Introduction

Research on peer relations has revealed two characterizations of children’s and adolescents’ moral development in the context of youth peer groups. On the one hand, peer groups form negative influences that encourage antisocial behaviors (Piehler and Dishion 2007). For instance, children and adolescents who play sports as part of groups that hold aggressive norms are more likely to engage in aggression (Malete et al. 2013) and youth who are part of peer groups that hold norms condoning relational aggression are more likely to engage in relational aggression (Brendgen et al. 2013). However, individuals also often reject negative moral norms from groups and resist negative motivations or goals (Killen et al. 2013; Mulvey and Killen 2015). What is not known is how youth and adolescents resolve this tension between loyalty to the peer group and their moral principles in the context of aggressive behavior. When adolescents are part of groups that engage in antisocial aggressive behaviors, talking about these behaviors appears to increase instances of these antisocial behaviors (Piehler and Dishion 2007). This indicates that something like “deviancy training” occurs among adolescents as part of their conversations in their peer groups (Piehler and Dishion 2007). Further, adolescents who engage in bullying behavior expect that their peers also engage in aggression and that they support such aggression (Hinduja and Patchin 2013). Thus, knowing more about the contexts of compliance with or resistance to peer group aggression is important, particularly as children and adolescents are aware of the harmful consequences of physical and relational aggression (Smetana 2006).

Although research has not yet focused on how children evaluate aggressive group norms, research has shown that children and adolescents like peers who challenge group norms that are morally problematic, such as ones that advocate for unequal allocation of resources (Killen et al. 2013), encourage race-based humor (Mulvey et al., in press), and perpetuate gender stereotypes (Mulvey and Killen 2015; Mulvey et al. 2015). Thus, what is not known is how children and adolescents evaluate group norms that condone aggression and whether or not they support challenges to such norms. This was a central aim of the current study. Recently, there has been a call for additional research on peer group processes surrounding aggressive behavior (Ostrov and Kamper 2015), thus this research aims to increase our understanding of peer group dynamics surrounding aggression.

Importantly, group norms involving aggression are multi-faceted. Groups can hold aggression norms involving either physical aggression, such as pushing and shoving, or relational aggression, such as teasing or gossiping. In fact, peer groups that condone and engage in physical or relational aggression may be quite common. Indeed, a recent meta-analysis noted that approximately 35 % of youth are involved in bullying, including physical and relational forms of bullying (Modecki et al. 2014). Less is known, however, about whether children and adolescents view these forms of aggression as equally problematic. For instance, not all forms of aggression actually entail bullying because they may not actually involve power imbalances (Ostrov and Kamper 2015). Moreover, when adolescents are deciding whether to challenge aggressive behaviors, they may weigh multiple factors, including the seriousness of the aggression and the repercussions for challenging different types of aggression. For instance, when deciding how to respond to aggression, adolescents consider how supportive their school, friends and parents will be if they engage in a non-violent reaction, their parents’ values regarding fighting, and their parents’ disciplinary norms (Farrell et al. 2010). Moreover, adolescents also cite their concerns over negative outcomes if they enact a non-violent response and their concerns about their self-efficacy for non-violent responses and how effective or ineffective such responses would be (Farrell et al. 2008). While research with children and adolescents has been limited in this area, research with adults using physical aggression in a sports context and relational aggression in a social context documented no age-related differences, but did document that intention and willingness to apologize did matter when evaluating different types of aggression (Gauché and Mullet 2005). Therefore, new research is needed to understand whether group norms involving physical and relational forms of aggression are evaluated differently by children and adolescents.

Group norms related to physical and relational aggression have been shown to be gender stereotypic, with physical aggression expected to be observed mostly in boys, and relational aggression shown mostly for girls (Crick and Grotpeter 1995). In contrast to stereotypic expectations, current research indicates that actual patterns of aggressive behavior do not reflect gender-associated forms of aggression. For instance, findings suggest that boys engage in both types of aggression (Ostrov and Godleski 2010). Other research indicates that gender differences in aggression are minor, and that there are few gender differences for relational aggression (Card et al. 2008; Lansford et al. 2012). Thus, it appears that associations between gender and particular forms of aggression may be stereotypic and not rooted in actual between-group differences. However, little is known about whether children and adolescents use these stereotypes, or expect their peers to use these stereotypes, in making decisions about accepting or rejecting aggressive group norms or whether they differ in their evaluations of physical and relational aggression norms.

Challenging aggressive group norms may be particularly difficult for children and adolescents. This is in part because children who display certain types of aggressive behavior are perceived as leaders among their peers (even if this is not actually the case) (LaFontana and Cillessen 2002; Parkhurst and Hopmeyer 1998; Rose et al. 2004). Thus, children and adolescents may fear that rejecting aggressive group norms will result in negative consequences for the individual who challenges the norm. For instance, children and adolescents may be concerned that challenging aggression will lead to the aggression being redirected to themselves. Research on bystander interventions indicates that peers rarely intervene when they observe instances of aggression (Salmivalli and Voeten 2004; Unnever and Cornell 2003), perhaps because of fears of retaliation or being victimized (Mulvey et al., in press; Thornberg et al. 2012). For instance, findings suggest that bystanders intervene only 25 % of the time when they observe bullying (Hawkins et al. 2001). On the other hand, prior research documents that children and adolescents are able to differentiate their own individual judgments about deviating from group norms from their expectations for the group’s response (Mulvey et al. 2014b), which suggests that children and adolescents may recognize the importance of challenging the group and intervening. However, this has not yet been tested in relation to aggressive group norms. Thus, what is not yet known is if children support standing up to aggressive behavior when their own group is engaging in the aggression.

Additionally, aggressive group norms can be complex and children and adolescents may struggle with deciding if they are morally unacceptable or not. For instance, having a group norm surrounding physical assertiveness in an athletic context may be viewed as legitimate or even desirable. In contrast, aggressive norms that lead to excessive or unwanted physical harm may be viewed as moral transgressions. Group members have to weigh such norms carefully to determine when assertiveness crosses the line to inappropriate forms of physical harm. Similarly, a group norm condoning sharing news or information about your peers may be viewed as helpful and benevolent. However, peers would probably view this as a moral violation when this information is of a negative nature and is shared in order to cause psychological harm. Thus, group norms about relational harm reflect a continuum from friendly gossiping to the harmful spreading of rumors. While children understand that aggression causes harm, they may also feel pressure to support their group and the norms of the group given the ambiguity that can be reflected in these types of norms.

The current study investigated how children resolve this conflict: would they support someone who challenges peer group aggression? Does it matter whether the group norm involves physical or relational aggression? Understanding whether children and adolescents would support challenges to group aggressive norms is important as research indicates that peers may be able positively influence the behavior of their friends. For instance, research indicates that those who do challenge morally unacceptable behaviors, such as prejudice, can positively influence their close friends’ attitudes and judgments (Paluck 2011). Thus, children and adolescents may impact positive change in their peer group if they do choose to challenge aggressive group norms that have the potential to inflict harm on others. Addressing this topic is particularly important given the pervasive influence of peers (Brechwald and Prinstein 2011), on the one hand, and the positive impact that resisting aggression can have on future instances of aggression (Salmivalli et al. 2011).

Moreover, the current study addresses developmental differences in responding to peer group norms involving aggression, with a specific focus on determining how children and adolescents reason about these types of encounters. It is important to understand the attributions and motivations that children express for why they act or refrain from intervention. This will help to explain the developmental differences previously documented for bystander intervention: compared to children, adolescents are less likely to intervene when they observe aggression (Bellmore et al. 2012; Mulvey et al., in press). The different reasons pertain to preserving group functioning and demonstrating loyalty to their group (Horn 2006; Rutland et al. 2015) as well as expectations that other peers do not want members of “outgroups” to intervene (Hitti and Killen 2015). Further, research has demonstrated that, between the ages of 10 and 14 years, the association between status in one’s peer group and aggressive behavior decreases (Cillessen and Mayeux 2004). Finally, developmental differences in aggression from childhood to adolescence have also been documented, with varying rates of increases and decreases in aggressive behavior over this age period (Olson et al. 2013; Xie et al. 2011). Given that evaluations of aggressive peer group norms may differ developmentally, it is important to examine these evaluations using a sample that includes both children and adolescents.

Current Study

The current study aimed to identify how children and adolescents evaluate challenges to aggressive group norms and their expectations surrounding such challenges. Prior research from developmental subjective group dynamics indicates that children do not support deviant or dissenting members of a group who disagree with or act against a group’s norms (Abrams and Rutland 2008). However, recent research from social reasoning developmental approaches indicates that children show strong support for members who dissent from unfair or morally unacceptable group norms (Killen et al. 2013; Mulvey and Killen 2015; Mulvey et al., in press). Less is known, however, about challenging aggressive peer norms, or the relationship between children’s moral judgments and their expectations regarding challenging the peer group. Challenging an aggressive group norm may prove more difficult for children than challenging other types of group norms. This is because children and adolescents may be concerned that standing up to the group may lead the group to focus the aggression on the member who resisted the group norm. Thus, concerns about repercussions, including social exclusion (Hitti et al. 2014; Mulvey and Killen 2015), may prove to be a barrier that leads to inaction on the part of group members who do not agree with aggressive group norms.

Further, it is possible that stereotypes may also prove to be a barrier in resisting peer group norms surrounding aggression. As reviewed above, there have been mixed findings regarding aggressive behaviors and gender, with early research indicating that boys engage in more physical aggression, while girls engage in more relational aggression (Crick and Grotpeter 1995) and more recent research indicating that these patterns may simply reflect gender stereotypes and that boys engage in more of both types of aggression, and girls engage in more relational aggression (Ostrov and Godleski 2010). In order to account for possible differences, gender of the group was varied in the current study, with some participants hearing about gender stereotypic forms of aggression (for instance, a girls’ group that engages in relational aggression) while others heard about gender non-stereotypic forms of aggression (for instance, a girls’ group that engages in physical aggression). In each case, participants heard about a single member of the group who wanted to challenge the norm. For example, if the group norm was to gossip (relational aggression), a single member wished for the group to remain impartial. If the group norm was to push others during a soccer game (physical aggression), a single member wished for the group to play nicely.

In the current study, participants’ moral judgments surrounding acts of resistance to aggressive group norms and their expectations for group judgments of such acts were measured. Participants included 4th graders (9–10 year olds) and 8th graders (13–14 year olds), as prior research indicates that children and adolescents often evaluate peer group dynamics in different ways (Killen et al. 2013; Mulvey and Killen 2015; Mulvey et al., in press). Further, participants’ social reasoning about their judgments and expectations was also assessed, drawing on social reasoning developmental theory (Rutland et al. 2010). Measuring reasoning is particularly important, as it helps to identify why children and adolescents make the decisions they do and what is driving such decisions.

It was expected that participants would positively evaluate resisting group norms involving physical and relational aggression and that they would expect groups to have much more negative evaluations of such challenges. Prior research has shown that, with age, individuals recognize that groups would not support challenges to their norms, even if participants individually evaluate such challenges positively (Mulvey et al. 2014b). Further, it was expected that participants would differentiate between challenging physical and relational aggression. It was expected that children and adolescents would be more likely to support challenging relational aggression than physical aggression because challenging physical aggression could incur more personal cost (the risk of redirecting the aggression towards the challenger) than would relational aggression. It was an open question of whether there would be differences based on the gender of the participant and gender of the target group. On the one hand, there are strong stereotypes associated with different types of aggression for boys and girls (Crick and Grotpeter 1995), and children show strong gender ingroup preferences (Leman and Lam 2008; Mulvey et al. 2014a). However, prior research on deviating from morally relevant group norms when the groups are defined by gender found that children and adolescents focused on the moral nature of the norm and not the gender of the group when making evaluations (Killen et al. 2013). Therefore, we tested for differences based on the gender of the participant and the gender of the group, but did not have specific expectations for our findings.

Central to this study, it was expected that different forms of reasoning would be used by participants who support challenges to group norms and those who reject such challenges. Specifically, it was expected that participants who thought that groups should challenge group norms would use reasons based on others’ welfare and autonomy, while those who believed that groups would not support such challenges would focus on group functioning and gender stereotypes. With age, it was expected that adolescents would judge resistance to group norms less positively than would children based on the high salience of group loyalty in adolescence (Horn 2003).

Methods

Participants (N = 292) included 90 9–10 year olds (M = 9.63 SD = 2.99, Range 9.40–11.61 years, N females  = 52), and 202 13–14 year olds (M = 13.95 SD = .43, Range 13.05–15.88 years, N females  = 100) from public elementary and middle schools in the Mid-Atlantic region. Participants were approximately evenly divided by gender (52.4 % female), and were ethnically representative of the United States (school demographic information identified approximately 30 % ethnic minority students in the schools). Further, school demographic information indicates that participants were from low to low-middle income schools. All students in 4th grade and 8th grade were invited to participate. Only students receiving parental consent (9–10 year olds), and providing student assent (all participants) completed the tasks. Parental consent return rates across the schools averaged 65–70 %.

Procedure

The tasks were administered by a trained researcher in a quiet room at each school. Participants were told that there are no right or wrong answers and that all responses are anonymous and confidential. Additionally, participants were told that their participation is voluntary and that they may choose to stop the assessment at any time. Participants were also given a warm-up task, which involved practicing using the Likert scale to be used in the survey. For 9–10 year old participants, the survey was read aloud by a trained researcher to small groups (3–4 participants) of participants of the same gender. For 13–14 year old participants, the survey was administered by a trained researcher to larger groups (25–30 participants). The necessity to read the survey aloud to the younger participants accounts for the difference in sample size between the 9–10 year old participants and the 13–14 year old participants. For both age groups, participants recorded their answers. Any questions the participants had were answered by the researcher. The survey took about 40 minutes to complete.

Measures

Participants completed the Challenging Aggression Task that was modified from Killen et al. (2013). The scenarios included were pilot-tested, with findings suggesting that both scenarios were viewed as wrong (and equally wrong). The Challenging Aggression Task consisted of hypothetical scenarios where groups held either norms of physical aggression (PA) or relational aggression (RA). In each scenario, participants were told about two groups, their own group and an outgroup, and the norm for each group. Each child heard two scenarios and there were two versions, which varied according to whether the group norm was stereotype-consistent or inconsistent (see Table 1). For instance, in the stereotype-consistent version, participants heard about a girls’ group that engaged in relational aggression and a boys’ group that engaged in physical aggression. In the stereotype-inconsistent version, participants heard about a girls’ group that engaged in physical aggression and a boys’ group that engaged in relational aggression. Approximately equal numbers of male and female participants from each age group completed each version. Participants were introduced to their group and the other group (i.e., “This group of girls is your group of friends.”) and completed three items to enhance their identification with their group (choosing a group color, symbol and group activity). Females received surveys with the girls’ groups labeled as “your group” and the boys’ groups labeled as “their group.” Males received surveys with the boys’ groups labeled as “your group” and the girls’ groups labeled as “their group.”

Table 1 Study design
Full size table

An example of the relational aggression scenario received by a male participant is:

Let’s say there are some kids at school who always sit alone, and act differently from the other kids. Their group, the girls’ group, says: ‘It’s okay to gossip all the time, because we don’t know why those kids act that way.’ Betsy, who is also in this group, wants to be different from the other members of their group. She thinks: ‘Even though you don’t know why those kids act that way, you shouldn’t always gossip about those kids.’

In this scenario about gossiping (relational aggression), the girls’ group supports gossiping and the deviant in-group member rejects it and prefers the group to be impartial.

An example of the physical aggression scenario received by a male participant is:

When playing soccer, your group, the boys’ group, says ‘It’s okay to push and shove just to score points.’ Michael, who is also in this group, wants to be different from the other members of his group. He thinks ‘Even though you want to score points, you shouldn’t always push and shove.’

In this scenario about pushing (physical harm), the boys’ group supports pushing and the deviant in-group member rejects it and prefers the group to be non-aggressive.

Dependent Measures

For each scenario, the same assessments were given. Participants were told that the dissenting member chose to tell the group his/her thoughts. Participants then rated: (1) Group judgment: How okay or not okay will they [the group] think what she says is? This assessment was measured with a Likert-type scale with ratings from 1 = really not okay to 6 = really okay. This measure assessed how acceptable participants think the group will judge someone who disagrees with the group to be. They also assessed (2) Group reasoning: Why? This was an open-ended social reasoning assessment.

Finally, they assessed: (3) Individual judgment: When you hear her/him, how okay or not okay do you think what she/he says is? This item was measured with a Likert-type scale from 1 = really not okay to 6 = really okay.

Coding Categories for Justifications

Participant responses to the open-ended reasoning question were coded using a coding system that was established drawing on prior research (Killen et al. 2013) and based on pilot-testing. The coding system included three categories based on social domain theory (Smetana 2006): Moral, societal and psychological. The top four codes used by participants were included in analyses. The subcategory for moral was others’ welfare (e.g., “That could really hurt someone’s feelings.” or “They will get hurt if you push.”). The subcategories for societal were: group functioning (e.g., “Groups work better when everyone agrees.” or “They all want to play one way. He needs to go along with them.”), and gender stereotypes (e.g., “Girls don’t usually push and hit.” or “Boys always play rough.”). In the psychological domain, participants cited autonomy (e.g., “It is up to her. If she doesn’t agree with them, she can tell them.” or “It is important to say what you believe.”). Justifications were coded as 1.0 = full use of the category; 0.5 = partial use; 0.0 = no use of the category and analyses were conducted on proportional usage. Partial use of a category indicated that participants referred to 2 different codes, so each code was weighted at 0.5. Less than 5 % of participants used more than one code and less than 10 % either did not provide reasoning or provided un-codable reasoning (for instance, just saying the group would think it was okay, but not providing a reason). Surveys (25 %) were coded by 3 trained research assistants and inter-rater reliability was high, Cohen’s κ = .92.

Results

The data were analyzed using repeated measures Analysis of Variance (ANOVA). Follow up tests were conducted using the Bonferroni Correction to control for Type I errors. Justifications were proportions of responses for each respective coding category, with the top four justifications analyzed for each question.

Analyses for reasoning were conducted using repeated measures ANOVAs, as reasoning was coded for no (0), partial (0.5) and full (1.0) use of the codes. Because of the nature of these data, which are proportional, but include many empty cells given that participants can choose to use or not use any of the forms of reasoning, it is effectively analyzed using ANOVAs because ANOVAs are robust to the problem of empty cells (see Posada and Wainryb 2008, for a fuller explanation and justification of this data analytic approach). Further, a review of analytic procedures for these types of data (covering 10 years in APA psychology journals) indicated that linear models with repeated procedures (particularly ANOVA) are appropriate compared to log-linear analysis for this type of within-subjects design (see Wainryb et al. 2001, footnote 4).

Physical and Relational Aggression: Individual and Group Judgments

In order to examine differences by type of judgment (physical and relational aggression and both individual and group judgments), age group, gender of the participant and gender of the group, a 2 (Age Group: 4th, 8th) × 2 (Participant Gender: Male, Female) × 2 (Group: Girls, Boys) × 4 (Type of judgment: Individual—PA, Group—PA, Individual—RA, Group—RA) ANOVA with repeated measures on the last factor was conducted. There was an overall effect for judgment, F(3,843) = 110.21, p < .001, ω 2p  = .278. Follow-up tests revealed that participants individually rated challenges to relational aggression (M = 4.98, SD = 1.45), as more okay than challenges to physical aggression (M = 4.59, SD = 1.64), t(298) = 3.92, p < .001, d = .25. There were no differences for the group judgments: expectations for challenges to group physical (M = 3.46, SD = 1.57) and relational aggression norms (M = 3.27, SD = 1.58) did not differ (p = .887). As expected, however, for both physical and relational aggression, participants were individually more supportive of challenges to the aggressive norms than they expected groups to be (PA: t(289) = 9.67, p < .001, d = 1.14, RA: t(291) = 15.43, p < .001, d = 1.80).

Additionally, there was a judgment (Individual—PA, Group—PA, Individual—RA, Group—RA) by age group interaction, F(3,843) = 8.21, p = .007, ω 2p  = .025. Pairwise comparisons revealed that children were individually more positive about someone challenging a group’s aggression than were adolescents for both the physical aggression condition (M children  = 4.92, SD children  = 1.58, M adolescents  = 4.44, SD adolescents  = 1.64), t(298) = 2.25, p = .025, d = .29, and the relational aggression condition (M children  = 5.33, SD children  = 1.31, M adolescents  = 4.82, SD adolescents  = 1.54), t(298) = 2.93, p = .004, d = .36. However, children and adolescents did not differ on their expectations for the group’s moral judgments for either physical (M children  = 3.29, SD children  = 1.50, M adolescents  = 3.53, SD adolescents  = 1.60, p = .315) or relational aggression (M children  = 3.26, SD children  = 1.47, M adolescents  = 3.27, SD adolescents  = 1.63, p = .991).

There was also a judgment (Individual—PA, Group—PA, Individual—RA, Group—RA) by gender effect, F(3,843) = 4.11, p = .008, ω 2p  = .011, which revealed that female participants (M = 4.91, SD = 1.47) were more supportive of challenges to physical aggression than were male participants (M = 4.23, SD = 1.73), t(298) = 2.50, p = .013, d = .42. There were no differences between male and female participants on any of the other questions. Finally, there was a judgment by group effect, F(3,843) = 5.43, p = .001, ω 2p  = .015. This revealed that participants expected that girls’ groups (M = 2.98, SD = 1.58) would respond more negatively to challenges rejecting the group’s relational aggression norm than would boys’ groups (M = 3.54, SD = 1.53), t(298) = 2.85, p = .005, d = .36.

Reasoning: Group Judgment for Physical Aggression

In order to test the hypothesis that participants would use different forms of reasoning if they expected that groups would perceive challenging physical aggression as okay than if they expected the groups would think such a challenge was not okay, a 2 (Age group: 4th, 8th) × 2 (Group Judgment: Okay, Not Okay) × 4 (Reasoning: Group functioning, Others’ Welfare, Autonomy, Gender Stereotypes) ANOVA was conducted, with repeated measures on the last factor. There was a main effect for reasoning, F(3, 861) = 60.77, p < .001, ω 2p  = .17, which revealed that, overall, participants made much greater reference to group functioning than to autonomy [t(291) = 10.98, p < .001, d = 1.01], stereotypes [t(291) = 9.62, p < .001, d = .92] and welfare [t(291) = 8.28, p < .001, d = .80], and that they made more reference to autonomy than to welfare [t(291) = 2.50, p = .013, d = .20], see Table 2. There was also an interaction between type of reasoning and moral judgment, F(3, 861) = 14.57, p < .001, ω 2p  = .044. Pairwise comparisons revealed that participants who thought groups would support a member who challenged the group’s physical aggression were much more likely to reference autonomy [t(291) = 3.72, p < .001, d = .46] and other’s welfare [t(298) = 42.78, p = .006, d = .24] and less likely to reference group functioning [t(298) = 4.51, p < .001, d = .46] or gender stereotypes [t(298) = 2.39, p = .016, d = .34] than those who thought the group would reject the challenge.

Table 2 Proportional use of reasoning: group judgment, physical aggression
Full size table

Reasoning: Group Judgment for Relational Aggression

In order to test the hypothesis that participants would use different forms of reasoning if they expected groups to judge challenging relational aggression as okay or not okay, a 2 (Age group: 4th, 8th) × 2 (Group Judgment: Okay, Not Okay) × 4 (Reasoning: Group functioning, Others’ Welfare, Autonomy, Gender Stereotypes) ANOVA was conducted, with repeated measures on the last factor. There was a main effect for reasoning, F(3,864) = 40.86, p < .001, ω 2p  = .12, which revealed that overall, participants made much greater reference to group functioning than to autonomy [t(291) = 2.73, p = .007, d = .94], welfare [t(291) = 9.73, p < .001, d = .69] or gender stereotypes [t(298) = 5.14, p < .001, d = .84] and made more references to welfare than autonomy [t(291) = 2.68, p = .008, d = .31], see Table 3. There was also an interaction between type of reasoning and moral judgment, F(3,864) = 20.70, p < .001, ω 2p  = .064. Pairwise comparisons revealed that participants who thought groups would not support a member who challenged their relational aggression norm were much more likely to reference group functioning [t(291) = 5.91, p < .001, d = .59] and less likely to reference autonomy [t(291) = 3.94, p < .001, d = .44] or welfare [t(291) = 3.32, p < .001, d = .37] than those who asserted that groups would think challenging relational aggression was okay. There were no differences in use of gender stereotypes between participants who thought groups would or would not support challenging the relational aggression norm.

Table 3 Proportional use of reasoning: group judgment, relational aggression
Full size table

Discussion

Youth experience high rates of aggression and bullying in peer contexts and victims of peer aggression experience serious negative outcomes including school truancy, academic failure, depression, anxiety, and self-inflicted harm (CDC 2012). While extensive research has documented factors associated with bullying and victimization (Ostrov and Kamper 2015; Rappaport and Thomas 2004), little research has focused on social-cognition and reasoning about challenging aggressive peer group norms. The current study reveals that peers do recognize the importance of challenging aggressive group norms, but also understand the pressures to conform to group norms and demonstrate loyalty to the group. In short, the findings reveal the complexity of children’s and adolescents’ evaluations of aggressive peer group norms.

Importantly, the novel findings were that youth and adolescents differed in their evaluations of physical and relational aggression. Participants were more positive about challenges to relational aggression than about challenges to physical aggression and the effect size for this finding was quite large. This suggests that they recognize the very harmful nature of relational aggression and may conceptualize this type of psychological harm as more problematic than physical aggression in a sports context. Furthermore, these findings suggest that children and adolescents perceive challenging relational aggression as important: perhaps they are more likely to expect that they will effect change by challenging relational aggression than they would when challenging physical aggression. This is interesting because relational aggression is more psychological and less overt than physical harm.

There are other possible explanations for the differences between expectations of challenging physical aggression and relational aggression. On the one hand, children and adolescents may have perceived physical aggression as more complex. Aggression is frequently condoned in sports contexts, but the damage incurred by physical aggression during sports is also irreversible. It is possible that children may have been more conflicted when weighing challenges to physical aggression because of the multi-faceted nature of this form of aggression. This is not to say that relational aggression is not also complex, as sharing information about others is also frequently condoned by society and drawing the line between sharing information and relational aggression is not always easy. However, this does suggest that future research should test a range of types of both relational and physical aggression in order to better understand which types of aggression, exactly, are particularly difficult to challenge.

An additional reason why participants may have expected physical aggression to be challenged less than relational aggression may involve concerns over potential repercussions of challenging physical aggression. They may have been concerned that peers who engage in physical aggression might, in turn, redirect the aggression and inflict physical harm on those who stand up to the group norm. This is in concert with prior research that suggests that concerns over victimization may lead adolescents to inaction (Thornberg et al. 2012). Prior research has shown that children and adolescents believe that challenging group norms may lead to exclusion (Hitti et al. 2014; Mulvey and Killen 2015; Mulvey et al., in press); however, the previous research did not examine challenges to group norms pertaining to aggressive behavior. It may be that participants expected that those who challenge physical aggression would actually become the targets of that physical aggression. Future research should examine whether children and adolescents expect that challenging different forms of aggression leads to negative outcomes, such as becoming a victim, for the challenger.

Further, it is important to examine whether children and adolescents expect different reactions (including repercussions) to challenging different types of group norms. This is especially important to understand as research documents that bystanders play a very important role in ending peer aggression and bullying (Abbott and Cameron 2014; Palmer et al. 2015; Salmivalli et al. 2011). However, intervening when you are a member of the group that holds a norm condoning the aggressive behavior may, actually, be a more difficult form of intervention than intervening when you are a true bystander who is not affiliated with those who are engaging in the aggressive behavior. Thus, more work is needed that examines repercussions for challenging group norms and which focuses on repercussions for members of those groups as well as bystanders.

Interestingly, analyses revealed differences between evaluations for girls’ versus boys’ groups and between male and female participants. Although the effect sizes for these findings were generally small, they were similar to effect sizes documented in other social domain research that examines gender differences (Killen et al. 2002; Killen and Stangor 2001; Mulvey and Killen 2015). Specifically, female participants were more supportive of challenges to physical aggression than were male participants, in line with stereotypes which suggest that girls are less likely to engage in physical aggression (Crick and Grotpeter 1995). This is also in concert with findings on aggression in sports-contexts, which revealed that females rated sports-related aggression as less acceptable than did males (Gardner and Janelle 2002). It may also be that female participants were more supportive of this challenge because of underlying stereotypes about sports in general. While both boys and girls play soccer at equivalent rates in the United States (U.S. Youth Soccer 2015), some research does suggest there are still stereotypes that suggest that boys are more capable in sports in general than are girls (Rowley et al. 2007). Therefore, future research should try to untangle the impact of different gender stereotypes by testing physical aggression scenarios in different contexts.

Participants also expected that girls’ groups would respond more negatively to challenges rejecting the group’s relational aggression norm than would boys’ groups. Together these findings suggest that gender stereotypes may exert an indirect influence on peer group dynamics involving aggressive group norms. While recent research has indicated that there may not be large differences between the aggressive behaviors of girls and boys (Card et al. 2008; Lansford et al. 2012; Ostrov and Godleski 2010), the present results suggest that these underlying stereotypes may still influence the decisions and evaluations that children and adolescents make about both physical and relational aggression. The results also indicate that the influences may be somewhat implicit: while some participants did use explicit gender stereotypes about aggression in reasoning about their evaluations, the rates of explicit stereotype use were quite low. As reflected in their moral judgments and reasoning, participants were primarily focused on issues of harm, autonomy and group functioning and showed strong support for challenging aggressive behaviors in both the physical and relational aggression conditions.

These results reveal the nuanced manner in which group norms function. Findings from developmental subjective group dynamics indicate that individuals dislike group members who are disloyal and deviate from group norms (Abrams and Rutland 2008). However, children and adolescents support group members who challenge group norms that violate expectations about unfair allocation of resources (Killen et al. 2013). The current study extends these findings to a new moral context: aggression. Importantly, the results indicate that children and adolescents recognize that while groups have norms that help define the group, there are contexts in which deviating from the norms is legitimate and acceptable. Results for both physical and relational aggression revealed that participants were likely to judge resistance to group aggression as acceptable, even though they also recognized that groups would not agree. These findings support prior research that has documented that children and adolescents are quite capable of distinguishing between group and individual perspectives on group norms and adherence to those norms (Mulvey et al. 2014b). However, the current findings demonstrate that children and adolescents can distinguish individual and group perspectives even when the norm involves peer group aggression.

Intervention research has examined whether individuals who are trained to challenge harmful peer behaviors, specifically intergroup prejudice, can influence their peers to also reject such behaviors (Paluck 2011). This research found that individuals who stand up for socially-just actions do play an important role, especially in influencing the attitudes and behavior of their close friends (Paluck 2011). Thus, peers who do challenge aggressive peer group norms are likely to make a difference. The results of the current study suggest that children and adolescents do want to challenge aggressive norms and that they expect that they will do so. However, findings did reveal a decline (with a small to medium effect size), with age, in support for challenging aggressive norms.

It may be that, as children enter adolescence, the social pressure of the peer group becomes stronger, resulting in adolescents adhering more rigidly to their peer group norms (Brechwald and Prinstein 2011). What these age-related differences suggest, however, is that future research should identify individual factors related to challenging harmful peer behavior in order to better understand how parents, educators and group leaders can encourage adolescents to stand up to aggressive and harmful peer norms. These findings are supported by the reasoning results, which suggest that judgments that challenging peer group norms is okay are driven by concern for others’ welfare and a sense of autonomy, while judgments that challenging such norms is not okay are driven by an adherence to group functioning and, in the context of physical aggression, by gender stereotypes.

The current study provides new information about how children and adolescents evaluate peer group norms regarding aggression, but it does have limitations. First, the sample was relatively homogenous and yet responses to aggression may differ based on factors such as demographics, social-emotional adjustment, exposure to violence, and school connectedness as proposed by the social-ecological theory (Espelage 2014). Therefore, future research should aim to test for differences among samples that differ on these key factors. Additionally, the current study was experimental in design, and thus tested two types of aggression in a controlled situation that did not allow for consideration of additional contextual variables, such as prior history of aggression, power differentials or status differences among group members, etc. While this is, in some ways, a strength of the study, as it allowed us to eliminate the influence of extraneous variables, it is also a limitation as aggression in peer contexts is necessarily complex and multifaceted (Ostrov and Kamper 2015).

Related to the importance of measuring multiple variables, future research should also aim to test a range of aggression contexts with attention to different stereotypes that may be relevant in those different contexts in order to isolate the impact of gender stereotypes on evaluations of aggression. For instance, it would be helpful to clarify how stereotypes about the type of aggression may interact with stereotypes about the context within which the aggression occurs. Additionally, as children and adolescents can be part of different types of groups, future research should examine whether differences emerge based on the type of group presented (for instance, organized teams, school-sanctioned groups, afterschool clubs or groups of friends). In the present study, we presented the groups as groups of friends, but it is possible that participants may have evaluated the soccer context as involving an organized team. Testing for differences based on type of group would be helpful in clarifying how children and adolescents respond to group norms involving aggression. Finally, the current study only examined 4th and 8th graders. Future research should aim to capture developmental differences across a broader age-range in order to most accurately understand developmental patterns in social-cognition surrounding aggression in childhood and adolescence.

Despite these limitations, the findings have important policy implications for designing and implementing bullying prevention programs. First, the findings suggest that such programs should aim to focus on training youth to intervene when their group engages in aggressive behavior. This is because the results revealed that children and adolescents do recognize these aggressive behaviors as wrong and positively evaluate their peers who do want to stand up to aggression. As the bystander intervention literature demonstrates, peers who challenge bullying behavior can be very effective in ending aggressive behavior (Nansel et al. 2001; Salmivalli et al. 2011). However, other findings also indicate that peers may be wary of intervening because of concerns about possible repercussions for intervening (Mulvey and Killen 2015; Mulvey et al., in press; Thornberg et al. 2012).

Thus, it is important that bullying interventions attend to helping youth feel comfortable challenging and intervening when they observe aggression. Bullying prevention programs, especially in the United States, have not always been highly effective (Evans et al. 2014; Merrell et al. 2008). It may be important to look in new directions when developing programs and policies. The current findings suggested that empowering youth to challenge their friends may prove beneficial in reducing aggression. Research on the KiVa antibullying program has demonstrated that programs focused on changing responses when children and adolescents observe bullying have the potential to be very effective in reducing prevalence of bullying (Yang and Salmivalli 2015). Further, interventions aimed at changing school norms surrounding aggression also show promising results (Perkins et al. 2011). Thus, future intervention efforts should continue to focus on these elements of the peer social context. Finally, the results suggested that youth perceive physical and relational aggression differently, and thus, future interventions should attend to these differences when helping children and adolescents feel comfortable in challenging different types of aggressive behaviors in their peer group.

Conclusion

This study documented that children and adolescents do not evaluate responses to group norms of all types of aggression in the same way. Participants were more supportive of challenges to relational aggression than physical aggression and subtle gender differences emerged, which suggest the continued influence of gender stereotypes about aggression on children’s and adolescents’ evaluations. These findings reveal that, even though research has demonstrated that there may be only minor differences in gender-associated aggression (Lansford et al. 2012; Ostrov and Godleski 2010), stereotypes about gender and aggression do influence how children and adolescents expect to respond to peer-group aggression. Further, the current findings point to a sophistication in youth social-cognition about group norms: participants understood that groups may show loyalty to their group norm, even when that norm supported aggression and even if the individual participant recognized the aggression as violating a moral principle.

These findings suggest that intervention programs, such as the KiVa program (Yang and Salmivalli 2015), should continue to attend to the important role of the peer group and social group dynamics (Espelage et al. 2007) in identifying ways to support youth as they navigate aggression in peer groups. Moreover, participants’ reasoning reflected their understanding of the multifaceted nature of these decisions: they referenced multiple forms of reasoning including group functioning, others’ welfare, autonomy and gender stereotypes. In sum, the present study revealed that children and adolescents approach group norms about aggression with care: they recognize the problematic moral nature of these norms, but also understand the powerful influence of the group in perpetuating such norms.