Comparing Policies in a Globalizing World : methodological reflections

Comparing Policies in a Globalizing World: methodological reflections. This paper explores the methodological challenges in comparing education policies in a globalizing world. It will be developed in the following ways. We begin with sketching out the contours of the changes that have taken place in the governance of education systems as a result of global processes and the challenges to presents us with regarding how we study, and compare, education policies. We do this by way of four ‘isms’ which we problematise as litmusses of global educational change. We then raise the question of comparison, and point to two conflicting ways that it can be used in studying education policy. In the final section of the paper we offer three (not exhaustive) methodological reflections – each with a different dimension through which to explore global education processes – time, space, and logics of governing in education policymaking.


Introduction
This chapter explores the methodological challenges in comparing education policies in a globalizing world. We begin with the claim that, for the most part, education policies, programmes and practices have and continue to be located in national territorial spaces, though this did not mean that the global was absent. Rather it is possible to detect a 'thin' global policy regime historically, and arguably so in the years following World War 2 until the 1980s and the rise of neoliberal policies. With the advance of neoliberalism as a global political project, there is a thickening of global policy making activity, on the one, hand, and the transformation of national and regional education spaces, policies and outcomes that are in turn aligned with globally-oriented agendas more closely tied to, and productive of, new social and economic forms -global economic competitiveness, the creation of knowledge-based service economies, and so on. In short, the form, content and scales at which education policies had become more global. This in turn has generated important challenges for researchers of education largely as education policies are no longer 'national' or indeed made by national states yet comparative education has tended to take as its basic unit of comparison when it comes to education policy -that education is sub/national and that education policies are made by the state. This chapter will be developed in the following ways. We begin with sketching out the contours of the changes that have taken place in the governance of education systems as a result of global processes and the challenges to presents us with regarding how we study, and compare, education policies. We do this by way of four 'isms' which we problematise as litmusses of global educational change. We then raise the question of comparison, and point to two conflicting ways that it can be used in studying education policy In the final section of the paper we offer three (not exhaustive) methodological reflections -each with a different dimension through which to explore global education processes -time, space, and logics of governing in education policymaking.

'Isms'
We've argued that in order to study and compare global education policies, we need to be very mindful of the conceptual categories that we use -in large part because though the name of the category might remain the same, the meaning of that category -such as the state, or nation, or indeed what we understand education to be, has changed. We have referred to the practice of deploying these same categories without asking questions about the meaning of that category, as methodological isms.
The basis of the way we understand and seek to use the term 'isms' comes from the coiner of the term 'methodological nationalism', Herminio Martins. He sees it as representative of 'a general presumption (in sociological analysis)…that the 'total' or 'inclusive' society, in effect the nation-state, be deemed to be the standard, optimal or even maximal 'isolate' for social analysis ' (Martins (1974, 276) quoted in Chernilo 2006. The idea of a 'general presumption' about the nature of the field of captures the essence of what we mean by the 'isms'. They can be seen as 'pre-theoretical', too obvious in their (assumed unchanging and unchanged) form and importance to require explicit theorizing, or being addressed as objects of inquiry, to the point where they become 'ossified', for example, as current analyses of education policy tend to retain the same methodological and theoretical assumptions in massively changed circumstances. It is this that we refer to as 'isms'--fixed, frozen and taken for granted, representing and embodying significant forms of the distortion and possible understandings of education policies, through the restrictions they place on the scope and targets of investigation. The four 'isms' we will be discussing here are methodological nationalism; methodological statism; educationism; and spatial fetishism.
As we have noted, methodological nationalism, is based on a -frequently implicitset of assumptions that essentially equates 'society' with 'the nation'. It operates both about and for the nation-state to the point where the only reality we are able to comprehensively describe statistically is a national, or at best an international, one (Dale, 2005: 126). This is exacerbated by the tendency to juxtapose an unreconstructed methodological nationalism to underspecified conceptions of 'globalisation' in a zero-sum relationship; that is, as the global has taken on more functions and power this has been assumed to be at the expense of a new disempowered state. This is far from the case, in that in many cases the national state itself has been a major force in advancing regional and global projects.
There are close relationships between methodological nationalism and what we refer to as methodological statism, the tendency to assume that there is a particular form intrinsic to all states, is closely related to this. Methodological statism essentially takes the version of the 'state' as found in 'Western democracy' as 'the organizing principle of political modernity' (Fine, 2003, 460, quoted in Chernilo 12). For Chernilo this constitutes the 'rather mythical image of the nation state as the final and necessary form of social and political organization in modernity' (ibid). And one further, relevant, consequence of this is that it makes political, rather than economic or cultural boundaries, the dominant means of differentiating societies from each other, setting distinct limits to both the bases and the product of useful comparison, since the national has become the basis of the collection of statistics of all kinds, with 'the state' (the historical generator of 'statistics') typically seen as the major collector of such data. Hence, as "…public authority has been demarcated by discrete boundaries of national territory…so, too, has the articulation of societal interests and identities that both buttress and make demands upon this authority" (Ruggie, 1993: 8). However, in a globalising era, the particular combination of responsibilities and activities that nation-states have been assumed to be responsible for can now be seen as historically contingent rather than functionally necessary, or even optimal, to the point where the question can be raised about the "…implications of a world in which the mutually reinforcing relations of territory, authority and societal interests and identities can no longer be taken for granted" (ibid.: 9). The depth of the penetration of these kinds of assumptions on the social sciences is summed up by Ruggie as displaying; "…an extraordinarily impoverished mind-set…that is able to visualize long term challenges to the system of states only in terms of entities that are institutionally substitutable for the state" (1993: 143). The point here is not to suggest that the state as an actor is unimportant. It has, and continues to be, a very significant and powerful ensemble of institutions that is able to mobilize power and act. Rather it is to focus upon, first, the way the idea of the state represents itself as a universal form rather an a particular representation that has been universalised, and second, on the way the state itself, as both a project and container of power, has evaded close intellectual scrutiny. In relation to this first point, of the universalization of the form of the state, this has made investigations into, for example, the Europe Union, as also involving a different form of the state, difficult but important (see Shore, 2006). Difficult as an assumed form of the state is essentialized; important as it points to the need to develop new concepts that help identify and reveal the changing geometry of state power.
We can illustrate the points made above about methodological statism by recognizing that the national state can no longer be taken-for-granted as the only, or most important, actor in the area of education. If we look closely at the governance of education-that is the combinations and coordination of activities, actors/agents, and scales, through which 'education' is constructed and delivered in national societieswe can identify four categories of activity that collectively make up educational governance (that are for the sake of exposition taken to be mutually exclusive and collectively exhaustive), funding; provision or delivery; ownership; and regulation.
These activities may in principle be carried out independently of each other and by a range of agents other than the state -though the state remains a possible agent of educational governance and at a multiple set of scales, from the local to the global.
Educationism: This refers to the tendency to regard 'education' as a single category for purposes of analysis, with an unproblematically accepted set of common objectives, and a set of implicitly shared knowledge, practices, assumptions and outcomes. This results from the fact that Education has been possibly the central project in modernizing societies. Since the early nineteenth century, mass education has been a crucial element of the modern nation state in the interests of collective progress and in the interests of equality and justice. Educational systems are almost invariably seen as rationalizing social projects whose universal expansion necessarily brings improvement and emancipation. This results in education being treated as abstract, fixed, absolute, ahistorical and universal, when, for instance, no distinctions are made between its use to describe purpose, process, practice, and outcomes. It is this 'flattening' of 'education', and the reluctance to recognise that there are crucial relationships between different representations of education, that are being occluded or disguised by the failure to distinguish between them, that makes it so important to identify and seek to go beyond educationism.
While the term 'education' potentially embraces the whole range sets of practices, processes, institutions and outcomes, carried out in its name, in practice debate and discussion tends to take place within the category Education', rather than challenging it. There is little questioning of the principles of Education (even there may often be little agreement between them), and this often persists in the face of evidence to the contrary (see e g, Benavot, 2002). The label education is also normative in that education is invariably viewed as a good thing (and the more the better). However, and crucially, this enables the avoidance of the fact that education is always about the acquisition of particular knowledges, by particular groups of individuals, under different circumstances, with the result that how far and in what ways it may or may not empower an individual or group, will depend upon a range of features of their social location. The point here is that 'educationism' occludes and 'flattens' all these multiple forms, or reduces them to one particular set of understandings.
Fundamentally, 'educationism' is the product and instantiation of analyses based in examining definitions and examples, rather than in examining the range of what is done in the name of these definitions-which, crucially, may be unintended as well as intended. It is assumed to be ameliorative, with any questions to do with the forms and outcomes of the attempts at amelioration. So, the crucial point for us here is that 'education' requires explanation rather than being taken for granted.
Spatial fetishism: Brenner (2003: 38) describes spatial fetishism as '…a conception of social space that is timeless and static, and thus immune to the possibility of historical change'. Failing to problematize space, or to see that space itself is both constituted by, and constitutive of social relations and structures, is a problem for the analysis of education policy, more generally, and global education policy more specifically. The reason is partly contained in the phrase -'global education policies'. Education policies are always about change -even if by 'change' we mean containing events sufficiently so as to put a brake on those dynamics that might otherwise change things.
For instance, those with social class privileges are likely to try and contain particular social groups who might organize so as to create more opportunities for social mobility from the classes below. Education policies might be advanced to ensure that things remain the same. Put a different way, education policies are aimed at re/organizing and re/ordering social relations through structures and strategies.
Education polices are thus concerned with social relations which are always spatial in some way.
Some spatialized relations might not matter in their outcomes, but others will. For instance, global policies such as school choice typically do for they will have very different spatial implications for families; not all families will have the financial resources, time, or conditions of work, to move children across the city so as to access a school that might be the 'best choice' (Ball et al, 1995). Other families will face not having a choice as their village or town only has one school. Space also matters in the organization of learning. And indeed, some education policies might have, as their intended purpose and outcome, the separation of social groups -smart kids in science streams versus the not so smart kids in general streams; girls in girls only schoolsboys in boys only schools; leafy neighborhood schools versus schools in dense city spaces, and so on. These spatial differences matter as they shape social relations, on the one hand, and are often key dynamics in systems of social stratification, on the other.
If we now add the spatial category 'global' to our analysis, we need to ask what kind of category it is, and what work it enables/disables in relation to the national, or local -which are alternative scales from which strategic projects of rule are launched. In this case, we are interested in the education policy work that those actors operating with a global scale horizon advance. In some cases the global scale enables actors to act in rather more omnipotent ways -where the global scale is invoked as a higher form of authority and rule. In other cases, the global scale enables policy projects to advance quickly -unencumbered by institutions and other actors who might have different views about the probity or not of these policies. Rescaling is thus a useful spatial move for education policy makers. For the comparative theorist, the question to be asked is: what is the role of space in global policy, and how might we compare the different capacities of policy actors to engage spatially.

Critical Comparison
The importance of the use-and misuse-of comparison in education policy lies in the fact that it has become a significant consequence and outcome of the changing relationships between globalization and education policy. This development is important in a number of ways in understanding how particular understandings of the relationships between globalization and education policy have arisen, and the consequences of these for the ways that we might try to explain them.
There are two main ways that we can approach 'comparisons'. On the one hand, we between these two sets of questions are crucial in understanding the contributions of comparison in addressing issues of governance of education. In the first case, the comparison is used as a 'resource', a contribution to the achievement of particular ends. In the other case, the comparison itself becomes the 'topic' of inquiry. In the first case, the findings themselves are taken to provide the explanation; in the second case, they generate further sets of questions. The best example of the difference between these two approaches is the use of large, quantitative cross national data sets. Those using comparison as a resource take these data for granted, and ask what we can learn from them; the other asks on what bases these data were compiled, and hence what was being compared. This is a crucial issue in understanding global governance in and of education. In the first case, comparison itself becomes a crucial tool of governance, with comparison being seen as a resource whereas in the second case comparison itself becomes the topic of investigation, the means of -critically, in the sense of problematizing, not taking for granted-analyzing it. The first sees comparison as a tool for providing generalized solutions, the second as a means of generating explanations. While comparison as resource might be seen as offering a crucial form of 'lesson learning', one striking point about PISA is that it does not really offer particular models for emulation. For instance, the success of Finland in PISA tests led to the development of a mini-industry (see Sahlberg, 2011) of visitors eager to find the secret of educational success, but they found only that the 'Finnish' model departed in almost all ways from that implied by the OECD-well educated and paid teachers, no national testing, etc. This illustrates another aspect of the kinds of comparison carried out by PISA; what are compared are outputs, in the form of test scores generated by PISA itself, not the different educational processes that led to them, nor the criteria used to produce them.
In adopting this second 'topic' oriented approach regarding comparison of global education policies, we need to ask ourselves: what exactly are we comparing? In our view this is an ontological and epistemological, as well as a methodological, question.
By this we mean that in considering comparison methodologically, we are also making decisions about how we think the social world works, and what might count as a means of knowing that world. Does the social world operate according to a set of regularities, and in the world that global education policy is present in, are we able to This latter -broadly critical realist -approach is the one we favour, in that we take the view that social realities are socially stratified, and that the causal mechanisms and powers shaping events are not necessarily visible to the researcher though the outcomes are (Sayer, 2000). Working backwards to work out the relationship between outcomes, mechanisms and causal powers is an important procedure for a critical and it is our explanations of these processes at work that sits at the heart of comparison. We'll return to this in our conclusion. For now, let's explore what might be gained by using different dimensions -time, space, governing tools and power -as a basis for comparison.

Three Critical Methodological Reflections on Global Education Policies
In this section we develop three methodological reflections around different dimensions through which a comparison can be made. These are by no means exhaustive; rather they are meant to encourage you to think imaginatively, systematically, and critically, about the different ways in which comparative research on global education policies can be carried out.

Methodological Reflection 1: The value of temporal comparisons -global teacher policy -from thin to thick globalization of education
One way in which we can compare education policies is by asking: how might we compare a particular kind of education issue of problem using time as one of the variables. Of course comparisons using time can be developed either synchronically or diachronically. By synchronic we mean comparing, for instance, different global institutions, and how these organizations sought to shape an education policy issue had any role in shaping policy for teachers. After all, the development and growth of education systems has been tightly bound to the rise of the nation state. A second move would be to establish when and which international agencies took it upon themselves to shape teacher policy, and from there ask: why, and how?
Digging a little deeper, we begin to see that though national and sub-governments were the key shapers of teacher policy, this did not mean that the international agencies were silent or disinterested. Far from it. Indeed two international agencies over the period 1950-2000 were very interested in teacher policy at the global level; UNESCO and the ILO (Robertson, 2012). Whilst respecting the right of Member States of the UN system to determine teacher policy, they nevertheless published an ILO/UNESCO Recommendation in October 1966 setting out the rights and responsibilities of teachers, including international standards for their initial preparation and further education, recruitment, employment and so on. As a Recommendation, it did not have the weight of authority or legal 'bite' that one might see behind a sub/national teacher policy. This meant that governments in national settings could ignore this guidance. Connell (2009) suggests that this resulted in a broad range of approaches in national settings as to how teachers were prepared, what power and autonomy they might have in their schools, and so on. As we have shown elsewhere (see Robertson, 2012;2013), the OECD and the World Bank, have entered into the teacher policy space -legitimating their presence and concern by arguing that teachers and teaching matter to pupil performance, and pupil performance matters to developing globally competitive economies. This line of argument has been given weight by other global actors who have become very prominent in the education policy field -including the giant education firm -Pearson Education, the global consulting firm McKinsey and Co., (Barber and Mourshed, 2007;Mourshed et al 2010) and foundations that include the Bill and Melinda Gates Foundation. Paradoxically this global conversation has not involved teachers in any significant way, and it might have via their global union, Education International.
Instead, the OECD seems to have acquired quite a strong voice in framing the nature of the problem and its solution, and is currently attempting to speak directly to teachers through its specially designed toolkit for teachers. Now, rather than a conversation that teachers might have with their sub/ national unions and sub/national Ministries of Education based on respecting their expertise and professional autonomy, teachers are given a one size firs all set of guidance notes to operationalise by the OECD (2014), whilst at the same time that teachers are made aware that they have also been ranked as part of a system of comparative performance.
This short account exploring how to compare a global education policy -such as teacher policy -using diachronic and synchronic time as dimensions reveals the insights that can be had from using such a lens. Time, of course, is always linked to space -and actors are always located in time and space. But by looking at global education over in this way -we are able to appreciate that even in the post war years (for instance with the Declaration of Human Rights established in 1948 -with a strong statement there in the right to free public education) there was what amounts to as a 'thin' global governance. By 'thin' we mean that though there were actors and guidelines, the weight of authority and legitimacy to shape education policy continued to rest with the nation states and sub/national governments. However, from 2000 onwards UNESCO and the ILO have been joined by a newer set of global players (such as OECD, World Bank, McKinsey and Co., Pearson Education, Bill and Melinda Gates Foundation) actively shaping teacher policy, and indeed crowding out UNESCO and the ILO as far as power, authority and influence goes. We might call this current period 'thick' global governance in that the global actors now dominate the framing of teacher policy. Elsewhere we have developed this analysis more fully (see Robertson, 2012Robertson, . 2013, However, for this purpose here -it is sufficient to point to the ways in which we are able to see shifts in what scales (local, national, global) become more dominant in shaping particular education policies and to propose reasons for what we think might be explanations.

Methodological Reflection 2: The value of comparing spatial recalibrationsrescaling education policymaking
We argued earlier that space matters in education policy analysis (Robertson, 2010).
We also argued that it particularly matters in thinking about global education policies in that it helps to focus attention on what kind of spaces we are comparing, and how is space itself used to advance policies at a global scale that might have been more difficult at the local or indeed national. Put a different way, the comparativist is being asked to compare the difference that space itself makes -in this case the global -in being able to advance a particular set of education policies.
At this point it is helpful to introduce the idea of scale; a spatial term which refers to a level at which particular kinds of institutions and actors concentrate -and from there seek to organize or govern social, political and economic activity (Smith, 2000).
In the previous reflection we argued that the global scale does house institutions who engage in framing and shaping education policy -but until recently their influence was fairly weak largely as these institutions lacked authority and legitimacy. Education policy was regarded as the preserve of nation states. However from the 1980s onwards, major changes began to take place within and between nations -as a new political project -informed by neoliberal ideas -began to be rolled out in countries like the UK, USA, Chile, New Zealand, and Australia.
The globalizing of neoliberalism has had major consequences for the form, scope and purpose of much education policy -as it was used to bring in what Peter Hall (1989) calls a third order change. By third order change he means a radical rupture in the ideational base that informs the what, who and how of policy. In this case, introducing neoliberalism into what were mostly Keynesian inspired social orders meant setting policies to work on aligning education more closely with the economy, making education into a competitive services sector, introducing policies that aimed to encourage a more competitive entrepreneurial identity, and rewarding institutions for acting in more economically efficient ways (Dale and Robertson, 2013). Bringing in a new social and economic order, however, is itself not straightforward. Previous ways of organizing social life, and the norms that ensured these ways of life are embedded and reproduced, are thus challenged and transformed into new practices with rather different logics, forms of reason and outcomes. In doing so, education space and its constitutive social relations are also reworked in new ways.
One strategic way forward for governments and aligned actors was to rescale (Robertson et al, 2002). By this we mean that strategic actors relocated themselves, or ceded some of their authority, to a new scale -above or below the nodal scale that had been a key passage point, or site of authority, for governing in order to drive these new political initiatives forward. Decentralisation policies became a favoured set of education policies. They aimed to use the local scale to advance initiatives that were likely to be blocked at the higher level. In this case local communities were asked to take on the responsibility for education policies aimed at generating competition and efficiencies, oversighting the work of teachers, aligning the school with the needs of local business, and so on.
However policymaking capacity -or some element of sovereignty--was also moved above the nation-state, to the regional and the global, so as to advance particular projects with rather different interests. A good example here is the Bologna Process -an education policy that emerged in 1999 which has had a huge impact on reshaping the degree architecture of higher education sector in Europe, as well as bringing in a new competency approach to learning. The initial move to rescale came from the French Minister for Education, Claude Allegre -who was particularly frustrated with the difficulties of getting change in the French academy. With universities being turned to as a means of generating a competitive knowledge economy, dealing with 'recalcitrant' academics willing to take to the barracks is quite a challenge (Ravinet, 2005). In 1998, Allegre used the celebrations surrounding the 800 years since the establishment of the Sorbonne, to announce the launch of a European Higher Education Space. At this point -only four countries were part of this agreement -Italy, Germany, the UK and France. In 1999, the Sorbonne Declaration morphed into the Bologna Process -an agreement ratified by 29 European countries. Since signing, the Bologna Process has grown rapidly in membership and reach -with 47 members and others who declare themselves Bologna compliant.
There are many interesting angles that the comparativist can be interested in here as this new regional space is created. Questions might include: how does this scalar project -the European Higher Education Area, compare with other similar regional projects where the capacity to govern higher education is being rescaled. What is the relationship new between this new regional scale and its capabilities, and those that remain at the institutional and national levels? Who gets to operate on which scale, and what are the outcomes for these different actors of any differences? Are all members of the European Higher Education area treated the same, and how might differences be accounted for? Are there convergences across the European Higher Education as a result, and new points of divergence, and what underlying mechanisms account for these outcomes? Has the ceding of power upwards generated the kind of outcomes Allegre was envisioning, and what explanations might we consider for what dynamics seem to be in play? In suggesting this array of questions, in our view any one of these would offer a fascinating exploration for the comparative scholar in bringing scalar processes into view and using space as an entry point into comparing the changing strategies, structures and social relations and arise from a particular kind of education policy.

Methodological Reflection 3: The value of comparing governing logics -the OECD and its global indicators
As we have noted elsewhere in this chapter, 'comparison' can be used in a range of ways in looking at education policy. At its simplest and least useful, it entails looking at two different entities, say countries, and asking how they are different from each other. This can provide fascinating contrasts, but it tells us little if anything about the nature of those differences and what, if anything, we can learn from them. More recently and relevantly, considerable importance has been placed on what we have referred to as 'competitive comparison', using comparison to construct a ranking of particular entities across particular qualities; PISA is the obvious example here.
However, here again, the use of comparisons for ranking tells us little about the substance of the policies and practices that produced them; they cannot, for instance, identify the reasons for the differences of importance. One key explanation of this is that 'simple' comparison between practices does not enable us to recognize the different rationales for which they are carried out, or what we will refer to here as the 'logics' informing them.
The way we will do this is to consider the nature of the interventions into the education field made by the OECD over the past 60 years or so. We will point very briefly to the nature of the intervention and then to the logic(s) that seem to be informing it. We should also note that in seeking logics, we also have to consider the purposes of programmes, since the two are closely linked, though rarely explicitly so.
However, in doing this it is important to identify the OECD's main priorities for education. Very broadly, following Rizvi and Lingard's analysis of Papadopoulos' (1994) history of OECD work in education, we can distinguish three main periods; in the 1960s, the main emphasis was on the 'quality' of education; in the 1970s, equality of opportunities and democratizing education, while in the 19980s the focus shifted to alignment with economic policy (and it is important to note that the term 'globalisation' is not mentioned in Papdopoulos' 1994 book). This brief history indicates which issues were of greatest concern to OECD members (and it is important to note that the OECD's agenda is formally determined by its members, whose voting strength tends to be related to the size of their contribution, with USA and Japan contributing over half the total funding).
One of the first OECD programme that concerned education was manpower forecasting. This arose in the context of postwar recovery, and its logic was based on an assumed link between the level of qualified manpower and economic growth. This was underpinned by a pervasive attachment to the logic of 'human capital formation', based on the inference that labour could be treated as a form of capital, and that its output could be enhanced through education.
A second logic deployed by the OECD can be found in the method of peer review that became quite prominent in the 1970s. This was based on a logic of 'lesson learning' from one's peers (countries). Reviews of national systems were carried out by experts from other countries, and fed back to those in the reviewed countries, with the idea that they would point to practices elsewhere that might be usefully adopted.
More recently-and following the creation of a separate Directorate for Education in 2002 -logics have tended to go into three, related, directions. The first was an 'ideological vocabulary of reform', which followed the success of neoliberal politics in USA and UK in particular; this emphasized the need to limit government intervention and to base governance on what was known as New Public Management (which essentially meant that states should be governed as far as possible on the basis of market principles); growing technical expertise in monitoring (taking the place of the earlier logic of peer review); and an increasing emphasis on quantification in shaping education policy, via the INES project (see Bottani, 1996). It is interesting to note that these might be seen in different ways as key elements that came together to form PISA. They represent a common logic of suspicion of 'politics' and the need to provide accurate information for the organization of not just the economy, but the whole field of pubic administration.
It should be evident now that PISA did not appear from nowhere, but emerged on ground already well prepared for it through programmes such as INES, but it goes beyond them in a number of ways, which are based on distinctive logics. What PISA provides, in a nutshell, is a tool for evaluating education systems worldwide by testing the skills and knowledge of 15-year-old students (emphasis added). REF??It claims to involve 400,000 students in 57 countries, making up 90% of the world economy, with nationally representative samples representing 20 million 15 year olds from 30 OECD and 27 partner countries (OECD, 2006).
Most importantly, PISA rests on what might be called a 'nominalist' logic-that all entities called 'national education systems' must necessarily have sufficient in common for it to be possible to diagnose them with the same set of tools, and to offer advice based on those findings. From a comparative policy perspective, the most basic flaw in the use of the PISA data arises directly from this, because, as we noted in the 'critical comparison' section above, the findings are themselves taken as sufficient explanations of national educational states of affairs; they constitute explanations rather than themselves requiring explanation, and this in itself shifts the focus away from their causes and on to the consequences that are manufactured through the PISA instruments.
Three other significant features of the logics informing PISA may be discerned: it rests on a logic of (especially statistical) expertise, and education policy is no longer best served by deliberations between variously informed and interested parties: in itself, it provides an accurate account of the 'health' of education systems, and offers remedies that will be universally valuable-such as evaluation systems; and it paves the way for the logic of competitive comparison, across particular dimensions of education.

Conclusions
Comparison is, above all, about problematising, rather than taking for granted and in this case problematizing phenomena that we have come to talk global education policies. As we have argued, through comparing things that are familiar to us with things with the same name in other places, we learn that there are different understandings of the same things in different places, at different times, with different origins and meanings. More briefly, comparing produces the possibility of difference, and it might be hoped, a desire to understand and explain those differences. So, a major issue for comparative approaches is to examine the relationships between nominally similar phenomena, and here, too, a comparative approach to global policies is valuable, not only indistinguishing meanings and uses, but seeking to explain those differences. And if this second step may not always be possible, the recognition that the same names are given to different phenomena in itself helps us to problematize those phenomena, rather than take them for granted, or assume that they 'must' somehow be comparable.
Finally, we suggest that it is also very useful in thinking about global education policy to ask: who compares and for what purposes? Recognising Novoa and Yariv-Mashal's (2005) excellent account of the possible purposes of comparison, which contrasts its use as a form of enquiry, it does require us to recognize that there are sides to be taken, and these have consequences for our analyses. All these issues are important when we are considering 'global' education policies---and this is especially the case when we recognize that a failure to problematize does not result in the status quo being taken for granted, but for it to be taken as an acknowledged fact.