Schools and their local religious contexts: building a framework of negotiations through qualitative meta-synthesis

ABSTRACT In this article, we consider the relationships between schools and their local religious contexts and develop a new empirically-informed middle-level theory for analysing these relationships as a complex framework of negotiations. This framework is based on a meta-synthesis of qualitative data from forty-five case-studies of schools in Great Britain and Northern Ireland, which were conducted by researchers at University of Warwick between 2009 and 2015 as part of three major projects. The synthesis design was in three stages. First, two main overarching descriptive nodes were identified, each with two sub-nodes: school – comprising school ethos and religious education; local context – comprising local patterns of belief and pupil religiosity. Second, these four sub-nodes were paired with each other to form six dyads of negotiation: four illustrated ‘direct contextualisation’, and two illustrated ‘indirect contextualisation’. The data were re-analysed through this six-fold framework. Examples of each dyad are set out, illustrating both straightforward and challenging circumstances. The value of this dyadic framework for negotiation is considered, in relation to research, policy and practice, especially in theorising contextualisation. Finally, we underline the value of qualitative meta-synthesis in theory generation.


Introduction
One of the most long-standing debates in education around the world is on the place of religions in schools, including questions of the status of 'faith' schooling, 1 responsiveness to pupils' religiosity, and whether and how religions are to be studied (e.g. Gardner, Lawton, and Cairns 2005;Hunter-Henin 2012;Russo 2012;Davies 2014). This debate is often framed around constitutional questions of formal religious establishment, secularity, and secularisation (e.g. Sullivan and Beaman 2013;Cooling 2010), which have four main features. First, the current place of religions in schools is often born of historically-rooted relationships between the state and religious institutions (see Jackson et al. 2007;Sullivan and Beaman 2013): those with a constitutional establishment, as in the United Kingdom or Spain; those with strong forms of separation, as in the USA or France; those with an intermediate position, as in Turkey or Italy. Second, however, in many places these relationships have altered in recent decades, whether in response to shifting patterns of religiosity, through immigration or a decline in traditional religious observance, as in Canada (Beyer 2013), in response to major political shifts, notably in post-Soviet states, such as Estonia (Valk 2007), or in response to the development of human rights law, especially freedom of belief (Hunter-Henin 2012). Third, more recent concerns with international terrorism have added political weight and urgency, both at policy level (Ghosh et al. 2016) and in schools (Arthur 2015), especially in relation to concerns about extremism and radicalisation. Finally, these changing issues about the place of religions in society are themselves played out across wider educational policies and practicesabout the nature of schooling, admissions, curriculum and assessmentwhich are themselves both highly politicised and in flux (Morris 1998;Walford 2008;Fancourt 2013Fancourt , 2015Fancourt , 2017b.

Grand theories and local contexts
In making sense of these structures and the changes, there is a creative tension between two processes. On the one hand, appropriate conceptual tools need to be developed and refined for placing education within general theories and philosophies about societies and religions. Previously, the secularisation hypothesisthat societies were becoming less religiouswas a commonplace sociological assumption, and it has often been deployed in educational research (e.g. Copley 2005;Mayrl 2011). However within sociology, this theory has become insufficient at explaining the complexity of religions and non-belief, both in the UK and internationally. Distinctions have come to be drawn between public and private secularisation (e.g. Casanova 1994), and between religion and spirituality (e.g. Spalek and Imtoual 2008;Beyer 2013), reflecting both changing complexities and developing methodological precision. The temporal assumptions of the secularisation thesis as a unidirectional decline in religion have also been critiqued, whether in noting 'resurgent religion' (Berger 1999), or a dialectical process, such as Martin's (2005) argument for 'successive Christianisations followed or accompanied by recoils' (8). Thus, recently Lee (2017) has suggested that: 'scholars have increasingly expressed the need to negotiate religionthat is, to provide better accounts of ways of engaging with and recommendations regarding the nature and role of religion in contemporary, often culturally diverse societies' (1), as religious institutions and actors engage with otheroften stateinstitutions and actors in a constant repositioning of their relationships. The secularisation hypothesis in education therefore also needs to be reviewed.
On the other hand, as Skeie (2013) suggests, 'researchers. . .have for a long time been aware of the fact that teaching about religion in schools does not happen in isolation from the surrounding world' (250) . Indeed, it almost goes without saying that schools are embedded within their localities, and therefore much educational research has considered the complex contextual realities of schools, investigating how headteachers, teachers and pupils make sense of their everyday connections to the local religious landscape in which they both play a part and to which they respond. Unsurprisingly, schools have been identified as key institutions within this landscape (e.g. Pomson and Deitcher 2009), and their role in fostering community or social cohesion has been explored (e.g. Osler 2007;Flint and Robinson 2009), notably the contested question of whether faith schools are cohesive or divisive (e.g. Dwyer and Parutis 2013;Hewstone et al. 2018). Other researchers have focused on the inter-relationship between pupils' beliefs and school ethos. For example, Green's (2012) ethnographic study of a Christian free school showed how its ostensibly Christian character was rejected or ignored by its largely agnostic or atheist pupils, whereas Pike (2010) suggested that the Christian ethos of a new academy school was a key element in raising pupil attainment. By contrast, others consider the range of complex religious identity choices by pupils in school (Casson 2011;Moulin 2011Moulin , 2014, particularly in relation to local inter-religious understanding (e.g. Casson 2013;Wilson 2015). In these discussions, the concept of context is itself theorised and indeed the European Network for Religious Education through Contextual Approaches has specifically focused on this question, in considering the shifting local patterns of religiosity and schooling within different countries Rothgangel et al. 2017).
To illustrate the entwining of these issues in schools (while at the same time demonstrating the rich, multi-layered nature of our data), here is a vignette from one (anonymised) school (see Jackson et al. 2010a).

Vignette -John Hawkins infant school
John Hawkins Infant School, in East Anglia, is a community school situated on an estate of mixed private and social housing on the edge of a small town. At the time of the research, it included children from local RAF and US Air Force bases and increasingly children from Portuguese, Polish, Latvian and Lithuanian migrant families. Children with English as an additional language constituted 20% to 25% of each class, and various languages and accents were heard as parents brought children to school in the morning. The local authority recognised it as a supportive environment for children with special educational needs, and it received several pupils with physical disabilities and learning difficulties.
Although it was not a church school, it had strong church links; the Anglican parish priest and independent evangelical church pastor were both governors, regularly leading assemblies, and there was a carol service in one of the local churches. As the headteacher commented: We do like to make sure that we do bring God and Christ into school in a way that we feel that we need to acknowledge the children perhaps not having that experience outside school, [but] people aren't evangelical in any way.
Each class had a discrete weekly religious education lesson following the local Agreed Syllabus. The main religion studied was Christianity, with Judaism as a topic in Year 2 religious education lessons, because, as the school's RE co-ordinator expressed it, 'Judaism is definitely very Key Stage 1 friendly, because of all the artefacts; the celebrations are quite child-and family-centred so children really do identify [with it]'. In a lesson on Shabbat, the teacher planned a balance between 'hands-on' direct sensory engagement and a sense of the special, and even the sacred. A calm atmosphere was maintained, and the uncovering of the challah bread and the passing round of the spice box and bread as the children sat in a circle had an element of ritual and reverence. The teacher related Shabbat observance and practices to underlying meanings: Teacher: They believe God created the world in six days and on the seventh God . . . Child: Rested Teacher: They smell the spices in the hope that the week will be sweet and fragrant.
Other religions were introduced in cross-curricular topics in the school's integrated curriculum, and through the celebration of festivals in assemblies.
Pupils were often enthusiastic about religious representations in books; a Muslim student of Indian heritage, Ali, immediately related to familiar sub-continental Hindu culture represented in a book on Diwali, and to the image of Rama: That's how one of our gods looks like. I've seen God's head before on a computerhis face is all blue.
Pupils openly identified with different religions: on discovering that a book was about a Muslim child, Ali, straightaway declared, 'I am a Muslim' and beginning to read it avidly. The African Caribbean boy, Benjie, and his Polish classmate, Leon, responded in the same way to a book on Christianity: Leon: I am a Christian. You go to church to celebrate. Benjie: Well, I'm going to be baptised next time I go to church.
Various features of this vignette are noteworthy. It is a state school, but with the considerable involvement of two local churches. Judaism is chosen for study because it offers accessible artefacts, not because there are Jewish pupils, and teaching about it could be dramatically ritualistic. Pupils openly identify as being religious, and a Muslim pupil identified a Hindu deity as 'one of our gods' (emphasis added).
However, on each of these issues, the school could be criticised or praised, depending on one's theological, philosophical or pedagogical perspective, and it almost goes without saying that these perspectives can represent deeply normative voices, such as the Church of England (2016) or the National Secular Association (2017). One might argue that its 'soft' Christian ethos was inclusiveor implicitly indoctrinatory. One might argue that pupils from minority religions were made to feel welcomeor were lamentably blurring their own religion with others. One might argue that the choice of Judaism was well made for pedagogical accessibilityor that this was a naive misrepresentation of a complex theological tradition. One might argue that the enactment of Shabbat was pedagogically engagingor an inappropriate imitation of sacred worship. As noted above, these different perspectives are often framed by and play into the debates about religious establishment, secularism and secularisation (e.g. Cooling 2010; Hunter-Henin 2012; Conroy et al. 2013).

Approaches to developing an analytical model through meta-synthesis
Rather than directly addressing these older debates, we seek here to develop a new understanding of contextualisation by analysing how religion isto use Lee's (2017) expression -'negotiated' (1), especially in asking what kinds of 'negotiation processes are underway?' (3). As she points out, the term negotiation can make 'ambiguities and ambivalences visible [but] it does not assume simple or unilinear exertions of power' (7). We here deliberately use the term to include both its original meaning of 'coming to an agreement' (e.g. negotiating with people) and its newer meaning of 'navigate' (e.g. negotiating obstacles or identities) in order to encompass a range of processes, but without seeking or assuming coercion or influence in any direction. However, this concept is not well developed in research on education: the collection of studies containing Lee's work (Guesnet, Labord, and Lee 2017) only includes two studies on education, and both are from a legal perspective (Hunter-Henin 2017; Vickers 2017); consideration of education is also missing from other recent inter-disciplinary works on shifting constitutional patterns, such as Sullivan and Beaman (2013). Thus, genuinely educational research is lacking, even though education is a prime site of such negotiations.
Moreover, if education is sometimes missing from the study of religions, educational research has been accused of not taking religion seriously (e.g. Grace 2004) or adopting unrefined theoretical stances: thus Green (2012) suggested that 'any analytical framework. . .needs to avoid the current polarisations' (392), such as overreliance on secularisation theory to argue for secular indoctrination (Copley 2005). If this relationship is to be researched, the approach should be sensitive and nuanced. Indeed, the balance between impartial description and normative exemplification can be hard to tread, and one researcher's example of impartial analysis is another's example of confusion or error. For example, one misapprehension is that to research the 'context' of religious education is thereby to argue normatively for a contextual pedagogy for religious education (see Wright 2008;critiqued by Skeie 2013).
An exploration of 'negotiation' can mitigate this risk. Following Lee, we conceptualise these interactions as being between religion and the secular, rather than between religion and non-religion, but to do so raises some definitional questions. Defining religion is not straightforward, and Lee had previously followed Knott (2005) and Towler (1984) in identifying 'conventional religion [as] religious institutions, their traditions, beliefs (which, in this context, certainly includes theism), and practices, and the people who adhere to them' (Lee 2015, 9). We also broadly adopt her definition of the secular as 'phenomenaobjects, spaces, people, and practicesfor which religion is no more than a secondary concern, reference point, or authority' (Lee 2015, 39); this is because the main purposes of schools are not religious, but instead more broadly educational. Clearly in faith schools, religious purposes would also be important, and here we would modify Lee's definition to argue for competing primary concerns rather than considering that their religious purposes are no more than secondary. However, this does not mean that other schools are automatically nonreligious, as the vignette shows, or indeed anti-religious, areligious, atheist or indifferent (Lee 2015); all these positions are also played out against schools' secular prioritiesas will be shown. Our focus is on how religion and non-religion shape and are shaped by educational practices through overlapping concerns, values and motivations, which may conflict or coincide. We seek therefore to answer the question: how can local negotiations between religions and schools be analytically framed? In doing so we seek to produce a more refined model of the patterns of causality rather than seeking to identify particular directions of causation.
There are two broad approaches to the generation of any new analysis. A common approach is to draw on an external 'grand theory' (Kettley 2010, 9), to address contextual religious and educational issues. This approach has allowed researchers to fine-tune established theories, giving fresh insights to contextual complexities. Obvious examples are the use of secularisation theory; thus Mayrl (2011) applied a macro-level policy theory of secularisation (Martin 1978(Martin , 2005 to explain changes in schools in New South Wales, and Bourdieu's (1986) theories of capital and habitus were adapted by Grace (2004) and Green (2012) to examine the dynamics of schools' religious practices. However, this approach runs the risk of carrying over any latent flaws in the external theory, and not doing justice to educational priorities, particularly when these theories emerge from sociology.
An alternative is to draw inductively on a wide range of existing data from the field, to construct a more empirically-rooted account, as 'middle range' theory building (Merton 1968, 39), with the 'analytical task of identifying those social mechanisms which give rise to specific behaviour in given contexts' (Kettley 2010, 17). Clearly this approach has its own potential short-comings, notably that it could lack reliability, as a few studies might over-dominate any such theorisation. To avoid this risk, one would at least require sufficient data from a variety of field sites to allow more generalisable patterns to be elaborated, but this could be complex and time-consuming, unless there are existing data. Here we have both drawn on Lee's sociological notion of 'negotiation' as a loose descriptive conceptualisation of contextualisation, and then explored the issue in more fine-grained detail through a meta-synthesis of a considerable bank of earlier research in order to explain how schools respond to their local religious context. Indeed, meta-synthesis is particularly promising as a way of developing middle range theories, as the depth of data allows for nuanced theorisation though crisp engagement with grand theory. It can also engage with more evaluative approaches to these issues (e.g. McMullen 2018).

The WRERU research tradition
This article uses findings from qualitative research carried out in forty-six UK schools by one research unitthe Warwick Religions and Education Research Unit (WRERU)in three major research projects over a decade, between 2006 and 2016. By way of explanatory background, since its establishment in 1994, WRERU has built up a tradition of and international reputation for a combination of empirical research and pedagogical theory, which have challenged generalising assumptions about religions by being alert to the complexities of religious identity, belief and practice and to the varieties of individual and group positions within broader faith categories. Young people's own perspectives on religion and religious learning have been the primary focus (in research interviews as in classrooms) and their views are brought into conversation with contextual factors and the wider traditions (faith, cultural, societal and educational) to which they relate.
The work of WRERU began with ethnographic research into the lives of young people within their religious communities, most notably studies of children from a Hindu background in Coventry (Jackson and Nesbitt 1993;Nesbitt 1998). These studies analysed the relationship between young people's religion and the religiosity of local community. The research then informed the production of classroom materials designed to enable school children to engage with the lives of other contemporary young people of different faith traditions (Barratt 1994;Barratt and Price 1996), thus connection was made between the research findings and another concern, the religious education curriculum (Jackson 1997(Jackson , 2004. In more recent years (post 9/11) this pedagogical interest of WRERU has been incorporated in European-level guidance promoting greater understanding of a plurality of religions among school pupils, notably through Jackson's contributions to the 'Toledo Guiding Principles' (OSCE/ OHIDR 2007) and 'Signposts' (Jackson 2014). This approach has not been without its critics, but this is not the place to describe or address them.
Partly as a result of these pedagogical developments, WRERU's research energies moved in another direction, from the lives of young people in their faith communities to religious education lessons and the school setting. This research emphasised the importance of hearing the voices and observing the experiences of pupils and teachers, and of allowing these to test and challenge prior assumptions (including the researchers' own) and simplistic categorisations of schools and religious education. The diversity of schools participating in the various WRERU projects (primary and secondary; with and without designated religious character; independent and state-maintained; selective and comprehensive) drew attention to a fourth issue, that of school ethos, and its response to, and impact upon, pupils' religion and belief, local religiosity and religious education.
The use of comparative case studies became the hallmark of this approach, and WRERU conducted three substantial mixed methods projects with significant qualitative school-based elements. First, the 'REDCo' project (Religion in Education. Dialogue or Conflict) was a collaborative project across several universities, led by Hamburg, and funded by the European Commission (Weisse 2007). Running from March 2006 to February 2009, it combined studies from eight countries: Estonia, Russia, the Netherlands, Germany, Norway, England, Spain and France. A qualitative strand employed questionnaires and group interviews to explore the importance of religion to students aged 14-16, their attitudes towards religious diversity, their experience of and responses to religious education. In England, this research was carried out with over one hundred young people in four secondary schools with contrasting religious and ethnic demographics (Ipgrave and McKenna 2008).
Second, in 2008, WRERU was commissioned by the then Department for Children, Schools and Families to investigate 'materials used in schools to teach world religions' (Jackson et al. 2010a(Jackson et al. , 2010b)the 'Materials' project. Exploration of contextual factors influencing the selection, use and reception of these materials entailed case studies in ten primary and ten secondary schools, including maintained, faith and independent schools. Methods for the case studies included documentary analysis, lesson observations, semi-structured interviews with staff and group interviews with pupils.
Third, the 'Young People's Attitudes to Religious Diversity' project was financed by the UK Government's research bodies, the AHRC and ESRC, under the 'Religion and Society' programme and conducted between October 2009 and December 2012 (Arts and Humanities Research Council 2013;Arweck and Jackson 2012;Arweck 2017). The qualitative research (Arweck 2017) involved thirteen to sixteen-year old students in twenty-one schools spread across the four nations of the UK. The selection ensured a diversity of localised patterns of religion and a variety of school types within each nation. Group interviews explored young people's views on religious diversity in their own localities (school, family and community) and in national and international contexts.
All three projects employed mixed quantitative and qualitative methods. In the REDCo project, the qualitative findings for each of the countries were collected together (Ipgrave and McKenna 2008), as were the quantitative findings (McKenna, Neill, and Jackson 2009). The other two projects were each presented in one publication (Jackson et al. 2010a;Arweck 2017), though the nature of the combination of methods varied. Quantitative data from the three projects would be impossible to combine in a metaanalysis because different data were collected, but the qualitative data are collectively valuable for understanding the nature and place of religions in schools. Each research project was subject to the university's ethical approval mechanisms and followed the BERA guidelines on consent and anonymity; many schools were concerned about these safeguards in the light of increased scrutiny of religion in schools, and we have been concerned in ensuring anonymity here as these issues have subsequently become even more sensitive. 2 We were engaged in the qualitative research element in one or more of these projects and have then drawn on this research to explore or develop new theoretical perspectives. We jointly led the case studies section of the Materials project (Jackson et al. 2010a, 121-175), and Fancourt commented on questions of dialogue and pedagogy for both the REDCo project and the Materials project (Fancourt 2009(Fancourt , 2016. Ipgrave has been involved in and drawn on all three projects (Ipgrave 2010(Ipgrave , 2011(Ipgrave , 2012a(Ipgrave , 2012bIpgrave, Miller, and Hopkins 2010;Ipgrave 2013Ipgrave , 2016Ipgrave , 2017aIpgrave , 2017bIpgrave , 2017cIpgrave , 2017dArweck and Ipgrave 2017).

Principles of meta-synthesis
For this article, we conducted a fresh meta-synthesis of the forty-five case studies, drawing on other approaches to the synthesis of qualitative research (e.g. Barnett-Page and Thomas 2009;Nag, Snowling, and Asfaha 2016), though a key difference was that we returned to WRERU's original data, rather than different authors' publications, as a form of secondary analysis. We are not aware of any other research in this field, either nationally or internationally, which draws on so large a sample of schools.
We adopted a three-stage process to our meta-synthesis. First, to go beyond a simplistic binary of local religiosity and school ethos, we returned to the themes in the three underlying projects, and from them two broad overarching nodes were developed, 'school' and 'local context', each with two sub-nodes: the overarching 'school' node comprised the two sub-nodes of 'school ethos' and 'religious education'; the overarching node of 'context' comprised the two sub-nodes of 'pupil religiosity' and 'local patterns of belief'. We then treated these four sub-nodes as an interconnected 'network' (Miles, Huberman, and Saldaña 2014, 111), and the fact that they were not all institutions or actors meant that negotiation could be viewed holistically.
Second, we recognised that these nodes and sub-nodes were not truly emic because they were broadly the main themes of the three underlying research projects, and it would be analytically disingenuous to claim otherwise. We therefore decided to take advantage of the richness of the amassed data by systematically focussing on the nodes' interactions with each other, as a secondary interpretive process. This entailed the sequencing of the processes of data categorisation and data connection into two distinct analytical phases (Maxwell and Miller 2008). We labelled these nodal interconnections as 'dyads of negotiation' since it was the nature of these connections that was of interest, and not the overarching nodes or sub-nodes per se.
For example, in John Hawkins Schoolin the vignettethe dyad between school ethos and local patterns of belief was significant, with two local Christian ministers as governors, and it was open to the school and the ministers to decide what was appropriate in terms of their attendance and involvement. The dyad is not of itself secularising or sacralising, but the site for such effects to be worked out: indeed here, arguably, an otherwise secular space is sacralised. By contrast, Ali, Muslim pupil who identified with a blue god, highlighted the dyad between pupil religiosity and religious education, as a form of identity negotiation, in that Ali claimed a degree of affiliation with a different religion from that of his family; school here is a site of inter-religious identity negotiation, and not simply either secularisation or sacralisation.

The six dyads of negotiationoverview
Given the four sub-nodes, six dyads of negotiation were generated. Four of them exemplified direct contextualisation, in that the negotiations were between the two overarching nodes: • School ethos and pupil religiosityschools' pastoral responsiveness to pupils' beliefs, values and traditions, such as whether school uniform accommodated pupils' religious symbols. • Pupil religiosity and religious educationschools' pedagogical responsiveness to pupils' beliefs, values and traditions, such as how pupils' own beliefs were treated in religious education lessons. • Local patterns of belief and school ethosschool/community relationships, notably the porosity of the school to local religious institutions, such as the involvement of local religious leaders in assemblies or lessons. • Local patterns of belief and religious education -Curricular responsiveness, considered as how responsive the school curriculum could be to surrounding religiosity and diversity, such as whether a school's curriculum would include the common religions or denominations, or religious issues, of the locality.
Two further dyads exemplified indirect contextualisation because they described negotiations within each of the overarching nodes; • Within the school node: School ethos and religious educationschool's curriculum design and choices, in relation to their constitution and wider statutory obligations, such as the curriculum in a church school. • Within the local context node: local patterns of belief and pupil religiositythe patterns of local religious nurture for pupils outside school, such as whether and how pupils were receiving madrassah education, or Christian formation outside school, and how they negotiated this as learners.
Whilst these two dyads do not involve negotiations between the school and the locality in themselves, they affect and are affected by the four directly contextualising dyads. Third, we then deductively sought to saturate each of these dyads with examples from the data, and this saturation confirmed the value of this recursive process (Schofield 2002), as a 'parsimonious, testable, and logically coherent' approach (Eisenhardt 2002, 31). The dyads are shown diagrammatically in Figure 1.
This figure shows how the process of contextualisation between the two overarching themes of school and local context is played out in the four sub-nodes (shown through the dotted lines). The sub-nodes are then inter-connected, as a framework of dyads, and the four dyads of direct contextualisation are shown in thin solid lines, whereas the two dyads of indirect contextualisation are shown by dash-dotted lines. Although we drew on network analysis in labelling four 'nodes', we decided not to term this model a 'network' both because network analysis typically only identifies the emergent causal connections between nodes (Miles, Huberman, and Saldaña 2014) whereas we had deductively created the dyads out of the first level of analysis, and because a 'network of negotiations' might suggest a group of inter-connected individuals or organisations. The diagram itself needs to be treated with caution, as the central 'cross-over' between the dyads of school/  community relations and pedagogical responsiveness may imply a new node; this 'cross-over' however is simply a product of visualising the relationships diagrammatically and is discussed further below. In the following sections we have selected various examples for each dyad, identifying both straightforward and more problematic situations of negotiation. As with the John Hawkins vignette above, each group of examples reflects the complexities and nuances in practice that challenge simple conclusions about religion in education, and about schools and localities: examples from boarding schools, where pupils have almost no local connections, were instructive in challenging our own assumptions. The cases selected as illustrations are not representative of the whole but exemplary, intended to demonstrate the potential of the dyadic framework for analysing research data. Our own research ethics of interpretation was therefore guided by the principle that we were neither judging these circumstances nor the agents involved. The approach seeks to be impartially descriptive rather than critical, unlike for instance Conroy et al. (2013), who in their study of twenty-four secondary schools largely focus on what they judge to be the loss of meaning and purpose in religious education. Furthermore, the dyads might also serve as practical guides for reflexive deliberation by educators and education leaders on the role of religions in their institutions, and we did not want the framework to appear normative. First, we set out the four dyads of direct contextualisation, and then the two dyads of indirect contextualisation.

Pastoral responsiveness (school ethos and pupils' beliefs)
The negotiation between school ethos and pupil religiosity is here termed pastoral responsiveness. All the schools involved thought that they should respond institutionally to pupils' beliefs, though their response ranged from welcome to caution. Faith primary schools were the most welcoming, where there were often concerns about the inadequate response to young people's religion in wider society or other forms of schooling. Thus, the head of Al-Hikmah Islamic School in North London proposed that her school was: Creating an environment that doesn't put the child in a situation where they feel a conflict of values. . .or it takes a longer time to get to the point of self-awareness and identity, or they're given mixed messages, or they're not even affirmed.
In this school, being 'faithful' was recognised and supported.
However, the way in which schools engaged with the religion of their pupils was not determined solely by the designated character (religious or otherwise) of the school. Some 'faith' schools, while supporting their pupils' religiosity, preferred different theological positions. The head of religious education at Christchurch Girls Church of England School in South London, with a largely practising Christian student population from the Black African community, commented that, The school chaplain finds that some of the students' religious positions [e.g. on Biblical literalism, creation and science, or the position of women] challenge her own theology and she would like to show them that there is another way of being Christian.
To have a faith was welcomed, but pupils' own doctrinal views were not expressly reinforced by the school. Here, the dyad is the site for negotiation between different denominational views, not between the religious and the secular.
The approach at Headley, a state secondary school in Oxford with a plurality of religions represented in its student body, was even more muted, and the deputy felt that: While we have the absolute regard for the individual faiths of the individual, we find the collective energies about Religion and Faith things that we want to be watchful and careful of, because they can result in anxieties and tensions.
Pupils as autonomous individuals were to be respected, but 'Religion and Faith' as reified categories were problematic: pupil religiosity was something to be managed rather than welcomed in school.

Pedagogical responsiveness (pupils' beliefs and religious education)
The second dyad considers negotiations between religious education and pupil religiosity: the degree to which and how pupils' own religion (or non-religion) is considered in the delivery of the curriculum. Sometimes, pupils' lack of explicit religiosity led to a rationale for a curriculum based on generic intellectual skills. At Eden Academy, part of a Christian academy chain serving a deprived area of the North East, the head of religious education explained that 'the pupils we have do not know much about religion. . .and they are not very interested in just learning about religions'. This judgement led him to focus on issues-based approaches and ethical dilemmas that would help the students to 'develop their thinking skills so that they can look at any kind of evidence and make opinions for themselves'. Religion had been largely removed from religious education.
Trent Vale, a state secondary school in the East Midlands with pupils from a range of religions, adopted a multi-faith approach with a responsive twist. At GCSE, all students studied Christianity but for their second religion could choose from Hinduism, Islam or Sikhism; most students studied their own religion. Supporting learning about three different religions concurrently was challenging for teachers, a way of delivering the curriculum that probably would not have been adopted had it not been for the plural religious affiliations of the pupils. In a Year 10 lesson on 'God and humans', for example, the teacher led the class through three parallel questions:

Explain how Varnashramadharma influences the life of a Hindu
Explain the importance of Sewa and how it influences the actions of a Sikh Explain how the key beliefs of a Muslim affect the way they live their lives Common themes were drawn out before the students worked in self-selecting pairs on their chosen religion from different resources.
One of the tensions in this dyad was that the representation of religions might not match the pupils' own experience of their religion. For example, at a secondary state comprehensive in the North East, a Sikh pupil commented on his learning: Some things I don't even know that much, but like I didn't know that Sikhs are meant to wear like baggy pants, like all them comb things and like your hair tied and stuff. But I'm not going to wear a turban. My father doesn't wear a turbanand he shaves.
Denominational differences and degrees of adherence could lead to significant misalignments between the image presented in lessons and pupils' own understandings.

School/community relationships (local patterns of belief and school ethos)
Local patterns of religiosity varied considerably for the schools researched, and so did the school responses to them as the management negotiated where religion figured in relations between school and community. In some cases, there was continuity and unity of purpose between community religion and school ethos. A rural secondary school serving a strongly church-going, protestant community in County Tyrone, included in its Mission statement: Central to the creation of this environment is a commitment to Christian values [and local ministers] support the spiritual life of the school in taking a weekly assembly This school, which was located in a religiously observant community, did not have an officially designated religious character, but was more formally explicit about its Christian influences than some church schools elsewhere. Another pattern observed is where church schools (for example St Bonaventure's Catholic School, Edinburgh) gave great weight to Christian values and teaching but are not aligned with their immediate local community, instead serving a more geographically dispersed faith community of Catholic pupils travelling from slightly further afield.
Some schools went to considerable lengths to understand and engage with the religious perspectives of the communities they serve. Senior managers of a secondary school in Bradford periodically went to local mosques to talk to the congregation and organise special events promoting the school within the community, and some even visited Mirpur (Kashmir), whence many of their Muslim pupils' families originated. Twelve teachers took part in a continuing professional development course involving visiting and interviewing local people to increase their knowledge of 'different groups in the community', and of their lives as Muslims including the tensions they faced 'post 9/ 11 and 7/7ʹ. While both these examples show schools engaging with the religions in the locality, a third example shows a contrasting approach seeking to compensate for a lack of religious practice in the community. It has already been observed how, even though it was not a church school, the head teacher of John Hawkins Infant School encouraged links with local churches, through local clergy participation on the governing body and in assemblies, and school visits to and services in local churches. Because the children did not have experience of religion in the community owing to low levels of practice, the headteacher wanted to make sure they gained that experience through school. By involving parents in celebrations of religious festivals, including those held in church, the school was not so much responding to as changing the shape of religious participation in the wider community; indeed, the school was also the agent of interdenominational collaboration, as the site where the two local ministers worked together as governors and in coordinating these activities.

Curricular responsiveness (local patterns of belief and religious education)
Where pupil religion and local community religion are aligned, then responsiveness to pupil religion in the curriculum may be considered as a responsive negotiation with community religion; the parallel GCSE studies at Trent Vale, for example, reflect the multi-religious character of that school's catchment area. Some teaching, however, moves beyond pupils' religion to facilitate their engagement with groups in the local community not represented in the school population and to incorporate in the curriculum the issues and impulses of the wider community. The Jewish religious identity of pupils at Nitzanah primary school in North London did not reflect the multi-religious and largely Asian-heritage local population; however, the school sought to develop links with other local groups as part of the pupils' religious education and in the interest of community relations. The school's schemes of work were designed so that learning about other religions was built around reciprocal visits between Nitzanah pupils and Sikh pupils from one local school and Hindu pupils from another.
The greater flexibility of the primary curriculum, with more opportunity for crosscurricular themes, was exploited at Sunnyside, a primary school in a northern town, to respond to the interests of residents of the socially deprived (largely indigenous white) estate it served. The head was keen to create a story that would bring the community together and help build a sense of shared identity. With little formal religious engagement in the area, he chose to develop with his pupils a project that combined history and literacy around the inclusive theme of 'remembrance', drawing together memories of local families and tapping a rich seam of experience and reflection, both religious and non-religious.
In secondary education, however, a curriculum tied to the themes and requirements of national examinations can limit the flexibility and local responsiveness in secondary education. The head of religious education in a school in the Outer Hebrides shared her views on Sunday ferries, the big subject for debate on the island in recent years. Allowing these ferries and the opening of shops and tourist facilities on Sundays proved problematic for a community with a strong Sabbatarian tradition. The teacher admitted that, were it not for the constraints of the examination syllabus, she would have liked to engage pupils with this community issue and use the film 'Chariots of Fire' as stimulus for ethical debate on Sabbath Day observance instead of using 'Million Dollar Baby' as prescribed by the examination board for a discussion on euthanasia. In the negotiations between local patterns of belief and curriculum the immediately local gave way to a topic of broader interest to wider society.

Indirect contextualisation
School's curriculum design (school ethos and religious education) The first of the indirect dyads considers the inter-relationship between religious education and school ethos, termed school's curriculum design. Consistency between the two was often apparent in faith schools. In Rambam School, a Jewish Secondary in the South East, the explicit rationale of the school was to nurture young Jews. The Department of Jewish Studies within the school was responsible for this, covering study of Tenach, Talmud, Ivrit, Jewish history, services and tsedakah. The school did not have multi-faith religious education, though other religions were sometimes introduced in Jewish history lessons, and under the term 'love of the world' a series of sessions for older pupils on interreligious dialogue involved engaging with Christian and Muslim visitors.
Nevertheless, ostensibly faith schools could be juggling with other priorities, cutting across their religious ethos. Erdingbury, a secondary independent school with a church foundation in the south of England, had recently shifted towards more philosophical approaches to religious education, which was deemed by staff to be particularly appropriate for a selective school with 'high ability' students, as well as a form of pedagogical responsiveness to the atheist or agnostic majority of pupils, like Eden Academy. However, the school was concurrently keen to reaffirm its Christian (High Anglican) identity, so inserted 'crash courses' in knowledge of the 'historic tradition' and the chapel liturgy into religious education: different demands led to curriculum compromises.
In state schools, wider concerns affected decisions about curriculum aims. At Lingard School, a multi-religious state primary in the East Midlands, a concern to emphasise a strong multi-cultural message led to a universalist theological position.
Here is the head teacher: As I keep saying to the children, religions are basically very similar . . . all the Gods' rules are basically saying the same thing: of how we must be patient and how we must be respectful, the more I go into religion the more I point this out to them.
However, this normative universalist position adopted in the context of diversity was not necessarily reflective of pupils' own different beliefs, and potentially in conflict with them, as few members of most world religions would suggest that all religions have precisely the same doctrines or ethics.

Local religious nurture (local patterns of belief and pupils' beliefs)
The second indirect dyad brings together local religiosity and pupils' beliefs, as a form of identity negotiation. This dyad is a reminder that the negotiations pertinent to our theme of religion and education do not just take place within school or between the school and other nodes, but there are other relevant extra-institutional negotiations taking place. The point of connection between local religiosity and belief is religious nurture. A Christian pupil in Christchurch School, a Church of England girls' secondary school in South London, explained: School religion is . . . like feeding information into your brain, but outside school it's experience, it's practical, you're actually living in that society.
Her comment repeats the idea that the lived religion of the pupils is in the community rather than school. This is especially so in her community where the (largely Africanheritage) population is 'very religious'. Where the levels of religious practice and general visibility of religion in the local community are low, however, school religion rather than lived community religion was often the young people's primary point of contact.
Further, teachers often understood their pupils as religious beings in relation to their belonging within a local faith community. At Lingard primary school, one teacher described the pupil population as 'living, breathing examples of different religions', particularly of Islam, Hinduism, Sikhismprominent community groupings in this area of the East Midlands. She was careful to emphasise the difference between religious nurture in the community and religion in school: School is a place, not where you leave your religion behind, but that you are learning in a different way that might require you to think slightly differently from how you are taught at home or how you are taught in your place of worship.
Other cases challenge this assumption of unity between local and pupil religion (or lack of religion), for example where there are local models of religiosity that the young people do not share. Pupils in Forres (Morayshire) and Birtley (Tyne & Wear) lived in proximity to close-knit minority religious communities, an alternative commune and a community of Haredi Jews respectively. In both cases they commonly viewed these local religious communities as 'weird', sharing jokes and myths about their practices: 'they apologise to carrots before eating them', wear 'odd' clothes or live in 'dirty' houses. Even where a school serves a multi-cultural catchment area, within that locality there may be smaller neighbourhoods where people of distinct religious identities congregate, pupils in a Glasgow school spoke of streets, and those in a Swansea school, of estates, where particular religions and ethnicities (such as Muslims or Poles) gathered. Thus, teenagers may have a negative relationship with a local religious scene, viewing their religious neighbours as 'other' and alien, and this affected their views of the world religions that they studied.
In the independent boarding schools where pupils came from further afield the young people were less engaged with local patterns of religious practice, nurturing or alienating. At Erdingbury School, students' interest in Dawkins' (2006) ideas and widespread adoption of New Atheism was evidence of teenagers setting up their own community of non-belief within school, influenced less by local religiosity and more by wider media-fed trends.

Discussion: a framework of negotiations as contextualisation
We have taken time to develop a conceptually nuanced middle-range theoretical framework for exploring the patterns of negotiation between religions and education in local contexts that is empirically rooted in a qualitative meta-synthesis of over forty casestudies. We consider that it is 'parsimonious, testable and logically coherent' (Eisenhardt 2002, 31), and here briefly outline three main potential contributions to wider debates and research. First, the examples given above offer an indication of the complexities within this field of research, especially in theorising what is meant by 'context'. With each dyad, neat correlations (whether of school type and place of religion within it) are challenged by the findings, for example in the variety of negotiations between school ethos or religious education with pupil religiosity or community religions. We noted earlier that the nodes and sub-nodes adopted for our model of analysis were not necessarily institutions or human actors, even though the actors (such as pupils, teachers, or head teachers) have frequently been referenced or cited in the examples. They are the negotiators of the relationships between the nodes, and their differing interpretations of those relationships help to explain the particularities (and occasional internal inconsistences) of different responses. The dyads are not just conceptual hooks but, in keeping with the inductive character of our analytical model, reflect the processes taking place in the contexts researched. This explains why some studies that previously emerged from this research can be situated in one or more dyads: school ethos and different approaches to Muslim religion of pupils (Ipgrave 2017c;Ipgrave, Miller, and Hopkins 2010), Jewish school ethos and curricular approaches to other faiths (Ipgrave 2016), or school ethos and curriculum (Fancourt 2016).
Second, we have set out how contextualisation is not simply a binary exchange between the school and the local context. We theorise contextualisation as being between the four sub-nodes, thereby identifying both direct and indirect contextualisation, so that a more complex framework of negotiations can be seen at work. Schools might respond at an institutional level to local religiosity, which has implications for their approach to religious education, and which then has to engage with pupils' religiosity, which is itself linked to local religiosity. A negotiation in one dyad has repercussions elsewhere, so that there is a constant process of re-contextualisation as the elements of different sub-nodes shift.
Further, the framework could be used as a way of exploring the contexts of other schools in the UK or other countries, and we consider that this model goes beyond its application to the underlying research. For example, the recent 'Trojan Horse' affair in Birmingham (Arthur 2015), when a few Muslim governors of a state school were alleged to have imposed Salafist requirements on the school (Clark 2014), is a good example of when greater conceptual clarity on the relationship between local religious context, pupils' beliefs and school ethos would be valuable, both for research and in policy debates. Indeed, it would be instructive to compare the contextual negotiations in those Birmingham schools with the porous affordances of John Hawkins School, to see if similar patterns of negotiation were in play, and therefore whether the public issue was the nature of the pervading religiosity -Islam not Christianity -and not the nature of the school-community relations.
Finally, the framework could serve as the starting-point for an analysis of the effect of policy at a local level. For example, how neoliberal policies on school governance (Arthur 2015), securitization (Mac an Ghaill and Hayward 2017) or assessment (Fancourt 2017a) affect these local negotiations. In this respect it can be modifiedindeed for research which focuses on one of the four sub-nodes more exclusively, these negotiations can be re-visualised around a particular sub-node. In Figure 2, school ethos becomes the focus: the other three sub-nodes sit around it, and the dyads connected to it flow out to them, while the other three dyads provide the wider context of negotiations within which this sub-node is located.
The framework is therefore a heuristic device, to provide a hermeneutical lens for other analyses that seek to explore these elements further, particularly by drawing attention to the interplay of the different factors involved.

Conclusion
As noted in the introduction, the place of religions in education is marked by a range of historic, well established patterns of establishment, alongside changing patterns of religiosity, more recent issues of radicalisation and anti-radicalisation, and increasing policy demands on schools and schooling, and remains highly contested. In addressing these we have focused on the micro-level of local context, rather than seeking to devise a model that would also cover the macro-level questions of the religious establishment or even how the two inter-relate, for example through policy implementation. The data from the case studies might have enabled some wider comments to be made in regard to these issues, but their richness lay more in enabling us to unpick the dynamics of context through the conceptual lens of negotiation. Religious (and nonreligious) beliefs, values and practices enter into a relationship with secular educational beliefs, values and practices which can support, contradict or constrain each other in nuanced and shifting patterns.
Moreover, we reiterate calls for greater empirical clarity and depth in these debates, to question an over-reliance on grand theory, notably about secularisation, or indoctrinationreligious or otherwise. As noted above, we have sought both to take religions seriously and to go beyond simplistic polarisation. In our account, where school ethos is not supportive of pupil religiosity, the effect of current negotiations might be secularising, even in faith schools, whereas the more porous affordances of John Hawkins could be a sacralisation of an ostensibly secular space. The framework also allows for interreligious and inter-denominational differences to be highlighted, as well as how agnosticism or atheism is played out.
Finally, we would argue not only for robust theoretical models, but also for a greater diversity of methods of theory-generation. Middle range frameworks such as ours can be placed alongside, and put in conversation with, those which are created from grand theory, including more generic sociological models of the place of religions in contemporary society, and as a result the process of theory generation is itself in oscillation between different approaches Notes 1. While it is common to talk and write of 'faith schools', the official term in England is 'schools of a designated religious character', this group of schools holding a variety of positions on the significance of religion within school ethos, student population and practice. Here, however we use the term 'faith schools' more loosely as we intend to include any such school, whether state funded or independent, both in the UK and internationally. 2. Due to the confidential nature of some of the research materials for some of the underlying research projects supporting this publication, not all of the data can be made accessible to other researchers. Please contact the corresponding author for more information.