Relations between school bullying, friendship processes, and school context

ABSTRACT Background Despite considerable anti–bullying efforts and greater awareness of the social processes underpinning bullying, bullying is still a serious problem across schools in many countries. In exploring the social processes that contribute to school bullying, research indicates complex relationships between bullying and the maintenance and building of friendships. While such findings provide important information about the social context of school bullying, more needs to be understood about the institutional context within which school bullying – and friendship – occur. Purpose The aim of this study is to better understand how school bullying relates to friendship processes, and how these are, in turn, influenced by the institutional constraints of the school context. Method The findings discussed draw on 3 months of ethnographic fieldwork conducted at one Swedish elementary school. The fieldwork involved participant observations, as well as semi–structured group interviews with 34 sixth-grade pupils (approx. 12 years of age), which were conducted towards the end of the fieldwork. Data were analysed thematically. Findings The analysis highlighted the importance of friendships to pupils but also identified the ways in which understandings of friendship relations were closely tied to the importance of social perceptions and the organisational constraints of the school context. Conclusions Taken together, the findings suggest that school bullying cannot be de–contextualised from the social and institutional contexts of school but may rather be connected to the perceived need for control in that particular arena. The study draws attention to how the complex relations between bullying, friendship and school context need to be better understood, in order to support efforts to prevent school bullying.


Introduction
Despite considerable anti-bullying efforts, the implementation of anti-bullying initiatives in schools and a greater awareness of the social processes underpinning bullying, recent findings have nonetheless highlighted an increasing prevalence of bullying in Swedish schools (Bjereld, Augustine, and Thornberg 2020), and elsewhere. While research has demonstrated that school pupils tend to perceive bullying to be a serious transgression Yoneyama and Naito 2003). These aspects, and the relations between them, form the basis of inquiry in the research reported in this paper. However, before describing the study in more detail, we situate our work further by considering the relevance of school context to the problem of school bullying.

Background
A number of researchers studying bullying from a sociological perspective have compared schools with other institutional arenas, including workplaces and prisons.
Questions have been posed about the lack of focus on the institutional context (Eriksson 2001;Eriksson et al. 2002;Yoneyama and Naito 2003). As Yoneyama and Naito (2003, 316), for example, have argued: Normally in sociology a wide prevalence of a problem is regarded as a social issue, worth examining as a potentially structural problem inherent in the particular social milieu. If it is a structural problem, there is a limit to what individuals can do to solve their everyday-life trouble: fundamental solution comes with the revision and change of the social structure itself [. . .]. In view of the extent of prevalence of the issue, the structural perspective appears underdeveloped in the discourse on bullying.
In considering the importance of the organisational structure of schooling, Eriksson (2001) and Eriksson et al. (2002) pointed to four aspects of the school context that might be important for understanding school bullying and may be summarised as follows: (1) pupils spend a long time, or an undefined amount of time, at school; (2) pupils are unable or limited in their ability to choose who they spend their time with at school; (3) pupils are not permitted to leave school; and (4) the number and presence of pupils is somewhat arbitrary, meaning not only that a school class may consist of one more or one less pupil and still function but also that individual pupils do not have officially defined roles and are thus not, in this sense, necessary for the class itself to function (Eriksson 2001;Eriksson et al. 2002).
In a similar way, Duncan (2013a) pointed to four characteristic features of schooling that can contribute to the unhappiness of pupils: (1) compulsion; (2) compression; (3) competition; and (4) control. The compulsory aspect of schooling is relevant for understanding school bullying not least because it draws attention to how pupils cannot easily leave school if they are being rejected, socially excluded, or bullied by their classmates and/or schoolmates (Duncan 2013a;Eriksson 2001). When combined with compulsion, compression also means that those who are subjected to bullying may not be able to find some space of their own, away from other pupils with whom they may or may not get along (Duncan 2013a;Horton 2018). As Yoneyama and Naito (2003, 325) put it, pupils are thus 'constantly forced to be involved in human relationships that are not of their own choosing, without being able to maintain a comfortable distance from others'.
While schools can take various forms and have different core values, they are often competitive contexts, wherein pupils are measured, judged, valued, sorted, and positioned as successful or unsuccessful (Duncan 2013a, 40;Horton 2018;Jacobson 2010;Langmann and Säfström 2018). As Duncan (2013a) points out, this serves to send messages that competition is more important than 'the humanistic qualities of collaboration and cooperation'. A number of researchers have pointed to the importance of the hidden curriculum and the learning of 'power-dominant human relationships' in schools (Yoneyama and Naito 2003, 325). Indeed, as Jacobson (2010, 276) argues, pupils 'learn the value-laden game of hierarchy and individuation, providing meaning to the experience of who counts and who does not within schooling'. This situation is reinforced through the use of control, in terms of determining where pupils sit, when they can leave class, and whether they can be inside or outdoors, for example.
The compulsory, compressed, competitive, and controlled aspects of schooling are relevant to the consideration of school bullying, since they point to the importance of self-presentation at school (Goffman 1959). While pupils may have little say over who their classmates or schoolmates are, they nonetheless organise their social relations within those social contexts (Eriksson 2001) through questions such as 'who belongs and why?' (Forsberg and Thornberg 2016, 20). This means that pupils may find it necessary to present themselves in particular ways (Goffman 1959(Goffman , 1990, not solely in order to increase their chances of forming or maintaining friendships, and thus fulfil their 'longing for belonging' (Søndergaard and Hansen 2018, 327), but in order as well to reduce the risk of being socially rejected or marginalised, and hence assuage their 'social exclusion anxiety' (Søndergaard 2012, 355;Søndergaard and Hansen 2018).
Taken together, then, schools may be perceived by some pupils as places where the fear and risk of being singled out and/or socially excluded are ever-present (Søndergaard 2012;Søndergaard and Hansen 2018), and such fears and risks may be exacerbated by the specific features of schooling discussed above. Indeed, as Gordon and Lahelma (1996, 303) have noted: The school as a physical space provides a context for the practices and processes that 'take place' there. But the physical school is more than a context; it is an aspect in the shaping of these practices and processes producing differentiation.

Purpose
Against this backdrop, the aim of the qualitative study reported in this paper is to better understand not only how school bullying relates to friendship processes but also how these are, in turn, influenced by the constraints of the institutional school context. The findings discussed in this article are part of a larger project (Strindberg 2021) exploring pupils' reflections on why pupils may bully or choose to refrain from intervening in defence of someone being bullied, despite understanding that bullying is wrong.

Ethical considerations
Prior to collecting any data, the study was ethically approved by the Swedish Ethical Review Authority. The study followed the Swedish Research Council's (2017) ethical guidelines for good research practice. The first author visited the school and the classes a couple of weeks before beginning the fieldwork and informed the pupils and teachers about the study and its implementation. Letters giving information about the purpose and implementation of the study were provided, together with consent forms, and informed consent was obtained from the principal, teachers, the participating pupils and their legal guardians. Rather than treating the issue of consent as a 'one-off event', we approached it as 'ongoing' by ensuring that pupils were able to take their time to decide, were able to pose questions about the study and were then able to make an informed decision as to whether they wanted to participate (Morrow 2008, 54). Pupils were also informed that they were free to withdraw their consent at any time, that they could stop the interview if they felt uncomfortable, that the information they provided would be treated confidentially, and that they would be assigned pseudonyms to facilitate anonymity. Pseudonyms were also assigned to the teachers, classes and schools, and for de-identification purposes certain identifying information has been omitted from the presentation of the findings.
While the first author adopted a non-judgemental approach during the fieldwork (meaning that pupils were not told off by the researcher if they broke school rules, for example), he was prepared to intervene in potentially distressful situations if the need arose to minimise potential harm. All the interviews were conducted in quiet rooms, separated from the rest of the class, to facilitate confidentiality. With pupils' comfort in mind, participating pupils were interviewed in friendship groups based on their own suggested groupings, rather than individual interviews or group interviews based on teacher-suggested groupings.

Data collection
The findings draw on 3 months of ethnographic fieldwork conducted by the first author at one Swedish elementary school. The school, to which we have given the pseudonym Birchwood School, was a rural public school located in a smaller Swedish municipality, which had somewhere between 300 and 400 pupils enrolled from preschool class (i.e. 5-6 years of age) to sixth grade (i.e. approx. 12 years of age) at the time of the study. The school setting was similar to many other rural schools in Sweden, and the school was selected because of its representativeness. The classrooms for grades four to six (i.e. approx. 10-12 years of age) were in a building opposite the school canteen. The classrooms for grades one to three (i.e. approx. 7-9 years of age) were in the main school building, which also housed other school elements, including the principal's office. The playground was in front of the main building and was laid out as an open space with a basketball court and marks on the ground for the game Four Square, 1 as well as a sandbox with a climbing frame, which was usually occupied by pupils from the lower grades (i.e. grades 1-3). The playground also had a concrete football pitch and another sandbox with swings and slides.
The data were collected during the autumn term and included participant observations and interviews with pupils from two sixth-grade classes (i.e. approx. age 12), which we refer to as classes 6C and 6D. In class 6C, 20 pupils consented to participate, and in class 6D, 14 pupils did so, giving a total of 34 pupils in total participating in the study. The first author spent 4 days a week at Birchwood School and approximately one and a half months in each of the two classes. The fieldwork was implemented in the form of a 'compressed time mode' (Jeffrey and Troman 2004, 538), during which the first author spent his time with the pupils, participated in their break activities when invited to do so, and engaged in informal conversations with them about things that interested them.
Observations were written down in the form of field notes. Whereas the early field notes were descriptive and focused on tangible social interactions and events, the latter notes were more reflective in style (Emerson, Fretz, and Shaw 2014, 353).
Towards the end of the fieldwork, pupils were invited to participate in group interviews consisting of 3-5 pupils, based on their friendship groupings (as explained above). Due to some pupils being absent at the time of interviews, or not wanting to be interviewed, 28 of the 34 participating pupils were interviewed. The interviews lasted about an hour. While most groups were interviewed on one occasion, two groups were interviewed twice because of their willingness to talk further about their school experiences. The interviews were semi-structured and organised according to a number of themes, including school and class, a day at school/breaks, friendships and loneliness. The interviews also included more general questions about the importance of school and about school bullying, such as why pupils might sometimes do things at school they know are wrong.

Data analysis
The interviews were audio-recorded and transcribed verbatim. In analysing the data, we used approaches based on constructivist grounded theory (Charmaz 2014). In the initial phase of coding, the data were read and re-read by the first author, who looked for recurring patterns in the data, and for aspects and segments to follow up and analyse further. The data were assigned different labels, then sorted and further reduced. At this stage of the coding, the first author tried to stay as close to the data as possible. Thus, rather than 'forcing data into preconceptions' (Thornberg, Perhamus, and Charmaz 2014, 410), the first author tried to stay open towards what the data seemed to suggest, and considered the codes as temporary, modifiable, and open for revision. More specifically, the analysis was driven by open questions, such as 'what do the pupils say?'; 'how can this be understood?', as well as more specific questions, such as, 'how, and in what ways, are friends important to pupils?'; 'what do the pupils say about the importance of status?'; 'how is the perceived importance of friends related to the institutional context of schooling?' By asking questions of the data and by further sorting and elaborating the data, a number of categories, such as the importance of friendships, were gradually formulated. Lastly, and in consultation with the second author, the findings were theoretically analysed, with a specific focus on the importance of the institutional school context.

Findings
Our fine-grained analysis of the observation and interview data yielded some interesting insights into the relations between bullying and friendship processes, and how these are influenced by the school context. In our presentation of findings, we use a number of case study examples drawn from the analysis of the participant observations and interviews to illuminate key themes. We have translated the anonymised excerpts used in this article from the original Swedish into English. The findings are structured in subsections that relate to the key themes that emerged from the data: (1) compulsion, compression, and the importance of friendships; (2) competition, status and the importance of perceptions; (3) control, bullying and the importance of social support.

Compulsion, compression, and the importance of friendships
In line with national requirements, pupils were required to remain at school from the time they arrived in the morning until the end of the school day (8:00 am to 2:20 pm most days). With the exception of physical education (PE), crafts, and religion lessons, pupils largely spent their days at school either in the same classroom with one or two allocated teachers or out in the playground during breaks. On Mondays, for example, pupils had three 50-minute lessons before lunch, interspersed by two shorter breaks (20 minutes and 10 minutes), an hour-long lunch break, and then three lessons after lunch interspersed by a 20-minute break. As is typical in most school systems, the pupils were allocated a class based on their age.
A perception of differentiation between classmates, friends, and best friends was made evident by one of the boys in Class 6D. He alluded to the fact that, while pupils were not free to select their classmates, they were free to decide who they spent their breaks together with: There are sort of two differences: if you sort of do stuff together sometimes, then you are friends, and if you do it sort of all the time, then you're like best friends, and if you sort of never do it . . . then you're just classmates.
The quotation, and the distinctions it implies, highlights how the relation of 'classmate' was regarded as more formally regulated than informally or socially grounded. It also illuminates how the nuances of friendship relations (i.e. being best friends, friends, or classmates) had implications for how pupils' days at school might unfold. One of the boys in the other class (6C), who often played basketball or Four Square during breaks, explained the importance of having friends at school by saying 'You can joke and think it's fun'. As he elaborated, 'school would be very boring without friends'. Such comments draw attention to the perceived enjoyment of having friends, and the idea that friends enable school to be more interesting.
Furthermore, a girl in Class 6D observed that not having friends at school would mean 'You would be all by yourself'. Whilst this girl did not explain in more detail what she thought it would be like to be alone, it was clear that she did not speak about it in a happy sense: rather, she seemed to perceive this to be a stressful and generally negative experience. In a similar way, a girl in 6C reflected that 'You might feel, like . . . a little outside the class or . . . I think it's important to have friends, because it's not the same thing to be alone during breaks'. The girls' comments point to the ways in which having friends may alleviate feelings of being left out, which was suggested by pupils as important for their overall school experiences and well-being. Their comments also highlight the relations between the perceived importance of friendships and the compulsory and compressed school context, wherein pupils were unable, or limited in their ability, to choose who they spent their time with. This notion was also evident in the observations of another girl in 6D, who remarked, 'Yeah, I mean, you have to be with each other for many, many years, so how you are perceived is important'. As well as illustrating the contextual significance of compulsion and compression, this also emphasises the significance placed on time spent at school (e.g. Duncan 2013a; Langmann and Säfström 2018).

Competition, status, and the importance of perceptions
It was evident from the analysis of data that being a pupil entailed being in a socially competitive and hierarchically ordered space, in which the social positioning of pupils was important for their friendship networks and overall school experiences. Indeed, while those perceived to be 'low status' peers could find themselves left out during breaks, those regarded as 'high status' could be distinguished in the playground by how other pupils positioned themselves in relation to them. For example, popular peers were never left out because other peers sought them out and sought to position themselves near them. This is illustrated by the following field note, which describes the activities of five pupils who all had socially central positions in their classes: It is break after lunch. [Four pupils] from 6D, [one pupil] from 6C, and a few of their classmates are kicking a ball against the wall outside the canteen. One of the girls in the group suddenly grabs the ball and shouts, "It's our turn to be on the football pitch!". "Wow, come on, it's football!", [the pupil from 6C] replies, and everyone runs to the pitch to play together.
As illustrated by this field note, status not only affected pupils' opportunities for participating, it also tended to cluster certain pupils together. This was also explained by another boy from 6D, who suggested that pupils interpreted the relative status of their peers and sought to align themselves in line with these interpretations. As he put it, 'They just want to hang out with people who are popular, and that is because they realize sort of that "you are popular" and then they follow after them'.
While the playground was a landscape where pupils could engage in various activities, such as meeting friends, talking about the past or coming weekend, or playing football, basketball, or Four Square, it was evident from the data that the possibilities for pupils to do so were related to their perceived social status. Pupils may have, thus, found it necessary to present themselves in particular ways, not least to increase their chances of being included and forming or maintaining friendships. This was highlighted by two boys (pupil 1 and pupil 2 in the excerpt below), when conversing about how pupils adopted a particular role at school: Pupil 1: Yeah, like . . . when you go to school you have to sort of have this, you go into your role Pupil 2: You think that you'll get more friends maybe, if you are something other than what you really are! Pupil 1: If you are considered cool, then you make more friends for some strange reason . . . it becomes sort of . . . it's really weird, but you don't get any friends if you're not popular . . . that's how it is! As these boys explained, pupils may have experienced social pressure to present themselves in socially acceptable ways in school, and to adopt particular roles or present themselves as something other than what they 'really are', not least in order to gain friends.
Understanding of role-pressures associated with friendship formation could also ultimately help in comprehending the ways in which bullying behaviours might be thought of as goal-directed (Volk, Dane, and Marini 2014) or strategic (Olthof et al. 2011), as followed up in the section below. Primarily, though, it illuminates the importance of the school context, and how the socially competitive nature of the context can further intensify a perceived need to present oneself in a socially acceptable manner (Goffman 1959(Goffman , 1990.

Control, bullying and the importance of social support
According to the analysis of data, it was evident that having close friends provided a form of social support for pupils, and a means of maintaining some degree of control in a school context where their days were otherwise largely controlled by others. This situation entailed, for example, the surety in advance that someone would probably be waiting for them at the start of the school day, that someone was going to keep them company during the day, and that they would have someone to accompany them to the canteen, to PE lessons, to the playground, and so on. One boy suggested that friendships were the most crucial aspect of school life. As he put it, 'The most important thing about being in school is probably that you have friends to be with'. While he did not go into much more detail about this, the pupil appeared not to have a broader network of friends beyond two pupils who were his best friends. As it was these friends who kept him company during breaks, he explained that he often felt left out on those occasions when these friends were absent. It is noteworthy, then, that while breaks are often regarded as the most enjoyable part of the school day for many pupils, this may not be the case for those who find themselves on their own in the playground, and may potentially lead to school avoidance strategies as a way of maintaining control.
A number of participants in the study suggested that being a school pupil entailed an ever-present risk of being bullied. They made clear that friendships were of particular importance, not least because of the social support and protection that friends could provide. One pupil in 6C, for example, suggested that friends 'support each other if something has happened . . . well . . . take care of each other'. While this pupil emphasised the importance of friends as caring, friendships were also understood as being protective against negative things happening in the first place. This was illustrated by one of the girls in 6D, for example, who explained that not having friends could lead to someone being teased more. As she put it, 'you sort of get teased a bit more then . . . when there isn't really anyone to support you'. As suggested here, not having friends implied being less likely to be supported after something had happened and less able to control the situation in the first place.
The relationship between friends and bullying was also raised by another girl in 6D, who suggested that pupils may bully as a means of reducing the risk of being bullied themselves, which may in turn improve their social standing: I think pupils might bully because they do not want to be bullied themselves! Sort of, "I may be afraid of being bullied, and then I bully someone else, so that it is harder for me to be bullied, and then maybe I get a little cool also, because I bully!" In the same interview, this pupil elaborated with another pupil on the links between popularity, friendship and bullying. Interestingly, they suggested that popularity provides more friends, who can then act as a social 'shield' to protect them against the risk of being targeted. Bringing into sharp relief the connections between popularity, friendship, and bullying, they explained that more popular pupils were less likely to be bullied than those who were unpopular. As they made clear, the perception was that popular pupils have more friends who 'care about them', who 'get involved', and who are willing to take their side in conflicts and thus help control the situation. Unpopular peers, on the other hand, do not enjoy the same level of protection, as they do not have peers who stand by them or socially 'shield them' in conflicts. Indeed, while pupils who are regarded as 'high social status' might be able to control the situation by fending off the threat of bullying, pupils thought of as having 'low social status' appeared to be socially perceived as easy or 'unshielded', targets.
While these reflections may suggest the importance of encouraging pupils to take on the role of 'defenders' in bullying situations (e.g. Sainio et al. 2010, 144), the perceived relation between friendships and control is also important for understanding why pupils might choose not to, despite understanding that bullying is wrong . For example, as three pupils (pupils 1, 2 and 3 in the except below) suggested: As illustrated by pupil 1 in the excerpt, there was an awareness of the risk of social escalation and the degraded social status that defending behaviour might entail. Pupils may have thus refrained from intervening in defence of their peers, due to the associated risk of the situation getting out of control. Indeed, as highlighted above, pupils' social status was not fixed or stable. Rather, pupils were well aware of how their actions may have been interpreted and socially judged by their peers and classmates. Thus, helping out a less popular or unpopular peer might have caused a shift in the status hierarchy and in turn lead to a loss of control due to an escalation of the situation and the emergence of 'social panic' (Søndergaard 2012, 356). This illustrates the social implications of asking pupils to intervene as defenders in bullying and, again, illustrates the importance of taking the school context into greater account when addressing pupils' behaviours in bullying situations.

Discussion
In this study, we have explored the ways in which school bullying relates to friendship processes, and how these are, in turn, influenced by the institutional constraints of the school context. In doing so, we have borne in mind a number of features that typify the nature of schooling in many countries and jurisdictions. These are: that schooling (up to a certain age or stage) is compulsory, which means that pupils are thus not permitted to leave the school setting; that pupils spend a considerable proportion of their time at school; that school is a compressed space where pupils are unable, or limited in their ability, to choose who they spend their time with; that school is a competitive context wherein pupils must compete not only scholastically but also socially; and that pupils' daily lives at school are largely controlled by others (Duncan 2013a;Eriksson 2001;Eriksson et al. 2002). As we have argued, these compulsory, compressed, competitive, and controlled aspects of schooling are relevant for gaining insight into school bullying, not least since they provide the context for the practices and processes that take place there (Andrews and Chen 2006;Duncan 2013a;Eriksson 2001;Eriksson et al. 2002;Gordon and Lahelma 1996;Horton 2018;Yoneyama and Naito 2003).
According to our in-depth analysis of rich data, it was clear that pupils felt that there were significant distinctions between best friends, friends, and classmates, with the understanding being that they would not be spending time with classmates who were not regarded as friends during their daily breaks (Eriksson 2001). As pupils are freer to decide who they spend their breaks with, it is evident that some pupils may risk spending their break times alone. While being socially excluded is difficult enough in itself, the compulsory, controlled, and compressed nature of schooling means that those who are alone during breaks cannot go elsewhere to find companionship: rather, they must remain in that setting during their time at school (Eriksson 2001;Eriksson et al. 2002;Langmann and Säfström 2018). In particular, the compressed nature of the school setting means that they also cannot escape the judgement of their peers, with whom they spatially and temporally share their schooled lives.
As pointed out by the participants in this study, friends are important because they allow pupils to attain some degree of control within a context where they otherwise have little control over their daily schooled lives. Friends provide a degree of control to pupils, not just in terms of knowing in advance that someone will be there to keep them company during the school day but, too, in terms of providing support and protection against negative interactions such as teasing and bullying. Indeed, having friends may provide a form of social 'shield', while not having friends entails vulnerability and potential exposure to the negative actions of peers.

Limitations
It should be noted that the findings of this study are based on a small sample of pupils from one elementary school in Sweden. This necessarily limits the extent to which it is possible to draw conclusions about the findings for other schools and peer groups. Moreover, we want to acknowledge that ethnographic fieldwork does not provide an exact picture or representation of any studied context. Rather, ethnographic fieldwork is 'inevitably selective' (Emerson, Fretz, and Shaw 2014, 353). Thus, some questions have been brought to the fore, while others have remained in the periphery. It is also important to note that the use of interviews only allows exploration and analysis of descriptions and reflections of pupils' reported behaviours, not actual behaviours. These aspects obviously affect the conclusions drawn, as well as the possibility of generalising the findings of this study. While we do not suggest that all schools, classes, or groups of pupils have the same conditions or are dealing with the same issues, there are, nevertheless, several shared patterns, or 'similarity between contexts' (Larsson 2009, 32), which, we argue, are central to schools in many countries and jurisdictions. As we have shown, this applies not least to compulsory attendance, competitiveness, limited opportunities to choose classmates, lack of pupil control, and the perceived importance of group belonging. Taken together, the findings and analyses discussed offer insight into the relations between school bullying, friendship processes and the school context that could help us better understand why bullying occurs, and support efforts to prevent it.

Conclusions
Our in-depth analysis of rich data can help us to better understand the complex relations between school bullying, friendship processes, and the school context. Overall, the findings suggest that school bullying cannot be de-contextualised from school's social and institutional context. They also draw attention to the importance of friendships to pupils. However, it is also evident that friendships may be contingent on how pupils are socially perceived: pupils tend to seek and favour the friendships of popular classmates and peers. In some situations, bullying can provide a means through which some pupils are able to improve their social standing by being positioned as popular, or 'cool' (Strindberg, Horton, and Thornberg 2020a). In this sense, bullying may serve to shield against the risk of bullying and thus assuage social exclusion anxiety (Søndergaard 2012;Søndergaard and Hansen 2018). In contrast, those who are perceived as unpopular may be less able to establish or maintain friendships and are thus at greater risk of being left out or targeted. As they are not socially shielded, they may be less likely to receive help when bullied because of the perceived risk of the situation escalating and getting out of control (Strindberg, Horton, and Thornberg 2020b). Our findings suggest that bullying and non-defending can be understood as protective strategies that relate to attempts to maintain some degree of control over social situations, through the maintenance of selfimage and the positive perceptions of peers (Jacobson 2010).
While school bullying and friendship processes are intertwined, they are also contextually situated within the specific school context. The compulsory, compressed and controlled aspects of many school contexts restrict the ability of pupils to step away and seek respite from potentially negative social situations (Yoneyama and Naito 2003). While different school contexts may be more or less scholastically competitive, they are commonly socially competitive contexts, wherein the social performances of pupils are constantly on show and open to the judgement of peers (Goffman 1959(Goffman , 1990. Indeed, the importance of friendship is writ large in school contexts, as pupils are aware that they must remain at school for many years and socially engage with large numbers of peers not of their choosing, in limited spaces, while at the same time navigating the social perceptions of those peers and the associated risks of bullying. An important implication here is the need for more research to focus on the significance of the institutional school context and in-depth exploration of how various aspects of schooling (such as compulsion, compression, control, and competition, but also potentially others) may contribute to the occurrence and perpetuation of school bullying (Duncan 2013a;Eriksson 2001;Horton 2018). In asking questions about the significance of the school context (Eriksson 2001;Eriksson et al. 2002;Langman and Säfström 2018), it may be particularly fruitful to conduct more ethnographic research in schools (Horton 2011) and consider the perspectives of pupils more thoroughly (Søndergaard and Hansen 2018). Indeed, as we have highlighted in this article, relations between school bullying, friendship processes and the institutional constraints of the school context are complex and worthy of further investigation, in order to better understand how bullying occurs and better support school bullying prevention efforts. Note 1. A common breaktime ball game where four pupils stand in a painted square divided into four quadrants.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by the Swedish Research Council under Grant (Vetenskapsrådet) D0775301.