Young Children’s Narrations of Relational Recovery: a School-Based Group for Children Who Have Experienced Domestic Violence

In the UK, domestic violence (DV) is one of the most common safeguarding concerns children and young people report (CAADA, 2014). However, little is known about how children experience participation in interventions that aim to support their recovery if they have been affected by DV. This study aims to understand children’s experiences of participating in a group programme facilitated by a DVorganisation in the UK. Interviews were conducted with four children (aged 7–10) using a flexible, creative and child-led approach. A thematic narrative analysis was used, using a small story approach to narrative data. Results indicate that issues of children’s agency, choice and intersecting identities are central to not only how children experience DV but also how they experience recovery. Findings highlight the experiential and relational aspect of therapeutic spaces that can enable children to form relationships and construct meaningful identities. Conclusions suggest that children need to be consulted in inclusive ways in order to contribute to the development and accessibility of services designed to support them when they have been affected by DV.


Introduction
Approximately 25% of children and young people in the UK experience DV during childhood (Radford et al. 2013). A vast amount of literature details the outcomes of children who live with DV, revealing the extent to which children who experience DV are at greater risk of developing mental health problems, emotional and behavioural difficulties, or experiencing other forms of abuse and neglect (Costa et al. 2015;Howarth et al. 2016;Radford et al. 2013). Whilst this literature is useful, our concern is that it does little to inform services and practitioners about what can help children to recover from these experiences.
The importance of involving children in research about issues which concern them is a central feature of much of the literature which cites children's rights (UNCRC 1989) as a framework which their methodological and theoretical approaches sit within (Cater and Øverlien 2014;Nolas 2011). Children's direct participation in research is generally accepted as a way of facilitating better representation of an otherwise marginalised group (Harcourt and Einarsdottir 2011). The implementation of rights-based research has meant that researchers have been able to document children's coexisting agency and vulnerability when they have experienced DV, as children describe themselves (Åkerlund 2017;Åkerlund and Sandberg 2017;Callaghan et al. 2015;Lapierre et al. 2017;Øverlien, 2017;Swanston et al. 2014). These findings significantly challenge what has been described as a discourse of 'damage' (Callaghan et al. 2015), by offering a relational and child centred picture of how children make sense of and cope with violence. However, despite evidence which identifies a range of ways in which children's lives are affected by DV, there is still little in the form of an understanding of how children experience the services that they may access.
We located one systematic review on the effectiveness of interventions for children, young people and families (Howarth et al. 2016). However, there was only one UK study included, and conclusions were too inconsistent to draw any Hazel James is currently an independant scholar. meaningful conclusions. We conducted our own review of the intervention literature and found a similar picture, only locating one intervention evaluation published in the UK in the last ten years (Smith et al. 2015) in which the Domestic Abuse Recovering Together (DART) Programme was evaluated. Similarly, results were inconsistent across and between parent, practitioner and child feedback and outcomes. Importantly, intervention evaluations do not necessarily seek to gather children's views about the experience of their participation; rather, questions are typically concerned about effectiveness, rather than experience. There is little qualitative research which shines light on how children themselves experience group treatment and programmes in DV service contexts. Our study contributes to the evidence by interviewing children and gathering their experiences and perspectives about their participation in a group programme.
With the above in mind, Houghton's (2015) research is noteworthy. She consulted young people about their participation in DV services, concluding that: young people need more intervention earlier, fun and friendship are important aspects of receiving support, and young people need their voices to be taken seriously by professionals. However, few researchers have consulted with children themselves, particularly younger children, about their own needs and experiences of services. Pernebo and Almqvist (2016) appear to be the only researchers who have consulted with children under the age of 12. They conducted a small-scale study across two services in Sweden interviewing nine 4-6 year olds about their experiences of participating in a group programme. They found children were enthusiastic to talk about the relationships they made with peers, challenging the assumption that children who experience trauma and DV are more likely to struggle to build meaningful peer relationships (Radford et al. 2013).
We argue that the absence of young children's voices in DV service research presents problems. We suggest this absence speaks to ethical dilemmas childhood researchers encounter. Due to risk-averse ethics procedures (Scott and Fonseca 2010) and additional ethical complexities of research about sensitive topics (DV) with populations considered to be vulnerable (children), children do not tend to be consulted themselves (Holt et al. 2017). Fears around protecting children from sensitive and traumatising issues still dominate the literature, despite arguments which seek to balance children's right to participation as well as their right to protection (Harcourt and Einarsdottir 2011) and despite researchers (Cater and Øverlien 2014;Eriksson and Näsman 2012) who argue that children are indeed competent research informants and are capable of articulating their experiences and perspectives in DV research.
We also suggest that due to children's epistemic status in society, age-based assumptions position them as less competent sources of meaningful knowledge. Consequently, despite rights-based frameworks, it is likely that age-based power inequalities mean that children's perspectives are less likely to be accounted for (Eriksson and Appel Nissen 2017;Iversen 2014). This is a particular problem in research which is considered sensitive and in service settings in which adult-centric views regarding outcomes generally dominate. Eriksson and Appel Nissen (2017) argue that adult 'expertise' presents a risk of overlooking children's own meaning-making, identities and expertise about their own lives, reproducing societal control by constructing and reproducing normative ways to 'be'.
In this paper, we use the term 'experienced' domestic violence (DV) to refer to children who might otherwise be described as being 'exposed' to, or 'witnesses' to DV (Callaghan et al. 2015;Överlien and Hydén 2009). We use the term 'experience', as a way of acknowledging that children do not simply witness violence, but DV affects the whole family (Cooper and Vetere 2005). Our reason for this is that we view language as a central feature of acknowledging children as active agents (Burman 2017). Our study draws on the work of Cater (2014) and Eriksson and Appel Nissen (2017) to support our argument that children's views about their treatment, intervention and service participation need to be taken seriously. We view children as able to contribute to these discussions in meaningful ways. We take a poststructural feminist position (Davies 1991), seeking to explore power relations through which stories are told. Further, we view children as active agents, simultaneously being shaped by, and shaping their environment, in and through power relations (Burman 2017).

Aims and Objectives
This study explores children's experiences of an 8-week group programme for children affected by DV. We aim to offer meaningful knowledge to service and clinical contexts that provide support to children and families affected by DV. Secondly, we aim to contribute to the work of others (Callaghan et al. 2015;Øverlien, 2017;Pernebo and Almqvist 2017) who challenge dominant discourses of damage and passivity in the context of DV. Thirdly, we aim to extend the work of Pernebo and Almqvist (2016) who highlight younger children's views of participating in a group programme. Our study extends Pernebo and Almqvist's (2016) work by offering a narrativeinformed analysis, accounting for children's psychosocial contexts. Further, we offer a UK-based perspective and an analysis of children's experiences of accessing a service delivered in a school-based context.
The group was an 8-week therapeutic programme for children who had experienced DV. The programme was developed and facilitated by two children's workers and delivered in a school. The programme adopted a strengths-based participatory approach, aiming to build resilience and empower children by supporting them to make sense of their previous experiences of DV and facilitating the creative exploration of emotions, feelings, relationships and needs (Lee et al. 2012). The programme was manual-based, but aimed to be flexible, enabling facilitators to respond to each child's needs and the group dynamic.

Participants
Four children aged 7-10 took part in this study (Table 1). All participants had completed the same programme at the same time. The programme was delivered between March-May 2016. Children were interviewed in May 2016. Five children participated in the programme, and all five children were invited to be interviewed, although one child did not attend the last few group sessions. Although he and his mother still provided consent for him to be interviewed for this study, the interview did not take place due to practical issues. This is important to note as his absence was noted by the children who had built what they described as positive relationships with each other. Therefore, although five children participated in the programme, four were interviewed. All children had experienced DV prior to their participation in the group for a significant period of their lives.

Data Gathering
One to one semi-structured interviews, using creative methods to facilitate discussion, were conducted one week after the completion of the programme. Interviews were conducted by the first author and they were 30-60 min in length. In an effort to adopt a child-led and collaborative approach, the interviewer consulted with the children about their preferred method of participation and their preferred location (we suggested a group interview, interviews in pairs/with a friend or individual interviews, and we suggested the school or the DV service as a location. We did not think the university premises would be appropriate). Each child chose to be interviewed individually at their school. Children were interviewed during class-time or break-time as chosen by them and agreed with the school. In the schools, it was usual for some children to leave classes or leave break-time to see another professional (for instance, a mentor, a counsellor or wellbeing worker). Therefore, taking some time out was not unusual for the school culture and children's every-day experience. However, as interviews require confidentiality, we were careful to ensure teachers were aware of this. We maintained contact with them to work with them to negotiate appropriate times accordingly, to ensure minimal risk of an uncomfortable situation for the child. Children were consulted in an on-going way to ensure their consent. As the children had chosen for the researcher to visit school to do the interview, they were generally enthusiastic to show the researcher around 'their' ground (including their classrooms, the corridors and the lunch hall)this was experienced as a way of inviting the researcher 'in'.
The interviewer used flexible topic guides as a loosely semistructured interview schedule ( Table 2). As we approached interviews in a child-led way, the topic guides were given to the children on printed out cards to choose which ones they wanted to talk about, which ones they wanted to leave, and there were some blank cards to add any other topics they felt were important. None of the children wrote their own topics; rather, they responded to prompts and further questions in interviews. The interviewer had a range of creative materials available for the children to use, to facilitate discussion, if they wanted to. The children had also created a range of materials from each group session (including drawings, pictures and objects). The children were invited to bring some of their favourite things from the programme to the interview. Each child selected a range of objects, drawings and photographs to bring and they were invited to talk about the importance of them in the interviews. It was intended for this to prompt discussion about the group in a way which centralised issues that were important to the child.
Methodological literature about research with children emphasises the importance of child-appropriate creative methods to facilitate participation (Darbyshire et al. 2005;Harris et al. 2015). Such methods are suggested because it is assumed that methods should be adjusted in order to accommodate for  This table provides an outline of the interview topic guide. It was used flexibly, in response to each individual participant children's age and developmental stages. However, these proposals have been challenged because of the age-based assumptions that underpin ideas that creative methods should be used for all children because they will be 'fun' (Punch 2002). We did not assume that the child's age was an immediate indicator of their preferred method of communication and participation. Drawing materials and toys were available during interviews, but they were not used in a structured way. Children chose to draw and talk (Einarsdottir et al. 2009) or play and talk (Christensen and James 2000) and interviewer used children's drawings or play to facilitate the interview. Due to the small number of participants, interviews were not piloted, although the researcher adopted a responsive approach to interviews by remaining flexible according to each child.

This research received ethical approval from York St John University's Health and Life Sciences Faculty Ethics
Committee. Ensuring the protection and safety of children is a well-documented concern of researchers who acknowledge children's right to participation, and their right to protection (Houghton 2015). We took an 'ethics as process' approach (Frank 2004) and considered ethics as on-going, relational and contextual. Following gatekeeper approval, we sought verbal consent to participate from the children and written consent from their parent/caregiver. The DV agency acted as gatekeepers, as participants were children who were already referred to the agency and on the waiting list to receive support. As usual practice, the agency conducted risk assessments to confirm that the child was living in safety and that the abusive parent did not live with them at the time of participating in the programme and this study. Therefore, for the child's participation in this study, we sought consent from the non-abusive parent (in all cases, the mother) and not the abusive parent. Seeking consent was supported through gatekeeper access to the parent and child upon initial referral to the DV service and assessment for the group programme.
Due to the potential sensitive nature of interviews, children only participated if the programme facilitators judged the risk of distress to be minimal. In general, children were enthusiastic to share their experiences of the programme, and it was considered unlikely for interviews to be distressing. Interviews were designed to enable children to choose how they participated and which topics they discussed. Due to the risks associated with the identifiability of children and their families, confidentiality was considered seriously. The interviewer remained in contact with the DV agency throughout the project, to ensure the safety of children and discuss any instances arising which may increase risk. This was particularly necessary during recruitment; however, the DV agency provided gatekeeper protection and used their usual assessment and safeguarding protocols, therefore at the point of the research interview, risk was considered to be low.

Analysis
A thematic narrative analysis was used, informed by poststructural feminism (an aim to centralise social structures and power relations) (Davies 1991). We used a thematic narrative analysis due to our concerns about the lack of contextualisation of children's voices in DV research (Åkerlund and Gottzén 2017). A narrative approach enabled us to ensure that our analysis directly explored the context of children's lives, addressing power relations as central (Andrews et al. 2013).
Our analysis consisted of two stages. Firstly, interview transcripts were coded, categorised and thematised for initial analysis (drawing on Braun and Clarke's (2006) thematic analysis). Two researchers analysed each transcript independently and compared for similarities and differences. There were similar themes across both researchers' analysis. Secondly, we used Georgakopoulou (2006) and Bamberg's (2006) theorisation of 'big' and 'small' stories to extend our analysis and apply a narrative lens to the initial themes we identified. We took a 'small story' approach, considering the data as narrative data. We analysed the small stories of the programme that children told, in relation to the social, cultural and ideological contexts in which they were told. By 'small stories' we use Georgakopoulou's (2006) definition: these are the stories that are told in a 'non-canonical' way. In other words, stories that might be non-linear, fragmented, and not told in the context of a life-story, but stories that still matter and can tell us something meaningful about (a) an experience and (b) context(s). Narrative theorists suggest that analysing the relationship(s) between stories and the social and ideological discourses drawn on in the telling of the stories, is meaningful (Andrews 2006;Bamberg and Andrews 2004). These small stories were identified through our initial thematic analysis by adopting an iterative approach and returning to transcripts based on initial themes, to understand the context of the telling of that story (i.e. what is the relationship between the stories the participant tells?). Our final themes were identified by analysing the relationship between the small stories told, and the social, cultural and ideological narratives which were present in the context of the children's lives.

Results
Four themes were identified: (a) having fun, (b) agency and choice in service and school contexts, (c) negotiating intersecting identities, and (d) relational recoveries. In our analysis, we make reference to a 'Top Tips' poster, created by the children, who were keen for us to share their ideas. The poster is property of the DV agency and therefore potentially risks revealing the location and identities of participants. Therefore, to protect the anonymity of participants we have presented the list the children made in a table (Table 3).

Having Fun
Fun is not a narrative resource assumed to be available to children who have experienced DV due to widespread assertions in academic and practice discourse that children who experience DV have greater challenges in peer relationships and are more likely to experience mental health difficulties (Radford et al. 2013). This produces the assumption that it is unlikely 'fun' is something these children can, and will, experience. Contrary to this, children told numerous small stories of 'fun'. It was one of the most frequently discussed concepts introduced by all the children. Often, when asked about the programme activities, children did not describe the particular activity in detail, rather, they spoke about the experiential aspect of being in the group which they described as 'fun'.
Liam stated 'it were very fun' and Jo felt the most important thing was everybody 'got to share ideas together', emphasising that she experienced this sense of being together, as fun and freeing. Within the theme of 'fun', participants told multiple small stories of relationships with each other (for instance, jokes they would tell, or the value of doing activities together rather than on their own). All participants spoke about the relationships formed in the group. For Jo, it was simply 'doing it [the group]' and doing it 'together' that was meaningful. It was not what they did, but how they did it that mattered most to her. Likewise, Sophie suggested that having fun was like a day without the 'blah blah': Interviewer: so what was it like to be able to make such a special friend in the group? Sophie: it was like a day without the blah blah Interviewer: a day without the blah blah? Sophie: yeah, I could forget about it all and just have fun.
Like Jo, Sophie drew on the concept of having 'fun', in order to articulate what making a new friend in the group, was like. Additionally, Sophie's use of the term 'blah blah' might seem relatively 'child-like'. This articulation of 'fun' could be understood as a negotiation and expectation of the interviewer's position as an adult who, based on the socio-political space of the school, she might have expected to be disinterested in what she had to say. However, when she was asked if she could explain what she meant, she said: 'It's all the talking [in school]no you shouldn't do this, yes you should do that, no you can't…'. Sophie re-negotiated her position as an active agent through the relational space of the interview. Firstly, this enabled her to articulate her sense of the constraining rules and boundaries of school and to describe the importance of a space which enabled fun. Secondly, this interaction with Sophie also points towards a relational negotiation with the interviewer. It could be said that the relational space enabled her to (re)negotiate her position as 'child' and articulate what she intended to communicate.
We view the telling of these small stories as a context-specific and relational (re)negotiation of discourses which propose children affected by DV may not be able to have fun in a meaningful way. It could be assumed the social and academic discourse of vulnerability of children exposed to DV means these types of programmes 'should' be filled with distress and risk. From a narrative lens (Bamberg and Andrews 2004), this canonical expectation is disrupted by the small stories of the participants. Our analysis does not minimise or mitigate these stories of vulnerability; however, our small story approach has identified many stories of fun and play, suggesting these themes are central to children's recoveries; specifically, the accessibility of fun in, and through, relationships.

Agency and Choice in Service and School Contexts
The theme of agency and choice was central to how children narrated their small stories of participation in the programme. Whilst children described many instances of having fun, there were also coexisting stories of choices that appeared to be constrained. These intersecting stories reveal the complexity of enjoying the group, whilst negotiating the constraints on the choices available to children in the context of school and a DV service. Children described having fun and making friends, highlighting multiple positive aspects of the group. However, they also highlighted that they did not wish to be treated 'like babies' (Table 1). This statement links to their consistent references in the interviews to choice and power. Not wishing to be treated like babies is indicative of the relationality of their agency and choice regarding how they experienced the intervention. Jo articulated that although she enjoyed the group, she was aware, particularly at the beginning about the lack of information made available to her, which was a source of worry. The scariest thing was meeting new people because all I was told was that there would be other children going, I didn't know who was going. All I knew was that it was you guys doing it but the only person I knew was [the children's worker] because she used to work with my cousin and she didn't [pause] but, but my cousin doesn't need it any more so she came to do it with me.
She followed this articulation of feeling uncertain, with 'I enjoyed it though, I enjoy new schools', before explaining that 'I felt happy that I had the choice about every time I wanted to come'. Jo articulated two contradictory small stories here; one of worry about meeting new people and not having access to information, and another of having choice and enjoying the very unknown newness she described feeling worried about. According to narrative theory (Georgakopoulou 2006), these stories are told in a non-canonical and fragmented way, but rather than presenting problems for analysis, the interaction between these stories is precisely our point of interest. We suggest this reveals the importance of children making active choices regarding their participation in services. It also reveals the relationality of agency and the complexity of articulating this to a researcher who is both part of a school (an adult in school context), part of the DV service (a researcher attached to the service she has participated in), yet also is not fully situated in both, and is simultaneously a curious person who wanted to hear her stories. Jo's participation in the group was a positive experience, but she still recognised her need for a sense of autonomy about how and when she participated. Sophie had a similar experience: Interviewer: do you think that you wanted to come [to the group] yourself? Sophie: my mum. My mum made me come… I just wanted to see what it was like yeah, but I wanted to go see some friends and stuff, after school and stuff. I wanted to skip some to see my friends, but my mum said no you have to go every week like every other person like every after school club, you've got to go. I was like uuuuurgh mummy Interviewer: That's tough? So OK [pause] so maybe sometimes there were times when you didn't really want to come because maybe you wanted to play with your friends, but your mum made you come Sophie: yeah Interviewer: so how did you feel when your mum made you come? Sophie: [uses an 'upset' toy bear to show the expression] Interestingly, Sophie was the most expressive about her perspective regarding her positive experiences of the group. She expressed fondness of the group and sadness at the group ending. However, she positioned these feelings alongside her sense of lacking choice about her participation. For Sophie, it was important she could come in her 'own time'. Her choice and agency were reoccurring topics of discussion in the interview. When asked about her ideas about a better way for it to be, she explained: If it was like in school times or something, and we could go when we wanted to go. If we were in lessons and we wanted to go, we could just say erm, I need to go to the group. Like so we could come in our own time… it isn't good when people force you to go, is it?
Likewise, Jo also discussed her initial worries. She said: We could have had a show around as a starter to see where we were going to be and who we were going to be with. Stuff like that… mhmm [on the first day] I didn't even know where I was going. I just arrived at [the school] and I just saw you and [the facilitators] and I was just like OK I'm really excited now but then I was nervous as well.
Although Jo and Sophie expressed their concerns, they communicated their anticipation and curiosity about the group, and suggested possible ways to improve the group. Jo's above articulation: 'I'm really excited now but then I was nervous as well' speaks to the complexity of the intersecting positions we describe here: these small stories of attending the programme are filled with newness, unknowns, relationships and fun. Notably, children actively negotiated their social positioning through the narration of their small stories. (New) Sociology of childhood scholars have critiqued the positioning of childhood as a time of development, dependency and freedom, arguing for the need to view children as active agents (Wyness 2012). Scholars who challenge the binary positioning of childhood and adulthood have also challenged the framing of childhood as a time of becoming rather than being (Burman 2017;Twamley et al. 2017). Power relations that underpin this categorical view of childhood and (lack of) agency, position children as less competent to offer views which are taken seriously (Burman 2017). Consequently, ideological discourses of dependency and becoming (rather than being) minimise the importance of children making choices. Our analysis of children's narrations of small stories suggests that the relationality of choice and agency are crucial to consider in the delivery of children's programmes.

Negotiating Intersecting Identities
DV is typically a non-normative childhood experience, in that it diverges from what children 'should' experience in childhood (O'Dell et al. 2018). Consequently, children and families affected by DV are typically positioned as 'other' in social discourse. As highlighted by Morrow (2011), a 'broken home' discourse can be profoundly impactful. This marginalisation through normative ideologies about risk, family life and vulnerability, was evident in the way children narrated their small stories of the programme and their everyday lives. In the interviews, children shifted the focus of their talk away from the programme and towards other aspects of their lives. This suggests that although the group was experienced as 'fun' and beneficial, it was also experienced as a space in which they negotiated their identities as children in families where there has been DV. Children's methods of shifting their talk to other topics speaks to the need to recognise children's other intersecting identities. Jack's references to his hobbies suggests it is important to consider all aspects of children's lives and identities, in spite of the DV, which led him to access the service.
Jack: but if [the group] was on a Thursday I'd have to rush home and then to the gym now Interviewer: ahh yeah, now you go to boxing after school then after school wouldn't really be that great? Do you think that it would have been better during school time or do you think after school still would be good? Jack: during school Interviewer: so if [the facilitators] did the group again for other children? Jack: during school, so I could play on my X Box for longer at home.
All children highlighted their need to be treated as individuals with lives that exist outside of DV. From a narrative small story perspective, the fragmentation and non-canonical way of telling these stories would seem like a disorganised structure of narrating experiences of the programme that children knew the interviewer was interested in. However, as we are interested in what this non-canonical storytelling can reveal by analysing the relation between the stories told and the context of their telling, there is something meaningful in acknowledging how children negotiate ideological discourses about childhood, vulnerability and family life. Children were keen to talk about the different things in their lives that were, and continued to be, crucial aspects of their identities. They did not want the group (and research interviews) to disrupt other things in their lives, such as lessons, seeing friends, and other activities (such as football, boxing, WhatsApp and the X Box). This can also be seen in their 'Top Tips' (Table 1) poster in which children suggested that DV professionals should 'talk (to children) about other things, not just things at home'. Therefore, we suggest that it is important for children's intersecting identities to be recognised and valued in DV service contexts.

Relational Recoveries
With little or no prompting, all children spoke about relationships in their lives, suggesting it is important to consider the significance of these relationships and children's psychosocial contexts, when they access services. First on Jack's agenda was to tell the interviewer about the significant relationships in his life: 'I've got lots of uncles… I've forgot them cos I don't see them very often'. However, unlike the other children, Sophie did not discuss her family members, apart from when she positioned her mother as 'forcing' her to attend the intervention. In fact, she positioned friendships as central, as she described the importance of her new friendship with Jo, and stated 'you can't break friendships can you?' It was clear that the children had made meaningful relationships with each other. Each child described the value they placed on their new friendships, again, highlighting the psychosocial and relational aspect of their experiences of the programme. Given that the children made friendships within the group, the group's ending was particularly significant to most of the children, who expressed their wish for the group to last longer.
Interviewer: we said that the best thing about the group was Liam and the worst thing for you was leaving [pause]. Do you think if the group were to carry on then you'd keep going? Jack: yeah Interviewer: Ah, how long do you think the group could have gone on for if it was the right way for you? Jack: Forever Jack also explained 'the best thing was meeting Liam… Leaving the group was the worst thing… Liam is my best friend now'. Likewise, Jo explained that the group ending was a celebration, but she also experienced conflicting emotions because she did not want to leave. 'When I graduated I felt happy and I didn't want to leave… but I have got my book and my pen and now I can just write.'. Sophie also attributed much of her meaningful changes to her new friendship with Jo: 'I feel like a different person… I didn't have any friends, but [Jo], she changed that…'.
These stories of relationships were central to each of the themes we identified, but it is also a theme that needs to be recognised itself. Our small story analysis recognises the relationality of both being in the group (relationships with their peers), and when relationships outside the group influenced the experience of the programme (for instance, families and others outside of the immediate group context). Children's narrations of these small stories reveals the ways in which children de-individualise their recovery trajectory by actively positioning themselves in multiple relationships. Social and practice discourse assumes that children who experience DV are likely to struggle to build meaningful relationships (Radford et al. 2013) and that recovery can be marked by the improvement of individual symptomology (Lee et al. 2012)therefore overlooking the relationality of recovery. We do not suggest that symptomology should be ignored, but we do propose an alternative extended picturethat recovery can be viewed as relational too.

Discussion
Our findings point to the experiential and relational aspect of therapeutic spaces that enabled children in this study to form relationships and construct meaningful identities. In our discussion, we explore the value of analysing the contexts in which children narrate stories of fun and friendships. We also highlight the relationality of children's agency and choice in DV service and school settings. We extend the theorisation of children's coping during DV as relational and agentic, and suggest that children's recoveries can be described as relational and agentic too. Our results have implications for a range of practitioners, services and researchers who aim to support and improve the lives of children and families who are affected by DV. Therefore, we do not tailor our discussion for a specific discipline or practice. We take an inter-disciplinary and crossdisciplinary approach, due to the range of ways in which our results contribute to research, practice and theory.
The small story approach (Bamberg 2 0 0 6 ; Georgakopoulou 2006) to analysing narrative data means that we have contextualised children's voices and understood them as relationally and contextually produced (Andrews 2006). This approach to narrative analysis has enabled us to identify children's (at times) fragmented, contradictory, non-canonical, but meaningful experiences and perspectives in relation to the programme they participated in. Narrative theorists have highlighted that analysing the relations between contradictory and fragmented stories is not only fruitful, but it is precisely within these spaces that meaning and stories are produced (Bamberg and Andrews 2004) Stories of fun and friendships co-existed with stories of constrained choices and complex negotiations of power. Coexisting relational positions and ideological discourses about childhood, recovery and family life, informed how children described their participation in this programme. We have identified the importance of choice, agency and enabling space to explore and value intersecting identities in a service context.
The group programme was delivered in a school, typically a space in which power relations are adult-dominated (Devine 2000). Children's articulation of friendships and fun can be understood as a negotiation of power relations within which children's lives are situated. Our interpretation of these small stories of fun and friendships is based on the assumption that schools are spaces in which children are often positioned as 'objects of schooling' (Rosen and Twamley 2018, p. xi), children may be 'infantilized ' (ibid. p. 245), and children are positioned in a 'needs' discourse (Devine 2000) dependent on adults. Our findings suggest that in school spaces or DV service spaces, such power relations can be (re)negotiated by children. Our findings support the notion that children play an active role in not only how they negotiate relational encounters and contexts during DV (Åkerlund and Sandberg 2017;Øverlien & Aas, 2016), but in their recoveries too. Children resisted the ideological and social discourse that children affected by DV cannot have 'fun', by co-constructing spaces and narrating small stories in which not only was 'fun' allowed, but fun could be had in, and through, meaningful relationships.
The theorisation of children as agentic in DV contexts challenges the outcomes evidence which is based on a deficit model, and positions children who experience DV as lacking capacity to cope effectively (Callaghan et al. 2015;Överlien 2017). Children in our study described experiences of having fun and making friends. This extends theorisations of children as agentic not only during their experiences of DV but during their recoveries too. Children asserted their positions as active agents, with identities that extend beyond that of being a child from a 'broken home'. This adds to Pernebo and Almqvist (2016)'s findings. Children they interviewed also highlighted having fun and developing peer relationships as a crucial aspect of their experiences of participating in a group programme. Our study extends Pernebo and Almqvist's (2016) findings: not only did children form meaningful peer relationships, but their experience of activities, their experience of beginning and ending, and their articulations of agency and choice was relational.
This emphasis on relationality also revealed the importance of recognising that although living in a family that has experienced DV has led children to access service, it is not the only thing that constitutes their sense of self. From a narrative lens (Bamberg and Andrews 2004), the vulnerability and brokenness associated with living in a family affected by DV is not the only narrative through which participants narrated the 'self'. We suggest that children narrated multiple intersecting identities through their numerous small stories: those directly related to the programme, and those which seemingly were not related to the programme (i.e. the X Box, the gym, their family relationships). There are two points in relation to the implications of this finding. Firstly, we suggest it is crucial to acknowledge the violence and abuse that children have experienced and offer a means through which to explore and express feelings around their experiences. Additionally, we argue that it is meaningful to offer children choices about how and when they talk about their experiences, and to explore and value other aspects of their identities too.
Choice and agency were central features of how children described their experiences of the group programme. They described positive experiences of peer relationships, and described the group as fun and something they would miss.
However, these experiences were also articulated alongside narrations of constrained choices: choice around participation, accessibility of information about the programme, and how and when they could attend. How and when children engage with services is a timely issue for discussion, given the prioritisation of early intervention in the UK (Howarth et al. 2016). This early intervention agenda has also been extensively critiqued because of the way in which it works to simultaneously centralise and marginalise parents and families who already live on the 'margins' (Featherstone et al. 2014;Macvarish et al. 2015). We suggest the early intervention agenda works on an individualistic level, and does not centralise the psychosocial and relational contexts children described as fundamental to their experiences of participation.
It has been suggested that children's readiness to talk and engage is one important consideration (Iversen 2014). Whilst readiness is important, it should be acknowledged that this assumption suggests 'failure' to engage might be due to lack of readiness. We argue that this is an individualising way to view engagement with services, and it lacks consistency with how the children in this study described their perspectives. Engagement must be considered contextually. None of the children in this study referred to feeling ready, but they all spoke of choice and power, suggesting that how they were approached and informed about the programme was central. Information about programmes and services, and treating choice and consent as relational, on-going and open to the possibility of change, is crucial. Stanley and Humphreys (2015) recognised that children typically have little choice about their referrals and involvement with services, whilst Cater (2014) argued that children's choice about participation in services could be more powerful than participation itself. Houghton's (2015) work with young people in Scotland also revealed that young people said that professionals should trust them to talk about DV how, and when, they wish. Our study supports Houghton (2015) and Cater's (2014) arguments. Participation, agency and choice are central themes. However, agency as a concept, is argued to be better understood relationally and contextually (Valentine 2011). We therefore suggest that readiness to engage should not be down to the child only; rather, considering children's psychosocial contexts is crucial. Sophie's explanation, that sometimes her mother 'forced' her to attend the group, demonstrates this. Whilst Sophie was keen to express the positive impact the programme had on her life, she still emphasised choice. She explained that choice meant not only choice about consenting to participate, but that choice and consent was on-going. She suggested a drop-in group would be best. Although drop-ins might be impractical for services, her message is important. Cater and Øverlien (2014) highlight that it is a rare experience for children to be taken seriously because they are rarely considered competent informants. We therefore take Sophie's message seriously, that choice, agency and consent are on-going, relational and dynamic. It is crucial for adults in children's lives to consider these issues and explore more, how they can use their epistemological positions as 'experts' to facilitate children's choices in contexts within which they typically have little choice.

Limitations and Further Research
We recognise that this study has limitations. We interviewed four children, and although we do not claim generalisability, this is still a small sample size. We also do not claim to shine light on the effectiveness of the programme, rather we were concerned about how children experienced it. Participants in this study were all white British children and do not represent a racial, cultural or ethnically diverse account of how different children experience a DV service programme. This was not a purposeful sampling decision, rather it was based on the children who were accessing service at the time of recruitment. The invisibility of minority group children, and the normativity of a Westernised culture is widely critiqued (Burman 2017;O'Dell et al. 2018). This normalising mechanism through which Eurocentric norms are established, and others are pathologised, is not sufficiently addressed in DV research. Future research should seek to include the views and experiences of a more diverse range of children affected by DV. There is a need to extend understandings about how children experience DV services and interventions. We suggest that researchers should seek to present contextualised analyses of children's perspectives by considering the contexts in which children's voices are produced. Finally, we argue children offer significant expertise about their own livestherefore greater participation of children in the design of research methods and support services is needed.

Conclusions
It is crucial to centralise children's choices about how and when they access services, participate in programmes, and talk about their experiences of DV. The peer relationships children form can be central to their experience of participating in services. Children in this study valued space to form meaningful relationships and have fun, emphasising that having fun 'together' was crucial. Further to this, we argue that not only are relationships important to children's recoveries, but that recovery can be described as relational. Although children described positive experiences, it is also the responsibility of researchers to hear the complexity of how positive experiences are articulated too. We have found a narrative analysis an appropriate methodology to support the identification of contradictory and complex speaking positions. Whilst all participants described fun, friendships and positive benefits of the group, they also articulated constrained choices and negotiations of intersecting identities which were situated in adult-child power relations, in contexts that do not often promote the epistemic status of children. We therefore suggest that children's services should consider children's psychosocial contexts and issues of relational power, agency and choice when designing and delivering programmes for children affected by DV. These are particularly relevant conclusions for two issues: firstly, the concept of relationality challenges readiness to engage discourses by offering an alternative view, that it is not only children who we should consider, but readiness should be considered relationally by exploring children's intersecting identities and the broader social contexts of their lives. Secondly, our findings point towards a relational view of recoverythis offers a different perspective to otherwise dominant individualising notions of recovery trajectories.
Our application of a small story analysis offers a substantially different way of viewing stories told by children in nonlinear and contradictory ways. Rather than assuming children's contributions lack consistency because of their age, we take this non-linearity as a point of interest for analysis. This approach has supported the tracing of meaningful power relations. Our analysis of these power relations has enabled us to engage more with the notion of negotiating relationality in contexts which are typically shaped by individualising notions of recovery and distress. We conclude that younger children can contribute in meaningful and significant ways to research such as this, and that narrative methodologies can support the analysis of complex (small) stories that are narrated in nonlinear ways.
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.