Victims, perpetrators, scapegoats and Russian dolls: Narrating violence within secure units for adolescents from a staff perspective

A rise of violent incidents at secure units for adolescents has been reported by the Swedish National Board of Institutional Care. Meanwhile, research aiming to understand how staff manage violence seems to be lacking. By examining an in-depth narrative by one staff member, “Meral”, this study aims to understand, on the one hand, how staff describe the violence they encounter in light of the context and situation, and on the other, how they describe their handling of violence from outside the immediate environment. Drawing on Georgakopoulou and Bamberg, identities are understood to be produced and performed within personal narratives from different positions in relation to one’s surroundings. The study shows how Meral’s professional identity is shaped and affected by violence. Of essential importance is the way Meral presents herself to herself: as “not afraid.” A narrative interpretation is that fear does not fit within the framework of the professional identity for staff. A key element of placing essential responsibility on staff to manage violence is keeping lines of communication open, which could be made clearer in policy documents, training and supervision. Therefore, studies like this one could result in the development of communication strategies for staff. This is important because emotional rules can generate emotional cultures that in the long run can be destructive for both staff and young people. Only when the emotional rules are identified can staff develop strategies for dealing with the violent incidents that are part of their professional life in a qualified way.


Introduction
On the one hand, secure units for adolescents can be described as violent settings (Alink et al., 2014;de Valk et al., 2015) for both staff and young people. On the other hand, they are also emotional settings (Andersson, 2021;Anglin, 2002;Biszczanik and Gruber, 2021;Warming, 2019) where the staff, by doing peoplechanging-work, are also doing emotional work (Andersson, 2020b;Degner et al., 2010). An essential aspect of a secure unit that it merges care and control, creating tense emotional zones in which both young people and staff often talk about their need for safety (Andersson, 2020b;Moore et al., 2017). As also stressed by Dockar-Drysdale (1998), inside these kinds of institutions, researchers and staff need to go beyond the question of "why": in other words, there is a need to shift the gaze from youth to staff regarding the question of responsibility for violent behavior. Organizations (such as secure units) have often been described as rational, unemotional, and neutral entities. However, recent research has also shown the importance of violence and emotions in and around organizations (Andersson, 2021;Biszczanik and Gruber, 2018;Flam et al., 2010).
Violence is a complex, culturally bound, and emotion-laden concept that most people prefer to ignore (Federn, 1990). But staff who work with young people at secure units inevitably encounter violence (Alink et al., 2014;Andersson, 2021). Violence often occurs in a panic, which implies the importance of communication (Dockar-Dyrsdale, 1998;Federn, 1990); moreover, staff are supposed to have the inner resources to handle violence (Braxton, 1995). Parkes (2007), in her article on the multiple meanings of violence, stresses that violence is a "slippery concept." Violence can take various forms (e.g. Å kerstr€ om, 2002;de Haan, 2008;Waddington et al., 2004), is subjective, and is both context-and situationdependent. All of this makes it an unsteady concept and hence problematic to define. However, de Haan (2008) argues that a more inclusive definition of violence has an important benefit: it could allow researchers to investigate personal narratives and subjective meanings of violence in a more flexible way, as I do here.
One function of narrative for social work is its ability to construct understandings of persons and groups (Riessman and Quinney, 2005). Nevertheless, narratives about violent acts can sometimes be told in a way that makes the violence disappear (Å kerstr€ om, 2002;Andersson, 2020aAndersson, , 2020bHyd en, 2008), since one important assumption of narrative is that individuals create their own meaning through experience and structuring experience into a story-like form (Riessman, 1993(Riessman, , 2008. Therefore, narrative creates identity, cohesion in groups, and a social reality. As my interviews for this project progressed, I discovered that staff perceptions of violence could not be understood separately from-were embedded in-the situation and context in which the violence took place, including coworkers, young people, and overall institutional rules. As Clandinin and Connelly (2000) stress, the experience is not just the "here and now situation" but must be understood as part of a continuum. My in-depth investigation of Meral's narrative offers a way to show how violence connects to these other aspects.
In this article, I view Meral as engaged in processes of producing her work identity through assembling various memories, experiences, and episodes within a narrative. By "narrative" I mean a "small story" (Bamberg, 2006) concerned with character, action and plot, of the kind that people use in everyday situations to create a sense of who they are, which always happens in dialogue with the environment (Clandinin and Connelly, 2000;Georgakopoulou, 2006a). By considering how work identities can be constructed in this way, my aim is to direct understanding to "narratives-in-interaction" (Georgakopoulou, 2006b). During my interviews with Meral, both she and I positioned ourselves in various ways in relation to one and other, showing that narrative is also a co-created process (Mishler, 1999), or as Clandinin and Connelly put it, to create narratives you have to negotiate relationships (2000). A position in a dialogue can be understood as a "metaphorical concept through reference to which a person's 'moral' and personal attributes as a speaker are compendiously collected" (Harr e and Van Langenhove, 1990). Given the limited research in this area, my aim here is, by using Bamberg's (2004Bamberg's ( , 1997 positioning analysis and Meral's narrative as the analytical unit, to demonstrate how Meral describes and handles violence within a secure youth institution, stressing the nature of personal narratives and their relevance to this particular social work context. Hence, rather than concentrating on performances of violence, I highlight the subjective experience of being violated. The study addresses the question: How is Meral's work identity negotiated and formed through her experiences of violence?

Context
A secure unit can be conceptualized as a "total institution" (Goffman, 1973). However, scholars have pointed out (e.g. Inderbitzin, 2006;Kivett and Warren, 2002) that youth institutions today more often are characterized by the relationship between staff and youth, with focus on rehabilitation and care. Sweden has a long history of using residential/secure institutions for young people (Levin, 1998). Secure units for adolescents are often closed off from the surrounding community and frequently situated in remote places. While not prisons per se, they have some prison-like characteristics: locked doors, barred windows, surveillance cameras, and sometimes high walls and fences. Overall, in the Nordic countries, young people are placed in institutions not by the criminal justice system, but by the child welfare system. Globally, youth institutions tend to be characterized by the constant presence of control and safety awareness; nevertheless, compared to the UK and US (cf. Hill et al., 2007), Nordic institutions principally exhibit a treatment tradition, rather than a retributive tradition.
Staff workers at secure units come into contact with youth from various backgrounds who exhibit complex problems (Andersson Vogel, 2012;Braxton, 1995;Denison et al., 2018;Dockar-Drysdale, 1998;Salln€ as, 2009). A key aspect of secure units is that they combine control and rehabilitation, which is the governing principle for forced treatment (Andersson and Johansson, 2008;Gharabaghi and Phelan, 2011;Pazaratz, 2000). Mosseri (1998) and Braxton (1995) stress a clash between two cultures: on the one hand, the institutional structure can intensify a young person's already severe psychological state; on the other hand, the institution represents an opportunity for the young person to grow. This can be understood as the paradox of placement. Secure unit staff workers are most often treatment staff with varying backgrounds and work experience; they may have worked with young people in schools or adults in prison. In Sweden, secure unit workers must hold a secondary school diploma, and an additional two-year vocational degree in social work is considered desirable. It is fairly common, however, for staff to lack the vocational degree (Ahonen and Degner, 2014). Most workers do receive continuing education in the areas of conflict management, suicidal prevention, and Motivated Interviews (MI). Treatment staff, the group in focus here, also make up the largest SiS worker group. In 2020, SiS employed approximately 2,200 treatment staff, mostly men (SiS, 2020). Individual institutions, however, are not required to provide supervision for staff, something Braxton (1995) has argued is vital to strengthen the workplace and avoid violence. Denison et al. (2018) stress that newer and less-educated workers are most likely to be involved in situations that could lead to injuries.
Every year approximately 1,100 young people (two-thirds of whom are boys) are placed at 22 different institutions in Sweden. The majority of these young people are aged 16-18, with the girls being somewhat younger. Placement may be the result of a court order or a municipal decision, based on troublesome behavior or difficult home conditions. 3 As a general indication of their psychological condition, one third of these young people reported being subjected to psychological or physical violence by a parental figure before institutionalization (SiS, 2019). They also reported a high degree of psychological vulnerability, including severe trust issues, depression, and suicidal thoughts (cf. Denison et al., 2018). In addition, as Hicks and Stein (2015) argue, staff at youth institutions have a key role preventing further destructive behavior.

Previous research
Internationally, multiple studies have focused on violence within social work practice (e.g. Koritsas et al., 2010;Macdonald and Sirotich, 2005;Shier et al., 2018;Zelnick et al., 2013) and some studies have focused explicitly on secure units/residential care institutions (e.g. Alink et al. 2014;Andersson and € Overlien, 2018;Harris and Leather, 2012;Winstanley and Hales, 2015). In her study of violence in nursing homes, Å kerstr€ om (2002) discusses the concept of violence using the term "boundary work," recognizing that violence involves a mix of images, feelings, and assumptions. Critically, Å kerstr€ om suggests that staff may fail to label aggressive behavior as violence because doing so would position them as victims and their clients as perpetrators, thus calling their own skills into question, since violence is out of bounds and violent clients out of place in a caring context. In her study of maltreatment by staff in residential care, Attar-Schwartz (2011) reports that one quarter of the young people studied had experienced physical assault and one third had experienced verbal assault from staff. She argues that this calls into question the quality of these institutions, since these young people are society's most vulnerable. Working with young people in care places great demands on staff. Luke et al. stress the importance of treating the young people as individuals and preserving their relationship (2014). A positive relationship between youth and workers may reduce difficult behavior and placement disruptions. Heron and Chakrabarti (2002) ask why staff working at residential youth institutions are viewed as a "key ingredient" when they frequently are unqualified and, outside the institution, seen as low status workers (cf. Silow Kallenberg, 2016 for a Swedish perspective). They note the inability of staff to undertake certain key tasks in a manner that meets the needs of young people and the problematic nature of placing young people at institutions whose workers are frequently disempowered and ill-equipped to meet their needs. Hicks and Stein (2015) point out a key tension for staff: that they must both handle young people's autonomy while also taking professional responsibility for their vulnerability. Furthermore, they point out that staff run the risk of reinforcing cycles of neglect because often the contact with the family of origin is destructive. Mosseri (1998) points out the importance of institutional staff having tools to deal with violent youth; otherwise they will be put in a position of fear, which will reflect the institutional culture (cf. Brown et al., 2018). Hicks and Stein (2015) stress violence as a form of neglect, which is associated with poor emotional well-being. Pazaratz (2000) stresses empathy, communication, and structure as a core element for staff, but also observes that many who enter a youth institution as staff members cannot do the work, as they lack the relevant education and work experience. Stein (2006) also stresses low levels of staff training as one component that could lead to violence, although as part of a wider issue of failure of management, not the individual. Denison et al. (2018) underline that the use of restraint and isolation is often due to lack of training and education, and dosReis and Davarya (2008) observe that staff differ in many ways when dealing with violence at youth institutions. In contrast, Å kerstr€ om (2006), in her study of juvenile institutions, found that staff members were skillful and both embraced and resisted new policy through actions that she conceptualizes as "doing ambivalence." In addition, Flam et al. (2010) highlight the paradox of organizations (i.e. secure units) created to reduce violence, as violence becomes an element in routine work practices.
Thus, scholarship on violence and secure units does exist. However, in-depth staff stories (i.e. qualitative accounts) about violence at secure units are hard to find. This paper aims to help fill that research gap.

Theoretical starting points
As Georgakopoulou argues (2006a), the main agenda in much of narrative research is the analysis of identity, as identities are understood to be produced and performed within personal narratives. It is in "the details of talk (including storytelling) that identities can be inflected, reworked, and more or less variably and subtly invoked" (Georgakopoulou, 2006a: 125). Narratives have a beginning, middle, and end, plots and sub-plots, and are peopled by various characters who move the plot along (Riessman, 2008). That is, narratives are affected by temporality, people and their actions, and the context (Clandinin and Connelly, 2000). Narratives thus have both explicit and implicit themes, especially relevant in social work contexts where tension between care and control is a feature. dosReis and Davarya (2008) stress that examining staff perspectives is vital to learning more about violence and aggressive behavior within youth institutions. Of course, staff perspectives at these kinds of youth institutions involve working with emotions. In this connection, Hochschild's (1979Hochschild's ( , 1983 pioneering work on emotional labor and feeling rules has been extended by important work regarding workplacerelated emotions and how to conceptualize them. Bolton (2005), Bolton and Boyd (2003), and Reddy (2001) argue that employees manage and control emotions in interaction with their organizations, so it is not solely the organization that defines the emotional agenda or how to manage emotions. Finally, as argued by Riessman and Speedy (2007), storytelling involves speaking for longer stretches than is customary. Meral's narrative is long and told as one long story, without breaks, disruptions, or digressions, capturing her unique voice, including her narrative signatures and the way she speaks to and for different audiences on different levels at different times. All of these factors shape the form of the narrative as I present it in the results section. Thus, a co-created process (Clandinin and Connelly, 2000). Like Bamberg (2006), I am not interested in a narrator who is self-reflecting or searching for who she/he really is, but rather in a narrator who is engaging in the activity of narrating, i.e. the activity of giving an account. Of course, this always takes place in a constant dialogue with the environment. Meral's narrative is co-created (Mishler, 1999) and her experience takes place on a continuum (Clandinin and Connelly, 2000).

Teller-focused interviews
The data discussed in this article were collected as part of a larger research project investigating staff perceptions of violence and emotions at secure units for young people. Within this larger study, 53 semi-structured interviews and five focus groups were conducted during January and May 2017 at three different secure units in Sweden. The interview guide included five themes: violence experienced in everyday work at the unit, impact of violence, relation to the youth, emotions, and coping strategies. These were addressed using open-ended questions (i.e., "Tell me about . . ."), forming a catalyst for the narrative process. The objective was to imitate an everyday conversation. Clandinin and Connelly (2000) stress this as the researcher's endeavor to create annals and chronicles. Each interview started with the question, "How do you define violence?" One influence was Hyd en's (2014) notion of "teller-focused interviews," with emphasis on the interviewerinterviewee relation as a means of facilitation and support, since violence is a sensitive topic and surrounded by taboos and silence (Hyd en, 2008). Here, the author's previous experience working both for SiS and as a licensed psychotherapist is worth noting, as it may have offered Meral some security, helping her to speak more openly about violence.
Meral's interview was chosen as the focus of this study partly for its narrative structure and partly for its explicit focus on violence and how Meral dealt with an emerging violent situation. Further, Meral's narrative is special for the way that she explicitly describes the violence that occurs in the workplace, and does so based both on her position as a staff member and on her position as someone exposed to violence by the youth. Furthermore, she talks about the emotional impact the violence had on her, which was not prominent in the rest of the material. In previous studies (Andersson, 2020a;Andersson and € Overlien, 2018), I have given comprehensive general descriptions of the relationship of staff at secure units to violence in their workplace. Focusing on one case offers an opportunity to really understand how staff at these institutions are affected by violence and how they handle the violence they experience. Meral, furthermore, had worked at both boys' and girls' institutions. As such, her narrative represents a seldom-told story in the dataset as a whole. While this might, perhaps, make her story less generally applicable, such claims are seldom made in connection to qualitative methodological approaches (Riessman, 1993). As Clandinin and Connelly (2000) stress, a tension exists around the aspect of certainty, because of the unconditional fact that different interpretations exist. The interview can be understood as coproduced: a "complex sequence of exchanges through which interviewer and interviewee negotiate some degree of agreement on what they will talk about, and how" (Mishler, 1999: xvi). If interviews are created in dialogue, they must be understood as situated retellings, i.e. they are not expressions of the one and only story the interviewee has about a specific personal experience. Both Meral and I also positioned ourselves in relation to one another in different ways throughout the course of the interview.
Throughout the project, research ethics were given the highest consideration. The project sought and received approval from the Local Ethical Committee in Stockholm 4 but also viewed ethics as an ongoing process, considering it at every stage, from the design of research questions to the dissemination of results.

Narrative analysis
Drawing on Bamberg's suggestion that analytical work can take place on different narrative levels (2004,1997), the present analysis circles around three different narrative levels. One level highlights the characters in the story, how they position themselves and others in regard to one another, and whether they are construed as active or passive. Another level involves how the narrator positions herself in relation to the listener. And a third level has to do with how the narrator positions herself in relation to normative beliefs: in this case, beliefs about violence in youth care institutions. By recounting her experiences and arranging them in a story (level 1) and by narrating them to a listener (level 2), Meral shapes a picture of herself (level 3). Levels 1-3 correspond to a personal level (how Meral shapes her story based on a violent incident), an interpersonal level (how the story of a violent incident is conveyed to the listener), and a societal level (how the story is shaped based on the current context, i.e. youth care units). The analytical idea is thus to combine form and content: to focus on both what is being told and how it is being told, creating the opportunity to study identity not as something we have, but as something we continuously construct.

Results
Meral was a young woman who had six years of experience at secure units working with both girls and boys. The average length of work experience of all of the interview subjects was four years, so in this context, Meral can be considered an experienced youth institutional worker. At the time of the interview she was working at a boys-only unit. Before the interview began, she said she was a little nervous because the subject-violence and emotions -was sensitive. I told her that during the interview there would be no right or wrong, and if we got stuck we would help each other. I observed in her body language that this calmed her down. The following excerpt is from the beginning of the interview, where, working on Bamberg's (2006)  Of immediate interest here is not the fact that various forms of violence occurthis seems obvious-but rather that Meral clearly states that violence is a difficult and sensitive topic to talk about, which should be seen as part of her emotional work (Bolton, 2005;Hochschild, 1983). Rather than being "slippery" (cf. Parkes, 2007), violence is a sensitive subject (Å kerstr€ om, 2002;Hyd en, 2008;Stein, 2006). One way to look at violence, therefore, is as a paradox: violence both exists and does not exist, and emotions exist, but should preferably not be shown or talked about (cf. Bolton, 2005). Although I use the word "role," Meral finds it difficult at this early stage to give a clear picture of the characters that presumably are discernible. Because violence is a sensitive subject, I try to help her and we continue.
Helping her is a clear aspect of co-construction, which is a rather new way of looking at narratives in contrast to the narrative canon (Georgakopoulou, 2006b). P: If you try to think of an actual situation, how might it look? M: I mean, it could be anything . . . a youth physically attacks a staff member, it starts with verbal threats and ends up with physical violence . . . or the youth throws stuff at the staff. That could happen.
Here, despite my help, Meral finds it difficult to begin her story right away. At this point, I am also wondering whether I am being too assertive; however, I trust the interview situation. The small story perspective, too, suggests that telling is not the only important thing; refusals and deferrals also become part of the analysis (Georgakopoulou, 2006a). Meral initially finds it difficult to talk about violence connected to her own agency; instead, in a seemingly distant way, she describes what the young people do. Both Meral and the other characters in her story are still, in a way, subject-less: she is vague and her story is, so far, without clear contours. True, the young people, at this level, are positioned as perpetrators, but Meral does not position herself as a victim (Bamberg, 1997). Thus, the young people seem to be the agents in control; implicitly, Meral appears helpless. I understand this as the first level in Bamberg's positioning analysis model (1997: 337). This is a seemingly common position regarding victims and perpetrators at youth care institutions (Andersson, 2021).
At this point in the interview, I tried to approach the violence again from Meral's perspective, but instead of asking for an actual situation, I asked her if she could give me an example. My intention was the same but I approached Meral in a different way, using other words. This made the story take shape and moved us to another level of positioning: M: Yes, there was a situation here with a youth who was very threatening, verbally threatening, and engaged in self-harm behavior. He just wanted to get away from the institution. Maybe he thought, if I self-harm and make a mess at the ward, maybe my social worker [from the social services agency] will put me in an open institution. He wanted a way out. Then he began to be . . . it started with verbal threats. "I want out. I want out." It ended up with him throwing things around. He threw a fruit bowl, and he threw things, objects at the television. He lifted a chair and threw it at the television. Then both the young people and the staff took action, because we did not want him to self-harm. It ended with one staff member trying to break up the situation, but then one of the boys jumped in, trying to protect the first boy by harming the staff worker. So he [the boy] jumped him [the staff member] from behind and punched him in the neck and head because he wanted to protect his roommate. This brings us closer to the violence. We can see the various building blocks of the violent situation that Meral will continue to reason about. Clear characters are emerging, although the boys are not definitively positioned as either victims or perpetrators. One interpretation is that Meral, with me as her audience, is seeking to nuance the image of the boys (cf. Andersson, 2020b). I see this as an important formative aspect regarding her work identity: not to lock the young people in the perpetrator position. Still missing, however, is Meral herself. It is as if she is describing the situation to me from the position of an observer. In a way, she resists presenting herself, but perhaps she does not yet really trust me. Precisely by not explicitly appearing in the story, however, she leads me to imagine that she is there, experiencing the situation. I position her as vulnerable, although this is not a position she herself has yet conveyed (cf. Bamberg, 1997). We continue: P: Mm, and you were in this situation? M: Yes, I was in this situation. It ended with three of my co-workers so injured that they had to go to the hospital. One got a concussion, one fractured a bone in their arm, and the third one had mild injuries. It is actually one of the worst incidents whatsoever I have ever experienced during my time working at these kinds of institutions. P: If we stop here -what was it like for you to be in this situation? M: It was messy. I think it was messy. I got an adrenaline rush and I have memory gaps today. If you ask me to give a detailed picture of the incident, I wouldn't remember everything in detail. I have a lot of memory gaps and I have also repressed a lot, because it was a very, very difficult incident. Now, Meral clearly presents herself as a participating subject in the violent situation. Perhaps that was what was initially difficult to put into words. Taylor (2011) points out that staff need to develop a language for this, as they do not have their own language for this relationship of tension (cf. Bolton, 2005). If the language is present, it is often pushed aside by practice. W€ asterfors (2019) has described staff journal entries about violent incidents as short and meaningless, lacking emotional color and information about the direction of the violence. Now Meral's position toward me as a listener becomes clear. She is moving to Bamberg's (1997: 337) level two, where the narrator clearly positions herself in relation to the listener (audience). However, Meral is also quite quick to move around and talk about colleagues. I also think this reflects a specific aspect of work identity within secure units: it is easier to talk about colleagues than about oneself. Perhaps it would have been easier for a male worker to talk about a violent situation, given that the environment is described as masculine (cf. Brown et al., 2018).
At this point in the interview, I felt it was difficult to pin Meral down and I needed to be explicit. I asked her to stay with the situation based on her position. However, she became vague and said that she had forgotten a lot. I suspected that it was extremely painful for her to recount this violent situation. Still, I tried to stay with the emotions she had expressed: M: There was no structure, it was like a riot, it was a violent situation. You had no control over the situation and nobody knew what they were doing and you let go of one boy because he said he couldn't breathe, and when you did that he jumped on another worker instead. So, no control over the whole situation. There was no controlled situation.
Here, I began to think that listening to Meral's story was like a game of Russian nesting dolls. Each question led to a new "doll" that opened up new considerations. The metaphor expressed for me how new stories and perspectives were constantly discerned and grew based on the central violent incident Meral took as her starting point. Another view of Meral's narrative is that the different positioning levels she describes do not follow a linear order (i. e. 1, 2, 3). Instead she jumps from one level to another, which can initially make it difficult for the listener to create a clear whole (Bamberg, 1997). Her jumping about is also an expression of an ongoing formative process in which her work identity is developed and described, which is common when emotional work is described (Bolton, 2005;Reddy, 2001). That is why it is important that the whole story is allowed to develop. As this happens, we get closer to both her different positions and the other characters in her story. Perhaps what Meral struggles to approach in front of me is the position of being helpless-out of control. Her job requires her to have control: over herself, her colleagues, and the young people she works with. But Meral describes a slightly chaotic situation. One way to create structure was to ask about her feelings in that situation, and I did so. Her reply gives us an idea of how Meral positions herself to herself: M: It is hard. I mean, I don't remember that much from the situation . . . actually I only have memory gaps . . . eh . . . everyone was shouting and doing things. The noise level was very high. I tried to stop two of the boys who were to help each other and at the same time I was trying to communicate to the boys that we weren't trying to hurt them. "We're holding you to protect you." But that information didn't seem to get through to them because they were in such an emotional state that they didn't respond to information. The only thing they had in their heads was to protect each other. They didn't see that we were trying to help them, and instead they started to hurt us, and they were very physical and it was hard. Afterwards it was even worse. I had to go home, it was so hard on me emotionally that I just broke down. It was too much. I was angry. I was sad. I was pissed off because we did not have the situation under control. And I felt that the whole situation was not handled properly.
Anger and sadness now become the affective framework for creating structure in the story with me. Following Bamberg (1997: 337), I view this as the third positioning level, because it is now clearer how Meral wishes to present herself towards me as a listener and audience. She now visibly talks about others and arranges them in a narrative structure, and she is talking to someone else (i.e. me) here and now and therefore creating a sense of self (Bamberg, 2004). However, in the violent situation she describes, no one seems to really understand each other. Meral wrestles with different positions: on the one hand, the boys position her as a perpetrator, but on the other hand, she is clear with me that she is trying to protect the young people, i.e. she takes a helping position. Again, she presents a form of duality regarding her professional identity. Perhaps these different positions initially also made it difficult for Meral to start talking about this violent situation. Curious about the last sentence in Meral's reply above, I asked: P: What was not handled properly? M: The whole situation, how we acted. We knew that one staff member was a trigger. Why keep him in the situation? Why didn't he just walk away? He just ended up deeper in the fight. I got frustrated about that, because it wouldn't have had to happen if we had had control over the situation. My co-workers shouldn't have had to go to the hospital and the boys shouldn't have had to get hurt and feel bad. One of them ended up in isolation. It was a lot at once.
Here we meet a new character: a scapegoat. In the context being examined here, it is common for the causes of violence to be explained with the help of static ideas (W€ asterfors, 2019)-in terms of one person, one direction, or one form (cf. Andersson, 2021). Finding a scapegoat is often about distancing oneself from violence or responsibility. Despite finding a scapegoat, however, Meral deviates from the usual explanatory model when describing the direction of violence in this context. She is clear that the staff lay behind the escalation of violence. Here, I think that Meral, in a way, challenges her own professional identity by actually being able to think and talk about colleagues who use violence against young people. For Meral, it is thus not a matter of finding a scapegoat in order to hide from responsibility; on the contrary, by identifying a scapegoat, she finds strength. She creates a narrative of resistance and places herself on the same side as the vulnerable young people, finding a definition of neglect which is linked to the young people (Hicks and Stein, 2015;Stein, 2006). In this way, her story is also an example of a resistance narrative with connected feeling rules that can break down a destructive emotional culture in which young people are portrayed as those who initiate violence (cf. Andersson, 2020a;Hochschild, 1979). Still, the previously described affective framework nagged at me. I felt like something was missing: P: Fear? M: Yes, absolutely, but from my point of view I did not feel any fear. I was not afraid. I was never afraid. I was not scared for a second in this situation. I was angry and frustrated afterwards, and sad because the whole thing could have ended up much better, I think. I notice today that I have repressed a whole lot, a whole lot of this situation.
Here an emotional paradox is revealed. Fear is spontaneously acknowledged at first, but then Meral quickly repeats three times that she was not afraid.
Emotions not placed in a narrative structure and context are only superficially described. In order for us to capture Meral's story of fear, therefore, we need to construct such a context. It is clear that Meral does not want me to see her as scared in this situation, which in a way would be adequate (cf. Hochschild, 1983). However, the important thing is how Meral chooses to position herself in her story, not how I position her (Bamberg, 1997)-although it is also possible to view me as a co-teller (Georgakopoulou, 2006b). Meral's work identity does not seem to contain fear or victim positioning; the emphasis is rather that one must cope with the situation no matter what, highlighting which feeling rules are adequate to display (Bolton, 2005;Hochschild, 1979). Previous studies have pointed to similar aspects of fear in youth care institutions (Brown et al., 2018). It seems that fear is a taboo feeling that is sometimes ascribed to weakness. As a result, staff rarely want to talk about their own fears, even though the situation clearly invites it (cf. Andersson, 2021). One way to talk about fear implicitly may be to talk about the importance of confident and strong colleagues: an emotional detour of sorts, which makes the concept of feeling rules blurry (cf. Andersson, 2020a;Bolton, 2005;Hochschild, 1983).

Discussion and implications
This paper aims to highlight how Meral negotiates her work identity by focusing on a violent incident. Meral's story captures contextual meaning in regards to violence on a number of different narrative levels. Using Georgakopoulous's (2006 b) notion of "narratives-in-interaction," we can identify five aspects of Meral's negotiation process: (1) it is difficult to talk about violence in the current context, (2) violence occurs in the kind of institutions she works in, (3) staff use violence, but do not see their actions as violence, (4) staff are exposed to violence, and (5) staff find it difficult to describe how they are emotionally affected by violence, especially in terms of fear. Based on these five aspects, which can be seen as a set of Russian nesting dolls, we can form a picture of how Meral's professional identity is shaped and affected by violence. Her narrative is not autonomous, but rather embedded, depending on how we transport her story and reshape it across time, space, and different contexts. The concept of the scapegoat is one such example, when Meral, in her narrative, gradually dares to depart from static ideas of staff versus young people. Moreover, it is possible to highlight a number of interactional positions. Based on Bamberg's (1997) framework, I believe that what is essential is how Meral presents herself to herself, which happens in relation to me pointing to the fact that her narrative is co-created within the current context. Staff at youth care institutions must not show fear or be afraid, and this has consequences for their emotional work, and naturally also emotional costs in the course of their daily work (Bolton, 2005;Hochschild, 1983). A narrative interpretation is that fear does not fit within the framework of the professional identity in youth care institutions, because a show of fear can be exploited by young people and make colleagues less willing to cooperate. This is also an example of how feeling rules are created and developed based on interactions with colleagues and young people beyond institutional rules (cf. Bolton, 2005;Hochschild, 1983). At the same time, it is important to remember that work identity is not a static concept, but rather, as we have seen with Meral, a dynamic process (cf. Bolton, 2005). Meral's professional identity is affected by how she negotiates rules about what counts as violence in the workplace, who is a victim, and who is a perpetrator. In addition, Meral's narrative shows that there is flexibility to emotional work: it is not a static phenomenon, and as Bolton and Boyd (2003) stress, emotional labor is multifaceted. Thus, it is important to emphasize that the understanding I attribute to Meral is context-dependent and changes over time (cf. Clandinin and Connelly, 2000).
Already in 1995, Braxton observed that a typical crisis situation in residential care consisted of youth losing control, hurting staff, and damaging institutional property. In 2006, Stein stressed an urgent need to improve quality. So how far have we come? Staff members need to be prepared to work with violent youth, but it seems that this particular notion is sometimes forgotten, perhaps due to not wanting to stigmatize young people further. Using the scapegoat concept, moreover, it is also possible to adequately shift focus to the staff when it comes to violence. Therefore, this paper is a story not only about Meral's work identity but also about her co-workers, her workplace, and the young people she works with. As such, it contributes to the professional debate by extending the discussion of work identity with practical implications for the field. A key element in giving staff the essential responsibility for managing violence is keeping lines of communication open, something that could be made clearer in policy documents and highlighted in training and supervision. To understand the complexity of violence, staff need to have access to reflexive rooms and places where violence can be examined based on current context, the kind of situation that caused the violence, the form the violence took, and the intention behind it. That is, the relational aspects of violence need to be made explicit (cf. Å kerstr€ om, 2002;Stein, 2006). Staff need support and help in identifying the different emotional rules that govern their work. This again shows the need for reflexive rooms, for example professional supervision or structured collegial supervision (cf. Hicks and Stein, 2010;Taylor, 2011). I believe research investigations like this could result in the development of communication strategies for staff. This is important because emotional rules can generate emotional cultures that in the long run can be destructive for both staff and young people (Bolton and Boyd, 2003;Brown et al., 2018). As identified here, these challenges, such as how to manage fear and anger, open up an emotional landscape in which the wider academic and professional literature can assist practitioners in making sense of the emotional work they do. Only when the emotional rules are identified can staff develop strategies for dealing with the violent incidents that are part of their professional life in a professional way.