1 INTRODUCTION

This paper focuses on the question of how OECD seeks to continue its dominance of education policy making globally, moving the discussion on from consideration of PISA to discuss its wider role in policy teaching and learning, or policy by example, as Simons succinctly puts it (^{SIMONS, 2015}). In making this argument, we foreground the importance of three key features of the contemporary governing of education: knowledge, the centrality of which is recognised in the term “epistemic governance” (^{NORMAND, 2016}); networks- the organisational form that is argued to be a more accurate way of representing contemporary governing than hierarchies or levels (^{BALL; JUNEMANN, 2012}), with particular attention to networks of expertise; and narratives - a key element, we suggest, in giving the appearance of direction and coherence to governing activity, including in education, and with specific reference to governing education in Scotland (^{ARNOTT; OZGA, 2016}; ^{OZGA, 2017}). In elaborating on these three features, we draw primarily on research carried out with a number of colleagues in Europe (including Scotland and England) over a period of years since 2000 (see, for example, ^{ARNOTT (2016}, ²⁰¹⁷); ^{GREK; LINDGREN (2015}); ^{GREK; OZGA (2010}); ^{OZGA; LAWN (2014}); ^{OZGA et al. (2011})).

The research reported in those, and other, publications, seeks to investigate education/schooling/learning as a key policy field in Europe, and a site of policy activity of the European Commission, to locate these investigations within a European Education Policy Space (EEPS), (^{LAWN; GREK, 2012}), to identify key characteristics of this space, especially those knowledge-based artefacts that construct it, for example indicators and benchmarks, practices and instruments of standardisation, and to analyse the growth of data in making and shaping this policy space. Finally, more recently, this research explores and analyses the work of policy actors in national and trans-national contexts, especially in doing “political work” that contributes to the making and sharing of meaning and the construction of shared agendas; in other words investigating the work actors do that: “both discursively and interactively seeks to change or reproduce institutions by mobilising values” (^{SMITH, 2009}, p. 13). We argue that in Scotland education policy has been key to shaping visions of society. As ^{Mazon (2018}, p. 1) argues, historical narratives: “introduce newcomers into a field. They legitimate a field’s intellectual territory to external parties. They also serve as political gatekeepers of a field, justifying change or preventing it.”

The UK is an increasingly heterogeneous entity where tensions between the policy directions of the UK state and the devolved nations of the UK have become more evident in recent years. There are tensions and challenges both between and within policy debates about responses to OECD. ^{Arnott (2017}, p. 9) argues that: As globalized policy influences in education have become increasing evident over recent years in such areas as international competitiveness and modernization, these globalized policy narratives sit alongside narratives which stress the need for education policy to maintain national integrity and quality.

A number of key trends in governing education are revealed through this research. Firstly, we find that changes in contemporary governing and changes in knowledge are interdependent; we suggest that as governing forms become more networked, and less hierarchical, so too does knowledge change; it moves more freely, uses more artefacts, and involves new actors in its production. Secondly, the knowledge that is now valued in governing is “actionable” (^{GRUNDMAN; STEHR, 2012}): that is, it identifies problems and solutions, often through analysis of data. Thirdly, new networks of experts and consultants in education (^{SHIROMA, 2014}), are responsible for translating this “coded” knowledge into governing practices and processes-the data do not “speak for themselves”. These trends have developed, and continue to develop, against a backdrop of “speeded up” or “fast” policy (^{PECK; THEODORE, 2015}). The speed of change, combined with a new landscape that is also shifting, creates the need for what we call “Governing narratives” that attempt, in some contexts, to build coherence in often chaotic, complex circumstances (^{ARNOTT; OZGA, 2016}; ^{ARNOTT, 2016}; ^{OZGA, 2017}; ^{OZGA et al., 2011}).

Our attempt to move beyond PISA does not imply that PISA-and performance data more generally-are unimportant. Indeed a clear characteristic of new governing arrangements is the involvement of new actors-especially transnational actors such as OECD, the EC and the World Bank and corporate actors like Pearson and McKinsey-in data production and use. In the governing of populations, data systems create governing assemblages that seek to shape individual conduct while apparently enabling autonomous, choice making activity supported by information. Local government and schools that used to be relatively closed to public and central government scrutiny are now rendered visible and calculable (^{OZGA et al., 2011}). Data expressed as public rankings, league tables and PISA results are both “official and popular” knowledge forms, and so, as ^{Piattoeva (2014}) argues, we can see them as doing political work-for example enabling and consolidating control over a wide network of actors and institutions-national governments, local authorities, schools and teachers included. However, as indicated earlier, data use in policy requires expertise which is often based in the capacity to “translate”, mediate or interpret policy-relevant or “actionable” knowledge from multiple sources of information and data (^{GRUNDMANN; STEHR, 2012}; ^{NEWMAN; CLARKE, 2009}).

As John Clarke puts it:

[…] new governance arrangements create the conditions in which new knowledges, skills and roles may flourish - ones that emphasise cross-boundary working. Transacting, translating, mediating and brokering characterise these new ways of working that are central to the forms of governance as partnerships, networks and collaborations (^{CLARKE, 2012}, p. 130).

Furthermore, as Sotiria Grek points out, although PISA naming and shaming is an indispensable part of OECD’s success, PISA “shock” has a temporal dimension: “spectacles come and quickly go” (^{GREK, 2017}, p. 298). These spectacles also tend to have diminishing impact, repetition dulls the effect. In looking further for explanations of OECD’s continuing influence, she develops an argument, supported by original interview data, that OECD sustains and builds its policy work through “the continuous crafting of its relationship with key education actors in other IOs and within national contexts” (^{GREK, 2017}, p. 298). The interaction of local, national and international actors in “iterative processes of collective learning” - for example in OECD Country Reviews-is key, she suggests, to OECD’s achievement of a paradigm shift in the thinking and framing of education. Our focus in this paper is on the knowledge content that supports these interactions, on the networks through which that content is produced, circulated and disseminated, and on the narratives that seek to provide coherence to the account.

In order to explore these issues, we look at the content and impact of the OECD’s Governing Complex Education Systems (GCES) project (2011-2016) as it entered the Scottish education policy space, examining the knowledge, networks and narratives that can be found in this report. That OECD project strongly informed the advice given to the Scottish government, following OECD’s Review of Scottish Education (^{OECD 2015}) and it also seems to us to encapsulate OECD’s search for a new formula to sustain its policy teaching alongside the performance-data driven approach of PISA, that has relevance beyond Scotland. Nationally embedded social and cultural constructions of policy remain significant to the policy field of education, however, in recent years, especially following the 2016 PISA results there have been increasing tensions in Scotland within the constructions of narratives informing education policy, and we elaborate on this point below.

2 OECD’S GOVERNING COMPLEX EDUCATION SYSTEMS (GCES) PROJECT (2011-2016)

As indicated above, our earlier research had found an increasingly strong relationship between knowledge and governing (^{GREK; OZGA, 2010}). Interestingly, GCES was preoccupied with exactly that relationship, especially the interdependence of knowledge and governing, one of its key questions being “How do governance and knowledge mutually constitute and impact on each other in complex education systems?” (^{FAZEKAS; BURNS, 2012}, emphasis added).

The answers that GCES developed from its six country case studies and extensive literature reviews, published in reports and working papers as “lessons learned” do not, in fact, constitute straightforward “actionable knowledge”, and do not translate easily into policy. Indeed, OECD seems to accept their limitations by describing these findings as “insights”. They include the conclusion that there is “no one right system of governance”, and that governments spend too much time focusing on structural reform, when it would be more productive to focus on processes. GCES also identifies tensions within the contemporary governing of education, though it does not describe them in this way, rather it talks about the need for systems to find a “balancing act” between accountability and trust, innovation and the avoidance of risk, and between consensus building and making difficult choices.

Effective governance, GCES claims, works through “building capacity, open dialogue, and stakeholder involvement”. However, while acknowledging the presence of an increased number and variety of actors in governing, GCES puts considerable emphasis on the continued importance of national or state levels in creating the strategic vision and the set of processes “to harness their ideas and input”, even in decentralised systems. Complexity features largely in the discussion: it is understood as describing both the context and the nature of policy making in education, and conceptualisations of systems that focus on top down or bottom up reform are dismissed as inadequate “to effectively address the rapidly evolving and sprawling ecosystems that are modern educational systems”. (^{SNYDER, 2013}, p. 6, quoted in GCES).

We find further echoes of our own research findings in the GCES argument that system complexity is related to data and the related spread of network forms. The horizontal logic that GCES sees as following from networks, and which is argued to be “indispensable in a world of knowledge, connection and complexity” is also understood as difficult to integrate with vertical systems of authority. Multi-level governance approaches, they suggest, may overcome this difficulty, but GCES is concerned about the potential blurring of lines of responsibility: “the central level needs important steering capacity if national or international standards are to be monitored and met” (GCES).The perceived failure of both top down and bottom up strategies leads GCES to focus on the “middle”. This middle proves extremely hard to define: according to GCES it can be the middle of the formal system i.e. districts, local authorities and so on, or a “meso” level that consists of “combinations of networks, chains, professional communities, initiatives, and groupings that are often invisible in the official charts of an education system” (GCES). It is imagined as an active, engaged and engaging resource developing collaboration, ideas, and expertise and “exercising collective responsibility for their students “success”.

The strength of the middle is to be developed through the use of data - GCES recognises the problems caused by increased data production, and sees their resolution as based in ensuring that local officers, leaders, educators and others “become highly competent to interpret and use such information.”

In relation to the three key themes of contemporary governing that we identified from our research-knowledge, networks and narratives-GCES foregrounds knowledge as central to governance, and extends the definition of knowledge from that based on performance data to professional and tacit knowledge, knowledge possessed by different actors, as well as research-based knowledge. In summary, the argument is that knowledge-based governance requires an evidence-based approach, but that the understanding of evidence needs to be widened and deepened. Knowledge systems, GCES states, combine descriptive system data (on achievement, graduation, and so on) with research findings on whether something is working, they include tacit knowledge transmitted informally (^{FAZEKAS; BURNS, 2012}). Knowledge is understood as a key medium of governance, and the effectiveness of knowledge-intensive governance “depends critically on the ability to learn” (GCES).

Networks are also central to the GCES message: they are seen as ending the stalemate of system reform caught in debates between top down and bottom up initiatives, and the strengthened “middle” is envisaged as a fluid, mobile, connected and collaborative networked form that supports the emphasis on process rather than structure that is threaded through the various GCES documents. Networks, interpreted in this way, are the ideal form for both the production and spread of knowledge, and, at the same time, they promote the learning that is essential in negotiations and dialogue that is “integral to collaborative progress in complex systems, [and] to enable those at the micro and meso levels to participate in nationwide strategy, using high quality knowledge to improve the quality of decision-making and practices.” (GCES).

Narratives appear as the responsibility of the state or national government to establish and disseminate strategic thinking: this focus on strategy is currently being further developed in OECD’s Strategic Education Governance project which provides “a forum for collective learning on how to govern education systems more strategically.” We say more about narratives in the section on Scotland below.

3 RE-READING GCES

Before moving to that discussion, however, we want to pause and reflect on the silences in GCES, especially those about politics and the political work that policy actors and artefacts do. Here it may be useful to say something briefly about our overarching methodological approaches and its informing theoretical resources.

We take critical policy sociology approach (^{OZGA, 2000}) to understanding governing that requires us to be attentive to history and cultural formations in different national contexts, while understanding that they are transnationally connected. We stress the importance of the “assumptive worlds” of policy makers (^{MCPHERSON; RAAB, 1988}) and their capacity in doing political work to carry agendas beyond the national context and broker new meanings as they translate policy, evidence and expertise between different sites (^{LENDVAI; STUBBS, 2007}) We are also attentive the capacity of policy actors to accommodate and mobilise national cultural capital within and beyond the national, and to strategies that protect or enhance their position in relations of ruling. This leads to greater emphasis on policy actors as mediators, brokers, translators and transactors.

We understand policy as discourse, that is, constructed and presented-in texts, speeches and other public forms-discursively. Policy texts carry definitions of problems, reference particular forms of evidence, and produce “knowledge” of particular kinds to guide the implementation of policy solutions. Critical Discourse Analysis enables the study of key policy texts, interviews and speeches (^{FAIRCLOUGH, 1995}; ²⁰⁰¹) with a focus on their interdiscursive features. That is, we study texts as persuasive, but also as referencing particular contexts and connections, to make visible relations between text, discursive practices and wider policy, especially where these connect to power relations between policy actors located in different policy spaces, where struggles take place over their meaning, interpretation and implementation.

Turning now to what Simons et al characterise as “re-reading” of policies-i.e. ways of approaching policy in education that have a critical orientation to issues bound up with policy, politics and power (^{SIMONS; OLSSEN; PETERS, 2009}) we now revisit the key concepts of knowledge, networks and narratives looking specifically at what is absent from the OECD GCES discussion. While we share OECD’s concern about the enormous growth of information and the speed with which it can be transferred across and within systems, we follow ^{Grundmann and Stehr (2012}) in suggesting that the difficulty of dealing with the massive expansion of information explains the growth of networks of experts who promote “cognitive consensus” (^{GRUNDMANN; STEHR, 2012}) as a way of reducing possibilities and making knowledge actionable: “The rapid growth of experts, advisers and consultants in education arises from the rapid expansion of knowledge/information, this provides opportunities for simplification of the problem of endless competing interpretation in order to provide a basis for action”. (^{GRUNDMANN; STEHR, 2012}, p. 20-21).

Cognitive consensus is a shorthand term for standardised policy agendas and their repertoire of benchmarks, indicators and competitive testing regimes (^{CZARNIAWSKA; SEVÓN, 2005}; ^{STEINER-KHAMSI, 2004}). The GCES report is an attempt to build such consensus around principles for the knowledge-based governing of education, but, as we have seen, it lacks the coherence and capacity for transfer of data-based agendas. We return to this point in the discussion of Scotland later in the paper. The point we wish to stress here is the growth in power of networks of experts, the kinds of knowledge that they promote, and their lack of accountability.

Networks seem to emerge in GCES as a natural form for the coordination and management of activity across different sectors, but a more critical reading sees their emergence not as natural but as political: key actors in networks-including experts-do governing work that mobilises or articulates political blocs; builds alliances, negotiates and reconciles interests, and assembles projects that define the direction and purpose of governing (^{BOSWELL, 2009}; ^{CLARKE, 2009}). The work that policy actors do in networks, whether as “chosen” experts, mediators and translators for OECD, or in private consultancy, strengthens the trend towards comparison and the search for trends and patterns in comparative date, and also increases the influence of analysts and gives considerable power to those who can interpret data and turn it into policy advice (^{GRUNDMANN; STEHR, 2012}, p. 20-21). Indeed the issue of interpretation becomes paramount, and experts are thus: “more than the diffusers of ideas; they develop conceptual knowledge in order to promote educational reforms, drawing on their substantial experience as policy advisers to governments and IOs”. Moreover, “their attributes as experts and consultants tend to obscure the ideological and political dimension of their activities of knowledge production for policy” (^{SHIROMA, 2014}, p. 2).

The political nature of that interpretation is often concealed, as Shiroma points out, because the label “expert” confers scientific status and authority. The growth of this form of expertise is recognised as a transnational phenomenon, with experts increasingly working between national and transnational arenas, and identified as a “new governing elite” (^{LAWN; GREK, 2012}, p. 75; ^{STONE, 2013}, p. 41) also described as a “magistracy of influence” (^{LAWN; LINGARD, 2002}, p. 292) and a new “European technocracy” (^{NORMAND, 2016}, p. 129). The authority of science is invoked to sustain their position, but, as Dale points out, the idea of science that is invoked here fails to acknowledge that “scientific authority” does not in itself ensure acceptance of models, without reference to “the set of political conditions” under which they are advanced (^{DALE, 2000}, p. 445). Nor is it attentive to the related recognition that scientific knowledge is produced, accepted and contested in specific contexts (^{DEMSZKY; NASSEHI, 2014}). As well as the emphasis on comparison, the use of experts promotes the idea of distance as a necessary precondition for the production of reliable knowledge. International organisations and the experts who work for them are not only informed about comparative performance, but can see the national as an entity from a distance more clearly than do those working within it:

[…] because when you sit up to your neck in the Scottish system, everything is Scottish. Everything is Scottish. [You feel] This is our system, we defend it as a fortress and all these influences from outside, they should be kept away. By sitting here and making comparative analysis, we identify what is specifically Scottish to the Scottish system. What is it that you should actually defend to keep these roots in national culture and national institutions […] We know it, we have the information, we have this distance that is necessary to do it. And we can compare and find out what is it that shines in the Scottish system.” (Senior EU Analyst)³.

At the same time, through these networks, there is a close alignment of the framing of governing problems with political priorities, so that knowledge production (by experts, analysts, consultants) is drawn into supporting the legitimacy and authority of governing knowledge forms (^{FENWICK et al, 2014}). In the next section, we elaborate on the idea of narratives interpreted as a governing resource, and exemplify this argument through discussion of the governing narrative in Scottish education policy, before looking at the impact of the OECD’s 2015 review and follow up discussions in that specific policy space.

4 NATIONAL NARRATIVES AND INTERNATIONAL INTERPRETATIONS

Governing narratives are constructed in a new globalising context, yet nationalism is increasingly present as a key, if volatile resource in governing narratives in Europe and beyond, where ethnic nationalism is a factor in the rise of right wing and populist politics (^{BONIKOWSKI, 2017}). In Scotland, a particular narrative of “modernised” or civic nationalism contributed to the rise of the Scottish National Party (SNP) and its success in forming successive governments in Scotland (^{ARNOTT; OZGA (2010}); ^{ARNOTT (2016}); ^{ARNOTT; OZGA (2016}); ^{ARNOTT; KELLY (2018})). It is important here to stress that the SNP narrative is one that seeks to both construct and draw from civic, rather than ethnic, nationalism, in pursuit of its overarching political project of independence for Scotland. Civic nationalism is usefully defined by ^{Ignatieff (1993}) as supporting “a community of equal, rights-bearing citizens, united in attachment to a shared set of practices and values”.

Since the late twentieth century, diverging policy narratives about welfare provision and social rights have reflected differing models of “education provision” in Scotland and England. The Scottish myth has been employed by the education policy community within “narratives” to support a “social democratic consensus”. In England, there has been a “historical narrative” of education as a private good (^{EXLEY; BALL, 2014}). Education and social justice are key policy arenas in which considerable tensions- between and within institutions, within and across Scotland and the UK have been evident in policy debates (^{ARNOTT; OZGA, 2016}; ²⁰¹⁸). Divergence in the delivery of welfare policies to address different needs as in the UK is not unfamiliar in decentralised governing arrangements. When and how government intervenes in society has featured prominently in prevailing policy narratives about addressing educational inequalities.

Devolved governing arrangements for Scotland within the UK and the associated legislation in 1999 and 2016 assumed the UK’s EU membership. For education policy in Scotland the European interconnections offered by UK’s membership of EU offered a different orientation towards social and education policies from that of the UK Government which favoured the US rather than European policies as a source of “policy learning” (^{ARNOTT; OZGA, 2016}; ²⁰¹⁸). Increasingly following the 2008 recession, UK Governments have placed European discourse within a global rather than European perspective and, as ^{Peck and Theodore (2015}) has argued, this has been placed in a neo-liberal form that stresses the primacy of the market in public policy.

In Scotland in recent years the pace of educational reforms has quickened markedly in relation to transition policies including policy on early school leavers. These education reforms must be located within wider political debates about how governments should handle not only educational challenges in relation to their international competitiveness but also with reference to poverty, social inequalities and sustainable economic growth. We have argued in our earlier research that national narratives in education policy had been framed in Scotland with Europe as a reference point, whereas in England policies have favoured comparisons with global economic policies (^{ARNOTT; OZGA, 2010}).

Neo-liberal tendencies and market - based assumptions underpin narratives and discursive strategies from the UK Government. The rhetoric of the Scottish Government differs from that of the UK Government in that it highlights societal challenges concerning inequalities and disadvantages for children and young people and sets out policy goals to address such challenges (^{SCOTTISH GOVERNMENT, 2016a}; ^2016b).

Over the past couple of years, alongside curricular and assessment reforms Scotland has reformed school funding and school governance. Re-engagement of early school leavers with learning and also training for employment has been a concern of educational reform in Scotland. The transition of young people at secondary education to the next stages of further or higher education alongside vocational opportunities available has been a common theme of educational reform in Scotland and also England. These reforms were spurned on following the publication of OECD’s PISA international ratings in November 2016. Internationally enhancing the educational engagement of early school leavers and social deprivation has been a consistent theme of OECD.

We have argued that different approaches by policy makers in England and Scotland tie with wider welfare and social policy issues and visions of “welfare regimes” (^{ARNOTT; OZGA, 2011}; ²⁰¹⁶). There are significant differences that in part stem from different “welfare traditions” that define education differently: ie as a “public” or “private” good. The SNP government”s work in “crafting the narrative” may be seen in their referencing “inwards” in policy discourse to shared myths and traditions that stress the public nature of schooling and of further and higher education in Scotland, and the role of all these publicly funded forms of education provision in the construction of a shared community; and in referencing “outwards”, especially to selected comparators, to education”s role in progress, social solidarity and collective development, as well as in economic growth.

These two forms of referencing of nationalism are combined in an overarching narrative of collective learning in which a “learning government” is enabled to lead a “learning nation” towards greater autonomy and self-reliance, and ultimately independence, within a Northern European frame of reference (^{ARNOTT; OZGA, 2016}). Indeed SNP political actors interviewed for our research saw education as contributing to the building of a distinctive polity: referencing the past, but also building commitment to collective and reciprocal learning. Importantly, this is discursively constructed as a “learning” project, in which politicians, professionals and publics may be constructively engaged.

This narrative draws on historical reference points: in the absence of a Scottish state, the key Scottish institutions (the law, the church and education) maintained a Scottish identity: education provision in Scotland contributed to the assertion of continued distinctiveness from England (^{MCCRONE, 1992}). Key elements of this narrative include relative uniformity in school provision (through a comprehensive school system), greater social mobility, meritocracy, especially the recognition of talent and its fostering regardless of social class, a broad curriculum uniting sciences, arts and humanities and public support for teachers and for education more general. These elements together came to form what has been termed the “Democratic Intellect” (^{DAVIE, 1961}). These elements can certainly be interrogated empirically: our point here is that they contribute to a mythology of Scottish education that offers resources in the construction of the contemporary governing narrative:

So it matters deeply to me personally that every young girl and boy growing up today - regardless of where they were born or what their family circumstances are - gets the same chances that I did. And of course it also matters to us as a nation. Scotland pioneered the idea of universal access to school education in the 17th and 18th centuries. (^{STURGEON, 2015}).

The SNP government tries to manage the considerable tensions between globalising imperatives of modernisation and economic competitiveness, alongside transnational prescriptions for educational excellence, while maintaining and sustaining ideas of national integrity and quality, and devotes considerable energy to seeking to combine these competing pressures. This somewhat contradictory narrative, described by critics as “speaking social democratic and acting neo-liberal” is threaded through the entire range of SNP government policy in education, which includes major curriculum reform in the programme “Curriculum for Excellence” (CfE), remodelling the relations between local authorities, schools and the inspectorate (^{GREK et al., 2013}), restructuring provision from early years through to the College sector, as well as seeking to draw the universities and colleges into their project, through reform of university governance. CfE marks a quite radical shift away from the traditional academic character of Scottish schooling (^{PATERSON, 2009}) that is being promoted through references to the need to extend opportunity and fairness, as well as ensure effective economic growth through education. The discursive shift is accomplished by foregrounding the social justice issue (i.e. we are well-schooled but the poorest pupils do very badly) and using that shared idea of Scottish education (as socially just and fair) to displace its meritocratic character and thus enable the dilution of academic curricula and the development of CfE. This discursive framing is not convincing to some educationalists (see, for example, ^{PRIESTLEY; HUMES, 2010}). An important element of the narrative is that which stresses the consensual nature of policy making in education in Scotland, an element that draws on the idea of collective learning and growing autonomy, and that also underlines (at least discursively) the importance of distributed leadership rather than central direction.

Yet the publication of the OECD review of Scottish Education-Improving Schools in Scotland (^{OECD 2015}) seems to mark a decisive break with that discourse of consensus-building. This policy review was commissioned by the Scottish Government “in order to inform the ongoing development of education policy, practice and leadership in Scotland, by providing an independent review of the direction of the Curriculum for Excellence (CfE) and emerging impacts seen in quality and equity in Scottish schooling” (^{OECD, 2015}, p. 3). The content of the review is too detailed for extended discussion here, but it is important to note that it contained extensive recommendations for change. Curriculum for Excellence was acknowledged as an ambitious reform, but its complexity and the plethora of initiatives associated with it, along with complex procedures and excessive workload, led the reviewers to identify a number of challenges facing Scottish education and to make twelve recommendations for action to improve Scotland”s education system, across areas such as leadership in schools, issues presented by existing data sources and, in particular, the complexities around Curriculum for Excellence. Montserrat Gomendio, Deputy Director, OECD Directorate for Education and Skills, said in introducing the report: “We applaud Scotland for having the foresight and patience to put such an ambitious reform as Curriculum for Excellence in place; we hope that our OECD review will help ensure that it will live up to its full potential and realise excellence and equity right across Scotland”.

The National Improvement Framework was developed by the government in 2016, and, in each iteration, has shown a narrowing of focus on attainment (^{ARNOTT, 2016}; ref from BELMAS). The 2015 OECD report seems to have also led directly to the Scottish Government”s revealing titled “Excellence and Equity in Scottish Education: a Delivery Plan” and the attempted introduction of compulsory standardised assessments, in reading, writing and numeracy, in P1, P4, P7 and S3 in 2016. Indeed tensions in the governing narrative of the SNP government seem to be increasing apparent in policy regarding standardised national assessments in primary schools. The introduction of national assessments in 2017 for primary 1 school children to assist in reducing attainment gaps was met by increasing criticism from professional associations, parent groups and also opposition parties in the Scottish Parliament. In October 2018, facing growing criticism the Scottish Government agreed to undertake an independent review of P1 assessments. The Cabinet secretary for Education in the Scottish Government, John Swinney argued:

The key measure the Scottish Government will use for assessing the standards of education is teacher professional judgment - not the outcomes of the standardised assessments. The standardised assessments are just one part of the range of evidence that a teacher will call on when assessing whether a child or young person has achieved the appropriate CfE Level. (^{SWINNEY, 2018}).

In May 2019, the Scottish Government was defeated by opposition parties in a vote to continue with primary 1 assessment. In 2016, the results of the PISA survey of 2015, in which Scotland participated, were released, revealing that Scotland”s schools recorded their worst ever performance, and that scores for maths, reading and science all declined. This was the first time since the tests began in 2000 that all three subject areas were classed as average, with none above average, and this was taken as evidence of the impact of establishing Curriculum for Excellence and changes to the first three years of secondary school.

The PISA shock-in a context where the governing discourse draws on ideas of established quality in education-was considerable. Perhaps a particularly wounding comparison was that which found that English pupils were “significantly above” Scots in science. The timing was also important, in that change in the political context had led the Scottish government to put more emphasis on education policy. After a decade in power, the intractability of the attainment gap between less affluent pupils and their more privileged peers sat uncomfortably with the SNP government”s claim to combine excellence with equity.

International comparisons by OECD in the PISA results in November 2016 from the UK, England and Scotland showed that the UK had not improved its international position over the three years since their previous PISA performance. The UK was placed 27th for Maths with international comparators. This was the worst UK performance in PISA results since 2000. In England the Schools Minister claimed the results for England would reaffirm the government commitment”s to expand the selective grammar school sector by increasing the number of grammar schools places (^{COUGHLAN, 2019}). For Scotland, the PISA ratings were its worst performance since participating in the PISA survey. The Depute First Minister and Cabinet Secretary for Education, John Swinney said “results underline the case for radical reform of Scotland’s education system” (^{COUGHLAN, 2019}). Swinney argued that there would need to be radical educational reforms to improve Scottish schools (^{SEITH, 2019}). The challenges of closing the attainment gap in Scotland alongside the 2016 PISA rating for Scotland resulted in a flurry of policy interventions for educational reform. However, in the summer of 2018, the Scottish Government pulled back its legislative proposals for significant reforms to the structure of schooling including enhanced delegation of powers to head teachers, and introduced some of the proposals by non-legislative means. The Scottish Government expert panel, the International Council of Education Advisers (ICEA), also recommended to the Scottish Government that non legislative routes were a preferred basis of policy interventions.

Political pressure was increased by the identification of Scotland’s First Minister with the intention of closing this gap: in this case using inward referencing to illuminate the potential gap between the national narrative and the reality:

My aim - to put it bluntly - is to close that attainment gap completely. It will not be done overnight - I accept that. But it must be done. After all, its existence is more than just an economic and social challenge for us all. It is a moral challenge. Indeed, I would argue that it goes to the very heart of who we are and how we see ourselves as a country. (^{STURGEON, 2015}).

The SNP governing narrative was further damaged by the performance of the party in the UK General Election in 2017 where it won 35 of the 59 Scottish constituencies - a fall of 21 seats from the 56 they won in 2015 (^{ARNOTT; KELLY, 2018}). The Conservatives secured 13 seats in Scotland - the party’s best performance in the country since 1983.

These factors help to explain the apparent permeability of Scottish education policy to recommendations from international organisations-the OECD-and from international experts-not only in the 2015 policy review, but also in a review of evaluation and assessment policy (^{OECD, 2013}) and through the GCES project (2011-2016), some of the results of which were reported to the Education and Skills Committee of the Scottish Parliament in November 2017, in order to “test the evidence base for the Scottish Government’s programme of reforms for school education” (^{BURNS; KOSTER, 2016}, p. 1). In addition, the Cabinet Secretary for Education, John Swinney, appointed an International Council of Education Advisers (ICEA) in 2016 to draw upon “world-leading education and business experts with a range of knowledge and extensive experience of advising educators and governments on education leadership, school improvement and reform” (ICEA no date). International experts included academics based in Canada, Malaya, Singapore, the US and Finland, along with three from Scotland. One of the external experts is Professor Andy Hargreaves, who is networked globally, holding visiting professorships in the US, the UK, Hong Kong, Japan and Singapore and Sweden and who describes himself in the report as “a researcher, writer, consultant and adviser [who] has delivered invited addresses and worked in 37 US states, 42 countries and all Australian states and Canadian provinces.”

Taken together, the various recommendations contained in these reports and in the minutes of the ICEA highlight the difficulty for the SNP of maintaining a distinctive, inward-referencing governing narrative in the face of underperformance as assessed by PISA. At the same time, it is interesting to see how fragile the cognitive consensus is around policy scripts as expressed in these various reports. A brief consideration of the recommendations of the 2015 report indicates that fragility:

a) Be rigorous about the [attainment] gaps to be closed and pursue relentlessly “closing the gap” and “raising the bar” simultaneously.

b) Ensure a consolidated and evidence-informed strategic approach to equity policies.

c)Develop metrics that do justice to the full range of CfE capacities informing a bold understanding of quality and equity.

d) Create a new narrative for the Curriculum for Excellence.

e) Develop an integrating framework for assessment and evaluation that encompasses all system levels.

f)Strike a more even balance between the formative focus of assessment and developing a robust evidence base on learning outcomes and progression. (^{OECD, 2015}, p. 12)

There is a considerable tension between the need for data-based rigorous evidence of attainment and the development of metrics that go beyond attainment levels; there are problems with constructing a new narrative for CfE that is not dominated by what can be measured, and the pursuit of an even balance between formative and summative assessment has eluded generations of policy makers and educationalists. Indeed a striking characteristic of these various reports is the gap between the confidence with which they are offered and any anchoring in practical planning.

Other recommendations-for example the creation of a “strengthened middle” which has been adopted in Scotland with the creation of the Regional Improvement Collaboratives (RICs), Local Authorities organised in collaborative groups-reflect tensions between centrally-directed reform and the idea of collaborative networks of shared expertise as outlined in the GCES report.

Tellingly, the key recommendation of the 2015 report for the creation of “a new narrative for CfE and to make it highly visible in Scotland”, while stressing that it should be evidence-based in order to shape the evidence agenda related to CfE and the new National Improvement Framework, places responsibility for its development squarely with national government, as a political act: We envisage narrative development to be an act of political leadership, then to be picked up and incorporated into the management of the system, and absorbed by the profession, schools, communities, parents, students, and the public at large” (^{OECD, 2015}, p. 19).

Uncertainties surrounding the UK’s relationship with EU, add to the tensions already existing in the differing views of the UK and Scottish governments about the extent to which Scotland can follow distinctive social and economic policies under existing devolved powers including policy making in relation to educational outcomes and social deprivation.

5 CONCLUSION

The paper attempts to illustrate the gap between the OECD’s ambitions for policy teaching through example and the messy and complex situations on the ground in specific national contexts. Whereas the PISA rankings can be interpreted simply and mobilised to support both transnational and national policy agendas, the difficulties of policy by example are exacerbated by the lack of clarity in the agenda being offered to national systems: much of the discussion of governance is mired in the recognition of complexity, while also attempting to extracts transferable rules. This produces severely decontextualized exemplification, and reduces the content of advice on restructuring to rather empty sloganeering, without substantial discussion of how - for example - “strengthening the middle” might be achieved without damaging existing arrangements, or without duplication, or increasing complexity. Nor do such transnational interventions necessarily work with existing governing narratives, indeed they may well contradict them, or, as in the example of Scotland discussed here, fail to locate the tension in governing narratives and its framing by changing politics. Poverty, attainment and participation in education are essential elements of policy debates in Scotland. The framing of education policy in Scotland and England has reflected tensions about whether education is a public or private good. The “inward referencing” to historical narratives in Scotland is significant in understanding the policy context there. While the nations in the devolved UK share narratives about the importance of their positioning in the international global economy, there are tensions regarding “social rights” and “inequalities” concerning delivery of education, that have consequences for policy work, and that may challenge OECD’s attempts to teach through example.