Technology and aesthetics in school excellence policies: the case of Through Growth to Achievement

Global policy development in schooling often looks to solutions that embrace technology and articulate a ‘what works’ approach to practice. This paper takes as a case study the Australian Government’s Through Growth to Achievement report on achieving school excellence. It analyses the report through the lens of ‘sublimes’, which provide insight into the way certain policy rhetoric can become reified. Through Growth to Achievement emphasises technological solutions which align with and reify an aesthetic notion of good teaching that draws on the ‘what works’ literature. This analysis argues that certain assumptions about technology, teacher practice and the ‘evidence’ for good policy are based on these reifications. A sublimes analysis makes these unspoken assumptions explicit and reveals the way they reinforce globalised policy approaches and foreclose others.


Introduction
This paper explores the way policy solutions are discursively constructed, taking as a case the Australian Government's Through Growth to Achievement report. This report, commissioned to find ways to improve school performance in Australia, focusses on technological solutions and aesthetic understandings of ideal teaching practices such as Hattie's Visible Learning (2009), which has come to occupy a privileged place in the 'what works' literature (McKnight & Morgan, 2019;Wrigley & McCusker, 2019). These logics perpetuate the idea that there are fast solutions to the 'problem' of education, many of which are used globally in education policy.
Underpinning these solutions sit a number of reifications which Flyvbjerg (2014) calls sublimes. These sublimes indicate various domains that make a project 'exciting' for stakeholders, offering insight into what makes certain policy prescriptions "attractive to decision-makers" (Gillett & Tennent, 2017, p. 97) and how particular approaches gain acceptance. This paper proposes the sublimes as a lens for analysing the motivations and discursive tactics mobilised within education policy. A sublimes analysis does not assume that a policy is intended to be evidence-based or 'rational' in an objective sense, but rather that policy is always a product of rationalisation, and that the actions a policy prescribes are consequences of these rationalisations.
After introducing the background to the report and a short introduction to the sublimes, this paper discusses three of the key recommendations in Through Growth to Achievement, arguing that the sublimes which dominate are the technological and aesthetic, and that these sublimes reinforce the 'evidence-based' approach to classroom practice. Finally, this paper notes some policy changes since the publication of the report, particularly the formation of the Australian Education Research Organisation (AERO), which supports the aesthetic sublime. This paper argues that 'excellence' is constructed within this aesthetic frame, further embedding the 'what works' logic in Australian schooling.

Excellence in global policy
Policy statements aimed at achieving school excellence are not a new feature of education systems and in some ways reflect the globalised nature of education policy reform (Rizvi & Lingard, 2010). Singapore's Thinking Schools Learning Nation initiative, launched in 1997, has contributed to the high performance of the Singaporean education system, with subsequent policies addressing technological and school system changes (McLay & Reyes Jr, 2019). Malaysia has expanded its Clusters of Excellence policy with a view to strengthening the human capital base and economic competitiveness of that nation through education reform (Noor & Crossley, 2013). In the United States, the Obama administration enacted a range of policies that is built on the Bush-era No Child Left Behind policy. These include the US Department of Education's A Blueprint for Reform (2010) and Transforming American Education: Learning Powered by Technology, National Education Technology Plan (2010). These policies articulated pathways to school improvement through the use of various administrative and accountability technologies, aiming to 'revolutionise' American education (Araya, 2015).
In 2016, the United Kingdom Department for Education released a white paper titled Educational Excellence Everywhere (2016). This white paper emphasised autonomy, accountability and teacher quality as mechanisms for school improvement, notably situating notions of teacher effectiveness in the 'what works' literature (Godfrey, 2017). The kinds of approaches put forward by these various policies represent particular imaginations of the act of teaching: some emphasising a technical, data-driven logic, and others that put forward an aesthetic notion of 'good teaching' grounded in decontextualised ideas of evidence. Through Growth to Achievement is an Australian expression of this global policy approach, and is used here as a case study in the way the what works logic is activated as an aesthetic sublime.

3
Technology and aesthetics in school excellence policies:…

Background to Through Growth to Achievement
In 2017, then Prime Minister Malcolm Turnbull commissioned a review into school excellence, which was delivered to the Australian Government in 2018 as Through Growth to Achievement (Gonski et al., 2018) by the lead author, businessman David Gonski. An earlier report (Gonski et al., 2011) to the previous Labor Government had led to a fractious debate about school funding in Australia. The focus of the first report was on creating equitable funding in the school system, seeking to mitigate the complexity of public funding arrangements in Australia. The central recommendation of the first Gonski Report was the creation of a national Schooling Resource Standard: a base amount of funding for all schools calculated for each student, with additional loadings for disadvantage (Gonski et al., 2011). Kenway (2013) argues that while the first Gonski report demonstrated a clear commitment to equity, it was constrained by then Prime Minister Julia Gillard's statement that "no school would lose a dollar of funding", meant to reassure the independent school sector. The subsequent National Plan for School Improvement, while committed to the Gonski funding model, diminished the focus on equity.
Debate about this funding model intensified in the years prior to Turnbull's leadership. In office, Turnbull sought to neutralise the issue by offering the funding model that the Labor party had proposed before losing the election (Balogh, 2017). As part of that offer, Turnbull announced a second Gonski Review, this time into excellence in schools, which implicitly tied the issue of school funding to performance and quality, rather than equity. When the report was released in 2018, Turnbull announced that "a quality education is the bedrock for success throughout life, we can and must do more [and] we now have the blueprint to do it" (Koziol, 2018). However, where the original Gonski report challenged the government of the day to comprehensively change Australia's school funding model, Through Growth to Achievement offered a selection of policy prescriptions reflecting broader trends in global education policy. Significantly, both Gonski reports demonstrated a trend towards standardised policy, an expression of what Savage (2020) calls "alignment thinking" in Australia's education policy landscape. Under these conditions, the vagaries of particular state systems and local circumstances are less important than the belief that there is one "best approach" and "making a standard guide to that" (Savage, 2020, p. 73).
Education policy development has increasingly been outsourced to private enterprise (Ball, 2012;Hogan, 2016), and the rationalities produced increasingly conform to profit-oriented logics. Slow and careful policy development is devalued, and governments emphasise 'fast policy', or quick, silver bullet solutions (Lewis & Hogan, 2019). This preference for fast policymaking was noted by former New South Wales Minister for Education, Adrian Piccoli, when explaining why the New South Wales Department of Education established the Centre for Education Statistics and Evaluation (CESE), an internal research unit: [CESE] was established to have an in-house research capability in the department, because out-sourcing it was expensive and took too long. [The department spent $15 million] on consultancies and research-very little of it was spent on universities, a lot of it was spent on consulting firms. Consulting firms, you pay them and you'll get roughly what you want and if you want to change a couple of sentences because politically they're dynamite, they'll change them. (Savage, 2018) Piccoli's statement sums up the nexus between political actors, the state bureaucracy, and the edu-businesses and think tanks that comprise what Ball (2012) describes as "Global Education Inc". Loughland and Thompson (2016) have noted the way think tanks produce digestible research, which can over-simplify complex problems and disregard significant data. These publications are cited extensively in Through Growth to Achievement, and Piccoli's remarks underscore the fact that policy development is never only responding to actual needs: it also responds to political and media imperatives. The actors that define the rationalities of education policy now encompass agencies like the OECD, which administers the PISA tests, which Gorur and Wu (2014) note have been integrated into political 'ambitions'. These tests produce what Rowe (2019) describes as "data-fication"-vast datasets that mimic the language of 'big data'. Rowe argues that this is a function of a 'third-wave' neoliberalism, an imaginary that is crystallised in Through Growth to Achievement.
The report itself is an amalgam of these trends in policy development. At a discursive level, Buchanan (2020) argues that "it proffers common-sense, logical solutions to the problem of Australian students' performance decline in both national and international assessment" (p. 5). It does this through a suite of recommendations, many of which draw on logics that emphasise technological solutions to problems, as well as evidence-based approaches that privilege particular forms of evidence.
The report identifies three priority areas: delivering at least one year's growth in learning for every student every year; equipping every student to be a creative, connected and engaged learner in a rapidly changing world; and cultivating an adaptive, innovative and continuously improving education system (Gonski et al., 2018). In support of these three priorities are recommendations across five areas addressing pedagogy, school system structure, curriculum and teacher professional standards. These areas are as follows: • Laying the foundations for learning • Equipping every student to grow and succeed in a changing world • Creating, supporting and valuing a profession of expert educators • Empowering and supporting school leaders • Raising and achieving aspirations through innovation and continuous improvement The final area of these five proposes significant changes, notably a Unique Student Identifier and the establishment of an independent research institute. The recommendations generally articulate an aesthetic understanding of education that rests on decontextualised notions of 'good teaching' and maximising the effect of teacher practice, which broadly aligns with (and cites) the 'what works' literature (for example Hattie, 2009Hattie, , 2015. Biesta (2010) has noted the shortcomings of the 'what works' approach, suggesting that it can set aside significant questions about educational values and broader social structures that influence educational outcomes. These are matters that are not addressed in Through Growth to Achievement: the report articulates a technicist approach to 'school excellence'. Before exploring the way these features appear in the report, I will first explain the sublime typology.

The Four Sublimes
The four sublimes-technological, economic, political and aesthetic-are elaborated by Flyvbjerg (Flyvbjerg, 2014(Flyvbjerg, , 2017 as a theoretical lens for policy analysis. Flyvbjerg's work draws on Foucauldian understandings of power and discourse, and this approach can be understood as a mode of discourse analysis informed by sublimes literature. These sublimes represent reifications of elements of a given policy often used to make it appealing to stakeholders, investors, decision makers and political figures. This theoretical lens informs researchers about how these sublimes may be active in policymaking. Flyvbjerg (2014) explains that the term 'sublime' was first used in relation to the positive reception of technology during the nineteenth and twentieth centuries. Flyvbjerg uses terms like 'rapture' and 'excitement' to characterise the sublime, and Frick (2008) explicitly relates these to the literary tradition of the sublime. In poetry, scenes of natural beauty-vast landscapes and soaring mountains-inspire a sense of awe, designed according to classical criticism "not to persuade the listeners, but to entrance them" (Longinus, 1965, p. 100). In the context of policymaking, this entrancement is part of the strategy required to gain political, financial or community investment. Frick (2008) argues that the sublime captures the personal, political, functional and aesthetic implications, dramatically influencing the shape and outcome of projects. A sublimes approach enables questions of motivation and relations to be explored by making clear the way these features interact, or even change, as a policy evolves.
The economic sublime represents the delight business people, trade unions and politicians get from generating jobs and money (Flyvbjerg, 2014). In schools, this sublime is often articulated in terms of increased social mobility or national productivity: education is positioned as a pathway to jobs and skills.
The political sublime is most often invoked where politicians can clearly articulate a political vision or signify a policy achievement. It is important to distinguish the political sublime from the mere presence of politics. The fact that policymakers exploit the four sublimes to 'sell' a policy is not in itself an expression of the political sublime, though it may be a political act. Rather, the political sublime refers to those features of a policy which give expression to political values, such as the discursive shape of the Obama administration's 'Race to the Top' or the Bush-era 'No Child Left Behind' policies.
The technological sublime refers to the excitement that comes from advances in technology-using technology to solve problems, invoking the 'latest and greatest' digital solutions. The trend towards data and analytics and eLearning, and a "faith in data as a solution to education problems" (Thompson & Cook, 2017, p. 742), is 1 3 one instance of this sublime. This sublime is tied to the excitement of engineers and technologists, looking for technological solutions, or opportunities to advance what is possible (Flyvbjerg, 2014).
Finally, the aesthetic sublime references the pleasure designers and people who appreciate design gain from creating something beautiful, iconic, or that exemplifies good design. In education, the aesthetic is best understood in the way particular models of learning are reified for their apparent elegance, ease of use and effectiveness. For example, Visible Learning (Hattie, 2009(Hattie, , 2012 has cache with policymakers and educators for its rigour, simplicity and systematicity-it occupies the aesthetic sublime when framed as the 'holy grail' of learning (Terhart, 2011).
For this analysis, the report was first read to make sense of the rationale and evidence for each of the recommendations. Recommendations were then coded where they exhibited a clear sense of one of the four sublimes, such as mentions of technological solutions or aesthetic approaches such as 'learning progressions'. Gillett and Tennent (2017) took a similar approach to coding in a study of a major government project. These coded sections were then re-read, looking both for recommendations that referred to specific actions taken by government or systems that also exhibited a sublime (rather than vague recommendations, such as those that began with the verb 'ensure' or 'review'), and asking who benefits from that arrangement and what the implications of a particular set of sublimes might be.
The table below is an adaptation of the approach employed by Gillett and Tennent (2017, p. 100). Characteristic code descriptions for each sublime are included, followed by recommendations that exhibited each sublime. Some recommendations did not exhibit any sublime, and are thus not included. Finally, the While this paper focusses on the recommendations, the sublimes were apparent in other parts of the report in support of these recommendations. Gillett and Tennent (2017) note that the sublimes are a dynamic framework, that they can be observed to "melt away" (p. 109) as policies and projects evolve, and it may be the case that subsequent policy changes exhibit different sublimes, making this analysis a snapshot of a policy arrangement at a particular time.
What became apparent through this analysis was the dominance of the technological and aesthetic sublimes within the recommendations made by the report. A given document may not demonstrate all the sublimes. In fact, part of the logic of these sublimes is that they represent competing motivations, and the presence of one to a greater degree than others offers some insight into 'optimistic' visions of the future that might drive a particular policy (Frick, 2008). This is notably the case in Through Growth to Achievement, with its minimal use of the political and economic sublimes. It is to the dominant technological and aesthetic sublimes that this paper now turns.

Technology and aesthetics in Through Growth to Achievement
The process of coding for this analysis drew attention to a central framing device used in the report, that of 'an innovative education system', which exhibited strong features of the technological and aesthetic sublimes. Supporting that proposed system were two recommendations that strongly reflected the presence of the technological sublime: a new online and on-demand assessment tool and a unique student identifier. This section argues that these elements exemplify these sublimes and seek to rationalise particular policy approaches.

An 'innovative education system'
The aesthetic sublime refers to the 'transcendent' quality of a project that is specific to its domain: its good design. This happens in education policy when particular practices are reified. In Through Growth to Achievement this includes its focus on an "innovative education system". Exhibit 27 from the Report outlines attributes of such a system: This description of a system is aesthetic: it describes a perfectly functioning process, an exciting and innovative system, while also referencing the role of data and technology. Reminiscent of the language of a start-up enterprise, this model builds on principles like experimentation, design and understanding 'old problems' with 'new mechanisms'. The language is saturated with ideas of the new. The diagram on the page-a circular process describing iteration and experimentation-is reminiscent of a design thinking model, popular in management training and learning design (Fig. 1).
The report hinges on the idea that students should achieve "one year of learning for each year in school". It includes three recommendations that outline a case for curricular reform and the implementation of learning progressions. The literature cited in the report in support of this draws substantially on Hattie's (2009) Visible Learning. Visible Learning has, in many ways, come to represent an aesthetic sublime in the space of education. The meta-analysis from which the Visible Learning literature has evolved is a synthesis of many quantitative studies which form what CESE (2016) describes as the "gold standard" in an evidence hierarchy. This literature centres the effect and impact of teachers through Hattie's prominent refrain: "teachers make the difference" (Hattie, 2003(Hattie, , 2015, suggesting that these "objective and rigorous" research findings can best guide teachers in their classrooms. Policymakers often invoke the work of Hattie, particularly in Australia where it is argued that "quality is the single most important in-school factor influencing student achievement" (For examples: NSW Education Standards Authority, 2018, p. 2; NSW Department of Education and Communities [DEC], 2011, p. 7; 2012, p. 2). Larsen (2010) has described this discourse as "teacher centrality", suggesting that not only is it an old conception given new expression, but that promoting school effectiveness research can ignore socio-cultural, economic and political contexts.
In fact, Hattie is one of the most frequently cited scholars in Through Growth to Achievement, reflecting a particular orientation towards this sublime. Terhart (2011) argues that elements of Visible Learning lead to a "modernised form of the romantic idea of the ideal, 'born' teacher" (p. 435). This ideal is an aesthetic understanding of the good teacher, and the systematicity of Hattie's effect sizes in the classroom is another, quantitative aesthetic that privileges a particular epistemological stance. Visible Learning is not unchallenged, and much critique has centred on what is left out of the analysis (McKnight & Whitburn, 2018;Rømer, 2019), or the appropriateness of its methodology (Bergeron & Rivard, 2017), though it has been argued that this is an unfair misreading of Hattie's work (Qvortrup, 2019). Nevertheless, it is a popular guide to classroom practice and is used by policymakers, school leaders and many teachers worldwide as a guide to 'what works' in the classroom.
The aesthetic sublime in education policy centres on the notion that there is an easy answer to the question of 'what works', despite this being contested (Biesta, 2007;Hattie, 2015;Mockler, 2011). 'What works' is a simplification, since classroom practice is messy. Focussing on effect size or describing 'evidence-based' practices gives policymakers a false sense of certainty about perfectly functioning education systems: believing that if everything is done properly no child will be left behind, gaps can be neatly closed. Through Growth to Achievement exhibits a neo-Taylorist mentality: a scientific approach to the management of education (Eacott, 2017). The appeal of this aesthetic to policymakers is understandable: out-of-school factors that impact schooling are complex and difficult. These factors are not addressed in the report, but it is indicative of the aesthetic sublime that it suggests a system might be created which mitigates these factors, raising questions about who is best placed to provide that solution. The enactment of that system is described in two of the technological solutions that follow.

A new online and on-demand assessment tool
Through Growth to Achievement is notable for its emphasis on technological solutions to the 'problem' of declining standards. It does this in two ways, both of which revolve around the collection of more data. The first of these is an "online and ondemand student learning assessment tool", adapted from a model developed by KPMG (Fig. 2): This tool aligns with Recommendation 11 of the report: "develop an online and on demand student learning assessment tool for teachers for the purposes of formative assessment and tailored teaching" (Gonski et al., 2018, p. 66). The underpinning logic suggests that teachers do not have sufficient time, expertise and capability to "assemble and assess data on student growth". The report claims that "few assessment tools or tests currently exist in Australia to measure an individual student's learning growth over time" (Gonski et al., 2018, p. 62). The discussion of these tools is framed within a broader need to develop 'tailored teaching', citing a blog post by John Hattie and a report from the Grattan Institute (see Goss et al., 2015;Hattie, 2017). In doing so, Through Growth to Achievement links the technological sublime with the aesthetic ideal of effective teacher practice.
This form of policy-merging a vision of teaching that privileges measurable impact with technological approaches to that measurement-is problematic. Thompson and Cook (2017) suggest that learning built on data collection and 'analytics' is not 'personalised' so much as algorithmically adapted-it is designed to prioritise engagement, based more on profiles than individual learners. Data made to function this way elide the role of the teacher and their moral agency, the significance of the classroom environment and the importance of belonging. Biesta (2016) has argued that numbers and measurement are "seductive", binding the technological with the 'what works' literature. It is noteworthy that all of the submissions the report cites in favour of a national computerised formative assessment tool were submitted by government departments and bureaucracies, and not by bodies representative of teachers or researchers (Gonski et al., 2018, p. 63).

A unique student identifier
In line with the importance the report places on data collection is Recommendation 22: "Accelerate the introduction of a national Unique Student Identifier for all students to be used throughout schooling" (p. 103). The report argues that "better data at a student level from a Unique Student Identifier will improve student and school outcomes". This implies that one of the core barriers to improving outcomes is a lack of data at a national level, despite education in Australia being delivered by the states. This shift in the locus of student data exhibits features of what Savage et al. (2021) describe as "scalecraft" within Australian schooling reform: shifting the boundaries of policy from state to national scale, performing a form of 'alignment thinking' in assessment practice. While the report suggests that the identifier is necessary for tracking student records as they move through the education system, there is no clear rationale to support this objective other than that schools should be aligned with tertiary institutions and "where there are calls to increase the relative priority of specific skills and subjects" (p. 102). The report argues that "the absence of a national, persistent USI is a barrier to creating national educational data sets that would assist in developing a comprehensive understanding of the impact of policy or partnership efforts" (Gonski et al., 2018, p. 102). The problem here is represented as a lack of data alignment at national scale, and the ability to trace students as they move through the system.
Education systems already have more data than ever on students: Thompson and Cook (2017) describe "an endless and ever-growing stream of data" (p. 742). Allen et al. (2018) have argued that the need to familiarise themselves with the data a USI would collect would further intensify teachers' already significant work. Policymakers often avoid reconsidering practices that have been introduced that may have unintended effects, such as intensified assessment regimes. The very concept of education as a competitive race, where nations might need to 'prioritise' particular skills, is itself a function of the economic and technological sublimes: the political idea of 'raising standards' imbricated with the technological imperatives of the early twenty-first century.

'Evidence-based' policymaking after Gonski 2.0
Since the report was delivered, the Australian Government has signed a new national agreement with State Governments which bear constitutional responsibility for schooling. The National School Reform Agreement (COAG, 2018) outlines a series of "reform activities" (pp. 8-9), many based on the recommendations of the report, including "enhancing the national evidence base" (Recommendation 23). This will be achieved through the Universal Student Identifier, the establishment of a national evidence institute to inform teacher practice, and "improving national data quality, consistency and collection to improve the national evidence base and inform policy development" (p. 9). In line with these goals, the Australian Education Research Organisation (AERO) was constituted in 2020, with a remit to generate "high quality evidence" by working with practitioners and partners (Australian Education Research Organisation, 2021b). The organisation's strategic plan notes that AERO's annual research agenda is "developed with AERO's independent Board and approved by Education Ministers" (p. 9), ensuring political oversight. There is no reference to universities or collaboration with education researchers. AERO occupies a similar space in Australia's federal landscape as that which Adrian Piccoli reported CESE plays in New South Wales: an in-house and fast research institute that synthesises vast troves of student data, making recommendations for 'what works' (CESE, 2020).
This analysis raises further questions about the formation of a body like AERO. For example, if the problem is framed as a lack of access to research to inform practice, it is unclear why a new body collecting more evidence and performing new analysis is necessary: such research already takes place in universities and state 1 3 bodies such as CESE. A more sensible allocation of resources would see existing research-much of which is paywalled and inaccessible to teachers-synthesised and made available through AERO. This strong presence of the aesthetic and technological sublimes in this policy assemblage offers one possible explanation: that AERO is an expression of the 'what works' phenomenon, and is constituted with the intention of entrenching decontextualised, data-driven approaches to classroom practice characteristic of this aesthetic of education.
Reports like Through Growth to Achievement are often used for political ends, regardless of the need for reform. Araya (2015) argues that the Obama-era Blueprint for Reform "lacks the evidentiary basis upon which to form an empirical critique" (p. 99), noting the absence of peer-reviewed literature in the policy. Lewis and Hogan (2019) explain that "the need for highly visible political action often tends to override the need for a comprehensive approach to reform and, importantly, a particularly nuanced understanding of what constitutes evidence" (p. 1). Like the Blueprint for Reform, Through Growth to Achievement largely draws on literature produced by think tanks, government reports and global organisations like the OECD, with research produced by experts in universities comparatively underrepresented. Ultimately, these policies orient 'excellence' towards notions of economic productivity and the development of human capital.
In looking at these reforms, it is clear how the twin sublimes of the aesthetic and technological are mutually reinforcing. The collection and aggregation of student data enables an organisation like AERO to produce more research supporting the evidence-based practice movement. Many of the goals in AERO's current research agenda reflect that objective: the use and collection of data, the need for evidence to support teachers engaging in 'best practice' (Australian Education Research Organisation, 2021a). I do not argue that these are unworthy of research, however such a specific vision of decontextualised practice-one that is already taken up by organisations like CESE-aligns with a narrative that teacher professional judgement is suspect (Mockler, 2018(Mockler, , 2020. They can also minimise non-classroom factors that have an impact on student learning; factors that governments may find too difficult or politically inconvenient to address. These messy narratives contradict the coherent sense of ordered policy associated with alignment thinking, which assumes that "progress will come through re-arranging diverse people, ideas and practices in line with common and apparently more efficient approaches, based on evidence about 'what works'" (Savage, 2020, p. 2).
This analysis notes the emphasis placed on technological solutions, something that Buchanan (2020) has observed in education policy in the United Kingdom, China, New Zealand and the United States. These solutions interlink with the privileging of certain notions of 'best practice'. Rømer (2019), for example, argues that the way Visible Learning traces the idea of feedback can create a "powerful centralizing effect", one that lends itself to intensified regimes of data collection within school systems. These mutually reinforcing sublimes obscure the possibility of complementary and important forms of practice that depend on teacher professional judgement and the personal relationship that happens in the classroom.
Through Growth to Achievement assumes a range of policy measures and approaches to education that follow predictably from a limited pool of evidence, the 1 3 Technology and aesthetics in school excellence policies:… selection of which reflects Easton's (1953) formulation of policy as the "authoritative allocation of values". Dominant logics-such as what works-are elevated as common-sense solutions, and contrary evidence is "overlooked or glossed over" (Flyvbjerg, 2014). Godfrey (2017) describes the risk with this approach, arguing that the United Kingdom's Educational Excellence Everywhere white paper privileges models of teaching drawn from particular research approaches: "while the government rhetoric supports professional autonomy, this message is in danger of being overwhelmed by its simultaneous support to a top-down model of knowledge production that promotes a disempowering prescription to practice" (p. 442). In addition, the actors who feed into the policy development process also shape what constitutes good evidence, whether that be a government agency such as CESE, or a global edu-business such as Pearson, further removing the notion of good practice from the context-laden environment of the classroom.

Conclusion
School excellence policies are sites of tension and struggle and competing interests are expressed through reference to the sublimes. A sublimes analysis helps us to understand the reifications, motivations and logics at work rhetorically and that are reflective of globalised policy interests. The heuristic of the sublimes brings into view the series of personal or institutional assumptions that drive the development of policies, pushing and pulling them in different directions, tracing lines of rationalisation. Through Growth to Achievement exhibits many of the features identified by Thomson et al. (2012) as an "emaciated assemblage of unimaginative top-down ideas" (p. 2). The focus on technological solutions, particular models of learning, and an 'innovative education system' further aligns education policymaking at a national level in Australia's federal system. These are characteristic of sublime visions of educational policy that are repeated in different countries around the globe.
The dominant logics in globalised education policy development have a political edge at the same time as a technocratic function seeking to impose decontextualised ideas from the top down. Flyvbjerg (2001) argues that social science is most powerful when it looks at what should be done in context-dependent, expert-mediated situations, rather than trying to be predictive. This invites us to pose questions about the report and the stakeholders that it represents, specifically "who wins and who loses?" (Flyvbjerg, 2012). In the case of Through Growth to Achievement, providers of platforms, systems and centralised approaches to learning are the winners. It is too early to say who will lose, but the privileging of logics that are often insensitive to context is not promising. This article is an initial attempt at using sublimes in policy analysis, and further work is needed to realise their value in understanding how and why policy takes the forms that it does. A sublimes analysis not only grapples with the complex motivations behind policy and its stakeholders but helps us to understand the strategic rationalisations that seek to make these policies 'acceptable'. Most significantly, the sublimes help us to understand the various ways that policy can be made to cohere.