Keywords

A Theoretical Background

Policy research has convincingly demonstrated the significance of the role of policy entrepreneurs, who bring in new ideas for how to define and solve policy problems, in policymaking processes (e.g., Verger, 2012). One of the resources that policy entrepreneurs draw upon is empirical research. However, entrepreneurs often advocate for certain ideologies and perspectives, using research in a strategic way to promote a particular solution. The identity and ideology of the knowledge entrepreneur also matter, including whether they are involved in a particular political party, a certain think tank, or a non-governmental organization or whether they work as a researcher at a university. According to McDonnell and Weatherford (2013), research-based evidence is likely to be used in combination with other evidence in line with the elected officials’ core values that evoke their positive response. For example, Kingdon (1995) noted that policymaking and agenda setting are not linear processes, starting with the problem formulation and ending with a policy decision for a solution. Instead, policymaking is characterized by parallel streams of politics and policies, with the implication that a solution can very well be preferred and suggested before a more specified policy problem is formulated.

This chapter draws on two bodies of related research from comparative and international education informing the research questions and making up the interpretative framework to examine the complexity of both layers and sequencing. The two bodies of research are the role of networks in the shift from government to governance and the study of “traveling reform” in diffusion versus reception studies.

The general shift from government to governance in Western countries is commonly seen as the result of new public management policies that most OECD countries introduced in the wake of the neoliberal reforms of the 1980s and 1990s. In this shift, policy networks play a prominent role (Lubienski, 2019). In the education sector, this shift created a new role for the state, new ways of regulating the education system, and new tools for generating, or alleviating, reform pressure. In turn, the outcomes of new public management reform projects triggered a proliferation of standardized student assessments. International comparisons (including international large-scale student assessments such as PISA, Trends in International Mathematics and Science Study [TIMSS], and The OECD Teaching and Learning Study [TALIS]) have since been referenced as key sources in policymaking for a variety of reasons. Governments have used them as primary monitoring tools to assess the quality of their teachers, schools, districts, and education systems, as well as to make policy decisions based on these assessments (Addey et al., 2017; Lingard & Sellar, 2016). The shift from government to governance has fueled a “governance by numbers” (Grek, 2008; Lindblad et al., 2018; Mølstad & Pettersson, 2019) approach, while also empowering non-state actors to participate in the new millennium as key policy actors. This has been interpreted as a clear sign of the “disarticulation and diversification of the state system,” the “destatalisation” of the policy process (Ball & Junemann, 2012, p. 24), and a changing role for the state where the state has handed over previous national influences and responsibilities to partly new arenas such as international organizations and private actors (Ozga & Lingard, 2007; Wahlström, 2014). Strikingly, the interplay between the state and local actors does not necessarily follow formal decision routes where the state delegates professional responsibility to its civil servants. Instead, at times, the state actively intervenes in the business of local practices through the use of evidence, formal standard setting, and sanctions (Wahlström & Sundberg, 2018).

We also draw on research on traveling reforms, which includes two different perspectives. Researchers who adopt a bird’s-eye view on the dissemination of international standards or global scripts in education take their point of departure from the global or the world perspective. This perspective enables an understanding of why and how some policies become scaled up and spread while others do not. Such a perspective allows for an understanding of the active role of international organizations in lending or disseminating specific policies and programs, coined as “best practices” or “international standards” (Bromley & Meyer, 2015; Krücken & Drori, 2010). Such a “perspective from above,” however, provides only one of several possible angles for understanding the spread of global education policy. By contrast, a “perspective from below” illuminates why, how, and when national or local policy actors selectively borrow global education policies (Anderson-Levitt, 2003; Sivesind & Wahlström, 2016; Steiner-Khamsi, 2012; Steiner-Khamsi et al., 2020; Wahlström, 2020). In these studies, the researchers examine which elements of international policy discourses will be of interest to national policy actors to interpret and reconstruct for the legitimization of national/local policy needs. The researchers examine the way global policies are used for legitimate and/or delegitimate purposes when recontextualized into national policy agendas.

In this study, the perspective is “from below.” From a national level, we explore how the Swedish government makes use of various knowledge sources of evidence for legitimating its reform recommendations. We are interested in the various international, regional, and national policy and knowledge actors that the government turns to for providing evidence for its recommendations for decisions.

Purpose and Research Questions

From this point of departure, the aim of this study is to advance the understanding of the mechanisms at stake in legitimating national school reforms. More precisely, we examine the evidence used by the Swedish government in the 2015/2018 Knowledge Achievement Reform as manifested in the reference lists of the green papers (GPs) and white papers (WPs) analyzed. For this purpose, the following research questions are relevant:

  • Whose knowledge is used to legitimate the 2015/2018 Knowledge Achievement Reform?

  • What types of knowledge sources are used to legitimate the 2015/2018 Knowledge Achievement Reform?

  • How much expert knowledge is translated upward into political knowledge?

Contextualizing the Swedish Decision-Making Process at the State Level

Over a long period of time, Sweden has appointed commissions as a way for governments to produce evidence for political decision-making and for the “anchoring” of policy recommendations in different segments of Swedish society. The Swedish system stands out because of its strong emphasis on the preparatory stages. Commissions of inquiry have two distinguishing features. First, they are appointed by the Cabinet and have a legal status similar to other government bodies. Second, commissions of inquiry are set up with a mandate to investigate a special thematic area or particular subject (Petersson, 2016). This institutional arrangement is not inscribed in Swedish law but has evolved as a praxis over a long period of time. Almost all pieces of legislation are prepared in this way, which means that around 200–300 governmental commissions are at work at any given time in the political system. By international standards, Swedish ministries are quite small; therefore, appointing commissions may also be a way to temporarily expand the staff of the ministry outside the ordinary budget. Once constituted, the commission has a considerable degree of autonomy. Although efforts have been made to limit both the size of the commissions and the time period for their work, it is still not unusual for their work to extend over several years (Trägårdh, 2007). Their work is finally presented and handed over to the responsible minister in the format of government official reports (green papers, known by the Swedish acronym SOU). Appointing commissions of inquiry has to do with the political relationship between the state and civil society. These commissions have served as a “linchpin in a system of democratic governance whose hallmark is deliberative political practices that involve a mix of civil servants, politicians, academics, experts, and representatives of relevant civil society organizations” (Trägårdh, 2007). As Trägårdh (2007) noted, this extended communicative law-making process also contributes to societal legitimacy and acceptance for a new law or policy and adds trust to the system.

Today, the timespan of the policymaking process has become shorter compared to the praxis during the mid-twentieth century. In the 2010s, the commissions of inquiry were more tightly controlled by the Cabinet and government ministries, while parliamentary representation declined. While the level of striving for consensus in the inquiry process has decreased, the level of conflicts has increased and the politics in parliament have become more polarized (Petersson, 2016). Moreover, in recent years, Swedes have demonstrated a decline in their trust of the political system as a whole, with a widened gap between the voters and the elected as a result (Amnå, 2006). In addition, Gunnarsson et al. (1998) have shown a shift in the way commissions were composed between 1988 and 1997, with a decrease in the number of larger commissions with a chairman in favor of one-person commissions with additional members and experts linked to the investigation. This development has continued. In a study analyzing the period 1990–2016, Dahlström et al. (2019) showed that 90% of the ongoing commissions of inquiry in 2016 were in the form of one-person commissions. At the same time, barely 3% consisted of parliamentary committees of inquiry. Dahlström et al. (2019) concluded that the kind of committees that once constituted the backbone of the committee system (i.e., broad parliamentary committees) have almost disappeared today. We understand this development as an expression of trying to rationalize and speed up the process of national policymaking in the context of increased global competition (see Kamens, 2013; Lewis & Hogan, 2019).

A History of Borrowing

Swedish school reforms have a well-established history of borrowing policy from other educational systems, particularly for policy domains related to curriculum and governing of schools (Hallsén & Nordin, 2018). The Governmental School Commission of 1946 talked about the Swedish system of whole-class teaching as being influenced by “German school life” (SOU 1948:27, p. 112). The same report also revealed that members of the commission had taken study visits to Denmark, Norway, Belgium, France, Switzerland, the United Kingdom, and the United States. During the period leading up to the realization of a nine-year comprehensive school for all children in 1962, Sweden also became an influential policy lender, with Stockholm as the leading marketizer of the new Swedish school system that attracted visitors from all over the world (Hallsén & Nordin, 2018). Educational reforms were thoroughly planned in terms of time and resources, and school experiments and investigations were often led by researchers and described as an expression of social engineering (Román et al., 2015). Similarly, in the SOU that preceded the curriculum reform of 1980, The School’s Inner Working (SOU 1974:53), an entire chapter is devoted to international examples of ways to improve a school’s work environment, with lessons drawn from England, Eastern Europe, and Denmark. The report School for Bildung (SOU 1992:94) featured a shift away from the country-specific references with talk in terms of a wider ongoing “internationalization,” of which Sweden is part, and references to specific countries as part of organizations such as the United Nations (UN) or the European Community. From 1995 on, following Sweden’s membership in the European Union (EU), education policy in the country has been elevated to a priority on the reform agenda. Even though Swedish policymakers have refrained from making explicit references to EU policy, developments in other EU countries have had a visible impact on school reform in Sweden, which has led scholars to talk about the Swedish strategy as a “silent policy borrowing” from the wider European education policy space (see Nordin, 2012; Waldow, 2009). In recent years, the OECD has emerged as an increasingly important source of expertise and authorization for the Swedish government when making reforms. The OECD today functions as an authoritative reference system in justifying borrowing of international standards created in PISA or national OECD expert reports (Grek, 2019; Ringarp & Waldow, 2016; Wahlström, 2018; Waldow, 2012).

The 2015/2018 Knowledge Achievement Reform

In 2006, the Swedish government decided to appoint a one-person commission led by a civil servant to investigate the implementation of the new goal-oriented curriculum launched in 1994. In the report that was published in 2007 (SOU 2007:28), the investigator concluded that the new public management of the Swedish school system along managerial principles and decentralized power had not worked out as hoped for. This finding led to the launch of a whole range of fundamental reforms in 2011, including a new school law, a new national teacher education, a new curriculum for the comprehensive school, and a new, more differentiated grading system, all to enhance monitoring and central control in a decentralized school system. In addition, Sweden at the time suffered from growing concern over continuously declining PISA results. When Sweden received its lowest PISA results ever in 2012, the national school crisis was a fact (Wahlström, 2018) and fueled the national crisis discourse and an emerging “scandalization” (Steiner-Khamsi, 2003) of the comprehensive school within politics and the media (Nordin, 2012, 2019). Against this backdrop, the Swedish government decided to take the unconventional measure to turn to the OECD for help (Pettersson et al., 2017). After a period of analyzing and visiting Sweden, the OECD published a final report in 2015. The OECD stated that more efforts were required to fulfill Sweden’s commitment to excellence and equity (OECD, 2015) if the country wanted to come to terms with its national problems. In the report, the OECD suggested a focus on the following three areas: (a) promoting quality with equity across Swedish schools, (b) building a high-quality teaching profession, and (c) steering policy and accountability focused on improvement. To address the OECD report and further refine the reforms of 2011, in April 2015, the Swedish government appointed a government commission called the 2015 School Commission, chaired by the Director-General for the Swedish National Agency for Education. During the last six months of the commission’s work, a researcher took over the chair position because the former chair was appointed Minister of Education. The 2015 School Commission was thus close to the government.

The commission presented its suggestions in SOU 2017:35 entitled A Gathering for School—A National Strategy for Knowledge and Equivalence (Source Document 2). In the white paper 2017/18:182 (Source Document 1) that paved the way for the 2015/2018 Knowledge Achievement Reform, the SOU plays an important role and is explicitly mentioned together with SOU 2016:94 (Source Document 7) and SOU 2017:51 (Source Document 8) as central to the law-making recommendations formulated in the white paper. The OECD thus played an important role as policy adviser (Lingard & Sellar, 2016) in setting the agenda for the Swedish reform, operating as a “boundary organisation” (Grek, 2019) that pointed out what policy areas to prioritize and indirectly what expert knowledge would be of most use.

Methodological Considerations

We have pursued the research questions by analyzing references made in published policy documents. The bibliometric network analysis is focused on “official knowledge” (see Baek et al., 2017) or texts produced by the Ministry of Education and Research (white papers) and by its appointed commissions (green papers). Since the white paper had no reference list, we began by searching through the entire volume, resulting in a total of 12 cited green papers of which one lacked a reference list. Of the 12 documents, two documents (SOU 2016:38, an interim document, and SOU 2016:38, the same document as the selected green paper [GP] but published two days earlier) were duplicates and therefore removed. Another two documents (SOU 2016:77 and SOU 2017:49) were not directly related to compulsory education. Thus, the delimitation was made due to (a) duplicate documents and (b) documents not related to compulsory education. We then ended up with eight green papers that served as source documents of which one lacked a reference list. In total, we have nine (eight green papers and one white paper) source documents for which we have first entered and analyzed a total of 1615 references. In the second step, we adjusted the number of references so that references cited by multiple sources were counted only once, resulting in a total of 1421 references. Table 8.1 illustrates the source documents and the distribution of references.

Table 8.1 Distribution of references in the policy documents of the 2015/2018 Knowledge Achievement Reform
Full size table

In addition to the quantitative analyses of citation frequency, we have classified the reference distribution according to location and type of document. The location category distinguishes between domestic, regional, and international references, where regional refers to the Nordic context. The type of document distinguishes between the following subcategories: reports, books, academic publications, governmental publications, and others.

Each reference has been given a unique identification number, creating a matrix for calculating in-degree centrality and co-citations in order to understand the reference network of the 2015/2018 Knowledge Achievement Reform. We used the software programs UCINET 6.627 and NetDraw for analysis and visualization of the data (Borgatti et al., 2002).Footnote 1

The Bibliometric Network Analysis: Findings

The bibliometric network analysis yields several interesting patterns related to the three research questions presented in the introduction of this chapter.

Reference Patterns in the Swedish School Reform

Bearing in mind the central role of the OECD in the launch of the 2015/2018 Knowledge Achievement Reform, it is interesting to see that as much as 80% of the references are domestic, mostly referring to various types of laws and regulations on how to govern the Swedish school system at large. Of the remaining references, 19% are international, and only 1% are references to other countries in the Nordic region. As shown in Table 8.2, Sweden thus distinguishes itself both in terms of the amount of domestic references and in the lack of references to the other Nordic countries.

Table 8.2 Reference distribution
Full size table

Sweden is an active member of the UN, so the UN Convention on the Rights of the Child has played an important role in Swedish policymaking ever since the parliament ratified it without any reservations in 1990. This importance is evident in the results where the UN Convention on the Rights of the Child is the most cited international document containing regulations. Recently, the parliament has decided to make the UN Convention on the Rights of the Child Swedish law from 2020, increasing its significance even more. Looking at the international references in total, the OECD stands out as the most frequently cited. The OECD is the fifth most cited publisher after the Swedish government, the National Agency of Education, the parliament, and the school inspectorate, in that order.

Another distinct feature of the reform is the limited use of academic publications; in fact, the white paper and three of the green papers do not contain a single academic reference. References to scientific literature are made in the format of synthesized knowledge. Hattie’s publication Visible Learning stands out as the most cited document of all, together with Swedish school law. Another research review that has had an impact is the book Excellent Teaching by Håkansson and Sundberg. Table 8.3 shows the distribution of the most cited documents in the Swedish reform.

Table 8.3 Most cited documents
Full size table

Looking at the level of authors, it is interesting to see that the most cited researcher is Professor Jan-Eric Gustafsson, who became the chair of the 2015 Swedish School Commission after the Director-General for the Swedish National Agency for Education (Skolverket) Anna Ekström, who was the original chair, was appointed Minister of Education and Research. Professor Jan-Eric Gustafsson has a total of 11 citations, followed by Linda Darling-Hammond with seven citations and John Hattie with six citations. The commission led by Professor Gustafsson wrote the green paper SOU 2017:35 Gathering for School—National Strategy for Knowledge and Equality: Final Report from the 2015 School Commission (Source Document 2), which was central to the 2015/2018 Knowledge Achievement Reform. Even though an entire commission is responsible for writing a green paper, this result shows the importance of the chair of a commission (together with its secretariat) in setting the direction and, in this case, filtering the kinds of references to be included (see Christensen & Holst, 2017). In the Swedish reform, where only 9% of the references were academic, even a small number of references makes a significant difference in the percentage of the total number of references in that category.

The Structure of the Evidence Base

The network design of the study also enables an examination of the relations between the expert commissions included in this study in terms of shared knowledge, reflected as explicit references in the source documents. As Fig. 8.1 shows, the overall picture is that the various commissions have produced highly specialized knowledge in their reports.

Fig. 8.1
A cluster of 8 network groups. Some are labeled the gathering for school, visible learning, and the U N Convention on the Rights of the Child.

Complete network structure. (Note: Regional = gray; domestic = white; international = black; source = circle; node size = in-degree centrality)

Full size image

The references in Fig. 8.1 are colored based on their domain, where domestic references are white, regional references are gray, and international references are black. The size of the nodes indicates the document’s in-degree centrality, that is, a document’s centrality in the network based on how many texts refer to that specific document.

The colors illustrate the dominance of domestic references and the lack of Nordic ones in the Swedish reform. The white paper (WP) (Source Document 1) is marked as number 1 in the middle of the network structure. The largest field, located above the central point constituted by the WP, is GP 2, Gathering for School (Source Document 2), which represents the commission of inquiry most directly responsible for preparing a succeeding recommendation for a reform and also bears the same name in Swedish as the WP. Among the shared references, most are domestic governmental references, in line with the overall picture. However, two international references stand out as cited by many. The most cited one is Hattie’s book Visible Learning, which is cited six times and marked as a black square with the number 1287. Its size also signals importance in terms of shared knowledge, as many cite this reference. The other international reference that stands out is the UN Convention on the Rights of the Child, marked as a black square with the number 1275 placed on the left side of Hattie in the network. The references to the UN Convention on the Rights of the Child are part of a larger discourse in Swedish school governing documents, where the UN Convention on the Rights of the Child has played an important role for a long time. Since January 2020, it has also become part of national law.

Also characteristic of the Swedish reform is the limited use of academic references. Out of eight source documents, four do not have a single reference to academic research. As previously mentioned, when referring to research, it is predominantly in the format of synthesized knowledge. In the books authored by Hattie and Håkansson and Sundberg, research knowledge is collected, interpreted, and presented in a supposedly more manageable and accessible format, making them “intermediaries” (Lubienski, 2019) between science and politics in providing policymakers with seemingly “useful” evidence.

Removing the white paper from the network structure shows the important role of GP 2 Gathering for School (Source Document 2). The field representing this GP is located in the middle of Fig. 8.2.

Fig. 8.2
A cluster of 8 network groups. The second cluster in the center is labeled the gathering for school.

Complete network structure without the white paper

Full size image

GP 2 has mutual references primarily with Source Documents 6, 7, and 8. Characteristic for these three GPs is the fact that they were all produced after the 2015 OECD report with themes in line with the OECD’s recommendations for improving quality and knowledge results in Swedish schools. However, the distribution of references does not reflect this thematic relationship. While GP 6 (Source Document 6) on the action guarantee for reading, writing, and mathematics has 32% international references, GP 7 (Source Document 7) on students’ absences has 81% domestic references and 18% international references. Likewise, GP 8 (Source Document 8) on the adaptation of rules for teacher certification has 92% domestic references, 8% international, and not a single reference to other Nordic countries. This result adds to the understanding of commissions operating first and foremost on a “pragmatic mandate” generating “highly specialized knowledge on the specific topic for which they are tasked” (Steiner-Khamsi et al., 2020, p. 128), which can result in radically different reference structures in SOUs, even when related as parts of a wider shared discursive context.

Upward Translation in the Swedish School Reform

Turning to the process of translating expert knowledge upward, the result shows a selective use of expert knowledge at the political level.

Figure 8.3 demonstrates that the white paper (Source Document 1) by the Swedish Ministry of Education and Research shares about 44% of the knowledge produced by the experts in the commissions of inquiry that it has appointed, while about 56% is unique to the white paper. As such, more than half of the number of references is lost in the political process of “upward translation” (Steiner-Khamsi et al., 2020). Additionally, this finding indicates that the Ministry of Education and Research looked for other sources of expert knowledge apart from the green papers to incorporate when writing the white paper (Source Document 1).

Fig. 8.3
An illustration has the components of political knowledge and expert knowledge. The expert knowledge components point to the political knowledge component under the white paper.

The political translation of scientific knowledge in the 2015/2018 Knowledge Achievement Reform. (Note: GP = Green paper)

Full size image

Not all green papers are equally central to a white paper. As signaled in white paper 2017/18:182 (Source Document 1), SOU 2017:35 (Source Document 2), SOU 2016:94 (Source Document 7), and SOU 2017:51 (Source Document 8) are of special importance to the Ministry of Education and Research and therefore hold a specific mandate visible in the number of shared references with the white paper. In this case, the process of upward translation thus consisted of two steps of selection, first by announcing that certain green papers would be given special importance and second by being selective in the use of expert knowledge provided by the green papers used.

Looking into the content of the three green papers mostly drawn upon by the Ministry of Education and Research, SOU 2017:35 (Source Document 2) and SOU 2017:51 (Source Document 8) both provide a wide array of expert knowledge suitable for an incremental reform that builds on and revises previous reforms. It is perhaps more noteworthy that the government also uses quite a lot of expert knowledge from SOU 2016:94 on student absenteeism (Source Document 7). In addition to those three green papers mentioned as especially central to the white paper, the Ministry of Education and Research also drew on expert knowledge from the green paper SOU 2016:59 On Good Grounds—An Action Guarantee for Reading, Writing and Math (Source Document 6). Launched as a result of the Swedish decline in PISA, this one-person commission led by an associate professor in education from Stockholm University focuses on developing methods for early interventions to improve children’s basic abilities in reading, writing, and math. The report uses Finland as a reference society and discusses how the early interventions that the commission found to be characteristic of Finland might explain its PISA success. Members of the commission also made study visits to Finland to learn about the Finnish system.

A Final Note on the Commissions in This Study

A brief look into the commissions that have produced the SOUs included in this study as source documents adds to the conclusion drawn by Dahlström et al. (2019) that there has been a significant shift in the Swedish committee system, away from broader parliamentary committees toward one-person commissions with groups of advisory experts. Of the eight committees in this study, six were one-person commissions. Although based on a limited number of cases, the result coincides with the conclusion drawn by Christensen and Holst (2017) that civil servants have been replaced by academics as chairpersons for government-appointed commissions. Out of the eight commissions in this study, five of them were led by academics employed at universities or persons holding a doctoral degree when handing over their reports to the Minister of Education and Research. The most influential commission for the 2015/2018 Knowledge Achievement Reform was the group behind GP 2 (Source Document 2), first led by civil servant Anna Ekström as chair before she was replaced by Professor Jan-Eric Gustafsson when she was appointed Minister of Education and Research.

Conclusions

In this conclusion, we will summarize the main findings of the bibliometric network analyses of the Swedish school reform in relation to the three research questions guiding the analyses. The first question concerned whose knowledge was used to legitimate the reform. Here, three characteristics stand out as most significant for the Swedish reform. The first is the high percentage of domestic, mainly governmental references. This finding shows the possibility for national politics to uphold a high level of self-referentiality even when the national political agenda to a large extent is dictated by international organizations such as the OECD. The bureaucratic machinery seems to act on its own institutional logic, relatively independent from external influence in the Swedish reform. The use of externalization as a way to legitimate the reform differs significantly between the different green papers and seems to relate to the topic. Some topics seem to be perceived as more national than others, which is illustrated through the structure of the reference list. In addition, the chair of a specific commission also has an important role in determining what knowledge is used. In the Swedish reform, it is interesting to notice that the chair of the most central green paper (GP 2) also stands out in the material as the most cited academic researcher. As researchers are increasingly replacing civil servants as chairs of government-appointed commissions in both Norway and Sweden (Christensen & Holst, 2017; Dahlström et al., 2019), what Steiner-Khamsi et al. (2020) described as an “expertisation of commissions” seems to create a new quasi-scientific policy space between science and politics for researchers to marketize their own research. The results of this study thus harmonize with previous research showing that context matters for what works in the local arena (e.g., Anderson-Levitt, 2003; Steiner-Khamsi, 2013). As shown by the results of this study, context itself has to be understood as a stratified concept, where aspects affecting what evidence is of the most worth and available for politicians to use in legitimating educational reforms are decided by many different interconnected, sometimes contradictory, factors. These factors range from how the bureaucratic system functions to which people occupy central positions, such as being chair in a commission.

The second question addressed the issue of what types of knowledge sources were used in legitimating the Swedish school reform. As hinted above, almost 60% of the references were governmental, which is almost twice as much as the other Nordic countries. Another distinct feature is the low number of academic references, similar to Finland but much less than Norway and Denmark. Of the eight source documents, four do not have any references to academic research, which is remarkably low in a context emphasizing the importance of evidence-based policymaking. When referring to academic research, both the most important references are systematic reviews presenting synthesized knowledge. As the production of scientific knowledge increases, the role of intermediaries becomes equally important. These intermediaries are people, networks, or organizations whose main task is to collect, select, interpret, and present academic knowledge for political and practical use. During the same period as the evolution of the 2015/2018 Knowledge Achievement Reform, there was a broad political consensus in Sweden on the importance of investing in useful knowledge. In 2015, the Swedish government even launched a new governmental institution called the Swedish Institute for Educational Research with the explicit directive to systematically collect and distribute research contributing to more evidence-based and effective teaching in Swedish schools to improve performance on national and international assessments. In a time of an overabundance of academic knowledge, the role of people and organizations operating as intermediaries thus becomes increasingly important as they provide academic research in a simplified and accessible way (Lubienski, 2019). However, when politicians start to promote such research financially, there is a long-term risk that academic research could lose some of its critical stance in order to be perceived as relevant to policy and practice.

The third and final question addressed the issue of what happens when expert knowledge is translated upward into political knowledge. The results show that, although the white paper drew on expert knowledge produced by the green papers, more than half of the references in the white paper were exclusive to that document. Expert knowledge is thus used very selectively at the political level, and politicians clearly import references of their own as well. Commissions offer a variety of evidence from which the Ministry of Education and Research can choose to include in the white paper. While these results confirm the idea of an overproduction of evidence (Lubienski, 2019; Steiner-Khamsi et al., 2020) where much expert knowledge is lost in the upward translation, it is reasonable to think that some expert knowledge is imported in a processed form as in-text references and therefore does not turn up in the reference list.

In summary, the results of this study show the importance of understanding context as a complex and dynamic concept where contextual mechanisms at different levels affect what knowledge is used in legitimating national educational reforms.