone that we like a whole lot less than others’, whereas the highest rating was accompanied by a very happy face, and described as appropriate for ‘someone we like most of all’. The number of ‘l’ ratings on a Likert scale of peer liking was tabulated and cross-referenced with classmates’ ratings to identify the number of MDDs for each participant. This is an indirect measure of mutual antipathy insofar as children are no asked explicitly if they consider themselves to have an ‘enemy’. As such, this approach is akin to what has been used in much of the reciprocated friendship literature, in which children typically are asked about those peers whom they like, or wish to associate with, rather than those peers whom they consider to be friends per se. There is evidence that most-liked and most-disliked nominations are highly concordant with corresponding ratings of liking and disliking (Vitaro, personal communication, November 9, 2006). These findings lend some confidence to the meaningfulness of assessing mutual antipathy with reciprocated low-liking ratings. Still, to be more consistent with the operationalization of the measure, we use the term mutual dislike, or MDDs, to refer to these dyadic arrangements.

A composite score for number of MDDs during the middle elementary years was computed by averaging the number of MDDs for second and third grade, r = .27, p < .001. Of the 453 participants included in the present study, 89 percent had mutual dislike data for both second and third grade; a single-year MDD score was used for the participants with only second or third grade data (11 percent). MDDs for second and third grade were each strongly associated with the composite MDD score (rs = .82 and .80, respectively). We utilized an average score representing MDDs in second and third grade to provide a more reliable estimate of exposure to MDDs during the middle elementary years than a single-year score. For descriptive purposes, means and standard deviations for MDDs in kindergarten through third grade are shown in Table 1.

Table 1. Correlations Among MDDs and Group-level Peer Disliking from Kindergarten through Third Grade.
1 2 3 4 5 6 7 8
Number of MDDs
1. Kindergarten MDDs —
2. First-grade MDDs .29** —
3. Second-grade MDDs .16* .21** —
4. Third-grade MDDs .25** .06 .27** —
Group-level peer dislike
5. Kindergarten nominations .39** .32** .21** .27** —
6. First-grade nominations .31** .42** .32** .33** .40** —
7. Second-grade nominations .16* .23** .23** .22** .35** .35** —
8. Third-grade nominations .27** .16* .19** .41** .38** .36** .41** —
Mean .81 .54 .70 .73 -.13 -.16 -.07 -.05
Standard deviation 1.20 1.06 1.18 1.17 .91 .91 .98 .93
Open in a new tab
*

p < .01

**

p < .001.

Note: Means and standard deviations for group-level peer dislike are z-scores.

MDDs = mutual dislike dyads.

Group-level Peer Dislike

The number of ‘like least’ nominations received from same-sex classmates, standardized within classrooms, were averaged across second and third grade to assess group-level peer disliking in the middle elementary years. Peer disliking scores in second and third grades were significantly correlated, r = .41, p < .001, and both scores were strongly associated with the composite score (rs = .87 and .85, respectively). Eighty-seven percent of participants had peer dislike data for both second and third grade; a single-year score was used for participants with peer dislike data for only second or third grade (13 percent). Means and standard deviations for group-level peer dislike in kindergarten through third grade are shown in Table 1.

Proportion of MDDs with Aggressive ‘Others’

The peer nomination inventory included three descriptors of aggression (‘Kids who start fights’, ‘Kids who say mean things’, ‘Kids who get mad easily’), which were summed and standardized within class (α = .89) to create an aggression score. Others in MDDs were classified as aggressive if their score on peer nominations of aggression was greater than .5 SD above their class mean. The aggression of others in MDDs was assessed with a count score representing the number of others with elevated aggression (>.5 SD), rather than with an average score, because it was not expected that the others in a child’s MDD network would necessarily have similar levels of aggression (in contrast to a child’s friendship network, within which behavioral similarity would be expected due to selection and socialization processes). Indeed, among children with more than one MDD, 30 percent had at least one MDD with a high level of aggression (>.5 SD above the mean) and at least one MDD with a low level of aggression (<.5 SD below the mean). Averaging aggression scores across others with divergent levels of aggression would misrepresent the extent to which these target children were exposed to others in their MDDs who were highly aggressive. Thus, a composite score was computed representing the total number of aggressive others for each participant across second and third grade. The proportion of MDDs with aggressive others to total MDDs was calculated and arcsine transformed, a common procedure used to linearize proportion scores (Cohen & Cohen, 1983).

Aggressive Behavior

Teachers completed the CBCL-TRF (Achenbach, 1991) each year. The CBCL-TRF is comprised of 112-items, rated on a three-point scale. The aggressive behavior subscale of the CBCL-TRF was used for the present study. Kindergarten, first, fourth, and fifth grade teacher reports of aggression were used in the present study. Teacher reports of aggression for kindergarten and first grade, r = .59, p < .001, were averaged to create a composite score for early elementary aggression (M = 5.02, SD = 7.00), and teacher reports for fourth and fifth grade, r = .63, p < .001, were averaged to create a composite score for late elementary aggression (M = 5.68, SD = 8.06). Again, composite scores were expected to provide more reliable estimates of aggressive behavior than single-year scores. Ninety-seven percent of participants had teacher report data for both kindergarten and first grade, and 83 percent of participants had teacher report data for both fourth and fifth grade; single-year scores were used for participants with only kindergarten or first grade data (3 percent) and for participants with only fourth- or fifth-grade data (17 percent).

Results

Preliminary Analyses

Means, standard deviations, and correlations for MDDs and group-level peer disliking from kindergarten through third grade are presented in Table 1. As shown in Table 1, correlations between MDDs contiguous years were moderate in magnitude, ranging from .21 to .29, and correlations across more than one year were modest to moderate, ranging from .06 to .25. Correlations among group-level peer disliking variables were moderate in magnitude, ranging from .35 to .41 for contiguous years and across more than one year. Correlations between MDDs and group-level peer dislike ranged from .16 to .42, contiguously and across years, and the middle elementary composite scores for MDDs and group-level peer dislike were also moderately correlated, r = .37, p < .001. On average, participants had between .54 and .81 MDDs in a given year of elementary school (including kindergarten through third grade), and participants had a mean of .71 (SD = .96) MDDs per year in second and third grade. Fifty-three percent to 69 percent of children were involved in no MDDs in a given year; 19 percent to 26 percent were involved in one MDD in a given year; and 12 percent to 21 percent were involved in more than one MDD in a given year. The average proportion of MDDs with aggressive others among children with at least one MDD in second or third grade was .39 (SD = .40).

Correlations among Primary Study Variables

As hypothesized, significant correlations emerged among group-level dislike nominations, MDDs, and aggression (see Table 2). Notably, the correlations linking middle elementary peer dislike and MDDs with late elementary aggression, rs = .42 and .29, respectively, were in the ‘medium’ to ‘large’ range of effect sizes (Cohen & Cohen, 1983). In addition, boys were more aggressive and less liked by the peer group, as compared with girls. Ethnic minorities were rated by teachers as more aggressive during late elementary school, as compared with European Americans. Children in lower SES families were rated by teachers as more aggressive in early and late elementary school, and also were less liked by the peer group (see Table 2). A higher proportion of MDDs with aggressive others was associated with fewer MDDs, lower group-level peer disliking, and lower late elementary aggression. Girls had a higher proportion of MDDs with aggressive others than boys. No demographic variables were correlated with MDDs, and the association between MDDs and later aggression did not differ by ethnicity.

Table 2. Correlations among Primary Study Variables (N = 453).

1 2 3 4 5 6 7
1. Gender —
2. Ethnicity .01 —
3. Socioeconomic status -.04 -.43*** —
4. Early elementary aggression -.20*** .04 -.23*** —
5. Peer dislike nominations -.11* .07 -.19*** .50*** —
6. Number of MDDs -.03 .08 -.07 .26*** .37*** —
7. Proportion MDDs with aggressive othersa .22*** -.00 .11+ -.12+ -.23*** -.26*** —
8. Late elementary aggression -.29*** .17*** -.26*** .61*** .42*** .29*** -.18***
Open in a new tab
+

p < .10

*

p <.05

***

p <.001.

a

Correlations including proportion of MDDs with aggressive others were computed using the subsample of participants with at least one MDD (N = 246).

For gender, boys conded as 0 and girls coded as 1. For ethnicity, European American coded as 0 and African American coded as 1.

MDDs = mutual dislike dyads.

Predicting Late Elementary Aggression

Hierarchical multiple regression analyses were conducted to evaluate the independent contribution of middle elementary MDDs to late elementary aggression. For this analysis, aggression in the late elementary years (i.e., fourth and fifth grades) was the dependent variable. All continuous variables were standardized. Demographic variables and early elementary (i.e., kindergarten and first grade) aggression were entered on the first step as control variables; peer dislike nominations and number of MDDs in the middle elementary years (i.e., second and third grade) were entered on the second step; the products of MDDs and gender, MDDs and early elementary aggression, and gender and early elementary aggression were entered on the third step to represent their interactions; and the three-way product of MDDs, gender, and early elementary aggression was entered on the fourth step. Standard procedures were utilized to test and interpret interactions, as recommended by Aiken and West (1991) and more recently described by Dearing and Hamilton (2006).

As shown in Table 3, each control variable was significantly associated with late elementary aggression. Specifically, at entry, gender, B = -3.03, SE = .59, ethnicity, B = 2.34, SE = .80, SES, B = -.71, SE = .33, and early elementary aggression, B = 4.39, SE = .30, each significantly predicted late elementary aggression. Consistent with correlation analyses, boys, ethnic minorities, and children from lower SES families were rated by teachers as more aggressive in late elementary school as compared with their demographic counterparts. As anticipated, group-level peer disliking during the middle elementary years also significantly predicted late elementary aggression, B = .89, SE = .34. In addition, MDDs significantly predicted late elementary aggression, B = .89, SE = .31, over and above demographic variables, early elementary aggression, and grup-level peer disliking.1 Gender moderate the association between MDDs and late elementary aggression, B = -1.30, SE = .60, such that MDDs predicted increased aggression among boys, B = 1.67, SE = .42, p < .001, but not among girls, B = .02, SE = .05, p > .10. The set of predictors accounted for 47 percent of variance in late elementary aggression.2

Table 3. Predicting Late Elementary Aggressive Behavior (N = 453).

Late elementary aggression (4th and 5th grades)
Predictors βstep ΔF ΔR2
Step 1 83.60*** 43
 Gender -.19***
 Ethnicity .12**
 Socioeconomic status -.09*
 Earlier aggression .55***
Step 2 10.61*** .03
 Peer dislike nominations .11**
 Number of MDDs .11**
Step 3 3.65* .01
 MDDs × earlier aggression -.00
 MDDs × gender -.11*
 Earlier aggression × gender -.08+
Step 4 .72 .00
 MDDs × earlier aggression × gender -.04
Open in a new tab
+

p < .10

*

p < .05

**

p < .01

***

p < .001.

Note: Earlier aggression refers to target child aggression assessed in kindergarten and first grade. For gender, boys coded as 0 and girls coded as 1. For ethnicity, European American coded as 0 and African American coded as 1.

MDDs = mutual dislike dyads.

Moderators among Children with at Least One MDD

A separate regression analysis was conducted to test interactions between MDDs and characteristics of both target children (i.e., gender, early elementary aggression) and the others in their MDDs (i.e., aggression) as predictors of late elementary aggression (see Table 4). This analysis included only participants with at least one MDD (N = 246) because the aggression of others in MDDs cannot moderate MDDs if a child has no MDDs. Demographic variables and early elementary aggression were entered on the first step as control variables; peer dislike nominations and number of MDDs in the middle elementary years were entered on the second step; and the proportion of MDDs with aggressive others was entered on the third step. The products of MDDs and MDDs with aggressive others, MDDs and gender, and MDDs and early elementary aggression were entered on the fourth step to represent their interactions; in addition, the products of MDDs with aggressive others and both gender and early elementary aggression were entered on the fourth step. The inclusion of these two-way interactions allowed us to enter three-way products that tested whether the hypothesized moderating role of MDDs with aggressive others varied by gender or target child aggression (fifth step).

Table 4.

Moderators among Children with at Least One MDD (N = 246)

Late elementary aggression (4th and 5th grades)
Predictors βstep ΔF ΔR2
Step 1 49.46*** .45
 Gender -.21***
 Ethnicity .12*
 Socioeconomic status -.12*
 Earlier aggression .53***
Step 2 5.65** .03
 Peer dislike nominations .09+
 Number of MDDs .12*
Step 3 .05 .00
 Aggressive others -.01
Step 4 2.23+ .02
 MDDs × earlier aggression -.02
 MDDs × gender -.14+
 MDDs × aggressive others -.12
 Earlier aggression × aggressive others -0.2
 Aggressive others × gender .01
Step 5 2.30 .01
 MDDs × gender × aggressive others .22*
 MDDs × aggression × aggressive others -.03
Open in a new tab
+

p ≤ .10

*

p < .05

**

p < .01

***

p < .001.

Note: All participants in this model had at least one MDD. ‘Earlier aggression’ refers to target child aggression assessed in kindergarten and first grade. ‘Aggressive others’ refers to proportion of MDDs with aggressive others. Boys coded as 0 and girls coded as 1. European American coded as 0 and African American coded as 1.

MDDs = mutual dislike dyads.

Like the regression analysis using the full sample, demographic variables, group-level peer dislike, and MDDs predicted increased late elementary aggression (see Table 4). Proportion of MDDs with aggressive others was not significantly associated with late elementary aggression. None of the two-way interactions significantly predicted late elementary aggression. However, the three-way interaction among MDDs, MDDs with aggressive others, and gender significantly predicted late elementary aggression, B = 3.33, SE = 1.63.

Follow-up calculation of simple intercepts, simple slopes, and regions of significance was conducted according to standard procedures (Aiken & West, 1991; Dearing & Hamilton, 2006), using the interaction utility described by Preacher, Curran, and Bauer (2006). These analyses yielded intercepts and slopes representing the relations between the independent variable (MDDs) and dependent variable (late elementary aggression) conditional upon levels of the moderator variables (gender and proportion of MDDs with aggressive others). Conditional levels of the proportion of MDDs with aggressive others were zero (i.e., zero proportion of MDDs with aggressive others) and one standard deviation above the mean, and conditional levels of gender were boys and girls. In addition, regions of significance were computed, to identify values of the moderators at which a significant association between the independent variable and dependent variable existed (Preacher et al., 2006). It is important to note that simple intercepts and simple slopes were computed with all control and predictor variables and interaction terms in the model, consistent with recommended procedures (Preacher et al., 2006).

As shown in Figure 1, MDDs significantly predicted increased late elementary aggression among boys with a zero proportion of MDDs with aggressive others, B = 3.16, SE = .79, p < .0001, but not among boys with a high (+1 SD) proportion of MDDs with aggressive others, B = -3.08, SE = 2.35, p > .10. Simple intercepts and slopes for boys are shown in Figure 1. MDDs were not associated with late elementary aggression among girls with a zero proportion of MDDs with aggressive others, B = -.00, SE = .91, p > .10, or a high proportion of MDDs with aggressive others, B = -.08, SE = 1.35, p > .10. Regions of significant analysis revealed that MDDs significantly predicted increased late elementary aggression among boys at less than .34 standard deviations below the mean of MDDs with aggressive others (i.e., approximately one-third or less of MDDs included aggressive others). The non-significant negative relation between MDDs and late elementary aggression among boys whose MDDs involved mostly aggressive others did not reach significance even when all MDDs included aggressive others. MDDs did not significantly predict late elementary aggression among girls at any proportion of MDDs with aggressive others.

Figure 1.

Figure 1

Open in a new tab

Relations Between MDDs and Late Elementary Aggression among Boys with at Least One MDD and a Higher (+1 SD) or Lower (zero) Proportion of MDDs with Aggressive Others. *** p < .001.

Discussion

Mutual dislike has received little research attention as compared with other dimensions of children’s peer relationships. Existing research indicates that MDDs are not uncommon in middle to late childhood and preadolescence (Hodges & Card, 2003), and some studies find that MDDs are uniquely linked with psychosocial maladjustment (Abecassis et al., 2002), whereas others do not (Parker & Gamm, 2003; Schwartz et al., 2003). Ours is the first study, however, to examine MDD involvement over the course of multiple years, controlling for earlier levels of the outcome variable (aggression) and group-level peer dislike, and simultaneously considering characteristics of target children and ‘others’ in their MDDs that may shape the influence that such relation-ships have on aggressive behavioral development.

We found that middle elementary MDDs both follow from earlier aggressive behavior and predict later aggressive behavior, controlling for earlier aggression and concurrent group-level peer disliking. However, moderation analyses revealed that MDDs predicted increased aggression only among boys whose MDDs involved mostly non-aggressive others. The magnitude of this association was high. In contrast, MDDs were not associated with increased aggression among girls or boys whose MDDs involved mostly aggressive others. Findings underscore the value of considering variability in associations under investigation, and suggest that MDDs may contribute to aggression (or not) depending on characteristics of target children and others in their MDDs.

Coie et al. (1999) explained that aggression might be fostered in the dyadic context due to expectations of success and incentives for aggression (e.g., access to resources, peer group reinforcement) that can evolve through a shared history of negative interactions. Other researchers have emphasized negative reinforcement as a mechanism for the development of aggressive behavior (Reid et al., 2002), such as when a disliked other submits to aggressive behavior in the context of a disagreement. MDDs may serve as a fertile context for the dyadic development of aggression due to the potential frequency and intensity of negative interactions that can result from reciprocated dislike (Card, 2007; Hartup, 2005). Interactions between mutual disliking pairs may be more volatile than interactions between pairs of children in which only one of the children dislikes the other due to the greater likelihood of negative engagement by both parties in an MDD. Indeed, preschool and school-age children respond with more negativity and anger to provocations by disliked peers as compared with well-liked peers (DeLawyer & Foster, 1986; Fabes, Eisenberg, & Smith, 1996), and peer provocations are viewed as more hostile when enacted by disliked peers as compared with other peers (Parker & Gamm, 2003; Ray & Cohen, 1997).

Although MDDs likely share some negative features in common, Hartup (2003) urged researchers to consider the heterogeneity of MDDs and different outcomes associated with various forms of MDDs. The present study takes an initial step toward understanding variability, by examining whether the influence of MDDs depends upon characteristics of target children (I.e., gender, aggression) and others involved in their MDDs (i.e., aggression). Analyses revealed that MDDs predicted later aggression among boys but not among girls. Likewise, controlling for group-level peer dislike, losing same-sex enemies over the course of a school year was associated with decreased aggression among school-age boys, but increased aggression among school-age girls (Rodkin et al., 2003). Schwartz et al. (2003) also reported that a higher number of MDDs and aggression were correlated only among boys, yet these correlations analyses did not control for group-level peer dislike. In contrast, Abecassis et al. (2002) found that involvement in same-sex mutual antipathies was associated with higher antisocial and aggressive behavior among boys and girls, controlling for peer rejection. Both Witkow et al. (2005) and Pope (2003) reported that same-sex mutual antipathies were not associated with aggression for boys or girls when group-level peer dislike was controlled. Further research is needed to clarify associations between same- or mixed-sex MDDs and aggression among boys and girls. Differentiating subtypes of aggressive behavior (i.e., physical vs. relational, proactive vs. reactive) might help accomplish this objective.

Examining additional characteristics of both target children and others in their MDDs should also help specify the conditions under which MDDs can promote aggression. Indeed, in the present study, moderation analyses revealed that growth in aggression occurred among boys whose MDDs included mostly non-aggressive others, but not among boys whose MDDs included mostly aggressive others. These results suggest that reinforcement of aggressive behavior is a potential mechanism by which boys’ MDDs foster increased aggression. For example, MDDs with non-aggressive peers might resemble bully-victim relationships in which bullies anticipate incentives for aggressive behavior and develop high expectations for success (Card & Hodges, 2007; Olweus, 1993). That is, in such relationships, bullies often elicit submission from their victims, or achieve the intended physical or emotional harm, and bystanders often provide further reward for such behavior (Schwartz et al., 1993). Boys, in particular, may be inclined to target other boys on the basis of behavioral, affective, or social vulnerabilities (Boulton, 1999; Erath, Flanagan, & Bierman, 2007; schwartz et al., 1993). Conversely, in the context of mutually aggressive MDDs, expectations of success and the likelihood of reinforcement for aggression might be diminished by a history of retaliatory aggression, or ‘punishment’ for aggressive behavior (Coie et al., 1999). Indeed, Coie et al. (1999) found that expectations of success for aggression were higher among random dyads of boys as compared with mutually aggressive dyads of boys.

An additional mechanism by which MDDs with non-aggressive others might promote growth in aggression, as compared with MDDs with aggressive others, involves the broader peer group context. For example, aggressing against non-aggressive disliked others, who are perhaps well-adjusted socially, might suggest that the aggressor lacks positive peer affiliations and thus lacks socialization opportunities that might correct aggressive tendencies and teach prosocial alternatives (F. Vitaro, personal communication, October 31, 2006). In addition, it is possible that some aggressors are unaffected by the punishments that would otherwise reduce aggression over the years of middle childhood. Of course, these findings must be interpreted with caution, and the characteristics of others in MDDs clearly require further investigation.

Results of the present study provide evidence for contributions of mutual dislike to increased aggression over time under certain circumstances (i.e., boys in MDDs with mostly non-aggressive others). However, it is likely that the relation is reciprocal, such that aggressive behavior also fosters involvement in MDDs, as suggested by the moderate correlation between early elementary aggression and middle elementary MDD involvement. Hartup (2003) asserted that aggressive children have a propensity for involvement in mutual dislike dyads, and these relationships, in turn, cultivate aggression. For example, aggressive children have trouble resolving conflict in the first place, potentially contributing to the initiation of MDDs, and peers who dislike one another may be less concerned with resolving conflict and perhaps more vested in perpetuating it through further aggression (Abecassis et al., 2002).

Several studies have shown that many preadolescents and early adolescents are involved in MDDs during a given year. Likewise, in the present study, 54 percent of children were involved in at least one MDD in second or third grade, and about 75 percent were involved in at least one MDD over the course of kindergarten through third grade. MDDs thus seem to be relatively common during the middle elementary years, and MDD involvement per se does not appear to signal a problem.3 Short-lived disliking relationships might even afford opportunities for refining social problem-solving skills (Shantz & Hartup, 1992). In contrast, only 12 percent to 21 percent of children in the present study were involved in more than one MDD in a given year (kindergarten through third grade). Likewise, 10 percent of the children were involved in at least one MDD for all four years of assessment, and 6 percent were involved in more than two MDDs per year, on average. The present study suggests that excessive MDD involvement during the middle elementary years is not only less common, but also can provide a context for development of aggressive behavior.

In contrast to such negative peer involvement, positive peer relationships in middle childhood allow children to meet increasing developmental needs for peer acceptance, and to gain social competencies that set the course for subsequent social adjustment (Masten, Burt, & Coatsworth, 2006). It is clear that children who are rejected during middle childhood are at risk for future maladjustment (Ladd, 2006), in part, because dislike by the peer group is not conducive to the promotion of prosocial behavior or inhibition of aggressive behavior (Bierman, 2004). The results of this study suggest that boys’ excessive involvement in MDDs with non-aggressive others during middle childhoood also fosters increased aggression during a developmental period when aggression typically (and importantly) declines.

It is important to emphasize that the present study did not assess actual interactions between children involved in MDDs, and thus we can only speculate about interactive behavior on the basis of individual characteristics (e.g., high peer ratings of aggression). Several discussion points are based on the assumption that at least some children involved in MDDs interact (aggressively) with one another, which is supported by recent research (Card, 2007). It will be important for future research to investigate variations on MDDs as well as interactions between children involved. In addition, we were not able to examine the extent to which children involved in MDDs with aggressive and non-aggressive others are representative of bully-victims and bullies, respectively. The relatively high rate of MDD involvement in the present study suggests that our operationalization of MDDs tapped a more common phenomenon than bully-victim dyads. It would be informative for future research to consider whether the aggression of ‘others’ in bully-victim dyads also moderates growth in aggression.

It is also possible that mutual low-liking does not reflect mutual disliking, and our operationalization of mutual dislike might have overestimated the number of children actually involved in MDDs. It may be more valid for researches to explicitly frame items to refer to mutual dislike or enemies, rather than assuming that mutual low-liking is indicative of mutual disliking (Hartup, 2003). In addition, we found that MDDs independently predicted later aggression when MDDs were assessed through reciprocated ratings but not when MDDs were assessed through reciprocated nominations. It is not clear to what extent this null finding reflects the strength of the association between MDDs and later aggression, or the restricted range of MDDs assessed through reciprocated nominations as compared with ratings (Pope, 2003). Finally, the present study was also limited by its exclusive reliance on same-sex mutual dislike data. As noted, other studies have shown that important differences may exist between same- and mixed-sex MDDs (e.g., Abecassis et al., 2002; Witkow et al., 2005), and it will be important for future research to differentiate these forms of MDDs.

Despite these limitations, strenghts of the present study include assessment of MDDs during middle childhood, over multiple years, and longitudinal analyses with statistical control of concurrent group-level peer dislike and earlier levels of aggression. Results indicate that MDDs during the middle elementary years predict increased aggression over time among boys, and provide the first evidence that the aggression of ‘others’ in chlidren’s MDDs moderates children’s subsequent aggression. Having mostly non-aggressive others appears to foster boys’ aggression. Future research can build on these findings by directly investigating patterns of interaction between children involved in MDDs (e.g., counter-aggression, withdrawal, problem-solving, etc.) that affect the course of these relationships and their outcomes.

Acknowledgments

The Child Development Project has been supported by grants from the National Institute of Mental Health (MH 42498, MH 56961, MH 57024, MH 57095), the National Institute of Child Health and Human Development (HD 30572), and the National Institute on Drug Abuse (DA 16903). Appreciation is expressed to David Cleary and Jared Lisonbee for their important early contributions to this work and to Robert Larid and David Schwartz for insightful comments.

Footnotes

1

We operationalized MDDs using reciprocated ratings of low-liking, and operationalized group-level disliking using like-least nominations. The use of separate methodologies to measure dyadic and group-level disliking was designed to reduce the influence of shared method variance, and therefore to provide a better estimate of the independent contribution of each variable to late elementary aggression. In addition, assessing MDDs through reciprocated ratings yielded a greater range of MDD scores, as compared with assessing MDDs through reciprocated nominations, which is an important consideration for constructs that require reciprocated responses. As an alternative, we classified MDDs according to reciprocated like-least nominations (M = .36 per year; SD = .37). MDDs classified according to ratings and nominations were moderately correlated (p < .001) for each grade (rs = .20, .25, .26, .30 for kindergarten, first-, second-, and third-grade, respectively). In addition, middle elementary MDDs classified according to reciprocated nominations were moderately correlated with late elementary aggression, r = .26, p < .001. However, MDDs classified according to reciprocated nominations did not significantly predict late elementary aggression, over and above group-level peer disliking, classified according to low-liking ratings, and other demographic variables.

2

An alternative analysis was conducted using MDDs and group-level peer disliking averaged across four years of elementary school (i.e., kindergarten through third grade) rather than two-year composites (i.e., second and third grade). This analysis yielded nearly identical results. That is, using the same set of control variables, including aggression assessed in kindergarten and first grade, and four-year composite scores for MDDs and group-level peer disliking, there was a significant main effect for MDDs, β = .09, p < .05, and a significant interaction effect for MDDs × gender, β = -.12, p < .05.

3

A follow-up, group-based analysis was conducted to determine whether children with at least one MDD in second or third grade were more aggressive in late elementary school than children with no MDDs in second or third grade. An ANOVA revealed no significant difference between children with at least one MDD and children with no MDDs, F(1, 446) = 2.12, p > .10. This finding suggests that the association between number of MDDs and late elementary aggression may be driven by increasing numbers of MDDs in a linear fashion, rather than by the presence or absence of a single MDD.

Contributor Information

Stephen A. Erath, Auburn University

Gregory S. Pettit, Auburn University

Kenneth A. Dodge, Duke University

John E. Bates, Indiana University

References

  1. Abecassis M. I hate you just the way you are: Exploring the formation, maintenance, and need for enemies. In: Hodges EV, Card N, editors. Enemies and the darker side of peer relation. New directions for child and adolescent development. Jossey-Bass; San Francisco, CA: 2003. pp. 5–22. [DOI] [PubMed] [Google Scholar]
  2. Abecassis M, Hartup WW, Haselager GJT, Scholte RHJ, Van Lieshout CFM. Mutual antipathies and their significance in middle childhood and adolescence. Child Development. 2002;73:1543–1556. doi: 10.1111/1467-8624.00489. [DOI] [PubMed] [Google Scholar]
  3. Aber JL, Brown JL, Jones SL. Developmental trajectories toward violence in middle childhood: Course, demographic differences, and response to school-based intervention. Developmental Psychology. 2003;39:324–348. doi: 10.1037//0012-1649.39.2.324. [DOI] [PubMed] [Google Scholar]
  4. Achenbach TM. Integrative guide for the 1991 CBCL/4-18, YSR, and CBCL-TRF profiles. University of Vermont Department of Psychiatry; Burlington, VT: 1991. [Google Scholar]
  5. Adams RE, Bukowski WM, Bagwell C. Stability of aggression during early adolescence as moderated by reciprocated friendship status and friend’s aggression. International Journal of Behavioral Development. 2005;29:139–145. [Google Scholar]
  6. Aiken LS, West SG. Multiple regression: Testing and interpreting interactions. Sage; Newbury Park, CA: 1991. [Google Scholar]
  7. Asher SR, Parker JG, Walker DL. Distinguishing friendship from acceptance: Implications for intervention and assessment. In: Bukowski WM, Newcomb AF, Hartup WW, editors. The company they keep: Friendships during childhood and adolescence. Cambridge University Press; New York: 1996. pp. 366–405. [Google Scholar]
  8. Bierman KL. Peer rejection: Developmental processes and intervention strategies. Guilford; New York: 2004. [Google Scholar]
  9. Boulton MJ. Concurrent and longitudinal relations between children’s playground behavior and social preference, victimization, and bullying. Child Development. 1999;70:944–954. doi: 10.1111/1467-8624.00068. [DOI] [PubMed] [Google Scholar]
  10. Bukowski WM, Adams R. Peer relationships and psychopathology: Markers, moderators, mediators, mechanisms, and meanings. Journl of Clinical Child and Adolescent Psychology. 2005;34:3–10. doi: 10.1207/s15374424jccp3401_1. [DOI] [PubMed] [Google Scholar]
  11. Bukowski WM, Hoza B. Popularity and friendship: Issues in theory, measurement, and outcome. In: Berndt TJ, Ladd GW, editors. Peer relations in child development. Wiley; New York: 1989. pp. 15–45. [Google Scholar]
  12. Card NA. ‘I hated her guts!’ Emerging adults’ recollection of the formation, maintenance, and termination of antipathetic relationships during high school. Journal of Adolescent Research. 2007;22:1–26. [Google Scholar]
  13. Card NA, Hodges EVE. Parent-child relationships and enmity with peers: The role of avoidant and preoccupied attachment. In: Hodges EVE, Card NA, editors. Enemies and the darker side of peer relations: New directions for child and adolescent development. Jossey-Bass; San Francisco, CA: 2003. pp. 23–37. [DOI] [PubMed] [Google Scholar]
  14. Card NA, Hodges EVE. Victimization within mutually antipathetic relationships. Social Development. 2007;16:479–496. [Google Scholar]
  15. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. second edition Lawrence Erlbaum Associates; Hillsdale, NJ: 1983. [Google Scholar]
  16. Coie JD, Cillessen AN, Dodge KA, Hubbardt JA, Schwartz D, Lemerise EA, et al. It takes two to fight: A test of relationl factors and a method for assessing aggressive dyads. Developmental Psychology. 1999;35:1179–1188. doi: 10.1037//0012-1649.35.5.1179. [DOI] [PubMed] [Google Scholar]
  17. Coie JD, Dodge KA, Coppotelli H. Dimensions and types of social status: A cross-age perspective. Developmental Psychology. 1982;18:557–571. [Google Scholar]
  18. Dearing E, Hamilton LC. McCartney K, Burchinal MR, Bub KL, editors. Contemporary advances and classic advice for analyzing mediating and moderating variables. Monographs of the Society for Research in Child Development. 2006;71:88–104. doi: 10.1111/j.1540-5834.2006.07103001.x. [DOI] [PubMed] [Google Scholar]
  19. DeLawyer DD, Foster SL. The effects of peer relationships on the functions of interpersonal behaviors of children. Journal of Clinical Child Psychology. 1986;15:127–133. [Google Scholar]
  20. Dodge KA, Lansford JE, Burks VS, Bates JE, Pettit GS, Fontaine R, et al. Peer rejection and social information-processing factors in the development of aggressive behavior problems in children. Child Development. 2003;74:374–393. doi: 10.1111/1467-8624.7402004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dodge KA, Pettit GS. A biopsychosocial model of the development of chronic conduct problems in adolescence. Developmental Psychology. 2003;39:349–371. doi: 10.1037//0012-1649.39.2.349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Egan SK, Monson TC, Perry DG. Social-cognitive influences on change in aggression over time. Developmental Psychology. 1998;34:996–1006. doi: 10.1037//0012-1649.34.5.996. [DOI] [PubMed] [Google Scholar]
  23. Erath SA, Flanagan KS, Bierman KL. Social anxiety and peer relations in early adolescence: Behavioral and cognitive factors. Journal of Abnormal Child Psychology. 2007;35:405–416. doi: 10.1007/s10802-007-9099-2. [DOI] [PubMed] [Google Scholar]
  24. Fabes RA, Eisenberg N, Smith MC. Getting angry at peers: Associations with liking of the provocateur. Child Development. 1996;67:942–956. [PubMed] [Google Scholar]
  25. Hartup WW. Toward understanding mutual antipathies in childhood and adolescence. In: Hodges EV, Card N, editors. Enemies and the darker side of peer relations. New directions for child and adolescent development. Jossey-Bass; San Francisco, CA: 2003. pp. 111–123. [DOI] [PubMed] [Google Scholar]
  26. Hartup WW. The development of aggression: Where do we stand? In: Tremblay RE, Hartup WW, Archer J, editors. Developmental origins of aggression. Guilford; New York: 2005. pp. 3–22. [Google Scholar]
  27. Hodges EVE, Card NA, editors. Enemies and the darker side of peer relations: New directions for child and adolescent development. Jossey-Bass; San Francisco, CA: 2003. [Google Scholar]
  28. Hodges EVE, Malone MJ, Perry DG. Individual risk and social risks as interacting determinants of victimization in the peer group. Developmental Psychology. 1997;33:1032–1039. doi: 10.1037//0012-1649.33.6.1032. [DOI] [PubMed] [Google Scholar]
  29. Hubbard JA, Dodge KA, Cillessen A, Coie JD, Schwartz D. The dyadic nature of social information-processing in boys’ reactive and proactive aggression. Journal of Personality and Social Psychology. 2001;80:268–280. doi: 10.1037/0022-3514.80.2.268. [DOI] [PubMed] [Google Scholar]
  30. Ladd GW. Children’s peer relations and social competence. Yale University Press; New Haven, CT: 2005. [Google Scholar]
  31. Ladd GW. Peer rejection, aggressive or withdrawn behavior, and psychological maladjustment from ages 5 to 12: An examination of four predictive models. Child Development. 2006;77:822–846. doi: 10.1111/j.1467-8624.2006.00905.x. [DOI] [PubMed] [Google Scholar]
  32. Ladd GW, Kochenderfer BJ, Coleman CC. Classroom peer acceptance, friedship, and victimization: Distinct relational systems that contribute uniquely to children’s school adjustment? Child Development. 1997;68:1181–1197. doi: 10.1111/j.1467-8624.1997.tb01993.x. [DOI] [PubMed] [Google Scholar]
  33. Masten AS. Peer relationships and psychopathology in developmental perspective: Reflections on progress and promise. Journal of Clinical Child and Adolescent Psychology. 2005;34:87–92. doi: 10.1207/s15374424jccp3401_8. [DOI] [PubMed] [Google Scholar]
  34. Masten AS, Burt KB, Coatsworth JD. Competence and psychopathology in development. In: Cicchetti D, Cohen DJ, editors. Developmental psychopathology: Vol. 1. Risk, disorder, and adaptation. second edition Wiley; Hoboken, NJ: 2006. pp. 696–738. [Google Scholar]
  35. Nansel TR, Overbeck M, Pilla RS, Ruan WJ, Simon-Morton B, Scheidt P. Bullying behaviors among US youth: Prevalance and association with psychosocial adjustment. Journal of the American Medical Association. 2001;285:2094–2100. doi: 10.1001/jama.285.16.2094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Newcomb A, Bagwell C. Children’s friendship relations: A meta-analytic review. Psychological Bulletin. 1995;117:306–347. [Google Scholar]
  37. Olweus D. Bullying in school: What we know and what we can do. Blackwell; Oxford: 1993. [Google Scholar]
  38. Parker JG, Gamm BK. Describing the dark side of preadolescents’ peer experiences: Four questions (and data) on preadolescents’ enemies. In: Hodges EV, Card N, editors. Enemies and the darker side of peer relations. New directions for child and adolescent development. Jossey-Bass; San Francisco, CA: 2003. pp. 55–72. [DOI] [PubMed] [Google Scholar]
  39. Parker JG, Rubin KH, Erath SA, Wojslawowicz JC, Buskirk AA. Peer relationships, child development, and adjustment: A developmental psychopathology perspective. In: Cicchetti D, Cohen DJ, editors. Developmental psychopathology: Vol. 3. Theory and method. second edition Wiley; Hoboken, NJ: 2006. pp. 419–493. [Google Scholar]
  40. Petti GS, Bates JE, Doge KA. Supportive parenting, ecological context, and children’s adjustment: A seven-year longitudinal study. Child Development. 1997;68:908–923. doi: 10.1111/j.1467-8624.1997.tb01970.x. [DOI] [PubMed] [Google Scholar]
  41. Pope AW. Developmental risk associated with mutual dislike in elementary school children. In: Hodges EV, Card N, editors. Enemies and the darker side of peer relations. New directions for child and adolescent development. Jossey-Bass; San Fracisco, CA: 2003. pp. 89–1100. [DOI] [PubMed] [Google Scholar]
  42. Preacher KJ, Curran PJ, Bauer DJ. Computational tools for probing interaction effects in multiple linear-regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics. 2006;31:437–448. [Google Scholar]
  43. Ray GE, Cohen R. Children’s evaluations of provocations between peers. Aggressive Behavior. 1997;23:417–431. [Google Scholar]
  44. Reid JB, Patterson GR, Snyder J. Antisocial behavior in children and adolescents. American Psychological Association; Washington, DC: 2002. [Google Scholar]
  45. Rodkin PC, Pearl R, Farmer TW, Van Acker R. Enemies in the gendered societies of middle childhood: Prevalence, stability, associations with social status, and aggression. In: Hodges EV, Card N, editors. Enemies and the darker side of peer relations. New directions for child and adolescent development. Jossey-Bass; San Fracisco, CA: 2003. pp. 73–88. [DOI] [PubMed] [Google Scholar]
  46. Schwartz D, Doge KA, Coie JD. The emergence of chronic victimization in boys’ peer groups. Child Development. 1993;64:1755–1772. doi: 10.1111/j.1467-8624.1993.tb04211.x. [DOI] [PubMed] [Google Scholar]
  47. Schwartz D, Hopmeyer-Gorman A, Toblin RL, Abou-ezzeddine T. Mutual antipathies in the peer group as a moderating factor in the association between community violence exposure and psychosocial maladjustment. In: Hodges EV, Card N, editors. Enemies and the draker side of peer relations. New directions for child and adolescent development. Jossey-Bass; San Franciso, CA: 2003. pp. 39–54. [DOI] [PubMed] [Google Scholar]
  48. Shantz CU, Hartup WW. Conflict in child and adolescent development. Cambridge University Press; New York: 1992. [Google Scholar]
  49. Tapper K, Boulton MJ. Victim and peer group responses to different forms of aggression among primary school children. Aggressive Behavior. 2005;31:238–253. [Google Scholar]
  50. Veenstra R, Lindenberg S, Oldehinkel AJ, De Winter AF, Verhulst FC, Ormel J. Bullying and victimization in elementary schools: A comparison of bullies, victims, bully/victims, and uninvolved preadolescents. Developmental Psychology. 2005;41:672–682. doi: 10.1037/0012-1649.41.4.672. [DOI] [PubMed] [Google Scholar]
  51. Vitaro F, Brendgen M, Wanner B. Patterns of affiliation with delinquent friends during late childhood and early adolescence: Correlates and consequences. Social Development. 2005;14:82–108. [Google Scholar]
  52. Witkow MR, Bellmore AD, Nishina A, Juvonen J, Graham S. Mutual antipathies during early adolescence: More than just rejetion. International Journal of Behavioral Development. 2005;29:209–218. [Google Scholar]

RESOURCES