Pedagogical devices as children’s social care levers: A study of social care workers’ attitudes towards boarding schools to care for and educate children in need

It has been proposed that boarding schools in England can be used to provide a stable education and care environment for vulnerable children in need, and the government is expanding their use. However, for vulnerable children to be placed in boarding schools, social workers will need to be willing to contemplate boarding as a viable care option. In this study we interviewed N = 21 social care practitioners including directors, senior and middle managers, frontline social workers, social worker-academics and family support workers who work with vulnerable children. Using thematic analysis of the transcribed interviews, seven major themes identified a range of issues and concerns held by social care workers about placing vulnerable children in boarding schools. We present these themes and consider the issues that will have to be addressed prior to changes in policy and practice. The study concludes that many of those within the social work profession are unlikely to consider boarding as an intervention for children in need. Further research in this area is a matter of urgency. further work was [sic] required to align the priorities of education and social care professionals working with young people and their families, and to challenge the view held by many social care professionals — as well as young people and parents — that boarding schools were only appropriate for those from more privileged backgrounds.


Introduction
The use of boarding schools for improving the attainment of socially disadvantaged youth is a matter of international education policy, with examples of its practice in the USA (Curto & Fryer, 2014), South America (Mexico) (de Janvry et al., 2012) and France (de Behaghel et al., 2017), and recently emerging in England (Boarding Schools Partnership, 2017). Families and young people unable to afford the fees but aspiring to attend a boarding school may compete for scholarships or bursaries. In some circumstances, social services might consider using a boarding school as an alternative to placing a vulnerable child in care. The number of children being placed in care is rising in England, and in Wales children are one and a half times more likely to be in care than children in England (Elliott, 2017). Positioning boarding schools within a broader package of care and intervention for vulnerable children is a contentious issue, yet the British government is currently supporting the expansion of the use of boarding schools (Boarding Schools Partnership, 2017). In order for this to be a success, social services will need to work in close partnership with educators and specifically schools. The success of such an initiative will depend largely on a collaborative attitude and supportive alliance between education and social service workers. In this article we present the findings from a qualitative research study that explored social and care worker attitudes in England towards the use of boarding schools for vulnerable children in need or those on the edge of care. The article is an attempt to clarify what work might need to precede any significant shift in policy towards the use of boarding for vulnerable children.

Literature review
Recently, there have been several attempts to test empirically for the academic and emotional benefits to vulnerable children of using boarding schools as part of a combined educational and social care intervention. Such interventions can be considered as using pedagogical devices (Bernstein, 1990) (i.e. a boarding school) as a 'social care lever' (mechanism to enhance the social prospects for children). The intent or aim is to move children in need from challenging family circumstances into proposed secure and stable socio-environmental conditions, in an attempt to create opportunities for vulnerable young people. However, such attempts have not hitherto been without their challenges. Two attempts at randomised control trials have resulted in failure to recruit a sufficient number of participants, and a third within-subjects design study also failed to recruit sufficient participants (Murphy et al., 2017). These studies assumed that boarding schools can raise the educational chances for vulnerable children and develop their emotional wellbeing and subsequent life chances (Boarding Schools Partnership, 2017). However, it appears from recent efforts to research this claim that a complex relationship between social care services and the boarding school sector exists. Social workers' concerns about inequality and poverty may be a factor in using boarding schools for vulnerable children, rather than mitigating the reproduction of such inequalities. Understanding the relationship between social care workers and the boarding school system is an important first step to explore prior to developing policy in the field.
For over 15 years, successive governments in England have experimented with the use of boarding schools in the provision of social care support and educational opportunities for vulnerable children (Department for Children, Schools and Families, 2007). In England and Wales, children are classed as vulnerable when they are recognised as a Child in Need through a social worker's assessment. More serious is a child that is at risk of serious harm and registered on a Child Protection Plan. A child can be considered on the 'edge of care' if there is a strong likelihood that removal from the home is being considered. Proposed guidance on using boarding schools as an alternative to statutory residential care, or as part of a residential care package, has been documented previously (Morgan, 2007). Morgan (2007) suggested that boarding school places could be considered by social care professionals for vulnerable children; nevertheless, use of boarding schools for children on a Child Protection Plan and Children in Need has remained low throughout this period (Nietmus, 2017). The low uptake of boarding places suggests social care services' concern regarding the potential for widespread use of boarding schools.
The Department for Education has advanced interest in using boarding school places for vulnerable children through the development of the Boarding Schools Partnership (2017), reflecting the government's commitment to 'enabling all children regardless of their backgroundto reach their full potential' (Department for Education, 2017). The aim of the Boarding Schools Partnership is specifically directed at increasing the use of boarding schools by local authorities as part of social care interventions to improve the educational outcomes for Children in Need. According to Nietmus (2017), there were (as of December 2017) only around 100 vulnerable children (i.e. on a Child Protection Plan or a Child in Need) in funded boarding school placements. The Department for Education, however, intends to increase this number to around 2,000 children in the next few years. Significantly, this commitment exists with very little available research about the costs, benefits and effectiveness in educational and social/emotional outcomes, or how longer-term life chances are improved. Such a dearth of available evidence begs the question of whether the key underlying driver for the Boarding School Partnership is an ideological rather than an evidence-based attempt to close the attainment gap.
The only available evidence is a small selection of pilot studies. For example, in the evaluation of Boarding Provision for Vulnerable Children Pathfinder (the Boarding Pathfinder), schools that 'signed up' to the project did so with 'the commitment to offer children and young people from the more difficult circumstances educational opportunities which they might otherwise not have had' (Maxwell et al., 2009, p. 5). The Boarding Pathfinder project worked with 10 local authorities and considered places for more than 75 young people, yet only 17 actually started at boarding school. In the period of the evaluation, just 11 remained at boarding school. Boarding provision included schools in both state and independent sectors. During the intervention, local authorities attempted to offer the boarding placement at the point of transmission from primary to secondary school. In particular, while there was goodwill between schools and local authorities involved in the project, incongruities in the expectations of schools and local authorities were discovered. Whereas schools were focused on the individual child, local authorities looked for 'a placement option that would support vulnerable families' (Maxwell et al., 2009, p. 7) (our italics). Nevertheless, the same report acknowledged a positive role for supporting children through the provision of local authority places in boarding schools; a number of issues remain unresolved, such as the identification of children, preparation for boarding, support (for children and their families) during school holidays, as well as funding for the placements. Similarly, whilst local authority managers might support the intervention, social workers-a key point of contact for vulnerable children and their families in local authorities-were not always aware or supportive of the boarding school placements (Maxwell et al., 2009;Lombard, 2011).
Where boarding school acts as a social care intervention, it intends to lead to better education outcomes. There is strong evidence for the low academic achievement of Children in Need (Sebba et al., 2015) and looked-after children (Mannay et al., 2017). The gap in educational performance between looked-after children and all others continues to grow throughout their school career (Sebba et al., 2015), which is a challenge for both educators and social workers. Some evidence to support the use of boarding comes from Studies of Early Education and Development (SEED) schools in the USA. Most SEED schools have a 5-day-a-week public boarding programme that offers a holistic education experience and curricula that address both academic and non-academic development. Findings suggest that socially disadvantaged children can benefit (Curto & Fryer, 2014). Using data from the District of Columbia Public Schools (DCPS) database lottery, lottery winners and losers were matched on a series of demographic variables, mathematics and reading scores. For each year a child is at a SEED school they achieved gains with effect sizes of around d = 0.211 in reading and d = 0.229 in mathematics. However, Curto & Fryer (2014) caution that these effects might be driven by gender, as in reading girls (d = 0.382) significantly outperformed boys (d = À0.138) and for mathematics girls (d = 0.265) again outperformed boys (d = 0.037), suggesting a cautionary interpretation of this result for boys.
Similar findings were reported in a study involving a treatment group of 244 children compared to a control group of 137 randomly assigned students in the French 'boarding schools of excellence' initiative. Disadvantaged students were in smaller classes and spent longer in class compared with the control group (de Behaghel et al., 2017). Large gains were reported in regression coefficients for mathematics tests (0.280), particularly for students who were already in the top third of students 2 years earlier. Other notable social effects on new boarders included a reduction in time watching television and poorer emotional outcomes at the end of the first year of boarding, although this was mitigated at the end of the second year, when students reported 'higher levels of motivation' and 'wellbeing had caught up with that in the control group' (de Behaghel et al., 2017, p. 37). There were different findings with respect to the effect of elite non-boarding high schools on vulnerable children in Mexico City, where students who were more vulnerable showed less academic progress and higher dropout rates than their more advantaged peers (de Janvry et al., 2012). This suggests that boarding may have some comparative benefits for vulnerable children.
One factor thought to contribute to better progression for Children in Need in boarding schools would be the stable care of the environment, thought to be important in supporting children's academic development (de Behaghel et al., 2017). Sebba et al. (2015) report significant variation in outcomes for looked-after children (three groups: those placed in care before the end of Key Stage 2; those placed in care after Key Stage 2; and those in short-term care for <12 months by the end of Key Stage 4), Children in Need/Child Protection Plan children at the end of Key Stage 4 but not in care, and a comparison group of children not in care and not in need. Those children who were placed in long-term care arrangements early in life made more progress than groups of looked-after children or Child in Need/Child Protection Plan children. For example, those entering care after Key Stage 2 performed worse in comparison to those who entered care before the end of Key Stage 2. The 'relative educational performance of children in need. . . showed a decline over time' (Sebba et al., 2015, p. 26), suggesting that Children in Need exposed to prolonged unstable care environments might be very significantly affected by this in terms of their education.
Advocates of boarding schools argue that the stability of the boarding environment could improve the Child in Need/Child Protection Plan child's chances academically. There is support for the provision of boarding (for disadvantaged, vulnerable, in-care children) from government and specific charities in the UK (e.g. the charitable education trusts including Buttle and Royal National Children's SpringBoard Foundation). One study of 52 looked-after children in Norfolk who attended a number of different boarding schools reported educational, academic and socio-emotional benefits to the children, with some outcomes being 'measurably enhanced' through this experience (Garrett et al., 2018, p. 6). Despite the financial costs of boarding that the local authority incurred, there were lower costs for the local authority associated with no longer being a child in need (Garrett et al., 2018). Conversely, removing a child from their home, family, attachment relationships and/or community environment may prove challenging to their emotional wellbeing (Curto & Fryer, 2014), and as for looked-after children, removal may also be a risk factor with regard to educational outcomes (Sebba et al., 2015). The experience of loss by separation, due to attending a boarding school, may lead to changes in identity that impact a child's sense of belonging-both at school and in their community. Curto & Fryer (2014) provide a review of 'loss' of identity. These factors might reduce social workers' motivation to recommend the use of boarding schools.
There is an absence of conclusive evidence supporting boarding for vulnerable children in England as an effective intervention. Based on previous efforts to trial such an intervention failing to recruit sufficient numbers of participants through social worker referrals, it is suggested that a complex relationship exists between social care service workers and the boarding school system. If boarding schools are to be made available as a social care intervention that provides stable education and care for vulnerable children, then it is first necessary to understand the attitudes of social workers towards boarding schools. For that reason, the present study sought to examine the attitudes of professionals involved in the provision of social care services to vulnerable children and those involved in the training of social workers towards boarding schools.

Design
We employed a qualitative method to explore the research questions in this smallscale study. We used semi-structured interviews that invited participants to talk about their attitudes and views on the use of boarding schools as a social care intervention. All of the interviews were audio-recorded using digital recording technology and were securely stored. The interviews were transcribed and prepared for analysis. The study received ethical approval from a university research ethics committee. In keeping with ethical approval, we removed the identifying features of our participants to maintain their anonymity. We elected to use numbers for the participants as the field associated with our study is narrow, and this seemed to offer the maximum level of anonymity (although it can limit the sense of personhood).
Pedagogical devices as children's social care levers 5

Recruitment procedure
We recruited our sample using a combination of approaches. First, we used purposive sampling by requesting access to details of the 157 local authorities previously contacted as part of the Boarding Chances for Children project funded by the Education Endowment Foundation. From this, we were able to identify local authorities that had (a) responded positively and engaged in the trial (n = 2), (b) responded positively, showed interest but not engaged in the trial (n = 49), or (c) not responded to the invitation to engage in the trial (n = 106). Having identified the appropriate named contact within each local authority, we sent invitations to participate via email to all local authorities listed in the groups above. The email included a brief introductory outline and an overview of the project, a more detailed information sheet and a consent form. We received a positive response showing interest in participating from five local authority areas in England. Of these, none had actually participated in the Boarding Chances for Children project, one had responded positively and showed interest but not engaged in the trial, and the remaining four had not responded to any prior contact about the trial. Respondents to our initial contact either became direct participants or passed us on to other staff members. Further to this convenience sample, we used a snowball sampling approach following leads and recommendations from participants already recruited. Recruitment took place between June and August 2017, and represented a range of local authorities in England.

Participants
We interviewed 21 participants using a combination of face-to-face and telephone interviews. A recording device failure meant the data for one participant were excluded. The remaining sample is N = 20. Participants spanned the full range of professionals in social work practice, researchers and academics involved in teaching and training social workers. We included directors of children's services (n = 2), social work managers (n = 3), experienced social workers (working in child protection, looked-after children services) (n = 2), newly qualified social workers (n = 4), family support workers (n = 5), virtual school head (n = 1) and qualified social worker-academics involved in teaching on social work degree programmes (n = 3). Most participants were female (n = 13, 65%) and half of the participants were in the East Midlands region (n = 10, 50%). See Table 1.

Data collection
Social care professionals' attitudes, views and perceptions were explored within the interviews by asking questions on the following areas of interest: (1) qualifications and experiences in a professional capacity; (2) social worker education and training; (3) prior knowledge of research studies into boarding for vulnerable children; (4) thoughts about placing a Child in Need/Child Protection Plan child in a boarding school; (5) approach to working with Children in Need/Child Protection Plan children; (6) considerations of boarding school as a positive intervention for a Child in Need/Child Protection Plan child; (7)  All interviews were anonymised and transcribed verbatim by two separate transcribers. These were then uploaded through a secure file portal for the manual coding of the data and identification of recurrent themes by one researcher. Interviews lasted on average 28 min (range 18-39 min), giving a total of 9 h, 20 min of interview and a total of 68,000 transcribed words.

Data analysis
We used a form of thematic analysis (Braun & Clarke, 2006) to code the transcripts, as this approach is useful, adaptable and flexible. Coding comprised five iterative phases: (1) interviews were transcribed and reviewed for initial coding; (2) field notes were drawn on to add codes to transcripts and sections of transcript were highlighted to identify relevant quotes supporting themes; (3) themes were cross-referenced with regard to our research objectives; (4) a full list of themes was prepared and crosschecked by the second coder, identifying the transcript themes and evidence; (5) related themes were clustered and unique themes separated, retaining the identity of their constituent subthemes. After initial coding, a second researcher independently audited the coding, to ensure data quality and to establish the trustworthiness and reliability of the themes identified. All coding was completed by hand and using word-processing software.

Research team
Our research team is made up of experienced educational researchers with an interest in children's academic, social and emotional development and wellbeing. None of the researchers had attended a boarding school and none had been a looked-after child.
One is a psychologist and academic, two were previously teachers and are now academics, one a research academic and one a postgraduate student. As a team, through open discussions we had established a balance in our views on the potential for boarding schools to be considered as a benign, helpful or harmful measure in the care of vulnerable children, leading to their educational and personal development.

Results
Below we present seven major themes identified within the data, with each major theme having several subthemes (see Table 2). Many of the subthemes are interrelated and create a complex network of connected factors. Each participant had a number, and this number follows quotes taken directly from their interviews.
The four 'Rs': Right child, right age, right circumstances and right school All social workers interviewed felt that many factors intersected in planning a child's journey from home to boarding school. Such were the complexities involved, that few of the children would be suitable, therefore seriously limiting the number of referrals made to a boarding school. Nevertheless, social workers could imagine the 'right child' and all of them could identify at least one child they knew within their caseloads they considered suitable. One participant said: I think for the right child, it definitely could work: it's choosing that right child, it's having those right children isn't it and what level they're involved with in need. . .
However, the right age might then rule out some of those children: I can imagine at 10, there's a great opportunity to change things around for them. If you get to 14, 16, you've already sort of lost them to a degree, whereas at 10, there's still that potential. . . Age also featured when participants considered boarding intervention as a decision about care: . . . depending on the age of the child: if you were looking at Year 6, you'd be more likely to be looking at long-term foster care or even adoption for those young people, and it's more likely to be successful because they are at the younger age range. Not as many young people of that age group tend to be placed in residential care: they tend to go into foster placements.
'Right circumstances' implied that many complex factors would have to be in place so that a vulnerable child could cope with the upheaval of moving from their family and/or local community into a boarding school. Access for visiting home, or for family to visit them, was important and the child would need to be capable of adapting to the new school environment: . . . I am sure for the right child, for the right circumstances, this might work. But I think that would be an exception rather than a rule.
'Right school' represents the approach that social care workers take when selecting placements in families and homes for children. There were strong feelings expressed about boarding schools and how these were thought to not be suitable for many children that social services work with. One participant said: . . . I think there's a perception of boarding schools as for posh people and clever people, and that perhaps a child in need living on a council estate in wherever, really . . . wouldn't fit in, and would be socially disadvantaged.
The same participant went on: Social workers, professionals working with children, are possibly unreasonably prejudiced against the private school system anyway . . . and some of us would like to see the private sector abolished! . . . There's a social disadvantage in mixing with children from very, very different backgrounds.
These quotes mirror the findings from research exploring the effects on workingclass children who transition into the higher education system, which may be inadequate in supporting the needs of the working classes (Reay, 2018). There was also a concern whether the school would be able to cope with the level of distress and disturbance often experienced by Children in Need/Child Protection Plan children. Social workers suggested that a more intense level of care might be required than the boarding school could provide. The attitude was that care is better provided within the environment of a family and community that matches their background rather than a boarding school environment.
Pedagogical devices as children's social care levers 9

Financial costs
There were concerns expressed about the financial costs associated with boarding schools for Children in Need. Whilst there could be a financial benefit to placing a child in boarding school over placing them within a children's home, doubts were expressed about financially investing so heavily in a child that might never become 'looked after': . . . There is an issue of budgets: there is an issue of money and austerity . . . this was an issue, because not only was it money, but it was child in need, child protection plan children, which really is an early intervention, so to ask them to fix that money in place for an early intervention is also very much against what they are doing. . .
Whilst early intervention was viewed positively, this created a tension, with concern over funding and investment between individual children and wider community benefit. For example, there was little value placed in the investment of sending a single child to boarding school when it might be better invested in the 'entire local community', perhaps 'investing in library resources or other focal points within the community' (P009).
In 2014/15 it was estimated (Curtis & Burns, 2015) to cost a little over £156,000 per year to place a child in a local authority care home, and just under that figure for a home run by a voluntary or private sector organisation. This is in comparison to foster care, which costs on average around £32,000 per year. The costs for a place in a boarding school could average around £10,000 per annum when grants and scholarships from the school and trusts might be obtained. Further, out-of-term costs would also be incurred, making the total figure a high spend in comparison to the average cost of intervention for a Child in Need-estimated to be around £3,000 per annum (Department for Education, 2013).

Positive environment, aspirations and benefits
Social workers suggested that boarding schools could provide a positive, beneficial environment for a Child in Need/Child Protection Plan child, supporting young people's aspirations: . . . actually the boarding environment, where you can focus on your education and you do get the care, and you can still be in touch with parents and go and see them-it's not ruining that relationship-I just think it still can be-and is-a much better place than, unfortunately, home is.
However, positive statements were often qualified with scepticism: I can see some sort of utopia where it could be absolutely brilliant, but how you provide the right level of support and convince the social workers and families of the benefits, I'm not sure.
Boarding schools were seen as being able to offer Children in Need/Child Protection Plan children stability (lifestyle, accommodation), to build character, develop resilience, provide routine, apply a structure, present discipline, supply nutrition, offer recreational activities and support aspirations: 10 D. Murphy et al.
I think it's got the potential to work at lots of different levels, really, because it's a safe way for the child to experience something new and that kind of cultural and social development that kind of gets forgotten about, because it's just about safeguarding and protecting, it's something that can build on that child's experience and confidence and self-esteem, which will hopefully come in good stead as they get a little bit older and may face challenges. And it's aspirational as well, isn't it?
It is clear that there were positive views about boarding schools as an intervention; however, below we show that there are also significant concerns that might seriously inhibit social workers from using boarding schools.

Social/cultural factors
Historically, social workers support families in communities and institutionalised care facilities for young children. Institutionalised care in the UK has been found to have enabled the systematic abuse of vulnerable children. Consequently, more than half (n = 13, 65%) of the social workers regarded the boarding school system as having an association with other forms of institutionalised childcare. For example, many participants referred to private boarding schools, even if they had experience of visiting children in state boarding schools, acknowledging that this would put them off using boarding as an intervention: I think that sense of it being a bit of an unknown quantity, as well as my own kind of personal views about the class nature of private education, would potentially prevent it as an option.
This 'unknown quantity' (P016) was paired with strong language and other forms of institution, such as the reform system known as 'borstal' (P001; P017), 'emotional behavioural difficulty schools' (P005; P008; P017), or a 'surrogate care system' (P002). The use of such terms suggests suspicion and prejudice against boarding schools.
Contrary to the fact that the numbers of children in care in England (Department for Education, 2018) are rising, participants spoke about keeping families together at all costs, indicating how boarding interventions would go against this value of social work practice: I think the other issue is that a lot of social workers are trained-and I don't know how in vogue it is at the moment . . . but there certainly has been a drive about keeping the family together, . . . and this thing about sending a child to boarding school, so again not seeing it as actually this is a way to support the family and child relationship in an environment that could actually give the parent and situation some respite, they're maybe seeing it as, 'Oh, we're pulling the children apart and that's against what we've been trained'. ) It was clear that boarding school has an image problem for many within the social care profession, and social workers were wary of boarding schools as a placement for a child. On the contrary, for some, the potential for boarding schools to offer respite and keep families together was also recognised: I think respite is key. I think boarding school offers the structure of time away from the family but then time back with the family, and contact in between, which is always very much encouraged and prioritised when it can be in social work. So yes, absolutely respite . . . A boarding school may Pedagogical devices as children's social care levers 11 be able to more closely work with a child than perhaps day school would be able to, whether more one-to-one or just by the nature of being a residential setting. (P002) Recognising the somewhat conflicting views about respite, the same participant suggested that respite may be considered less effective and does not consider important contemporary theoretical approaches such as those based on attachment theory, saying: I think the idea of, just removing a child, but making it better and that respite will make things better is quite an outdated idea in social work. Now, you know, modern-day social work is so much caught around attachment theory and understandings that the child flourishes best when the family is supported to care for them. (P002) The issue of attachment featured when contrasting residential care and boarding school, suggesting that a temporary attachment figure in the child's life can be disruptive to their development: . . . residential care is often criticised for the turnover of workers that leads to children not having permanent, stable one-to-one relationships: there's a real danger, I think, that the boarding school project just mirrors that turnover of very well-meaning and often very kind but not permanent figures in a child's life. . .
Social workers are aware that 'distressed, frightened children don't learn' (P017) and many children require emotional, therapeutic and pastoral support when entering a completely foreign environment such as a boarding school. One social worker questioned whether boarding schools would provide a safe space: I think emotional regulation is absolutely crucial in doing that, because what we know of children is that they will not learn unless certain building blocks are in place, and emotional regulation is one of those absolute cornerstones that allows us to learn. We don't learn when we're frightened. (P014) It is important to situate the views of social workers on boarding schools and the notion of keeping families together within the broader history of social work practice and the development of the social work profession, moving away from the idea of institutional care. Participants considered boarding schools to be like the children's homes of the past. A participant explored their attitude towards institutions as a barrier to referring children to a boarding school place: I think there is probably-again, it's probably prejudice rather than evidence-about well actually, if the child is troubled in some way, in what way will a boarding school wrap around them in terms of the needs for support that they've got?
Another said: One of the issues about social workers is that generally they don't believe in institutional care, and that probably 95%, if not higher, of social workers would always want children to be with families and to be placed in families. So I think there is always, this is a natural predisposition against any 12 D. Murphy et al.
sort of institutional care. So, institutional care would cover a boarding school, so that is a sort of inbuilt prejudice that they might have, or a sort of practice philosophy, you might say.
In the event a family cannot remain together, social workers claimed to be more likely to consider placements for children in which the child might feel at home or belong. In contrast, boarding schools were considered to represent a significant cultural difference between many of the children that social workers support, and the kinds of children that might 'fit in' at a boarding school. There was a tendency among participants to homogenise boarding school culture, limiting the possible variation of children that attend boarding school, resulting in the suggestion that moving a vulnerable child to a boarding school would be experienced as a kind of 'cultural shock' (P014).
Social workers might be deterred from referring children to boarding interventions because of their own views about how a child would fit in based on the class differences that are perceived to exist. One social worker said: . . . what boarding school stands for in public perception and perhaps concerns around the cultural clash and elements of concern around boarding schools representing quite an elite, upper-class, expensive institution, and generally when you're working with children from a socially disadvantaged background, how would that cultural experience be for them and how easy would it be for them to feel at home, to integrate, and to flourish in an environment where . . . boarding school does feel different from any other kind of school in the public perception. (P002) Another said: . . . Boarding school is theoretically-and I know it's only theoretically-a middle-class phenomenon, and therefore, social workers . . . don't believe in boarding schools, and think the best way to help children is to keep them with their families. . .
Class differences featured as a potential barrier when another social worker considered their own perceptions of boarding schools: I guess my own kind of association with boarding schools is that they would be upper-middle to upper-class young people, or young people from wealthy backgrounds, who probably have quite different socio-cultural experiences to the young people that are usually identified by local authorities.
Drawing this theme together, the participants' views should be set in the context that within England, the vast majority of boarding schools are attended by wealthy middle class and the richest in society. There are a small number of state schools with boarding facilities, yet the participants' views are likely representative of the majority view on boarding schools. The social/cultural values and beliefs that matter to social workers are concerned with keeping the child within a family environment and recognising that social class is a very important and dynamic feature in understanding children, and one for which boarding schools are not considered to provide a good match.
Pedagogical devices as children's social care levers 13

Outcomes
More than a decade ago the Maxwell et al. (2009) study reported that senior social workers identified the use of boarding schools as a 'positive placement choice' (p. 10) rather than for crisis cases or emergency situations. Several participants supported this: Certainly, I feel that some of the young people I worked with when I was in the care sector . . . had boarding been an option, and they had been able to enter boarding school, I really believe that the outcomes of their life would be absolutely radically different.
Three-quarters of the participants interviewed believed that boarding school chances can impact on attainment and raise aspiration. As one participant described: I think it's [boarding] got the potential to work at lots of different levels, really, because it's a safe way for the child to experience something new. (P015) Social workers viewed the boarding school intervention as an opportunity for Children in Need/Child Protection Plan children to experience a new environment which: . . . could offer them routine and that structure in their lives that they've not previously had, that they would take into adulthood, it'll help them mix and socialise and associate with other people that they may not have necessarily had the opportunity to do that with; and then there's all the academic opportunities, and even more so, it could improve their relationship with their family, because that's what they needed, some respite and space from each other. (P006)

Knowledge of boarding schools
Most perceptions and attitudes expressed were based on indirect experiences of family and friends, previous social work cases, and through press and media-related sources. There appeared to have been very little social worker education and training about boarding schools as a social care intervention. According to one respondent: It's not ever floated as an idea, it's not mentioned in any textbook, it's not on any whiteboard, it's not in any discussion, so it's not going to be in the psyche of a social worker to think, 'Ah, boarding school'.
Very few social workers had visited a boarding school, yet when they had, the associations were negative. For example, one social worker (P008) had visited a boarding school because of a safeguarding concern that had been raised about a child. Another (P004) had visited friends at a boarding school as a child and recalled that at the time they had felt the intensity of living together, and felt that the peer relationships would potentially be overwhelming for a young person with difficulties. For both these participants, contact with boarding school had left a significant impression as a difficult environment in which to live. When asked about their feelings about boarding schools, one participant responded by saying: 14 D. Murphy et al. However, one social worker quite explicitly questioned the extent to which social work practice can be truly child-centred, saying that often when a child has to be removed from home: It's not child-centred because we're not asking the child.
This raised interesting questions about whether all social workers are referring to the same thing when using terms such as 'child-centred' practice. Murphy et al. (2013) have pointed out that this is often the case, and that social work in the statutory setting is unable to offer the kind of unconditional and non-directive environment for children in a way that is consistent with the theory underpinning child-centred approaches.

Discussion
In this study we aimed to explore social worker attitudes towards the use of boarding schools as a social care intervention. We analysed the interviews of 20 social care workers who talked about their attitudes and beliefs about the use of boarding schools as a social care option for vulnerable children. This is an emotive topic, and naturally produced strong feelings and reactions from the participants. There were some positive attitudes amongst social workers to the outcomes and potential benefits of the use of boarding schools as a pedagogical device, acting as a social care lever. However, these views were often outweighed by misgivings, and reservations towards boarding schools. It is our interpretation of this major finding that issues other than educational privilege are more important to social workers. Whilst three-quarters of the social care professionals interviewed considered that boarding schools could have a positive impact on attainment and raise aspirations for some children, the issues of keeping families together, supporting local communities, social class differences and the perceived risks associated with any form of institutionalised care were major reasons for social workers to not consider boarding schools. Added to this were concerns about the costs of boarding placement as an early intervention for a Child in Need, whereas boarding as an alternative to residential care placements was a more palatable option.
Social work is adopting a more evidenced-based approach (e.g. Social Care Institute for Excellence, 2016), although our interviews suggest that practice is more emotionally driven than empirically guided. This was especially so for an emotive and contentious issue such as placing vulnerable children in boarding schools. More importantly, there was broad consensus that the aim of social work was to do everything possible to keep families together. So strongly was this expressed that we identified the goal of keeping families together or keeping children in families as a socio-cultural norm within the profession. This is to some extent, of course, a paradoxical finding. The social work profession is often regarded with some trepidation, feared by many as taking children away from the family (Prendergast, 2016), and at the same time the profession is often reviled by the media for their apparent failure to intervene sufficiently robustly to protect a child (Laville, 2015). And this at a time when the number of children taken into care is rising (Department for Education, 2018). 16 D. Murphy et al.
The Pathfinder evaluation report (Maxwell et al., 2009) identified 'gaps' that needed to be bridged between the school, social care, health and youth justice systems, and recommended to those considering boarding options that: further work was [sic] required to align the priorities of education and social care professionals working with young people and their families, and to challenge the view held by many social care professionals-as well as young people and parents-that boarding schools were only appropriate for those from more privileged backgrounds.
(p. 6) The perhaps apocryphal 'elite' aspect of boarding school provision was a concern to social workers, both with regard to distancing the child from their family of origin, both socially and culturally, and legitimising a socially divisive system of schooling.
In contrast to keeping a Child in Need in the family home, boarding schools were considered to potentially offer security and opportunities for improving educational attainment. However, as reported in the literature review, a study in France showed that it was only students already scoring highly in tests that made greater gains in mathematics compared with their lower-achieving peers, although this was not until the end of the second year (de Behaghel et al., 2017). Notably, the authors focused on the emotionally 'disruptive' nature of being placed in a boarding school (p. 31), and explored wellbeing, motivation and study habits over an extended period. Again, higher-achieving students fared better than lower-achieving students 2 years after the placement began. These findings were reflected in social workers' fears that boarding might be better for Children in Need that are already achieving highly academically, and leaves open the question of the impact on lower-achieving children. This also links to concerns raised that removing children from their communities might have a disruptive effect on their social/emotional development.
Summarising the findings in this study, social workers appear less than enthusiastic about using boarding schools as interventions for vulnerable children. Whilst social workers wanted to offer the best opportunities for vulnerable children, they were concerned about using boarding schools because they saw them as part of a pernicious social process of class apartheid and segregation. For, as Bass (2013) has noted, even when the boarding school environment is positive and gives access to social, cultural and educational capital, the quality of children's experience is dependent on the home environment that the child has left behind and their own individual motivation to succeed.
In light of the findings of this study, we suggest that if the DfE wants social workers to use boarding schools, then social workers will first need to be more fully persuaded that they can offer the appropriate setting, circumstances and opportunities for a child prior to placing them in a boarding school. Evidence is clearly needed, and the social work profession might first need to undergo significant transformation, which might not be consistent with the direction, intentions and future aims of the social work profession. Indeed, it is far from clear from the social workers in this study that there is an appetite for such transformation. In fact, some social workers might argue that the ideological basis to social work culture is antithetical to placing children in boarding school. It is likely that boarding schools have an image problem that they desperately need to address if the boarding school sector and social services are going to work Pedagogical devices as children's social care levers 17 more closely together. Both boarding schools and social work have evolved with long histories and traditions that seem to have counter-narratives-one premised on keeping families together and the other essentially separating children by breaking the bond of family attachment to foster independence (Poynting & Donaldson, 2005). Social workers would indeed find this difficult to support, unless the child was already being removed for child protection purposes. Evidence of not only the educational but also the social/emotional benefits of boarding for vulnerable children would need to be shown before any such shift in social work practice is likely to occur.
According to our analysis of the data, some social workers were open to the positive potential offered by boarding schools as a social care intervention. Similarly, if social workers become involved more closely and earlier in the care of Children in Need by working with schools, educational outcomes might be improved. But to what extent does social care share the same goals for children that contemporary educators might? Social care is increasingly concerned with the management of 'high-risk' cases, and their priority is often focused on the physical safety of a child rather than their academic attainment. Removal of a child from the home is seen as a last resort, and boarding school is currently perceived almost exclusively as a removal option. For this reason, social workers saw only a small number of children as potentially suitable to be placed in a boarding school, rendering boarding an unlikely option.
Many of the concerns raised by participants from the social care profession reflect the sentiments of a boarding school counter-narrative emerging in the UK (Duffell, 2014a(Duffell, ,b, 2015Schaverien, 2015;Duffell & Basset, 2016;Stibbe, 2016;Renton, 2017). This draws on some of the current attitudes and views towards boarding schools from outside both the boarding school sector and the social work profession. This counter-narrative is concerned with the significant harm and damage done to the emotional and psychological development of some children that attend boarding schools-separated from their family and often without their consent (Duffell, 2014a). Such disparate views from professionals in different sectors are unlikely to be quickly reconciled, and it is essential for the Boarding School Partnership to engage with all stakeholders if it is serious about succeeding without replicating the harm caused by other institutional care interventions.

Limitations
The current study has several limitations. First, the sample used in the present study was not geographically representative of the whole of England. This was largely due to the use of convenience and snowball sampling methods. Nevertheless, the sample included a wider range of professionals involved in the field of social care work, ranging from qualified and unqualified social workers to social workers, managers and senior managers. There is no reason to suspect that these professionals take a distinctly different view of boarding schools and the role they might play for vulnerable children to professionals from other parts of the country. Second, only some of the participants had actually been faced with the reality of making a decision to place a vulnerable child in a boarding school. Future studies in this area might well benefit from interviewing a range of professionals more directly involved in the decision-18 D. Murphy et al. making process, including those from the educational institutions the child might be faced with leaving and/or joining.

Conclusion
In conclusion, we suggest policy decisions regarding the use of pedagogical devices as social care leavers work with, engage and include a wide range of representatives from both the professionals 'on the ground' and researchers and academics across both the social care and higher education sectors. Moving to promote initiatives that are unsupported by high-quality research is inadvisable at best, and potentially damaging to young children and their families at worst. However, conducting research in this area is not easy. Conducting high-quality research trials that test the effectiveness of boarding schools is expensive, difficult to design ethically and can prove extremely challenging in terms of recruitment. Notwithstanding these difficulties, we suggest that more evidence is needed before significant decisions are made that will shape people's lives for ever. Currently, it seems that such initiatives are based on ideological grounds that might better serve the interests of some more than others. It is apparent that social workers will need further convincing if the yet to be established benefits of boarding schools are to be made available to vulnerable young children. Boarding schools themselves also need to look both inwards and outwards to consider what difficulties they need to address and overcome in order to make themselves a suitable environment for a wider number of children. If boarding schools want to make what they offer more consistent, with a wider range of society, and enable children from more socially and economically deprived backgrounds to study with them, then change is inevitable. Whether there is the will to change is another matter. There is much work needed before such a pedagogical device will be able to act as an ethical and effective social care lever.

Ethical guidelines
This research was carried out under the BERA Ethical Guidelines and ethical approval was gained from the School of Education, University of Nottingham ethics committee.