The Role of Defending Norms in Victims’ Classroom Climate Perceptions and Psychosocial Maladjustment in Secondary School

The healthy context paradox (Huitsing et al. 2019) states that strong defending descriptive norms work adversely for victims for three reasons. First, being victimized in an otherwise positive classroom may negatively impact victims’ causal attributions of their situation. Victims make attributions to explain why they are victimized (Graham and Juvonen 1998). These attributions include locus (whether the cause of is internal or external to the victim) and controllability (whether the cause can be changed). In classrooms where defending behaviors are widespread, it is likely that victimization is uncommon (Saarento et al. 2015) and being a victim of bullying is thus not normative. Consequently, victims are more likely to blame themselves (i.e., make an internal attribution) as they are the only one or one of the few who have this problem. They may also conclude that nothing can be done about it (uncontrollable), because the many defending efforts in the classrooms did not to stop the perpetrators (Kaufman et al. 2020). Victims who make internal and uncontrollable attributions have more psychosocial problems than victims who do not make these attributions (Schacter et al. 2015). Thus, defending descriptive norms may relate to more negative classroom climate perceptions and psychosocial maladjustment in victims.

Second, social comparison theory (Festinger 1954) states that people generally compare their situation with that of others to define their self-worth. In classrooms with strong defending descriptive norms, victims mainly have “upward social comparison material”: most peers are better off. These peers are seen to be able to defend victims against bullying and to have the resources to offer help (e.g., being confident about oneself, or being surrounded by supportive friends who can shield from a revenge of the bully; Hawley 2014). Having many classmates with such resources yields more opportunity for upward comparisons. These comparisons in turn emphasize victims’ own disadvantaged position, making them feel even worse about themselves and their situation in a in generally positive classroom.

Third, victims may perceive defending as less effective in classrooms where it is widespread (a so-called defending-inflation). Defending may be less effective in classrooms with high defending descriptive norms because in such classrooms, all students may try to defend a victim including those who may be unsuitable for it, such as unpopular or other victimized peers. Indeed, being helped by someone who is victimized as well may exacerbate victims’ problems (Guarneri-White et al. 2015; Schacter and Juvonen 2020). Therefore, defending in classrooms with high defending descriptive norms may be inefficient to stop the bully and instead escalate psychosocial problems and negative feelings about oneself and the classroom. Based on these arguments, we hypothesized that defending descriptive norms may amplify the association between victimization and negative classroom climate perceptions and psychosocial maladjustment (Hypothesis 1).

Are popularity norms also subject to the healthy context paradox? Contrasting hypotheses are possible. On the one hand, defending popularity norms – just like descriptive norms – may exacerbate the link between victimization and maladjustment or negative classroom experiences. This may again be due to the negative attributions victims may develop in such classrooms, or because the popular defending behaviors are not sufficiently effective for victims. Moreover, when defending is related to popularity, students may primarily defend to gain status for themselves (self-serving goal) rather than to help the victim (other-oriented goals). Indeed, defending can be motivated by personal gains, such as status gains – in particular when opposing a bully (Pronk et al. 2019). Importantly, the goals that underlie prosocial behavior may affect the quality of the help that is provided. For example, people with self-serving goals seem more likely to give dependency-enhancing help (Halabi et al. 2008; Jackson and Esses 2000). If this occurs in defending situations, victims may become dependent on their defender and may feel inferior and powerless as it may strengthen the idea that they are unable themselves to stand up against the bully. It can thus be hypothesized that defending popularity norms amplify the link between victimization, negative classroom climate perceptions, and psychosocial maladjustment (Hypothesis 2A).

On the other hand, defending popularity norms may not be susceptible to the healthy context paradox and actually decrease the adverse effects of victimization. This could work in two ways. First, as adolescents seek popularity, behaviors associated with popularity are valued (Dijkstra and Gest 2015). These behaviors become reputationally salient: a valuable tool to gain popularity in the peer group (Hartup 1996). Consequently, students approve defending behaviors and are unlikely to stand up against them (Henry et al. 2000; Laninga-Wijnen et al. 2020a). It will be comforting for the defenders (and the victims) to know that their peers have their back when they stand up against a bully, as defending can be risky (Garandeau et al. 2018). When popularity norms for defending are stronger, it is less likely to be disputed and it more effectively signals that people do care about the victim. This will be advantageous for victims’ adjustment and classroom climate perceptions.

There may be a second way in which popularity norms work positively for victims. High defending popularity norms imply that popular peers engage in defending and unpopular peers do not. Being defended by an unpopular (or) victimized peers may not be effective or exacerbate problems (Guarneri-White et al. 2015; Schacter and Juvonen 2020). But popular defenders have the social power and impact to deter bullies. This power may be indispensable, as bullies are often popular themselves (Garandeau et al. 2018) and look down on lower-status peers (Van Kleef et al. 2008). Popular defenders may decrease the status of bullies by standing up against them (Laninga-Wijnen et al. 2020a). Consequently, in classrooms with high defending popularity norms, victims may develop a cognitive anticipation and trust that the bullying will cease over time. Based on this reasoning, a contrasting hypothesis is that defending popularity norms will not be subject to the healthy context paradox but instead work out positively for victims’ classroom climate perceptions and adjustment (Hypothesis 2B).

The aim of this study was to clarify the role of descriptive and popularity norms for defending in the psychosocial adjustment and classroom climate perceptions of victimized youths. In line with the healthy context paradox, we expected that victims in classrooms with high descriptive norms for defending are worse off than victims in classrooms with low descriptive norms for defending (Hypothesis 1). We had two contrasting hypotheses for popularity norms. On the one hand, popularity norms may be susceptible to the healthy context paradox in similar ways as descriptive norms, hence enhance the plight of victims (Hypothesis 2A). On the other hand, victims may be better off in classrooms with high popularity norms than in classrooms with low popularity norms (Hypothesis 2B), because popular defenders may have the power to deter bullies and because defending is less likely to be disputed by others.

We tested our hypotheses for defending norms while controlling for classroom-levels of victimization (Dijkstra and Gest 2015). In this way, we extended prior work by disentangling the absence of negative aspects (e.g., low victimization) from the presence of positive aspects (defending norms). Based on prior work (Huitsing et al. 2012), we expected victims to do better in classrooms with higher average levels of victimization (Hypothesis 3).

Classroom-level factors may not operate in isolation to affect victims’ classroom climate perceptions and adjustment. Strong defending descriptive norms may be particularly harmful for victims when classroom-levels of victimization are low. At the same time, it may help victims to be in classrooms where defending is effectively shown by popular peers (popularity norm) and where they are not the only victim. Furthermore, a context characterized by both descriptive and popularity norms for defending may work out positively, because the defending is widespread and associated with popularity. Being in a classroom without any norm of defending may be indicative of no effort to stop the bullying or to make a victim feel better, which may work out negatively for victims’ classroom climate perceptions and adjustment. Therefore, in additional analyses, the interactions between classroom levels of victimization and norms were explored (i.e., descriptive or popularity norms).

We conducted our study in a sample of secondary school students, which adds to the literature in several ways. Most prior work on the role of classroom factors in victims’ adjustment focused on elementary school students. In the Netherlands, the classroom also constitutes an important context for secondary students, given that they spend most or even all of their time within the same classroom peers across an entire school year. Examining classroom norms, victimization, and adjustment in secondary schools clarifies whether the theoretical rationale for younger ages also applies at later ages. Victimization may occur more indirectly and may be less visible for teachers in secondary school (students have up to 15 teachers for varying subjects); this may enhance the adverse impact of victimization and therefore leaves less room for contextual factors to buffer against this. Moreover, compared to children, adolescents attach increasing value to peer popularity (Li and Wright 2014), hence, parsing out the relative impact of popularity norms versus descriptive norms is key to understand the role of the broader classroom context in this developmental period.

We conducted multi-level regression analyses in Mplus Version 8 (Muthén and Muthén 2016) to take the nested structure of the data into account. Classroom-level variables were grand-mean centered; individual-level variables were centered at the classroom mean. We used the MLR-estimator (Yuan and Bentler 2000) to account for the potential non-normal distribution of the residuals. We tested one model for each outcome separately (cohesion, cooperation, isolation, conflict; belonging, loneliness, self-esteem), as including all outcomes simultaneously resulted in too many parameters and model nonconvergence.

The analyses included four steps. First, we tested empty models and examined intraclass correlations (ICC). Second, we included the individual- and classroom-level predictors to explain variance in the outcome variable at these two levels (Main models, Model A.1, Appendix 1). These main models were used to assess the role of individual-level and classroom-level variables on our outcome variables of interest. Third, we included the random slope for the association between individual victimization and the outcome variable, to see whether there were statistically significant differences between classrooms in the effect of students’ victimization (Model B.1). Fourth, we simultaneously included three two-way cross-level interactions (i.e., defending popularity norm*individual victimization, defending descriptive norm*individual victimization, and classroom-level victimization*individual victimization) to test whether the variability in the association between students’ victimization and classroom climate perceptions or school adjustment was explained by the defending descriptive norms, popularity norms, and average levels of victimization in classrooms (Model C.1). We included these three two-way cross-level interactions simultaneously, to account for potential confounding and to parse out the relative effects of descriptive norms, defending norms, and victimization levels. We interpreted cross-level interactions only if 1) models containing these interactions (Models C.1) had a better fit (lower AIC) compared to Model A.1; and 2) at least one of the cross-level interactions was significant and predicted a relatively large part of the variance in the random slope (≥ 20%; Cohen 1988). We conducted simple slopes analyses with the Preacher and Hayes method for multi-level analyses (Preacher et al. 2006). A score of -1 was entered in the formula to indicate low victimization on the standardized victimization scale (thus referring to those scoring 1 SD below the mean of victimization). We also calculated coefficients for students who were moderate (0), high (1), and very high (2) on victimization. For descriptive norms, a score of -1 was entered in the formula to represent low norms (scoring 1 SD below standardized descriptive norms), whereas moderate norms received a '0’ and high norms received a ‘1’. For popularity norms, the value of ‘-0.6’ was used to refer to low norms, whereas values of '0’ and ‘0.6’ was used to refer to moderate and high norms, respectively (based on 95% interval of popularity norm scores).

Recent work has shown that the adjustment difficulties of victimized youth may be exacerbated in healthier classrooms (e.g., classrooms with low victimization rates; Huitsing et al. 2012; or low aggressive descriptive norms; Morrow et al. 2019). The current study examined whether the healthy context paradox phenomenon also emerges when classrooms are characterized by positive aspects that are specifically aimed at making victims to feel better (Cohen and Wills 1985), namely the defending norms. Moreover, we examined whether descriptive and popularity norms would be equally susceptible to this paradox. Findings for descriptive norms were mostly in line with the healthy context paradox: adolescents who experienced higher levels of victimization were worse off in classrooms with stronger defending descriptive norms, as they experienced less cooperation, cohesion, and feelings of belonging. This was particularly true for students who were severely victimized. Defending popularity norms did not seem susceptible for the healthy context paradox. Stronger defending popularity norms related to more positive classroom climate perceptions and higher feelings of belonging in general. The impact of defending popularity norms was also true for youth reporting higher levels of victimization (i.e., there were no random slopes of victimization and no cross-level interactions). Thus, defending descriptive norms seem to matter for victims specifically, whereas popularity norms seem to positively affect the classroom as a whole.

The current study has several strengths. First, whereas prior work mainly focused on how defending norms affect defenders, this study examined how these norms relate to victims’ adjustment. We need to know whether victims are helped by defending norms, as this is a primary aim of anti-bullying programs (Huitsing et al. 2019; Paluck et al. 2016). Second, we extended prior work on the healthy context paradox that, so far, mainly focused on how the absence of negative classroom aspects enhances the plight of victims. Future studies should further examine combinations of positive and negative classroom aspects. Prior work showed that prosocial and aggressive norms may co-occur within some classrooms (Laninga-Wijnen et al. 2020b). In such contexts, aggressive popularity norms overruled the positive ones by fostering aggressive friendships and behaviors, despite prosocial norms. In a similar way, the simultaneous presence of aggressive popularity norms may mitigate potential positive role of defending popularity norms in victims’ adjustment. Third, we demonstrated the importance of distinguishing between descriptive and popularity norms. Descriptive defending norms generated a healthy context paradox, whereas popularity defending norms promoted feelings of belonging and positive classroom perceptions among all students, including victims.

This study also had some limitations. First, we used cross-sectional data. Direction of effects therefore remain unknown. For instance, it could also be that students who do not belong to the classroom, are at particular risk of becoming victimized over time because they lack “social protection” (Hodges and Perry 1999). Longitudinal data could provide more insight in the temporal precedence of these aspects. Moreover, longitudinal data can provide insights in the theoretical explanations that are provided in the current manuscript as potential underlying reasons on why the link between victimization and adjustment may vary as a function of the classroom norm (e.g., self-blame, defending inflation). For instance, in a longitudinal study it can be tested whether the effect of being defended depends on whether it contributes to the ceasing of bullying over time. Person-centered analyses may provide insights in whether victims who are strongly being defended, but for whom the bullying does not stop, develop (more) severe feelings of hopelessness and consequently are worse off compared to victims who are not defended at all. Or whether severely victimized youth are particular kind of adolescents, who for instance have depressive predispositions and who regard the world in a negative and distrustful way, and consequently perceive defending attempts of others in a similar way. Moreover, it can be tested whether norms make it more likely that being defended works out adversely or not for such types of victims. The current study provided an important first step, by showing how defending norms work out concurrently for those who are being victimized despite these norms and forms the basis for future longitudinal studies on this new topic.

Second, defending was measured with peer nominations. Not everyone in the classroom may be aware of who is being bullied and who is defending these victims, and whether this is compatible with how victims experience it (Malamut et al. 2020). Moreover, we did not distinguish types of defending. Direct defending (confronting the bully) and indirect defending (consoling the victim) can work out differently for students’ classroom climate perceptions and psychosocial adjustment. For instance, when defenders take a public stance against bullying, it is more likely to affect students’ classroom climate perceptions than when defenders privately comfort victims. Recently, a reliable and valid scale for measuring various of defending was developed (Lambe et al. 2020), which will be valuable to investigate the impact of being defended on victim’s adjustment.

Third, we based our conclusions on seven analyses, for each outcome separately, which may enhance the chance of a Type-I error. At the same time, it should be noted that we tested a complex model on a sample with relatively few cases on the class-level (n = 45 classrooms), which limits statistical power to detect significant effects. Further, results were consistently in the same direction, also emerged when examining peer-nominated victimization (sensitivity analyses), and were in line with previous work on the healthy context paradox. Three out of the seven tested cross-level interactions for descriptive norms were significant and explained a relatively large part of the variance in the outcome variable (≥ 20.0%). Moreover, the complete models explained 10.5% to 31.3% of the variance across outcomes, which is comparable to or even larger than other work examining the healthy context paradox. For instance, in a study that conducted longitudinal multi-level regression models to examine the role of the KIVA intervention in victims’ depressive symptoms, social anxiety, self-esteem, and school wellbeing, 2.4% to 10.2% of the variance was explained in these outcomes (Huitsing et al. 2019). In another study that used multi-level regression analyses to examine the role of aggressive norms in the link between victimization and children’s self-views, about 0.5% to 9.8% of the variance was explained at the individual-level (Morrow et al. 2019). The relatively large variance explained in the current study strengthens our confidence that defending norms play a role in some aspects of victims’ adjustment – yet importantly, not in all aspects. We also tested seven interaction effects for classroom levels of victimization and popularity norms respectively. Only one of these was significant, which we did not further interpret to prevent over-interpretation. Replication studies are encouraged to examine whether our found pattern is also consistent across other (longitudinal) studies with larger samples.

Fourth, though investigating the moderating role of norms in the victimization-adjustment link provides important information about contextual differences in victims’ experiences, no insights are provided in more subtle, relational processes. That is, the role of defending on victims’ adjustment may depend on who defends whom in a particular context (Huitsing et al. 2014). For instance, victimized girls who were defended by victimized friends were found to experience increased internalizing problems whereas this was not the case for boys (Schacter and Juvonen 2020). Future studies are encouraged to dig into these questions, in order to better understand whether the effects of being defended on victims’ adjustment within a particular context are dependent on who defends whom.

In this study the association between victimization and maladjustment depended on the broader classroom context. In line with the healthy context paradox (Huitsing et al. 2019), victims in classrooms with defending descriptive norms were worse off than victims in classrooms without these norms. Defending popularity norms worked out positively for the general classroom climate perceptions and feelings of belonging of all students, including victims. Therefore, interventions may benefit more from encouraging defending popularity norms than from stimulating defending descriptive norms. Some interventions, such as Meaningful Roles Intervention (Ellis et al. 2016) and the Roots Intervention (Paluck et al. 2016) aim at encouraging defending popularity norms by rewarding prosocial behavior, for example by assigning prosocial leaders in a classroom (who are often popular students) and exchanging compliment cards. Our study provides preliminary evidence for the effectiveness of such interventions. To conclude, this study indicates that defending descriptive norms relate adversely to victims’ classroom climate perceptions and feelings of belongings, whereas defending popularity norms foster positive classroom environments for all students.