PISA as an ideational roadmap for policy change: exploring Germany and England in a comparative perspective

Education policy has undergone transformation in many countries over the last decade. In this article, we focus on the effects of the most significant international initiative in secondary education, which is the Organisation for Economic Co-operation and Development's (OECD) Programme for International Student Assessment (PISA). We analyse two countries that provide variation regarding the degree of change in their respective education policy-making due to this study; while Germany substantially reformed its education system in reaction to its mediocre PISA results, almost no change has been observed in England. As we show, alterations and shifts in ideas of education policy best account for such a change.


Introduction
Education policy has been in a state of transformation in many OECD countries over the last decade. In response to rising demands of the labour markets, growing costs of public education institutions, massification of academic education or demanding demographic challenges, education systems are continuously under pressure. In this context, international organisations (IOs) provide important incentives for national education policy reforms through soft forms of governance in this field. They frame ideological debates about educational goals, present policy solutions to common problems or offer expertise service in highly specialised areas.
Traditionally considered an exclusive national domain of western welfare states, education policy has become a highly contested area of international governance (Martens, Rusconi, and Leuze 2007). But what kind of effects do the activities of IOs have? Do they lead to greater harmonisation among national education policies by promoting uniform solutions for commonly shared problems? Or do national institutions continue to follow their own logic and not respond equally to IO governance? In this article, we explore how IOs influence national policy-making in the field of secondary education. Our argument is that a constructivist-inspired approach highlighting the importance of ideational factors in education policy provides useful insight into the mechanisms of IO influence. IOs can act as utterers of these educational ideas, and their educational ideas may provide a roadmap for policy change.
In this contribution, we focus on the effects of the OECD's Programme for International Student Assessment (PISA), currently the largest international comparative study in secondary education. PISA surveys the competencies and skills of 15-year-olds at the end of compulsory schooling in the key subjects: reading, mathematics and science. 1 Since the year 2000 it has been conducted triennially with the results being published the following year. PISA goes far beyond the 34 members of the OECD; it analyses and compares academic performances across 75 different countries and regions. 2 For the purposes of this paper, we analyse two countries that provide variation regarding the degree of change in response to PISA in their education policy-making; while Germany substantially reformed its education system due to its mediocre PISA results, almost no change has been observed in England. In this article, our focus is on alterations and shifts in the ideas of education policy. Since the 1980s the paramount idea in England regarding education can be characterised as a means to boost productivity, i.e., as human capital, and thus it is relatively close to the OECD economic paradigm of education. In contrast, German education was traditionally framed primarily as a subject of individual self-fulfilment, which did not meet the orientational framework of education as promoted by the OECD. We claim that by changing the German ideational priorities the OECD paved the way for comprehensive reforms.
Our work presented here on the impact of an IO, such as the OECD, on education systems and ideas relates to the growing research on globalisation and internationalisation of education. While many studies focus on the economic aspects of education or its effects on markets and liberalisation processes, such as General Agreement on Trade in Services (GATS) (Knight 2009;Verger 2010), others examine the impact of international agreements, global institutions and transnational processes on the design, reform and further development of education policies (Mundy 2007;Robertson and Keeling 2008). With our contribution, we emphasise how international actors shape not only education systems but also the underlying conceptualisations of education.
The empirical parts of this article are based on a qualitative approach consisting of elite or expert interviews (Tansey 2007;Bogner, Littig, and Menz 2009). Between the years 2007 and 2008, we conducted a total of 30 semistructured elite interviews in Germany and England. 3 We interviewed policy makers and representatives of important stakeholders in the education sector. In order to gain an in-depth understanding of our cases, we supplemented the interview findings with policy documents and statements made in newspaper articles.

IOs as purveyors of ideas in education policy
Although today IOs play a central role in almost every policy field, most IOs have limited direct governance capacities at their disposal. From the perspective of international relations theory such soft governance by IOs has convincingly been conceptualised from a constructivist perspective. Scholars look at how IOs exert influence over their member states in the absence of legal means or material coercion as 'teachers of norms' (Finnemore 1993(Finnemore , 1996; see also Barnett and Finnemore 2004). To measure the impact of IOs on national education policy, they examine how IOs set norms, form opinions, engage in blame and shame policies, provide financial incentives or technical assistance, coordinate activities and offer consulting services (McNeely and Cha 1994;Nagel, Martens, and Windzio 2010).
The OECD is such a case where an IO exerts soft governance; it pushes incentive compliance by idea production and promotion, policy evaluation and data production (Martens and Jakobi 2010). It is only equipped with modest legal or financial instruments and cannot enforce or implement policies in its member states. In contrast, it relies on its good arguments and persuasion (Marcussen 2004). Particularly, in the policy field of education, the OECD is an actor with limited institutionalised competences (Henry et al. 2001). As regards the OECD's PISA study, it is a voluntary process; the participation of each state as well as the continuation of the programme as a whole needs to be renewed every couple of years. Although PISA is not the first large international education study, it is the most encompassing comparative assessment on secondary schooling to date. With PISA, the OECD has become a successful policy entrepreneur and a potential driving force of transformations in the realm of education policy (Nagel, Martens, and Windzio 2010).
Having the ability to govern is a necessary but not a sufficient condition for IOs to influence national policy-making. Furthermore, whether an IO's soft governance is considered or not in national political life crucially depends on its overall reputation. Those IOs with a good reputation are accepted as sources of information because they feature apolitical and technocratic expertise (Barnett 2002, 113). As Sharman stated, the impact of IO outputs such as published reports 'is inseparably bound up with judgments about the reputation of that institution' (Sharman 2007, 30). While outputs of IOs with bad reputations can easily be ignored by states and domestic actors, the outputs of well-accepted IOs can generate repercussions on national policy discourses, such as generating pressure and a push for reforms. The OECD is such a case of an IO backed by institutional reputation (Sharman 2007, 31Á32). With its PISA study, it even advanced to become an 'eminence grise' (Rinne, Kallo, and Hokka 2004) in education policy-making; its analyses are therefore highly regarded.
To analyse how and to what extent the PISA study is received at the national level, one has to take into account a country's domestic factors that determine if, how and to what extent IOs are in fact able to affect the political dimensions in a country. Scholars often therefore look at the institutional set-up of a country in order to analyse its capacities for change. However, such an approach has been criticised as being too static to convincingly explain change (Béland and Cox 2011a). As a consequence of this criticism, the turn towards ideas became a fruitful endeavour in comparative politics as well as in international relation theory to overcome the limits of institutionalist approaches in explaining policy change (Blyth 2003). In this article, we thus concentrate on the ideational 'misfit' between the ideas of education policy as proposed by the OECD with its orientational framework of PISA and the existing domestic ideas of education that accounts for changes in national education policy. In this regard, IO frameworks shape the discourse in a country and influence policy-making in a policy field at the domestic level.
Since the term 'ideas' is used in multiple ways and has multiple notions, an understanding of how it is used in the context of our study must be addressed. Basically, we follow Béland and Cox (2011b, 4) in conceiving ideas as causal beliefs that 'provide guides for action [and help] to think about ways to address problems'. Hence, ideas are a cognitive frame for interpreting an issue, identifying a problem and choosing appropriate solution strategies. However, ideas are not just tools in the hands of strategic actors (Lieberman 2002, 699), they need agents to be disseminated. However, they can indeed serve as instruments for actors to make claims. By this way, we accept the premise that the perception of reality is a product of an actor's constructivist interpretation. Referring to Wendt, what an actor wants he/she wants because of how he/she thinks about it (Wendt 1999, 119). How to think about something is strongly influenced by the ideas that serve as 'cognitive filters through which actors come to [. . .] conceive of their own interests' (Hay 2011, 69). As a consequence, the change in ideas can also cause a reassessment of policies and policy goals. In this regard, ideas matter in the process of policy-making insofar as 'policy actors may define Á and redefine Á their interests and preferences in terms of cognitive and normative factors, thereby linking ideational and institutional change' (Baldi 2012(Baldi , 1003. A central contribution of IOs, such as the OECD, is to initiate and promote an orientational framework. In this regard, the OECD plays a role as 'purveyor of ideas' (Mahon and McBride 2008, 3). In other words, the OECD acquires ideas, embraces them as its own orientational framework, cultivates them within the organisation and disseminates them to the member countries through diverse channels, such as reports, studies, symposia and so forth. Additionally, the educational values the OECD lobbies for are embedded in the wider ideational framework of the IO. Given that the OECD was founded to achieve economic growth and employment in the member countries, its understanding of education appears to be obvious: education 'plays a critical role in enhancing economic competitiveness and growth, facilitating personal development and building strong and healthy societies' (OECD 2006, 5).
Overall, the OECD fosters the connection between the economic and educational spheres by drawing from human capital theory (Rubenson 2008). As regards secondary education, this economic paradigm is reflected in OECD recommendations about investments and evaluation in education, flexible school degrees, school autonomy and individual support for students (Henry et al. 2001;Popp 2010;OECD 2010OECD /2011. Ideas about education can generally be distinguished into two major ideal types: from a neo-liberal economic perspective, education provides a helpful means to improve the competitiveness of a national economy. Increasing human capital serves as a tool to further productivity and to generate or increase wealth for a nation. From a societal perspective, education is crucial for providing social cohesion, as it is a means to create a national collective identity and consolidate democratic participation by providing for a national curriculum, promoting a common language or disseminating a common history. Education is also helpful in overcoming the reproduction of social inequalities as it increases each individual's chance to participate in society. In addition, it can be a means of self-fulfilment and personal refinement as a part of a collective cultural enterprise (Nagel, Martens, and Windzio 2010, 15).
For our basic hypothesis, we expect national education policies to change in accordance with the OECD's orientational framework when there is, first, a misfit between national principles and ideas of education and orientational frameworks as promoted by IOs (with a good reputation); and, second, a high problem pressure as initiator for policy change. In this regard, pressures from problems open the window of opportunity for policy change, while the ideational misfit between the member state and the IO assesses the direction of change. Thus, change takes place when the internationally promoted model Á such as the conceptualisation of the PISA study Á and the nationally predominant understanding of education do not fit with each other. Such a misfit creates opportunities for political learning and redistribution of power resources (Börzel and Risse 2000, 5). Concerning our case studies, we show that the German understanding of education did not fit within the OECD orientational framework, whereas the English ideas of education and the one promoted by the PISA study were more congruent.
Germany Á educational reorganisation through PISA 'Profound' and 'tremendous' are two striking adjectives to describe the changes that took place in Germany's secondary education sector during the last decade. In a policy field that did not experience any major reform undertakings since the mid-1970s, the transformation in German education policy-making by the turn of the millennium was both astonishing and severe. In a nutshell, multifaceted changes resembled a general turn towards evidence-based policy-making and output orientation. However, the ignition spark for the new reform enthusiasm came from outside; the OECD's PISA study impressively highlighted the German weaknesses and mediocrity in educational matters compared to its peer countries and thus stimulated comprehensive reform efforts (Niemann 2010). With its PISA study, the OECD placed a massive reform pressure on Germany's education system and substantially impacted the German ideas of the purpose of education. The direction the reforms headed for was geared towards the IO's orientational framework. Therefore, the German education reforms can be characterised as a direct response to the intervention of the OECD's PISA study.
Generally, the German education system features rather reform-resistant structures. Since education policy is a core competence of the Lä nder and all 16 Lä nder have to reach unanimous decisions in almost all major education issue areas, it is not likely to enact or implement comprehensive education reforms rapidly. 4 However, recently this has not been the case. Comprehensive reforms took place and the German Lä nder acted in concordance to reform secondary education. As argued in this paper, the comprehensive reforms were achieved because the OECD created a large adaptive pressure and managed to shift the emphasis in German education ideas from a social cohesion to an economic-focused framework. In combination with the identified weaknesses of German education performances, the misfit between German traditional educational ideas and the orientational framework of the OECD created huge adaptive pressure for Germany to reform its secondary education system according to the proposed economic idea of education, which was interpreted as more promising to produce better outcomes.
The PISA study as the decisive watershed in German education policy Before PISA, Germany's education system was characterised by a general backlog of reforms, a reform-hindering political environment and an almost non-existent culture of (outside) evaluation. Structurally, German education politics is a classic example of 'joint decision-making' (Scharpf 1985). Education policy is a matter of subsidiarity in Germany; the 16 Lä nder have the ability to exercise governmental powers and to fulfil governmental responsibilities in matters of education. Hence, the authority of each Land to shape its own education system makes coherent reform approaches especially difficult (Gruber 2006, 200). 5 In German education policy, coordinating institutions between the Lä nder prevented a strong fragmentation between the 16 education sub-systems. Certainly, differences between the Lä nders' education systems exist, but they are not completely incompatible since huge differences are not tolerated under a partial pressure for uniformity (Wolf 2008, 21Á22). In the need for horizontal coordination, the Standing Conference of the Ministers of Education and Cultural Affairs (KMK) serves as the main coordination committee between the Lä nder. 6 As far as the school sector is concerned, the crucial deep-rooted feature of the German tripartite system is the division of all children at the age of about 10 between the different school types depending on their ability (Ertl and Phillips 2000, 393). Although this very principle was fundamentally criticised and contested in the past, no crucial effort has been made to modernise or even abolish it. Furthermore, the German education sector was traditionally highly input-oriented (Gruber 2006, 202), driven by bureaucratic supervision and ex-ante regulations (Herrmann 2009, 59). This basically meant that there was no prevailing tradition that controlled the outcomes of what the education system actually produced.
The dissolution of the old structures in German education policy was increasingly accelerated by the emergence of the OECD's PISA study in 2001. Overall, participation in comparative studies, other than TIMSS 7 in 1995, was quite a new phenomenon for German education policy-making (Bos and Postlethwaite 2002, 254). Without being sensitised by continuous performance evaluations PISA shook Germany's educational self-perception as a leading knowledge society to the core and caused a landslide of reforms. Essentially, PISA revealed that Germany's performance in education was, compared to other industrialised countries, mediocre at best. This was completely opposite of what was expected: Germany perceived itself as a top performer in education matters. This sophism was ultimately revealed by PISA. The first PISA report publicised in 2001 showed that, in all three tested areas of academic competence, the performance of German students at the age of 15 was significantly below the OECD average (Baumert, Stanat, and Demmrich 2001). In the following PISA studies of 2003, 2006 and 2009, the performance of German school students steadily improved (see PISA-Konsortium Deutschland 2004;Prenzel et al. 2007;Klieme et al. 2010), but it was still below the target goals. Besides the poor results in general, it also became obvious that in all four PISA studies Germany was among the OECD countries with the highest variation of performance among students. In no other industrialised country was academic success as much determined by socio-economic background or migrant status as in Germany, and the education system seemed to be unable to reduce the existing social inequality. Thus, social exclusion through the education system became another aspect of the German post-PISA discourse in which education was addressed as a social policy (Allmendinger and Leibfried 2003). Additionally, the discourse on education reform also increasingly addressed the stratified German school system as a potential cause for the large performance variations in PISA (Allmendinger et al. 2009;Schleicher 2009).
Accordingly, the public reactions were surprisingly severe. Against the background of the famous and often-cited 'PISA shock' in late 2001, a process of re-evaluating central principles in education was initiated. While in Germany education was long considered a topic discussed mainly by experts and elites, it eventually became a central concern in public discourse and was catapulted into nearly everybody's perception (Lundahl and Waldow 2009, 377). The media assessment of newspaper articles dealing with matters of PISA underscores this diagnosis: until 2010 in no other country did PISA produce as many related articles as in Germany (Martens and Niemann 2013). PISA has become a brand in Germany that serves as a publicly perceived indicator for the quality of the education system. Confronted with massive pressure caused by both the negative results and the strong public reactions, political stakeholders began to introduce comprehensive reform schemes in education.
The ideational turn: from social to economic education principles While prior to PISA the rather holistic German education idea was to educate in order to create social cohesion and enhance individual prosperity, the emphasis following PISA has been increasingly put on an economic aspect to underline the importance of educating in order to boost national and individual productivity. In the aftermath of PISA, education became a topic discussed in terms of activating human capital and making it available to the national economy and the welfare state.
The emergence of PISA can be identified as a decisive watershed in altering German ideas regarding education. This is not to say that an economic understanding of education was completely absent or irrelevant in the German education discourse prior to PISA (Allmendinger et al. 2009, 49), but it was generally overshadowed by the 'old' idea that education should serve as a means to guarantee social cohesion and to cultivate self-refinement. Currently, the German discussion about educational ideas is characterised by an increased evaluation of academic performance under the lens of economic usability.
While traditionally the idea of social cohesion (including internalising democratic values) was predominant, the ideas of competition and performance became focal points in recent years. Generally, it became more significant that education is not exclusively something to refine people's lives but is instead seen as an integral part of creating a positive economic feedback for the whole society (Interview GER11). Even the topic of social cohesion is increasingly termed as an economic factor. As a German policy maker puts it, students who are not well educated do not have a chance to participate in working life and thus do not contribute to national development and welfare (Interview GER12); consequently, severe rifts could emerge and divide German society. This does not mean that the traditional idea became obsolete. Though facets of education apart from the economic ones are still regarded as important (Interview GER07), they are no longer seen as predominant. Generally, it is widely accepted in the public discourse of different social and political actors that education serves decisively to activate human resources and enable national prosperity (Interviews GER02, GER04, GER08, GER09 and GER11). On the whole, the replacement of the old idea was a rather abrupt and rapid process. The precondition for this shift was provided foremost by the OECD. As we will show in the subsequent section, the OECD managed to promote its economic orientational framework for education on the German national level and thus enabled and facilitated substantial policy change.
All-encompassing school reforms after the PISA shock Subsequent to the fundamental reform processes in the late 1960s and early 1970s when 'education policy came from the dark' (Allmendinger and Leibfried 2003, 63) but before PISA, the German education system was characterised by a massive reform backlog. Even German re-unification in the early 1990s constituted a major reform undertaking for only the Eastern Lä nder (Wilde 2002). However, since the end of the 1990s, the field of education has been characterised by multiple changes.
The slowly spreading awareness of the need for education reform, which took off in the late 1990s (Baumert et al. 2003, 137), was finally boosted by PISA, 'after which the educational policy-making discourse and patterns of educational policy-making [. . .] were altered fundamentally' (Lundahl and Waldow 2009, 377). In the aftermath of the negative PISA results in late 2001, the responsible representatives of the Lä nder agreed within the KMK on an action plan that provided a long-range framework for substantial education reforms. The main emphasis was placed on early education in order to create a better basis for academic performance and counterweigh the socio-economic background of children (Carey 2008, 17Á21). Furthermore, the catalogue of actions included the advancement of unprivileged students (especially those with an immigrant background). It also stipulated the expansion of quality assurance, improvements in the methodological and diagnostic skills of teachers, and increased opportunities of education and advancement by the extension of all-day service offers (KMK and BMBF 2008). In order to improve education quality, the focus shifted to empirical evaluations. In 2002, the KMK agreed to the introduction of educational standards and the establishment of a central agency to monitor compliance with them (Nieke 2003, 201). Since 2009, the Lä nder have continuously reviewed the achievement of the educational standards in comparison with one another and established procedures for reviewing individual schools by external experts. Even after the fourth PISA round which was published in 2010 these developments are still in motion, and additional standards were successively enacted for further school subjects. 8 In the light of shifting of procedures in education policy-making, educational research now plays an important role for policy makers in shaping and justifying education policy (Interview GER01). In this regard, measures of 'evidence-based policy-making' have been integrated: political stakeholders increasingly incorporate expertise by consulting institutions and scientific advisors in the decision-making process in order to identify crucial problems and figure out how to deal with them effectively. Several institutions were created to develop, operationalise and review education standards. The implementation of quality assurance, education standards and evaluation mechanisms can be subsumed under the term 'empirical turn', which reflects a major twist in German education policy-making.
Taken together, the German state pulled back from detailed input steering in favour of providing a framework in secondary education. While prior to the reforms the state authority regulated curricula and made prescriptions on what had to be taught, it now prescribes what competences a student should exhibit on the basis of standardised measurements and educational standards. In this regard, the shift towards output orientation became manifest.

Promoting ideas: the OECD and the economic turn in German education
The role of the OECD in the process of restructuring the education scene in Germany was a decisive one. Besides enabling the reform movement by showing the weaknesses of the German education system and, thus, opening a window of opportunity, the OECD had a crucial influence on the understanding of the purpose of education in Germany in terms of the breadth of ideas in education. First and foremost, the OECD framed the discussion in terms of mobilising idle human resources by providing better education. The interrelation between education and economic policy became a central focal point, and PISA was able to tie it into, 'the debate about Germany's economic performance and the educational system's ability to provide education according to the economy's perceived needs' (Lundahl and Waldow 2009, 378). This interpretation was contrary to the prior non-economic understanding that was dominant in German education. As shown in the assessment of German education ideas, recent debates in Germany mostly accept the definition of education as an economic factor. In this context, the OECD was considerably successful in disseminating its economic understanding of education and opened the corridor for integrating its orientational framework into further reform measures in German secondary education.
Since the OECD generally had no impact in terms of legal governance, it notably affected German secondary education policy by influencing public debates and creating informal pressure to convert to its orientational framework (Interview GER03). The pressure for secondary education reform triggered by PISA also bypassed the usual German veto points in advance and led to concordant reform efforts across all Lä nder. As the OECD has only soft governance at its disposal, its 'strategy' to influence the German idea of education, which enables the structural reforms, took place through the back door and proceeded in several interrelated steps.
First, with the PISA study it was highlighted that the German secondary education sector was not producing satisfying outcomes; students' performances were below the average. It was acknowledged by German policy makers that the OECD linked the PISA results to predictions about the future German economic performance by showing that it does not suffice to have a well-educated elite while a huge percentage of students are left behind by the education system (Interviews GER02 and GER11). Overall, the OECD framed an understanding of education as a fundamental precondition for economic prosperity.
Second, although it does not give direct recommendations, PISA provides hints about best practices. The 'winners' in the league table with similar cultural and societal structures (e.g., Finland or Canada) serve as blueprints or role models for a successful school system. Without referring to them directly, the OECD recommends features from school models of the PISA winners worth adopting. Taken together, with PISA the OECD created an 'air of competition' regarding educational matters that were translated into overall economic welfare issues, and this competition rests on economic aspects.
Through this focal shift, German political forces favouring an economic idea of education were provided with sound arguments stemming from a wellacknowledged institution outside the country. Hence, actors expressed their educational demands based on economic reasoning. In the light of the PISA results, informal actors from the economic sector Á such as interest groups and foundations Á also intervened increasingly in issues of education. Government and employer representatives promoted better education performance as a means to secure the economic performance by hiring qualified employees (Interviews GER04 and GER09). By addressing the sensibilities of economic performance, they gained influence in the process of education policy-making and hence produced a feedback loop changing the principal ideas of German education.
In essence, the OECD not only promoted its economic orientational framework regarding education but also compelled Germany to adapt its instruments for assessing school performance. By doing so, the IO in turn also shaped how education policy-making in Germany was conducted. Furthermore, by means of PISA, the OECD supported already existing efforts by the Lä nder to reform secondary education (Interview GER04). Compiled reform schemes that did not make it onto the political agenda but were still in the pipeline received increased attention after PISA and provided a basis for the introduction of reforms.
England Á no need for a PISA study 9 While Germany discovered the need for education reforms just recently, England started over a decade earlier to modernise its education system Á without the help of the OECD. The 1988 Education Reform Act is a milestone in English education policy, changing the whole secondary education landscape and bringing about a policy shift towards marketisation. As key elements, the Act introduced a national curriculum which set minimal standards for skills and content, control over budgets was delegated to schools and a system for testing students and measuring schools' performance was established (Eurydice 2006). The main effect of the act was the close link between the provision of state education with market mechanisms: parents were able to compare schools' performances; new types of schools, such as the grant-maintained schools, provided alternative choices; and the option of publishing testing results increased the pressure on poorly performing schools. In short, power shifted from local education authorities to the central government, schools and consumers (John 1989). In the following years, a number of reforms established what could best be described by the concept of 'quasi-markets' (for an overview of the reforms see Whitty 2002). In this regard, England was an early prototype of the neoliberal education model.
England's education policy of the last 20 years can be characterised by an overall continuity. Although the education system has been constantly reformed since the 1980s, there are distinct consistencies between Conservative and New Labour education policy (Power and Whitty 1999), and these main elements have remained important within current education policy. Throughout these reform processes, international comparative assessments did not play a major role Á the English system had already gone through fundamental reforms before the PISA study. Moreover, these reforms were already very close to the OECD's later idea of education policy.

PISA as a non-issue in England
Already the 1999 assessment of the International Association for the Evaluation of Educational Achievement (IEA) showed weaknesses of the English education system: when compared with other countries, English students scored below average . The expectations to achieve good results in the first PISA study were therefore limited. With the publication of the results in 2001, however, the study suddenly gained attention: English students scored better than expected in being eighth in literacy and maths, and even fourth in science. As The Times (2001) put it, students 'have been transformed from global dunces into worldbeaters'. When it came to the second and third rounds of PISA, the average reading score dropped in 2006, and in maths the results even fell below the OECD average while they were still well ahead in science. Even though the results have not dropped significantly in 2009, the English performance stagnated at a low level compared to other participating countries. However, the interest of the public and policy actors in this international assessment decreased despite the poorer English performance (see OECD 2006).
Overall, there have not been any institutional changes in England in response to the country's scores in the PISA study. PISA did not bring about anything completely new due to a strong tradition of testing in English education. In contrast to many other European countries, the outcomes of the education system have continuously been assessed and monitored in England. For many, participation in the PISA study was just another burden: 'We do the PISA, another morning off, we do another test' (Interview ENG09).

Economic ideas of education
Generally speaking, the notion of a liberal education has predominated in England for centuries (Miller 2007, 185). Students are taught a curriculum with subject disciplines to cultivate the self. The principled idea of education was 'self-justifying, an end in itself' (Sanderson 1993, 189); a liberally educated mind, trained in an abstract discipline, could apply itself to any subject matter. Moreover, during the late nineteenth century and early twentieth century, a liberal education was associated with views on social class. These Victorian ideas of education were preserved in public schools and Civil Service and were transmitted throughout the twentieth century (Sanderson 1993, 190).
Concepts of justice and equality in the post-1945 welfare-state settlement have also infused education practice. With the 1944 Butler Act, education was made free for all students. This was an integral part of preventing Britain from returning to 'the stagnant, class-ridden depressing society of the 1930s' (Simon 1991, 35). Under the 1944 Butler Act system, schools and teachers received great autonomy, but the majority of local education authorities in England failed to exercise control over the curriculum and other policies (Chitty 2002, 263). In the 1970s concerns grew about the state of England's education system in general and the autonomy given to teachers in particular (Woods 2002, 121). In a famous speech at Ruskin College, Oxford in 1976 the then Labour Prime Minster James Callaghan responded to these concerns. He set out a number of key concepts that would resonate in the education policy over the next decades, for example, the need to question teacher autonomy and the importance of value-for-money in education expenditure (Kelly 2004, 174Á176).
Thatcher's Conservative Party returned to the Victorian ideas of education in the 1980s, emphasising the individual's freedom of choice and competition. The New Right fundamentally criticised the collectivism of the welfare state and reinvented a 'form of Victorian laissez-faire individualism' (Ball 2008, 75). In general, such an economic understanding of education is also part of New Labour's political rhetoric. Former Prime Minister Blair's education policy principally followed that of his Conservative predecessors (Power and Whitty 1999;Kelly 2009). There were, of course, some deviations from the Tories' education policy (Paterson 2003), yet the overall direction has remained the same. The English ideas of education thus come very close to the OECD's economic orientational framework of education.

Dynamics in England: emerging discourse and competition
With regard to the PISA study and other international comparisons, the government employs a 'pick-and-choose' strategy. The reception of the fourth PISA round in England shows that the executive uses positive results as a proof of good governance, while ignoring or trying to ignore failing results. In the English case, the PISA results were handled in a specific way. In England, Wales, Northern Ireland and Scotland, separate surveys have been conducted in all three rounds. In 2000, the response rate among the schools initially selected for the English survey fell below the 65% level required by the OECD. Nevertheless, the OECD accepted the data as having met the required technical standards and included Great Britain in its PISA publications and statistics (Office for National Statistics 2001).
Unlike 2001, the PISA results of 2003 (published in 2004) became a politicised issue in England. Although the response rate was only marginally lower than in the previous survey, the government opposed being included in the PISA statistics. The results were much worse than in the previous PISA study, and it was argued that the data could not be compared reliably with the performance scores for England from the study of 2000 or with any other country. The government insisted on being left out Á ironically with reference to the low response rate (The Economist 2004). As a consequence, its results from the PISA study in 2003 were not published, and England was not included in any OECD publication of this PISA round. In hindsight, the concerns about biases in the PISA results were unsubstantiated and the response rates only differed slightly from many other participating countries. Further analysis shows that the data would have been comparable to other countries since it is possible to reliably assess the magnitude and direction of response biases (Micklewright, Schnepf, and Skinner 2010).
In 2006, the response rate was sufficient for the English scores to be compared to other national-level data. In contrast to the first PISA round, the average reading scores dropped in 2006. While England's results in science were still above the OECD average, the results in mathematics fell below it. Considering that this happened within a political period in which Prime Minister Blair's three main priorities for government were 'education, education, education', it is rather remarkable that declining results in PISA 2006 generated only slightly increased media attention in England (Interviews ENG12 and ENG02; see also Grek 2008). The third survey confirms an overall decline in the English PISA performance (OECD 2006). The results of the fourth PISA study were published in December 2010. Compared to other participating countries, England's performance had not improved. In the words of Andreas Schleicher, head of the PISA group: while 'many other countries have seen quite significant improvement', the situation in England is 'stagnant at best' (The Guardian 2010).
However, in the meantime, PISA initiated a political discourse that has led to a growing recognition of other countries' education policy. The English political dialogue, which has never been very outward-looking, has begun to change 'because of the concern to be world class' (Interview ENG06). As a member of a government department reported, political actors in England have started looking at other countries that scored higher in terms of participation and performance such as, for example, Finland (Interview ENG12). To put it more drastically, 'in the war to create the world's best schools, the fight's to the Finnish' (The Times 2009). In recent years, political actors in England have become aware of the fact that they need to compare themselves to other countries (Interview ENG05). The intention of English policy actors is to understand the reasons for the success of other countries, and at best, to import them into domestic education policy: 'What are these countries doing that is different from us?' (Interview ENG11). Thus, for politicians, one reason to engage with multinational bodies is 'to find out what others are doing and learn from it' (Interview ENG03).
The awareness of the impact of international comparisons did not change overnight; rather, it has been a 'gradual progressive realization' (Interview ENG06). What has changed is the attitude of political actors towards the work of the OECD; nowadays, the studies and publications of the OECD are being taken more seriously (Interview ENG08). Political actors have noticed that PISA has become a 'brand' in England, well marketed by the OECD (Interview ENG12). Although the slow-growing pressure has not actually translated into concrete initiatives, political actors have started outlining concepts to improve England's performance in the PISA study (Interview ENG11). In the words of a department official: 'we are incorporating that into our policies' (Interview ENG12). This evidence certainly should not be overestimated since there have not yet been concrete initiatives in English secondary education due to PISA. The emerging discourse, however, can be interpreted as part of a more general debate about globalisation, knowledge economies and education policy (Ball 2008).
In sum, because of the relative congruence between the OECD's understanding of education and English secondary education, PISA did not trigger major institutional reforms. While the national ideas of education in Germany were at odds with those of the OECD, English ideas were much more congruent with the orientational framework as promoted by the OECD, particularly because of the primarily economic perspective. This minimised the impact of PISA at first. However, the political discourse regarding PISA and English education has slowly emerged. PISA revealed and named other countries, such as Finland, which have a similar understanding of education. These countries have implemented their ideas of education more efficiently than the English (Knodel and Walkenhorst 2010). In terms of the OECD study, they score better in PISA. Following former Prime Minister Gordon Brown's speech on education, the aims were clear: '. . . by 2015, we will be in the top three in science and the top five in maths out of all OECD countries. We will systematically assess our performance against other countries, reporting annually on progress' (Brown 2010). Today, Gordon Brown and the Labour party are no longer in office and there is not much time left until 2015. After the disappointing performance in PISA 2009, it is now the Coalition government that is in charge. The Conservatives' School Secretary Michael Gove emphasised the importance of the OECD's PISA study in a variety of speeches. Yet, only the publication of PISA 2012 will show if England is able to climb up the league tables.

Conclusion
Ideas about education offer explanatory value in assessing policy change in education policy-making processes. Particularly in the field of education, it is valuable to explore how the orientational framework of an IO, such as the OECD, can impact domestic policy-making in this field. Both Germany and England faced sub-optimal PISA results. Yet, comprehensive reforms only occurred in Germany, whereas England was in a business-as-usual mode. First, PISA catapulted education back onto the German policy agenda, because it revealed the shortcomings of the secondary education system compared to its peers. Thus PISA created massive reform pressure in this regard. Second, the misfit of national ideas about education compared to the orientational framework promoted by the OECD moderated the direction for the encompassing adaptations of the German secondary education system. In the emerging public education discourse after PISA the OECD sustainably disseminated the economic understanding of education within Germany, and reforms were conducted according to it. Subsumed under the terms 'evidencebased policy-making' and the 'empirical turn' the German education landscape was altered in accordance with the OECD's orientational framework focusing intensively on economic aspects of education. Even more strikingly, the changes displayed an ideational shift from a focus on social cohesion to an OECD-like understanding of education as a primary economic factor for the whole nation and the welfare state.
England, in contrast, had reformed its education system before the 1990s and established a 'quasi-market' in this field. The ideas behind these reforms matched the orientational framework of the OECD well. Although the English results were also below average, policy makers did not initiate a fundamental debate about comprehensive reforms. From the point of ideational misfit, drastic change was not to be expected; in England, education was already seen foremost as an economic enterprise. However, England's policy makers identified several minor misfits that became objects of an emerging discourse about education reform. Furthermore, through PISA English policy makers have turned towards other education systems and started to compare their performance to countries with higher PISA scores.
In this article, we have argued that including ideas in research on education policy allows understanding the IO influence on domestic policy-making. Education policy today is no longer the unique business of a nation or country; international actors as well shape and influence the ideas and goals of education policies from the outside. Further research should address the issue of how ideas matter, precisely how national actors pick up and use certain ideas. How ideas change within an actor's perception and in relation to other actors has to be analysed. How do actors use ideas as instruments to increase their influence in education policy-making? A closer look at both England and Germany would allow for an assessment of which actors succeeded in entering the arena of education policy-making.