The importance of studying the consequences of peer victimization is hardly disputable. From the purely theoretical point of view, such research helps us better understand the human nature and our fundamental needs, such as the longing for affection (Baumeister and Leary 1995) and status (Anderson et al. 2015), and the impact of different developmental contexts – in this case, the peer context – for individuals’ development. Another important justification for such research is that it helps motivate prevention and intervention efforts by showing that if we do not put effort in helping victimized children and youth, the costs may be immense – both in terms of human suffering and in terms of monetary expenditure to societies. Finally, understanding the mechanisms leading to negative outcomes and the potential moderating factors helps us design more effective interventions and target them better.

The adjustment of youth aggressing against their peers needs attention as well – much for the same reasons as mentioned above. This topic might be even a more challenging one than research on the victimized, considering the heterogeneity of the aggressors’ characteristics and motivations (Peeters et al. 2010; see also Salmivalli and Peets 2018) – aggressive behavior may vary from dysregulated, non-selective emotional outbursts to skillful, goal-directed bullying of a less powerful peer. Although the special section focuses on victimization, two of the studies include aggression, looking either at the main effects of victimization and aggression, as separate from one another (Kretschmer et al. 2017) or the interactive effects of the two (Schwartz et al. 2017).

The four studies deserve to be applauded, for many reasons. Still today there are relatively few studies taking such a long-term perspective on peer victimization and adjustment, and even fewer studies that adequately control for potential confounders, utilize highly sophisticated techniques such as latent transition analysis and, importantly, investigate the mediators and moderators of the long-term effects of victimization. The studies provide lots of material that could be discussed, and for sure lots of inspiration for future research in the field. Each one of the studies either addressed moderation directly or stimulated thoughts surrounding moderation. Thus, most issues addressed in my commentary revolve around moderation, in one way or another. I will also discuss the importance of context as a potential moderator of the effects of victimization, and extend my comments to some practical implications and general conclusions.

What are the Aspects of Victimization that Matter?

Studies on peer victimization have roughly speaking followed two research lines, or traditions. One is research on “victimization in general” (Perry et al. 1988); the other is research on “bullying victimization”, or “being bullied” (Olweus 1978). The corresponding traditions from the perpetrator perspective are those of studying aggression vs. bullying.

With respect to the adjustment of the targets, does it make a difference whether an individual is targeted by aggression in general, or by a specific subtype of aggression called bullying? The studies at hand do not, and were not designed to, provide an answer to this question. However, it is good to keep in mind that conceptually, bullying victimization can be considered a subtype of victimization - a subtype in which the perpetrator is more powerful and abuses his/her (physical or social) power to repeatedly humiliate the target. Moreover, research on bullying victimization has, more than research on victimization in general, emphasized the role of context and of bystander responses (e.g., Pozzoli and Gini 2012; Salmivalli et al. 2011These features (unequal power, repeated attacks, involvement of the peer group) might moderate the victimization-adjustment -link, both concurrently and longitudinally.

Haltigan and Vaillancourt (2017) started off by explicitly measuring ‘being bullied’, providing the participating youth with a definition (following Olweus 1996) that clearly differentiates bullying from fighting, aggression, and teasing. However, identifying students targeted by bullying with Olweus’ measure is typically based on the cut-off ‘two or three times a month’ or more often (Solberg and Olweus 2003). Haltigan and Vaillancourt ended up dichotomizing the responses as to reflect whether the individual had experienced 1) no victimization at all or 2) at least some victimization. Without taking a stand on whether this was a good decision or not, it certainly led to extremely high frequencies of youth in the elevated victimization category. As many as 74% of youth in grade 5 and still 46% of students in grade 10 – the prevalence decreases as a function of grade level, but remains high – belonged to the elevated victimization category, according to the online supplementary materials provided by the authors. This creates an intriguing controversy with the measurement strategy originally chosen – what implications it might have for the interpretation of the findings is unclear.

It is known that the intensity of victimization is related to its outcomes, whether operationalized as frequency, as the multiplicity (being victimized in many different ways) or as the number of peers being involved in bullying the targeted student (Van der Ploeg et al. 2015). There is also evidence indicating that power imbalance between the target and the perpetrator adds to the adjustment problems associated with victimization (Card and Hodges 2005; Malecki et al. 2015), and so does the intention of the perpetrator, as perceived by the target (Malecki et al. 2015). These findings, although based on concurrent data, illustrate the usefulness of assessing several aspects of victimization to better understand its impact on adjustment. What are the aspects that matter most is an important question of moderation: they may be the aspects included in the definition of bullying (e.g., power imbalance), but they may be something else as well. Some good candidates are the developmental context (the period when victimization is experienced), the chronicity of victimization (how long it lasts), and the social context (classroom norms, bystander responses).

Being Victimized in Adolescence

Being victimized by peers is likely to be especially onerous during adolescence, a developmental period when peer relations are increasingly important. Peers are sought for emotional support and friendship, more time is spent with them, and peer status becomes more relevant than it was ever before (Fontana and Cillessen 2010; Lam et al. 2014). There are also other reasons why victimization in adolescence would be extremely hurtful. As evidenced in two studies included in the special section, victimization tends to be stable. Haltigan and Vaillancourt (2017) showed that grade 5 victimization status was predictive of grade 10 victimization status, whereas Brendgen and Poulin (2017) provided evidence of victimization by peers at school increasing the risk for workplace victimization, both through a direct effect and an indirect one (via depressive symptomology in young adulthood). Besides such ‘forward-stability’, or the likelihood that the adolescent targets will also be victimized in a later time point, it is important to keep in mind that many victimized adolescents were already victimized before, surprisingly many of them for several years.

Based on our (unpublished) large-scale data collected annually in Finland since 2009, the prevalence of students who experience being bullied by their peers steadily decreases as a function of age. (This echoes the current finding of Haltigan and Vaillancourt who found a grade-by-grade decreasing trend of students in the elevated victimization category, as well as other previous literature.) In the same time, the relative proportion of those who have been bullied for a year, or for several years, increases. For instance, more than 70% of grade 9 (the last year in middle school; 15 years of age) students who are being bullied say this has been going on for a year or more. It should be noted that until grade 5, such responses are very rare. In grades 1 to 4, the most common responses to the question about the length of the negative treatment (asked from those who have reported being repeatedly bullied) are that it has lasted for one or two weeks, or for one month. Thus, quite many students experience some victimization in elementary school, and many, but not all of them, escape this plight during the years to come. Victimization beginning in middle school is rare; many of the students who are targeted in adolescence have been in that situation for a long time. This has been well illustrated also by previous studies of some of the present authors (Brendgen et al. 2016; Haltigan and Vaillancourt 2014) examining the trajectories of victimization from elementary to middle school. The impact of adolescent victimization may last into adulthood not only because adolescence is closer in time to adulthood than middle childhood is, but also because among adolescent victims, a long history of victimization is more likely.

Do the Effects of Victimization Last, and How Wide They Extend?

Is victimization by peers a unique predictor of long-lasting difficulties in life? According to some studies, yes. According to others, not really, once you control for key confounders.

Kretschmer and colleagues (2017) found, somewhat surprisingly, very little effect of victimization and aggression/bullying on mastering various normative developmental tasks in early adulthood. Especially when several confounders were controlled for, but even when they were not, victimization had no impact on most outcomes assessed. One interpretation is that the consequences of bullying involvement are quite specific, related to internalizing problems, or mental health problems more widely, than on the overall life success of those targeted. However, some past research provides evidence of robust effects of victimization on a wide variety of outcomes including (but not limited to) physical health, attainment of educational goals, and wealth, as described by Brendgen (2017). The question proposed by Brendgen whether the lasting effects of victimization are explained by continued victimization in other contexts is indeed interesting. Perhaps the data sets utilized in the Kretschmer study would enable testing this, but also other theoretically justified hypotheses on moderation.

Haltigan and Vaillancourt (2017) found that chronically victimized youth (those staying in the elevated victimization latent class) were seen by their parents as having more internalizing problems in grade 11 than the nonvictimized ones (who were consistently in the low victimization class). This, of course, makes sense. Perhaps a bit surprisingly, and sadly, the authors also found that those who had moved from the elevated victimization class to the low victimization (or perhaps ‘no victimization’) class did not differ with respect to their parent-reported internalizing problems in grade 11 from the chronically victimized. This suggests that those escaping the victim role were not recovered from this negative experience by grade 11, as compared with those who were still victimized. This is not in line with the proposition that the prospective negative effects of victimization are actually signaling continuing victimization. The finding raises questions such as how long the ‘movers’ had already been victimized, and when did they move from elevated to low victimization class? Or could it be that parental reports on their youth’s internalizing problems are increasingly based on their stabilized view of their adolescents’ characteristics, rather than the true situation in grade 11? We need to put continuous effort in better understanding the circumstances under which past victimization continues to affect adjustment vs. can be left behind.

The study by Schwartz and colleagues (2017) takes into account the interactive effect of aggression and victimization, an approach that has been recognized as important in previous studies by the authors themselves as well as others. Although aggressive victims (or bully/victims) have been described as the most maladjusted group in many studies with a wide variety of adjustment indicators (e.g., Schwartz 2000), the present findings indicated that they were not overrepresented among the individuals arrested in youth or in adulthood. In fact, peer-reported victimization in middle childhood decreased the likelihood of adult arrests among aggressors, besides having a negative main effect on the likelihood of juvenile arrest. According to these, and some previous preliminary findings described by the authors, some characteristics that make an individual vulnerable for victimization may make them less likely to join antisocial gangs, or to engage in premeditated, goal-oriented crime. In the Kretschmer study (2017), victimization was unrelated to law-abidance in both data sets, whereas bullying perpetration seemed to increase the likelihood of delinquent behavior. Interestingly, the data sets used by Kretschmer would enable to test the interactive effect that Schwartz and colleagues focused on, namely the one between aggression and victimization.

The Importance of Context

The social context in which victimization occurs is important in many ways. Whether having friends (Hodges et al. 1997), the characteristics of friends (Hodges et al. 1999), or friendship quality (Hodges et al. 1999; Kendrick et al. 2012) may diminish the risk for victimization and, more relevant for the present discussion, alleviate the consequences of victimization has been the focus of much attention. Some studies, but not all, suggest that friends serve a protective role. The findings are not unanimous regarding which friendship characteristics or qualities serve this function best, and which might even exacerbate problems.

The study by Brendgen and Poulin (2017) did not find friend support alleviating the consequences of victimization. There was a main effect on depression, as well as on workplace victimization later on, but friend support did not protect against later problems among those who were victimized. Clearly, we do not know yet what kind of support from peers makes a difference, and via which mechanisms. For instance, the measure used by Brendgen and Poulin might tap the closeness, sharing and intimacy in a friendship rather than support or protection received from a friend. Intimacy and closeness may bring along co-rumination, which is not likely to be helpful (Rose 2002). Indeed, Rose et al. 2017 recently reported (2017) that co-rumination with a best friend increased the likelihood that depressive symptoms led to greater peer stress among adolescent girls. Even though it seems that merely having a friend protects against ending up as victimized in the first place (Hodges et al. 1997, 1999), having a victimized friend might increase the risk of being victimized (especially in the case of relational victimization – see Sentse et al. 2013).

It could be expected that for those already victimized, friends can be most protective when they mitigate self-blame, give helpful advice, or provide concrete support. In studies testing the friendship protection hypothesis, we should probably take into account the presence vs. absence of friends, the characteristics of friends, friendship quality, but also assess the actual interaction between friends (e.g., co-rumination, encouragement and validation). Besides friends, also other peers may serve a protective function: victimized youth who are supported and defended by at least one classmate are better adjusted than the ones not receiving such support (Sainio et al. 2011). Such classmates may be empathic, prosocial peers who have the courage to stand up for the vulnerable. Finally, Brendgen and Poulin brought up an important question of whose support (e.g., parents vs. friends vs. spouses) matters in which developmental period. This is related to another question: what is the impact of social support received at the time when victimization happened (e.g., in childhood or adolescence) vs. support at the time when we measure the outcomes (e.g., early adulthood)?

Victimization is often embedded in the larger peer context – which is emphasized especially in the literature on bullying victimization. In a Finnish qualitative study (Teräsahjo 1997) the most negative and traumatic memories of adults who had been bullied during their peers at school were not the mean words or hurtful acts from part of the bullies. Instead, what the former victims remembered as most distressing was the perception that ‘no-one cared’: they described how peers witnessing their plight were laughing, or just completely overlooking what was happening. Thus, the whole classroom context may be more or less protective against the negative consequences of victimization. The situation of a victimized student is fundamentally different in a classroom where rewarding and reinforcing the bully is normative, as compared with a classroom where peers tend to show that they do not approve of bullying (Salmivalli 2010). Interestingly, a recent experimental study using the Cyberball paradigm (Sandstrom et al. 2017) echoes this by showing how the ratio of (virtual) peers involved/not involved in ostracizing a young person in a game situation is predictive of immediate negative outcomes (feelings of happiness, need fulfillment, and self-esteem).

The ‘Healthy Context Paradox’

The context may have some seemingly perplexing implications as well. It has been found that the negative impact of victimization may be especially damaging in contexts where the overall level of aggression and/or victimization is low. In such contexts, victimization may not appear as a serious problem; however, it might be serious in a way we do not immediately come to think about. I call this phenomenon ‘the healthy context paradox’.

Bellmore et al. (2004) found that the association between victimization and anxiety was stronger in classrooms with low social disorder. In their study, social disorder was operationalized as the classroom level of aggression and victimization. Nishina and Juvonen (2005) found that observed victimization buffered against the negative effects of experienced victimization: the negative effects were mitigated for those victimized students who also saw others being harassed in their environment. Huitsing et al. (2012) found that victimization was related to maladjustment especially in classrooms where 1) the average level of victimization was low or 2) victimization was highly centralized, in other words targeted at few specific students rather than “a little bit on everyone”. The findings of the above studies were interpreted in terms of social misfit and attribution theories. In other words, being victimized in a context where very few others share this plight, one is a social misfit deviating from what is normative. In such a context, attributing the cause of victimization to oneself (“It must be me” – there is something wrong with me that causes this treatment) is likely. Seeing others being victimized, too, enables to attribute victimization to external factors, perhaps something in the individual perpetrators, in the teacher’s ability to manage the classroom, or classroom characteristics in general.

Along the same lines, a recent longitudinal study by Garandeau et al. (2016) compared the adjustment of stable victims (those who were victimized by peers at time 1 and time 2, a year apart) in two types of classrooms: those where the proportion of victimized students had decreased across the one-year period and those where it had either remained the same or increased. Stable victims felt more depressed, more socially anxious, and were also less liked at the later time point in the former type of classrooms, where the overall situation had changed into better. Thus, an improved social environment can be detrimental for some children – those who remain to be victimized.

Practical Implications

The most obvious practical implication of the studies included in this special section is that peer victimization needs to be countered. It tends to be stable, and it has negative consequences that can be long-lasting. Stability is more likely for some youth, such as early maturing girls or late maturing boys (Haltigan and Vaillancourt 2017) and it is partly mediated by depressive symptoms (Brendgen and Poulin 2017). Such pieces of information, as they accumulate, help to pay attention to those most at risk for continuing victimization and to prevent the process of victimization becoming chronic.

Haltigan and Vaillancourt (2017) showed that the degree, rather than form of victimization differentiates between youth. The online supplementary material provided by the authors show that the most common form of victimization experienced is verbal, followed by social exclusion, followed by physical victimization, followed by electronic victimization - a very universal pattern indeed, found also in our large-scale data collected in Finland. Educators often pay attention to a single form of victimization, seeking to find out how that specific form should be intervened. Especially in the case of electronic victimization, they also tend to see it as an out-of-school experience that cannot be tackled by means of the school. However, most youth who are victimized online are victimized offline (which goes to other forms, too – most students who are victimized tend to be victimized in many ways). I would argue that a common fault in intervening victimization is that adults intervene in a specific incident of a specific type (student X did this specific mean act to student Y this morning in the school hallway) whereas attention should be paid to the overall situation of students who are repeatedly tormented by multiple means, in multiple occasions, even by multiple perpetrators. Even when a single incident does not seem severe enough to be intervened, such a single event is often part of a much larger picture – a situation that can be understood as detrimental only when seen as a whole. It is the overall situation that needs to be intervened, besides the single incidents as they occur.

Even if not based on the present findings, I would like to point out some implications of the ‘healthy context paradox’ for prevention/intervention work. The good news about overall rate of victimization going down in a school or in a classroom is not necessarily good news for every student in this context. Decreasing the number of victims is, and should remain to be, the ultimate goal of policy-makers and school practitioners. However, we should not be satisfied simply with achieving significant reductions in rates of victimization. Also the evaluations of the effectiveness of anti-bullying interventions should not be limited to declines in victimization. It is critical that both researchers and practitioners address the specific difficulties encountered by children who remain victimized despite interventions.

Conclusions

The question of moderators of the victimization- adjustment link is important. The moderators may be individual (sex, characteristics of the targeted child), they may be interpersonal (having friends/supportive classmates, or the power imbalance between the bully and the victim) or they may be related to the larger context (classroom or school characteristics). The moderators may be related to the nature of victimization itself, such as its intensity and chronicity.

It is conceivable that the consequences of victimization depend on the developmental period during which victimization takes place, as well as its chronicity. It is also important to tease apart the effects of these two factors. When we start our data collection in an adolescent sample, it is good to keep in mind that many of the participants were already victimized at the time when we were writing the grant application for our study.

An important message from this special section as a whole is that we need to take a dynamic approach to studying victimization across development – a direction towards which the papers in this special section take important steps. The outcomes of victimization may be specific to the age when victimization took place, as well as the age when these outcomes are measured, but also the key moderators and mediators might be dynamic, varying across time.

Even when providing strong evidence pointing to a conclusion, correlational studies always leave us with some ‘reasonable doubt’ concerning cause and effect. Even controlling for several potential confounders of the association between victimization in childhood/ adolescence and adult outcomes cannot wipe away this problem. Therefore, in addition to longitudinal correlational studies like the ones in this special section, we need more well-designed intervention studies. In addition to analyzing pretest-posttest changes in the primary outcome variables, such studies should include a variety of carefully selected ‘secondary’ outcome variables, to shed more light on the impact of victimization and its (dis)continuation over time, across developmental periods.