Power, authority and expertise: policy making about relationships and sex education in English primary schools

ABSTRACT This paper analyses the experiences of English primary school leaders who perform critical roles in interpreting government policy and navigating the landscape to design relationship and sex education (RSE) for their pupils. It considers schools as sites of political contestation and educators as policy actors, the voices of whom are often absent in the literature about RSE. Drawing on Peter Morriss’ theory of power, the paper considers how primary school decision makers utilise their epistemic abilities to advance their policy preferences and how structures of authority and legitimacy energize and constrain them in their policy-making work. The findings suggest a paradox in power: while the national government has delegated decision-making about RSE to individual schools, it has simultaneously failed to uniformly equip schools to make appropriate policies about RSE and enabled parents to deny schools’ credibility and authority as leaders in RSE.


Introduction
This research was inspired by a curiosity and concern about the precarious predicament of front-line providers responsible for children's formal learning about relationships, sexuality, gender, the development of their bodies, and themselves as individuals, in relation to people around them. At the time of the data generation (2015-16), relationships and sex education (RSE) was not yet statutory in England (though this was on the horizon) and the government had tasked each school with developing its own relationships and sex education policy (a requirement that remains in the current legislation, which came into force in September 2020). The study from which this paper derives aimed to offer a new perspective on RSE that re-defined schools as sites of political struggle and teachers as policy actors. The results illuminate complex and diverse stories of school-based policy deliberations and negotiations. This paper draws on Peter Morriss' (2002) theory of power to analyse the capacities, freedoms and constraints experienced by school decision makers in policy making about RSE. Morriss' theoretical approach is well-suited for this research because it addresses the substantial literature that attributes poor quality RSE to poor leadership for RSE in English schools, including low priority for RSE in senior leadership teams and a lack of qualified or experienced RSE educators (e.g. Formby et al. (2011); House of Commons Education Select Committee 2015 ;Ofsted 2013;Macdonald 2009;Alldred and David 2007). Considered in relation to RSE policy, Morriss' theory (2002) helps to analyse the productive capacities of school leaders to create and enact suitable RSE policies, within their contexts.
The central argument of this paper is that England's policy for RSE has produced a paradox in power: while the government has delegated decision-making about RSE to individual schools, it has simultaneously failed to uniformly equip schools to make appropriate policies about RSE and authorised parents to deny schools' credibility and authority as leaders in RSE. In addition, the government guidance on RSE (at the time of data generation) was ambiguous and inconsistent with recognized good practice (the International Sexuality and HIV Curriculum Working Group 2011; Cullen and Sandy 2009;Mason 2010;Alldred and David 2007). While this study was conducted in England, schools and governments elsewhere are and have grappled with questions about how to design relationships and sex education policy, and therefore the discussion and insights offered here are widely relevant.
I use 'relationships and sex education' (RSE) to refer to the subject, as this is the term used by most study participants, however it is known by other names. The national government policy at the time labelled the subject 'sex and relationship education' (SRE) (Department for Education and Employment [DfEE] 2000) and internationally 'life skills education' and 'comprehensive sexuality education' are used (e.g. Breuner et al. 2016;UNESCO 2018). In current legislation, 'relationships education' is statutory (with sex education remaining optional) in all English primary schools, and 'relationships and sex education' is statutory in all English secondary schools (Children and Social Work Act 2017, sections 34 and 35; Department for Education [DfE] 2019). In England, RSE is part of a broader, umbrella subject known as personal, social, health and economic education (PSHE).

Morriss' theory of power as a tool for analysing RSE policy making
Morriss proposes that power is the capacity to realize one's aims, which means that the holder effects specific and deliberate changes, rather than merely influencing outcomes or affecting trajectories (2002). Applied to RSE policy making, Morriss' theory is a tool for analysing the capacity of school leaders to achieve coherence across policy intentions, that is policy vision, objectives and values, and policy enactment, the practices that make policy visible and experiential (Fenwick and Edwards 2011). Further, one might interpret Morriss' emphasis on 'effects' to translate not only to policy enactment in education but to specific learning outcomes. To be clear, the data that informed this paper comprise the accounts of school-based policy makers alone: head teachers/principals, subject lead teachers for RSE, other teachers, pastoral care and administrative staff. The data addresses only their experiences of RSE policies and policy-making processes -not the delivery or outcomes -and therefore while I explore power in terms of the capacity to realize one's aims, I cannot comment on the effects produced among pupils. Nonetheless, this in-depth reading of policy makers' intentions and accounts of policy enactments contributes insight into their experiences of leadership for RSE and is indicative of RSE provision.
While I aim to analyse RSE policy makers' power, I nonetheless agree with the substantial literature and theorisation of policy making and enactment as complex, ambiguous and contradictory, which are driven by, and undone by, human and nonhuman subjects and concerns, with local policy enactments as disparate from each other as from the national policies upon which they are based (Fenwick and Edwards 2011;Ball et al. 2011;Ball 1994). Ball has called policy manifestation a 'wild profusion' and a 'bricolage' -images to represent the disorderliness of policy making (Ball 1994). Against this backdrop, Morriss' concept of power, the possibility of one effecting specific outcomes, does seem too tidy to apply to educational policy making.
Morriss' theory of power is a dispositional one, however: it recognizes that power is conditioned by 'what [people] can do given the circumstances they do find themselves in ' (2002, 80), further that one may not choose to exercise power. There is value in exploring what policy makers can do and where they experience constraints. In a study of school policy making about inclusion, Meo (2015) found that while policy makers interpreted and enacted policy in creative ways, they did so within discursive boundaries that limited what they could imagine, think and do. While this example relates to the influence of dominant discourses in the educational context, 'circumstances' include localised contexts. Morriss uses the term 'ableness' to describe things that people 'can do in the conditions that happen to obtain at those times ' (2002, 85). Ablenesses include things like having the necessary qualifications, formal authority and procedural competence.
While situating school policy makers' within their structural, discursive, material and relational contexts, I centre school policy makers as agents of change and envisage their power as productive and positive, in line with Foucault's conception of a force that 'traverses and produces things, induces pleasure, forms knowledge, produces discourse' (1980,119). I am interested in the power they wield in spite of or because of personal aptitudes and skills, professional roles, cultural norms, codified regimes of behaviour, standards of knowledge, and local, community and national webs of relationships.
With regard to personal leadership qualities that may have a bearing on experiences of power, Morriss' concept of epistemic abilities has particular relevance for RSE policy making. Epistemic ability means possessing the knowledge that is necessary to enable an agent to make decisions with intention and reasonable confidence in the outcomes they will produce (2002). This distinctive aspect of Morriss' theory of power is importantand I refer to it throughout the discussion below -because, as mentioned earlier, lowquality RSE in English schools has been repeatedly attributed, in part, to low levels of expertise and qualification among RSE educators and policy makers (Formby et al. (2011);House of Commons Education Select Committee 2015;Ofsted 2013;Macdonald 2009). This paper further interrogates the suggestion that policy makers must have some special knowledge about the subject in order to exercise power, which in this case is about making the choices needed to bring about specific changes to achieve an RSE vision.

Where are the children?
Before proceeding, I want to recognize that the desirability of school leaders alone articulating an RSE policy is contested. Scholars and practitioners argue that learning about relationships and sexuality is a unique subject area and given its nature children must be involved in defining RSE policy and providing feedback to refine RSE delivery (Renold and McGeeney 2017;Pound, Langford, and Campbell 2015). I sympathise with this perspective. I asked my contacts at each school to identify anyone who contributed to the RSE policy -adult or child -with a view to inviting them to participate. At each school, it was adults who decided the policy and therefore their data inform this paper.
The paper begins with a brief introduction to the history of RSE in England and a review of the relevant literature. These sections contextualise the RSE task assigned to schools involved in this study. The subsequent section outlines the research design and data analysis processes. I then discuss the major themes that emerged from the data that illuminate how power is experienced and exercised by policy makers at each of the three schools in their efforts to effect RSE policy.

Context
Relationships and sex education (RSE) has a tumultuous history in England: heated public debates and political decisions about RSE have repeatedly been entwined with conflicting ideas about the distribution of responsibility (schools, parents, health professionals) and the 'right' to control children's education as it pertains to sex, sexuality and relationships (Pilcher 2004(Pilcher , 2005Thomson 1994Thomson , 1997DfEE 2000). While RSE has been part of state education for the best part of a century, it has always occupied a peripheral position and, in contrast to academic subjects, the national government has often adopted a 'laissez-faire' approach, preferring to delegate policy decisions by commissioning or delegating to other organisations (e.g. civil society organisations, arms-reach organisations, schools). The contentious nature of RSE is one of the reasons why successive governments have been reluctant to take responsibility and provide leadership in RSE. However, while there are some studies on government policy-making processes in relation to RSE (Wilkinson 2017;Lewis and Knijn 2003;Thomson 1997), there is limited research about how schools have implemented government RSE policy and how they have encountered contentious issues.
Although primary schools in England have been responsible for developing their own policy for RSE 1 since 1986, there is a paucity of knowledge on how schools performed this duty. This gap in the literature persists in spite of a number of reviews over the last 13 years that have suggested that RSE was inadequate in many primary schools, in part due to poor leadership ( Educationalists such as Ball (2003Ball ( , 2009) and Wilkins (2015) have suggested that as managerial responsibilities have been devolved to local levels, school leaders are being incentivised to develop business and management practices. Wilkins suggests that school leaders' value is based on their ability to make the school 'intelligible and responsive as a cost-cutting, profit-making business ' (2015, 12). Relationships and sex education do not align well with these financial and political shifts and bring to the fore questions about how this subject might fare in the hierarchy of school policy priorities.
Given the contentious history of RSE, evidence of poor RSE quality and shifting emphasis for leadership in schools, there is a compelling rationale for researching how school leaders, and others involved in deciding RSE policy, exercise agency and employ their knowledge and abilities to determine their school RSE policy.
It is not the intention of this paper to engage in an in-depth discussion about what characterises an 'appropriate' RSE policy or programme, however this analysis could not discuss power in RSE policy making without making some judgments about RSE quality. There is scant literature on the precise competencies that RSE policy makers or educators should possess, so I point instead to the literature about good-quality RSE itself. Drawing on expert frameworks and academic literature from Agenda (2019), UNESCO (2018), Renold and McGeeney (2017), the International Sexuality and HIV Curriculum Working Group (2011), Cullen and Sandy (2009); DePalma and Atkinson (2009) and Mason (2010), -all of which consider a range of issues related to quality RSE at the primary level specifically -I offer the following key principles for good quality RSE in primary schools: • RSE should be grounded in human rights, including human dignity, safety, equal treatment, the right to health and opportunities for participation; • RSE should advance gender equality; • RSE should promote reflection and critical thinking for example, about the influence of social norms, values and beliefs on behaviours, emotions and body image; • RSE should be responsive to children's lived experiences, culture and questions; • RSE should be comprehensive, addressing a wide range of issues to a suitable level of depth; • RSE should be inclusive with regard to identities, choices and family structures; and • RSE should be informed by evidence, for example, the science of human development, physiology and sexuality, curricula, pedagogy, experiences and outcomes of RSE.
This is not an exhaustive list of recommendations for good-quality RSE, but comprise standards that are consistently associated with good-quality RSE.

Research design 2
The study was conducted with three primary schools in the same mid-sized English city. The schools and participants included in this paper are: • Latimer Primary School 3 : a popular, ethnically diverse local authority-managed school near the city centre. The respondents are Laurence, a member of the senior leadership team; Sarah, a member of the senior leadership team; Amy, a member of the pastoral care staff; and Iris, a member of the pastoral care staff. • Fleming Primary School: a small private school in an urban, affluent area. Most pupils are white and have English as a first language. The respondents are Simone, a member of the senior leadership team, and Lena, a parent and member of the administrative staff.
• Wingfield Academy: a large academy located in a predominantly British, white, middle-class neighbourhood. The respondents are Kirsten, a teacher and PSHE lead, and Patrick, a member of the senior leadership team.
I was introduced to the head teachers of these schools by a contact at the local authority, a contact at another primary school from a separate research project and my doctoral supervisor. I conducted two semi-structured, in-depth interviews with most of the respondents. Building rapport over two interviews helped to build trust (Vincent 2013), leading to purposeful conversations that yielded 'rich insights into people's biographies, experiences, opinions, values, aspirations, attitudes and feelings' (May 2011, 120). The interviews, lasting about 1 hour each, were conducted at the school sites during school hours. Data generation began in June 2015 and was completed in December 2016. My analytical approach included data immersion, coding and meaning making through abduction. I reflected on social dynamics among the agents, how they each contributed to decision-making and their perceptions. I noticed repetitions, tensions and inconsistencies. I coded the data using NVivo, a qualitative data analysis software application. My initial codes mirrored terms that I found in the theory, including terms from Morriss' theory (2002) such as 'able', 'expertise', 'abilities', 'knowledge' and values from the literature about RSE, including human rights, gender equality, diversity, community/parental engagement, pupil involvement and scientifically accurate information (Agenda 2019; UNESCO 2018; Renold and McGeeney 2017; the International Sexuality and HIV Curriculum Working Group 2011). I re-read the transcripts and wrote short observations and reflections on each participant and school, identifying themes and ideas that diverged from the theory and literature. Further analysis drew on theoretical frameworks that had not been considered prior to data generation.

Data and analysis
Policy makers at all three schools exercised some level of power -the capacity to effect their aims -by successfully undertaking the task of revising the written RSE policy and implementing it. However, the data suggested that all policy makers aspired to more than simply adopting and enacting a RSE policy: they intended to implement a new, improved policy that would support pupils' learning in RSE.
I focus the rest of this paper on the nature and appropriateness of policy makers' decisions about RSE policy, with regard to their epistemic abilities around RSE and their ableness. Epistemic ability refers to special knowledge and expertise in the subject. Ableness includes how these individuals experienced their own, and others', authority within their school and the wider context, and other circumstances that affected their freedom and capacity to make appropriate choices for RSE policy.
I identified the following patterns through my reading and analysis of the data: • Policy makers for RSE possess varying levels of epistemic ability in RSE • Government authority on RSE is uniformly observed and respected • Policy makers' ableness is restricted by the legal provision that parents can withdraw their child(ren) from RSE Each of these themes influenced the capacity of decision makers for RSE to effect change in RSE policy. After addressing each of these themes I will discuss the insights that the paper offers for policy making about RSE.

Policy makers for RSE possess varying levels of epistemic ability in RSE
In his research about anti-bullying policy, Walton (2010) argues that definitions constrain not only understanding of the problem but pose limitations on how policies are implemented. He wrote, 'the ways in which problems are articulated and approached varies in accordance with degrees of power held by individuals. Whose voices are included in articulations of the "problem"? ' (2010, 136). Walton suggests that it is those in formal positions of authority for policy making who define the agenda and decide whose voices matter in policy enactment, which relates to Morriss' recognition of 'ablenesses' -as things that people can do, due to their position or qualification, at a point in time -as a part of power. In the case of RSE, while a policy maker may not hold the responsibility they do in relation to RSE policy as a result of their epistemic abilities in the subject, their epistemic abilities nonetheless contribute to shaping the narrative, the problem and the solution in RSE. For example, Abbott, Ellis, and Abbott (2016) noted how overarching policy language around public health used in discourses around RSE shaped educators' understanding and delivery of RSE: educators' failure to reflect and critically engage with RSE much beyond the dominant discourse suggests low epistemic ability. This section explores the epistemic abilities of the school-based policy makers as suggested through their definitions of RSE. At Latimer Primary, Sarah, a member of the senior leadership team, defined the school's approach to RSE in terms of what it should enable children to achieve: What we're trying to do here is giving children enough knowledge and empowering them to be able to make healthy choices as they grow up and being able to navigate the complex . . . relationships between friends . . . and as they grow older, between co-sexual partners . . . It's what we call 'caught and taught', so it's not just about . . . a letter goes home and we're going to have the 'sex lessons' . . . we're talking all of the time about permissions, and about consent, and about the way that we are, and we're living that through the values of the school.
Sarah makes clear that her commitment to teaching about consent, healthy relationships and about informed decision-making (within and beyond safeguarding concerns) are embedded in RSE and in the culture and ethos of the whole school. 'Caught and taught' is a pedagogical approach that refers to teaching lessons and values through everyday practice.
Wendy, a member of the pastoral care team, gave this example: [Laurence] just handed out some information about how to talk to girls. You know, you make sure you are inclusive and aren't setting yourself up, saying 'oh isn't she a pretty little girl' . . .. we are educating people all the time. We are constantly saying things like, 'better you rephrase that, reframe that question . . . we would like you to use these words instead', and that's the only way you can move forward . . . This passage illustrates how the staff promote values that underpin good RSE -such as gender equality and respectful interpersonal relations -in their everyday interactions.
The school is encouraging staff as well as pupils to critically reflect on their language and their behaviour with a view to promoting inclusive, caring relationships. This suggests a collective commitment to RSE values and reflects a level of planning and anticipation of results that Morriss suggests is an important aspect of epistemic ability (2002). Another principle of RSE at Latimer Primary was that lessons should be child-led. Iris said, 'at the beginning of it all, we try to find out what do [pupils] know, what do they want to know, and what they need to know . . . there is no point going and talking about masturbation if all of them know what that is'. Although much of the literature about children's participation derives from secondary schools, it is increasingly recognized that even at primary-level children's ideas can contribute to making RSE more relevant to their lives (Agenda 2019; Renold and McGeeney 2017;Pound, Langford, and Campbell 2015).
Among the respondents at Latimer, four of them had completed at least one comprehensive course related to RSE, and three of them had undertaken additional training on different aspects of social and emotional learning and/or pastoral care. These qualifications, as well as years of experience in RSE, were reflected in their epistemic abilities for RSE.
In contrast to Latimer, Wingfield Academy is a large, predominantly white, middle-class primary school. Patrick, a member of the senior leadership team, defined RSE as follows: '[RSE] is making children aware of kind of age-appropriate information that they should be aware of, just so there's not such a stigma around the subject, and it is taught to them in a factual way, in a mature way'. Patrick's understanding of RSE was focused on natural science, such as human physiology and reproduction. While Patrick signalled an emphasis on evidence-based RSE, he seemed unaware that sex and reproduction should be taught in the context of social situations, including discussions of emotions and considered decision-making, according to scholarly literature (e.g. Mason 2010). Research with children suggests that teaching about sex and reproduction is delivered too late and too often through biological, 'factual' discourses that are less meaningful for them (Pound et al. (2016), UK Youth Parliament 2007). Kirsten also used a positivist discourse and offered more detail about the content: [RSE] is . . . a science lesson . . . different parts of the body, the science behind how babies are made. We don't in year one, we're literally just naming body parts, and again that's more of the safeguarding. Touch wood nothing ever happens, but if it were to, they would know exactly what to say Kirsten situated RSE as part of safeguarding efforts and as part of science education. Like Patrick, Kirsten suggested pupils would learn to distinguish fact from fiction in RSE, and these distinctions were perceived as objective and uncomplicated. Kirsten acknowledged that she had adopted her narrative about RSE from a one-day training session on RSE and this was the framing she used to rationalise RSE, delivering it verbatim to educators charged with delivering RSE and to parents. Kirsten and Patrick, as the key decision makers for RSE at Wingfield Academy, demonstrate much less sophisticated epistemic abilities in RSE compared to respondents at Latimer.
As an independent school, Fleming Primary School was not obligated to design its RSE policy with reference to government guidance at the time, however the school's definition and understanding of RSE was comparable to that offered by respondents at Wingfield. Simone, a member of the senior leadership team, explained the school's approach to RSE: So we're looking at relationships in terms of interactions . . . what is appropriate, what is inappropriate. So it is looking at how to keep them safe . . . we relate [relationships] a lot to the physical side of things, with hormones. We blame an awful lot on hormones, actually, certainly further up the school, children seem to accept it as perfectly acceptable . . . We don't do [sex] in PSHE, we do in science . . . the sex in relationship begins with life cycles . . . we have the eggs and they hatch, and then we look at that, and we get chicks . . . it's not human reproduction, but this is how [babies are] made . . . it's something that happens between a mummy and daddy. We only use 'mummy and daddy', we don't use any other words because there is always a mummy and a daddy.
Like Patrick and Kirsten at Wingfield, Simone and other respondents at Fleming Primary defined RSE in terms of its contribution to safeguarding and as a vehicle for delivering sex education through a biological, science-based discourse. Among respondents at Fleming, Simone had the most comprehensive perspective of what RSE entailed and its objectives, she had completed a continuing professional development course on PSHE. However, some of her statements (e.g. 'we blame an awful lot on hormones', 'there is always a mummy and a daddy') conflict with arguments that RSE programmes should be honest, factual and inclusive (e.g. DePalma 2013).
The suggestion that 'there is always a mummy and a daddy' conflicts with government guidance to schools on the Equalities Act, which includes sexual and gender identities among the protected characteristics (Department for Education [DfE] 2014). From a human rights perspective, RSE should represent heterosexual and same-sex couples and relationships in non-discriminatory ways, thus including teaching about couples of two women or two men, and single people, who may create families by having children through assisted reproductive technologies or adoption. Respondents' reluctance to discuss different kinds of families and reproduction suggests discomfort with an important aspect of RSE.
Other respondents at Fleming Primary reinforced this traditional perspective. Speaking about the school culture, Lena, a parent and school administrator, said: Something that the school is very proud of, is allowing children to be children, and protecting that childhood as long as possible, from all the stuff that goes on in the outside world . . . part of that is the whole kind of biological side of sex and sexual relationships, and sexual body parts . . . that's not part of . . . fluffy teddy bears Lena's suggestion that the school needed to protect children from the 'outside world 'for instance, by not teaching them about their own bodies, and Simone's comfort with allowing children to believe things that were not based in evidence ('children seem to accept it as perfectly acceptable') suggest a lack of epistemic ability and an insincere approach to children's learning. These views represent a misplaced perception that children are somehow not 'in' the world and that by exposing the 'biological side of sex' and 'sexual body parts', their teachers might rupture a delicate, imagined membrane securing the integrity of their childhoods.
In contrast to Latimer, respondents at Wingfield Academic and Fleming Primary demonstrated low levels of epistemic ability. They portrayed RSE as primarily about teaching the biological 'facts' of puberty and reproduction, while references to human relationships were polarised: inappropriate and unsafe (i.e. RSE as safeguarding), or friendly, familial and uncomplicated. In reality, sexuality and relationships, perhaps particularly the relationship with oneself, is at times fraught with uncertainty and tentativeness, in turn bringing us joy, pleasure, dissatisfaction, frustration or pride, and RSE should be considered an opportunity to engage children in frank reflection and discussion about these wide-ranging aspects of social, emotional and physical experience. The epistemic inability of respondents at Wingfield Academy and Fleming Primary to conceptualise RSE in more complex ways, and to envision -let alone articulate -what they wanted their RSE policy to achieve for the children they delivered it to, translated into shallow RSE policies and practices that were unlikely to produce the kind of transformative, affirming life experiences that good RSE can deliver (e.g. DePalma and Atkinson 2009; Cullen and Sandy 2009).

Government authority is uniformly observed and respected in school RSE policy making
As Morriss argued (2002), authority is a key element in power and agents' exercise of power often operates within a system where there are different levers that distribute and mediate authority. As noted in the introduction, primary schools were tasked by government with developing their own RSE policy, but within certain parameters, and within the overarching context of a neoliberal education system. At Latimer Primary, Iris stated that whatever other criteria and resources the school considered in their policy-making process for RSE, 'ultimately, in terms of the school, whether we like it or not, the local authority and government tell us the line, and that's the line we don't cross'. This recognition of government as the ultimate authority on RSE was reflected by most respondents at Latimer and Wingfield, the two state schools in the sample. Even Fleming Primary, the independent school, used government guidance (DfEE 2000) as a benchmark for what they should be delivering in RSE.
At Wingfield Academy, Kirsten situated the responsibility of being accountable to government as central to her role: As PSHE lead, it is my job to measure the success of PSHE, how children are developing and their progress . . . when I first took over, my first thing . . . was to get a new scheme then, because [the previous scheme] was awful and it was really hard for me to measure progress when it seemed like most of the year groups were doing the same thing . . . we have an assessment criteria now, which we didn't have before, so now I can actually see which year groups are making the most progress, which year groups have more higher-achieving students, and who's been pushing things, which we never had for PSHE before . . . So hopefully by the end of the year we'll actually have some sort of graph or table of progress, which would be amazing.
Here Kirsten discusses how the selection of a new curriculum package enabled her to monitor and document progress in RSE. Kirsten's talk was embedded with neoliberal principles that have dominated UK education policy in recent years, for example with terms such as 'assessment criteria', 'higher-achieving students' and 'table of progress' (e.g. Ball 2003). Although Wingfield focused more on 'factual' content in RSE (see above), and therefore perhaps some conventional learning and assessment modalities could be applied, I would suggest that given the underpinning values and aims of RSE (e.g. Agenda 2019; UNESCO 2018), mainstream assessment methods are not suitable. My assertion that Wingfield Academy's emphasis on process and documentation is disproportionately dominant in their RSE policy making reflects critiques of other social and emotional wellbeing programmes that suggest the complexity of the subject matter is poorly served by an educational system that is dominated by targets and performance measures, in part because efforts to monitor indicators and document achievement displaces authentic engagement with the intended beneficiaries and programmatic intentions (e.g. Watson et al. 2012). While respondents at Wingfield also mentioned pupils and their experiences, it was clear that demonstrating their performance and achieving good outcomes themselves in government (e.g. Ofsted) inspections was at the forefront of their minds. The government, not the pupils, was their client.
Respondents at Latimer Primary were uniquely critical of the government's authority and guidance on RSE, even while they adhered to it. Laurence, a member of the senior leadership team, gave the example of contraception to illustrate how the government guidance limited what the school could teach in RSE, against the judgment of their own staff. He said: In primary school we're not supposed to talk about contraception . . . I personally think that if people are asking then I should answer that. But it has been something we have had to try to avoid . . . I think that really should be mentioned in year six. There are really young parents.
Although the 2000 government guidance does not explicitly prohibit primary schools from teaching about contraception, contraception is clearly identified as a topic for secondary schools (DfEE 2000). According to respondents, the local authority advisor for RSE had advised the school against teaching about contraception, reinforcing suggestions in the official guidance. Amy, the SENCO coordinator, also suggested that earlier, statutory teaching about sex and contraception would produce better outcomes around teenage pregnancy: When you're looking at things like teenage pregnancy . . . it's still . . . higher than anyone would expect . . . we're doing it [RSE] too late, we're targeting children too late . . . if schools were given less autonomy . . . and all had to deliver kind of statutory, the basic, from an earlier age, than I do think, ten years' time, we would see a difference.
Amy's complaints about the national RSE policy echoed critiques from studies and surveys involving secondary pupils (Alldred and David 2007;Teenage Pregnancy Independent Advisory Group 2009;UK Youth Parliament 2007). Reinforcing her colleagues' comments, Sarah noted that government guidance for teaching about reproduction does not reflect reality, including the reality of some of her pupils: There are things in the papers and in the news, and on the telly, that the children are hearing about, and then they are coming into school to do the relationships and sex education, and we're talking about, 'well, you need a man's sperm, and a lady's egg, and this is how a baby is made', well actually, no, there are so many other ways that babies are made now. And we have children in the school who have been made in lots of different ways, who know that they have . . . suppose there was an opportunity to be able to, to answer that question or to talk more generally about, 'well actually, Jim, there are lot's more ways, yes'.
By 'made in lots of different ways', Sarah refers to assisted conception methods, such as in vitro fertilisation. Sarah expresses her frustration that the government does not support a comprehensive factual representation of human reproduction, and at the centre of her concern is the pupils themselves. Sarah's view of government RSE policy contrasts sharply with that of respondents at Wingfield and Fleming, who perceived government policy on RSE, and their own RSE programmes, as factual and science-based (despite its narrow definition of human conception). In expressing a desire to represent reproduction more comprehensively, Sarah and other respondents at Latimer articulated their accountability to their pupils. As Iris noted above, RSE policy at Latimer is informed by pupil experience and aimed to meet pupils where they were at. The power of these RSE policy makers to make appropriate choices for their pupils -through their own professional competency and their knowledge of their pupils -was constrained by government policy.
While government policy delegated responsibility for RSE policy to individual schools, thus facilitating their 'ableness', using Morriss' term (2002), to be the authority in this area, at the same time the official government guidance was not informed by current evidence in the field about good-quality RSE and not aligned with guidance about implementing Equalities legislation (e.g. teaching about the range of possibilities for conceiving children, with particular relevance for diverse forms of families). This created a tension, posing limitations on the ableness of policy makers, particularly those with a high level of epistemic ability in RSE, to exercise their power to create appropriate, high-quality RSE policies. In addition, wider education policy priorities and culture worked against the values of high-quality RSE, leading those with low epistemic abilities (i.e. Patrick and Kirsten at Wingfield Academy) to apply inappropriate benchmarks and objectives to the subject.

Policy makers' ableness is restricted by the provision, in government policy, that parents can withdraw their child(ren) from RSE
At the time of the study, parents/carers were permitted to withdraw their child from RSE. 4 In making this provision, the UK government sought to satiate groups, including some religious bodies and conservative groups, who argued that education about relationships and sexuality was a private domain, to be governed by parents alone.
When I first met with respondents at Latimer Primary, they were about the embark on revising their RSE policy. They spoke to me about their current approach to RSE. Sarah, a member of the senior leadership team, explained: It is a belief of mine that it is our job to [be responsive to the community]. And I know [other schools] don't believe that it is. So, how far you change, how far you will accommodate different parents' beliefs, in order to provide some form of relationship education, and how far you say, 'no, that's what we're doing' . . . we will try to accommodate parents' own ideas of what their child should or shouldn't hear . . . [but] they don't tell us what we have to teach . . . Sarah and other policy makers at Latimer were motivated 'to provide some form of relationship education' to all of their pupils, and therefore they spent what they described as 'hours' working with parents/carers to secure their cooperation and/or to adapt their RSE programme where necessary. The data supported Alldred, Fox and Kulpa's (2016) finding that when primary school policy makers engaged parents in their RSE, parents became more engaged and supported their children's learning, and the school benefited from greater parental support for RSE. Parent engagement has been recommended by several reviews of RSE (e.g. House of Commons Education Committee 2015; Ofsted 2013). However, respondents at Latimer noted that this work was laborious and not always successful. Some families, they shared, were explicitly opposed to same-sex relationships and transgender identities. Sarah noted that the school would not avoid talking about same-sex relationships, but they also would not emphasize them. Due to parent objections, the school had to compromise on their efforts to uphold equalities legislation, in particular the protected characteristics of gender and sexual identity (Equalities Act 2010).
These examples suggest that Latimer Primary's policy makers' power to effect outcomes in RSE was intrinsically linked to them engaging in specific behaviours to secure their authority: specifically, demonstrating their expertise among parents and adapting RSE to be culturally acceptable. Their ableness to exercise power -that is, to enforce the decisions that they had made about their RSE policy -was tampered by the policy provision of the parental right to withdraw their child, and therefore they had to work to secure this ableness.
When I spoke to respondents at Latimer Primary again after they had revised their RSE policy, their enthusiasm to invest time and energy into tailoring RSE for every family with concerns had dampened. The school had adopted Andromeda, the comprehensive commercial curriculum package for PSHE (including RSE) that had been endorsed and subsidized by the local authority. Sarah and Wendy noted that the RSE components of Andromeda were less progressive than previous resources (now dropped because of their outdated aesthetic), and they suggested that relying on one curriculum made educators less likely to assess the cohort and make RSE lessons responsive to what pupils needed and wanted. However, respondents defended their reasons for adopting it. Sarah explained: Part of the feeling [among staff] was that there was safety in having something backed by City Council and Healthy Schools . . . for me that was fine, if that made people feel comfortable . . . to have a big battle in those PSHE resources was just not on the agenda . . . She noted that new Standard Assessment Tests (SATs), and other academic expectations established by the government were high priorities, and therefore there was less staff time available to focus on RSE. She continued, Most our [city] schools use [Andromeda], so it doesn't matter if you don't like it and go to another school, we're all using it. So there was a bit of safety in that . . . there were some resources to support reducing homophobia a few years ago, used in one of our local schools, and there was a big backlash from the community and it hit the press . . . members of staff were victimised and it was horrible, really horrible. And I think schools have been kind of anxious about resources and anxious as to the . . . potential backlash from parents. By suggesting 'it doesn't matter if you don't like it and go to another school', Sarah refers to a common approach -mentioned by a number of respondents at Latimer -by parents who are not happy with RSE: to threaten to (or actually) move their child(ren) to another school. This action is a clear and bold statement from a parent/carer that they do not recognize the school as a leader or authority on RSE. However, when all schools in an area are using the same curriculum, a parent/carer's 'exit' threat has significantly less traction.
Sarah also highlights a genuine concern about safety among staff, underpinned by recent memories of community backlash against progressive RSE resources. Although respondents at Latimer expressed motivation to promote equalitiesincluding equalities around gender and sexual identity -the data also suggests that the school and staff feel tentative and anxious in this area. Added to the current pressures on the school (e.g. new SATs), the policy makers for RSE decided that the effort required to secure the trust and support of school families to deliver more progressive RSE was too great.
The experience of Latimer Primary suggests that in practice, the government policy provision that allowed parents to withdraw their child from RSE in effect enlisted parents/carers as third-party adjudicators who were empowered to deny schools' credibility in RSE. As a result, from a position of deep commitment to good quality RSE, and a willingness to engage with their community in diverse ways to meet the needs of their pupils, Latimer's approach to RSE policy shifted to a more collegial one, driven by concerns about staff safety combined with the realities of limited staff resources and pressing priorities to which they would be held accountable by the central government. As suggested by Morriss, epistemic abilities are not sufficient to take actions to affect desired outcomes; power must be understood in relation to the circumstances in which decision makers find themselves (2002).

Conclusions
In these concluding paragraphs, I will revisit the original aim of this paper and I will also consider the value of Morriss' concept of power (2002) in analysing RSE policy-making power. I then offer some recommendations for government policy.
In employing Morriss' theory (2002), I have centred the productive power of school policy makers in enactments of school RSE policies, from how policy makers articulated their understanding, aspirations and expectations for their RSE policies to the circumstances that supported or hindered them in effecting their vision for RSE. One of the most significant contributions that this paper makes to literature in this field is a compelling argument, supported by detailed accounts from school-based policy makers for RSE, that the UK government's policy for RSE has created a paradox in power by first delegating RSE policy making to schools and then restricting their capacities to exercise this power by both failing to appropriately equip and qualify them, and by offering a legitimate mechanism for parents to deny schools' authority and credibility as leaders in RSE.

Epistemic abilities and ableness as central to RSE policy-making power
Despite all three primary schools developing their RSE policy with reference to the government guidance of the time, and even with some shared context given their location in the same English city, the RSE policies they enacted varied from each other in quality and scope. This analysis effectively suggests how the nature of school RSE policies may be attributed in part to the capacities of the policy makers themselves, in particular their epistemic abilities -or special knowledge and experience -of RSE. The data also signals that the ableness of policy makers -what they can do given the circumstances they find themselves in -must be understood with reference to place, people and time. Legacies of earlier RSE policies, including community backlash against RSE, endure and put a chill on bravery in RSE. Parental concern and objections also limit the potential of policy makers' power. While Morriss' tidy dimensions of power (2002) become interwoven and complicated by relationships, conditionalities and context in the policy makers' stories, this theory still succeeds in bringing a focus to the qualities of those in position to effect change in RSE policy and its relationship to good-quality RSE.

A delegated government policy for RSE that suits neither RSE experts nor RSE novices, resulting in mediocre RSE for everyone
To highlight how the government RSE policy has created a paradox in power, it may be helpful to synthesize and reiterate here the experiences of the three schools in relation to government RSE policy. While the government uniformly authorised schools to decide RSE policy, this paper illustrates disparities among schools with regard to epistemic abilities. All three schools appealed to government sources as the greatest authority in RSE policy, but whereas respondents at Latimer took a critical view of the content of government policy and could often draw on additional resources to compensate for its failings, respondents at Wingfield Academy and Fleming Primary did not have the epistemic abilities or resources to achieve an appropriate RSE policy.
Respondents at Latimer Primary, while complying with official regulations, shared an understanding that government guidance served to constrain their policy preferences, in effect limiting their ability to exercise their epistemic expertise. For instance, in spite of literature documenting the value of content and pedagogy around gender and sexual diversity in primary-level RSE (e.g. DePalma and Atkinson (2013); Cullen and Sandy 2009), the data from Latimer suggest that gender and sexual identities are still one of the most contentious issues in RSE and government policy failed to support them in upholding equalities legislation around these protected characteristics.
At the time of this study, the government also failed to compensate for varied epistemic abilities across schools by offering training and career development opportunities, or by responding to deficiencies with other resources or tools to improve RSE. Instead, as if in recognition of these deficiencies, government policy offered an instrument for parents to delegitimize schools' authority in RSE: the right to withdraw their child from RSE. The extensive labour and time that Latimer Primary invested in engaging with parents about RSE -a process that was necessary to continually defend their RSE expertise and re-assert their authority in RSE -was one of the greatest challenges that respondents at the school experienced, and ultimately informed their decision to step back from a more collaborative, rights-based and child-focused RSE to what they acknowledged was a more mediocre model: similar to the RSE offered by many other primary schools across the city. While the government has invested resources to upskill schools in RSE, as a result of its new statutory status (Children and Social Work Act 2017), this funding will not be sustained at this level beyond 2022 and the right to withdraw children from sex education remains.
In contrast to the experience at Latimer Primary, the government's leadership and authority was rarely problematised by respondents at Wingfield and Fleming, and parents at these schools, in turn, rarely challenged the authority or epistemic abilities of RSE policy makers at these schools. This was not coincidental: the cultural and socioeconomic diversity of the school community at Latimer suggests divergent values frameworks, and Latimer sought to provide a more progressive RSE than the government guidelines. However, Latimer also experienced the policy of the parental right to withdraw as a problem because of their aims for RSE (e.g. offering 'some form of relationship education' to every single pupil). Nonetheless, I am compelled to recognize a degree of irony in observing that at Latimer Primary School, where educators had the greatest epistemic ability and consequently the most to offer the pupils in RSE, parents were more likely to challenge the school's epistemic leadership and the quality of their RSE.

Coherence between policy aims and actions are necessary but inadequate, power requires a rich RSE vision
The appropriateness of choices made is, in order to achieve envisioned outcomes, is an important aspect of Morriss' theory of power (2002). My analysis underlines the need to understand power in this context as more than simply making appropriate implementation choices to achieve the agreed policy, but equally about defining an appropriate RSE vision in the first place. In this regard, Morriss' theory of power is limited.
Across the three primary schools, respondents represented a range of visions and expectations for RSE policy, as well as a range of suggested interventions or responses to enact them. Respondents at Latimer Primary expressed commitment to a clear, collective vision of a value-centred RSE that included a focus on human rights and gender equality, critical reflection and dialogue, a responsiveness to children's lives, interests and cultural backgrounds, and aspirations of aligning with expert knowledge about RSE. For those at Wingfield Academy, the policy aim articulated for RSE largely comprised producing effectiveness and documenting performance in RSE -in line with overarching neoliberal discourse in education -with little regard for what 'performance' might signify in RSE. At Fleming Primary, archaic conceptions of childhood integrity complicated and prevented a more appropriate policy vision for RSE. While Morriss' emphasis on appropriate choices to deliver desired ends might suggest that respondents at Wingfield and Fleming Primary expressed power in RSE policy making, I would argue that their aspirations for RSE were inadequate, compared against literature about good-quality RSE (Agenda 2019; UNESCO 2018; the International Sexuality and HIV Curriculum Working Group 2011).

Policy implications and recommendations
The accounts provided by key decision makers for RSE at each of the three schools are testimony to the complex, messy business of school-based RSE policy making. From the starting point of recognizing that each school has successfully revised its RSE policy, we see how policies on paper are continually worked on and adapted through implementation. Policy making as constant activity, movement of people around social problems, assessing and testing the conditions and possibilities before them, is a theme that has been explored in some depth (e.g. Fischer and Gottweis 2012;Ball (1994); Stewart 2009), and it is evident in these accounts. Policy making for RSE is an unending practice and alongside management and/or teaching identities, some educators are also policy makers. Government policy for RSE should acknowledge and respond to this reality.
This analysis offers some valuable insights and lessons that are applicable to a wide range of governments, national and regional. Although England adopted a statutory and arguably improved policy for RSE (Children and Social Work Act 2017) since this research was conducted, these analyses are still relevant. This paper suggests that while good national policy is necessary, much more is needed to raise the quality of RSE.
Notably, good quality expertise and resources for RSE, and consistent government leadership.
This paper adds to the literature demonstrating that schools vary enormously in terms of the RSE expertise across their staff (e.g. Formby et al. (2011)) and therefore they require different forms and levels of support from government. Schools with limited expertise may rely on government policy and guidance to inform them about good-quality RSE, to point them in the direction of goodquality resources and teaching aides, and to support them in accessing good quality, affordable continuing professional development to build RSE capacity across their staff. Where schools are rich in experienced, critical and informed RSE staff, national government policy needs to provide the flexibility for these providers to exercise their judgment to design and deliver RSE that is most appropriate for their pupils.
While epistemic ability -expert knowledge and skill -in RSE is critical, it is insufficient. Government bodies and schools -as well as parents and pupils -are partners in RSE. Governments have work to do to be respectful, supportive partners. My data and analysis suggest that partly because of the high variation of epistemic ability in schools, all schools need a high standard of government backing and leadership for RSE. It is not sufficient to invest resources for a year or two when the policy has changed (as appears to be the current strategy), funding must be sustained through time and governments must continue to champion and prioritise RSE. This builds on earlier scholarship by Alldred and David (2007), which found that a two-year period of support and intervention for RSE in secondary schools was not enough to raise the priority of RSE in schools. While they have delegated RSE policy to schools, governments at national and regional levels must stand behind schools and publicly declare their confidence that schools, adequately supported, are excellent places for children to learn about relationships and sexuality, as complementary to family settings. Governments must rebuild trust that has been lost so that schools know they will be supported, especially when their credibility and authority is challenged. While this final point is itself grounds for further research, governments and schools (or at least many of them) should find more collaborative and reciprocal ways of partnering with families and communities to increase the value placed on RSE and raise aspirations for what RSE can do and effect for and with children.

Notes
1. As noted above, since the new legislation came into effect in September 2020, primary schools are now only obliged to have a policy on relationships education, not relationships and sex education. 2. The proposed methodology for this study received ethical approval by the University of Bristol School for Policy Studies Ethics Committee in April 2015. 3. All school and participant names have been anonymised. 4. Under the current legislation, parents/carers can withdraw their primary school-aged child-(ren) from sex education, but not relationships or health education.

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