The critical challenge: Policy networks and market models for education

Abstract

This paper reviews a number of approaches to considering how policy transfers through advocacy networks, focusing on education issues in general, and market-based policies in particular. While policymakers and private funders are demanding evidence on the effectiveness of proposed interventions in education, it is not at all clear that they themselves consider evidence in promoting particular policies. Instead, as is apparent with policy proposals for market-based reforms, quite often it is not rigorous research that advances policies, but effective advocacy. Indeed, evidence indicates that an infrastructure of rapid production and dissemination of data has emerged through advocacy organizations and networks, often to obscure or produce alternative evidence. While critical theorists have noted for some time the expansion of neoliberal reform models in education, those analyses offer little in terms of understanding the mechanisms behind the expansion and proliferation of such policies, much less effective ways to challenge their growth. The paper highlights limitations of one of the most popular theoretical perspectives for understanding such networks, and notes how, in the marketplace of ideas represented by new policy advocacy networks, critical scholars are poor at packaging their message, especially compared to intermediary organizations. The paper concludes by introducing a framework of economic transaction for understanding policy transfer. In doing so, it offers a challenge for critical scholars hoping to influence education policymaking.

In much of the world, policymakers, philanthropists, and experts are demanding evidence on the effectiveness of proposed approaches for addressing issues, often as an indicator of the suitability of different interventions for receiving funding and support. For instance, in the United States, the federal No Child Left Behind Act of 2001 famously uses the term “scientifically based” over one hundred times, referring to research and practices that have an empirical basis and would be favored under the legislation. Similarly, the “effective philanthropy” movement promotes business-style performativity measures of impact and cost-effectiveness to evaluate and endorse humanitarian efforts across the globe. But in education policy in particular, there are serious questions not only about the degree to which policies are actually evidence based, but also how evidence is produced, whether it is useful, how policymakers access or use evidence on policy proposals, and how new forms of advocacy networks convey ideas across time and space, and perhaps—in doing so—re-shape those ideas.

As a case in point, education policies emerging from neoliberal economic models have proliferated rather rapidly around the world in recent decades. Interestingly enough, this expansion has occurred despite a paucity of evidence on their effectiveness in addressing their education goals. Instead, proponents often refer to the effectiveness of such approaches outside of education, and then—reflecting the general global trend of market penetration into traditionally non-market sectors—have argued that such market logic should be extended into state systems of mass education. Supporting this logic, an infrastructure of rapid production and dissemination of data has emerged through advocacy organizations and networks, research outfits, policy entrepreneurs, bloggers and other internet-based thought-shapers. These factors are both responsible for, and benefit from, removing traditional barriers to the rapid dissemination of information, but at the same time create a new information landscape that can both democratize and undercut access to quality information (Lubienski et al., 2014). Yet it is not at all clear that the information advanced though this vast infrastructure is even utilized in policymaking (DeBray et al., 2014; Jabbar et al., 2015; Lubienski et al., 2015; Scott et al., 2015).

Thus, even as policymakers pay lip service—if not actual attention—to the need for evidence-based decision-making, the sources for information have become more diverse, and the information itself more diffuse and disputed. And governments’ capacity to collect and weigh evidence that could illuminate policy proposals is often being diminished and contracted out to non-state actors. This is happening at the same time that advanced methods of knowledge production have become more nuanced, sophisticated, and precise, albeit also arcane and thus necessarily opaque. Into the chasm between research production and policymaking, we are seeing the entrance of new actors—intermediaries—that seek to collect, interpret, package, and promote evidence for policymakers to use in forming their decisions (Lubienski et al., 2011).

Critical theorists have noted for some time the expansion of neoliberal reform models in education (e.g. Buras et al., 2010; Klaf and Kwan, 2010; Lipman, 2011). But these analyses are often primarily descriptive, and thus have less to offer in terms of understanding the mechanisms behind expansion and proliferation of such policies, much less effective ways to challenge their growth. Other strands of policy research have considered the movement of policy ideas in a number of ways that help explain the patterns of knowledge production, transfer and use. But some are less helpful for understanding the emerging landscape of new policy networks, and few truly explain the drivers for information across those networks. In particular, the advocacy networks around market-oriented models of education policy represent a newer phenomenon with attributes that require new methodological approaches, and new perspectives for conceptualizing the relationships, pathways, motivations, and behaviors of participants in those networks (DeBray et al., 2014; Gulson et al., 2017).

In this article, I briefly review a number of approaches to considering policy transfer, focusing on education issues in general, and market-based policies in particular. In the next section, I outline the concept of advocacy networks, and highlight the emerging role of intermediaries within those networks. Then after considering some of the current approaches to understanding how policy ideas transfer across nodes, actors, and contexts, I describe an on-going, multi-site study that examines this issue through a mixed-methods investigation of actors working in policy networks. In reporting some of the Phase 1 findings from the study, I note a few of the limitations of one of the most popular theoretical perspectives for understanding such networks. The concluding discussion introduces some theoretical considerations for analyzing policy transfer through a lens of economic transaction, and offers a challenge—and the beginning of a roadmap—for critical researchers hoping to influence education policies.

The question of policy proliferation

The era of public policymaking that followed the decline of the Soviet Union marked a notable period of standardization of social policy in many places around the globe, as formerly centralized economies and social systems moved rapidly toward Western liberal ideals and neoliberal public policy models. The 1990s saw a global movement toward liberalization, privatization of state-dominated sectors, and encouragement of private ownership and enterprise not only in the former Soviet Bloc, but in South Asia, Africa, and Latin America, not to mention moves even further in this direction in established market systems such as the United States. Even before the fall of the Berlin Wall in 1989, liberal Western nations with lengthy traditions of state welfare policies, such as in New Zealand, England, and Sweden, had begun to move toward more decentralized, market-style arrangements for social services such as education. As a more recent continuation of this trend, we are seeing the spread of state-funded autonomous schools in many nations: for instance, Free Schools and Academies in England, Partnership Schools in New Zealand, and charter schools in Colombia, the US and Canada.

While critical researchers have observed and opposed such policies, they have typically been quite poorly situated for addressing a number of crucial questions that could enhance both the understanding and effectiveness of opposition to such policies. For instance, why do some such ideas, models, and practices multiply when they do, while others do not? Is it that evidence of their effectiveness recommends them across contexts? Is it that governing bodies (often supra-national ones) or institutional cultures promote or encourage the adoption of particular practices? Do policy entrepreneurs spread the gospel of a particular approach across contexts? Moreover, how do such ideas spread, in terms of routes, mechanisms, and drivers?

Other scholars have long noted the tendency of some policies—some more than others—to be reproduced, emulated, adapted, or co-opted, substantially or symbolically. Sometimes, of course, this happens through coercive measures, such as imperialism or the imposition of structural adjustment policies. But scholars have been more recently drawn to how policy ideas spread by more nuanced means.

For instance, many observers (and some policymakers) see idea diffusion as simply the natural proliferation of innovations or effective approaches that have generated a track record of proven effectiveness (Steiner-Khamsi, 2012). While there is a compelling internal logic to this view, analyses drawing on this perspective tend to slight the role of power, influence, and concerted efforts to promote ideas and agendas (Cresswell and Merriam, 2011). Some of the education policy literature highlights the “viral” spread of ideas, as a standardizing force, but also one whose effects can be localized (Anderson-Levitt, 2003).

World Culture Theory, drawing from neoinstitutionalist sociology, is more explanatory, positing that different systems are isomorphic to increasingly standardized models (Meyer and Rowan, 1992; Ramirez, 2012). According to this perspective, this is partly due to a liberal notion of progress, but also due to a yearning for legitimacy, as countries emulate practices, policies, and approaches from what are seen as leading models. But such more unitary perspectives have tended to take a satellite view of the issue, and fail to consider contextual issues driving changes within specific systems (Verger, 2014). Other approaches in the mobility literature have looked largely at the created spaces and networked pathways through which policy ideas move, and are sometimes refashioned.

Overall, much of this literature is useful for descriptive analyses of the patterns of proliferation of policy ideas. But with some notable exceptions, these perspectives do not always account for the complexity and agency involved in the actual transmission of those ideas. For the research described below, we drew on the Advocacy Coalition Framework (ACF) first outlined by political scientist Paul Sabatier and colleagues (Sabatier and Jenkins-Smith, 1999). The framework has particular usefulness when—as is the case with incentivist policies—there is apparent uncertainty around factual issues, and policy objectives are heavily contested by myriad actors in policy arenas. Such actors will find allies and build network relationships with other actors with whom they share core values, although they may differ on secondary policy and strategy issues; for instance, incentivist networks may form around school choice, but actors may disagree over emphasizing choice within or beyond the state sector.

The ACF has an advantage in that it goes beyond the traditional “iron-triangle” conception of actors (interest groups, legislators, and the bureaucracy) and incorporates not only multiple levels of government, but also non-state actors such as think tanks, university-based researchers, the media, and advocacy groups, into the policy equation. Thus, it conceptualizes networks as incorporating a range of actors working in a more or less concerted fashion, covering multiple functions, such as idea creation, political lobbying, and shaping public perceptions of an issue. Focusing on policy subsystems, or issue-based networks, ACF draws attention to the use of different strategies and instruments, which could include public opinion, litigation, or legislation, for instance, or demonization of opposing coalitions, over time (Sabatieret al., 1987). ACF is particularly useful for considering how advocacy of (and opposition to) policy ideas shapes the issue, question, and policy solution, and presents many advantages when considering an issue such as different strategies in the advocacy around incentivism.

Studying advocacy networks

The question of how policy ideas move and change is becoming more relevant in an age, and in places, where sophisticated, arcane and specialized research evidence is more accessible, but no more better understood, to broader audiences; where the policy process is less transparent and more open to interest groups, and where traditional forms of accountability between policymakers and their broader constituencies are diminished. Such conditions have become more prevalent in a number of places, particularly where economic inequality and the susceptibility of political systems to private resources has privileged the influence of private actors in public policymaking. In the wealthier nations, the United States has become one of the leading sites in this trend, due in a large part to policies that allow for the accumulation of massive wealth by a few people and philanthropies, and a political system that encourages private funding of individuals and ideas (Gilens and Page, 2014).

The political-institutional context in the US makes it a useful site for re-evaluating not only the emerging landscape around policymaking, but also for examining how evidence shapes and is shaped by that landscape. In particular, one of the inherent features of these landscapes is the increasing presence of policy advocacy networks, populated largely by intermediary organizations (IOs) that seek to assemble, interpret, and advance information for policymakers to utilize in the policymaking process. For our purposes, we define IOs as organizations that serve, at least in part, a function of connecting research producers with users, often by way of selecting, interpreting, packaging and promoting particular research, and can include players such as think tanks, new and old media, philanthropies, community organizations, bloggers, lobbyists, policy entrepreneurs, and other advocacy organizations. Importantly, this is happening in a context where research evidence on education policy (especially relative to other sectors) is frequently discounted for a number of reasons, including that education research is often extremely sophisticated or theoretical; that the field of education is inherently ideological (Aristotle, 1946), and less susceptible to empirical evidence; and that many people went to school and feel that experience gives them the necessary expertise and common sense to promote policy preferences. Consequently, to meet the demand—whether it be substantive or symbolic—for research to be “used” (however that may be defined) in policymaking, intermediaries have rapidly populated the education policy landscape in the US.

Since 2011, the Research on Intermediary Organizations (RIO) in Education Policy project has been examining these issues by investigating policy advocacy networks in and across several major American metropolitan areas (Lubienski et al., 2011). Drawing on an ACF, investigators have focused on the role of IOs as they network with policymakers, researchers, and with each other in packaging and promoting research evidence in and between New York City (in Phases 1 and 2), New Orleans (Phase 1), Denver (Phase 1), Los Angeles (Phase 2), as well as from Washington, DC (Phase 2). These cities were chosen because they have been leading sites for different versions of “incentivist” policies—that is, generally market-oriented reforms such as charter schools or merit pay for teachers that seek to align policy goals and institutional environments in order to encourage individuals and organizations toward particular desired behaviors. Researchers have so far conducted over 200 interviews with policymakers, researchers, and especially with actors in networks of IOs on their use and sources of—or audience for—evidence on incentivist policies, and their relationships with other actors and organizations in or outside the networks.

While this is an ongoing study now a year into its second phase, the longevity of the project at this point offers some useful insights into how advocacy networks are constituted and intersect, how they operate and, in doing so, shape research evidence and its use, and how they impact the political economy of evidence production and use. ¹ Some of the findings pertinent to present discussion include:

Policymakers either report little evidence of using research, or display diluted conceptions of research “evidence,” at least in this area of education policy.

The pace of a policy’s movement or expansion is not linked to evidence of its efficacy.

Sources cited tend to be based on relationships, reputation, and access.

Our approach of conceptualizing IOs as discrete actors is misguided. We conceived of IOs as actors operating in the space between researchers and policymakers. While this is true, the diversity of organizational forms in this space, acting in this regard, is not limited to discrete actors. Many organizations play larger roles that also span the intermediary space.

The most effective IOs operate in networks where they often play a convening role, are resourced to do research and/or advocacy, and often serve as agenda-setters on behalf of national organizations or policymakers.

Funders play a central role as connectors and facilitators, primarily through IOs.

In some cases, we are seeing what could be called privatized public policymaking, as private, non-state actors are ultimately making policy decisions (Layton, 2014; Lubienski et al., 2016a).

Local and meso-level networks typically do not include a capacity to produce or assess more than basic evidence on proposed or implemented policies, so that function is often left to outside actors.

The research that is valued tends to marginalize that produced by most university-based researchers.

There is a notable lack of non-partisan research brokers.

IO networks reflect and promote echo chambers (Goldie et al., 2014).

These findings together indicate that IOs operate in a space largely removed from more traditional forms of expertise, with increasingly obsolete forms of quality control to evaluate information claims. This is a crucial concern, since it suggests that funders are employing a strategy of using IOs in policy networks to establish new channels linking knowledge production, from a wider and less reliable range of producers, with (what are presumably) knowledge users in policymaking circles.

However, in addition to insights on IOs and their networks, a further finding from our study involves the recognition of the limits of the ACF for understanding these patterns. While the ACF is one of several theoretical frames for analyzing policy networks and policy implementation, it exhibits some significant shortcomings in several key areas when applied to the issues we study, including:

accounting for, or helping to examine, the over-production of “evidence,” and the variation in that over-production across contexts

the strong presence of policy brokers, which ACF conceptualizes as relatively scarce because of the need to establish trust

predicting behaviors in and of policy networks

addressing not just potential pathways by which information travels through policy networks, but also the drivers of movement through those pathways.

In fact, some of these concerns apply to many of the other theoretical approaches to understanding the movement of policy ideas across contexts. Drawing from the study described here, the next section outlines some theoretical considerations for conceptualizing policy movements in a way that may ultimately address some of these issues. Understanding how such policies emerge and emigrate is essential in policy research. Such an understanding helps researchers better theorize and analyze the rise of the market around education. Moreover, especially for those interested not just in critiquing, but in resisting the seemingly inexorable expansion of the market, understanding the strategies and tactics that have been successfully leveraged by advocacy organizations promoting that expansion can offer some insights into more effective opposition.

Policy networks as marketplace

In 1990, political economists John Chubb and Terry Moe championed the idea of markets as a metaphor for how education systems could be more effectively organized (Chubb and Moe, 1990). They argued, based on their analysis of a national US dataset, that schools set in more market-like institutional environments are driven by “decentralization, competition, and choice” (p. 67). Thus, lumping all three of these “basic features of markets” together under the banner of “choice,” they concluded: “Without being too literal about it, we think reformers would do well to entertain the notion that choice is a panacea” (p. 217, emphasis in original).

Of course, inasmuch as this was a metaphor (Henig, 1994), it was a thinly disguised one, and was commonly taken (as intended) as a prescriptive model for organizing school systems, under the logic that schools had been unnaturally shielded from market forces, despite the fact that they—like most individuals and organizations—are responsive to competitive incentives of the marketplace, according to this line of thinking.

Following that logic, and—as has been a global trend—extending it to previous non-market areas (Kuttner, 1997; Sandel, 2012), we might consider a “marketplace of ideas” in policy advocacy networks. As with some other markets, they can exhibit serious information problems, suggesting the possibility of market-oriented analytical frameworks for understanding these issues. That is, without taking the metaphor too far or too literally, we might conceive of the transmission of ideas across policy networks in transactional terms in order to better understand these dynamics and obstacles, including the “push-pull,” supply and demand forces that drive the movement of ideas (Lubienski et al., 2016b).

In this sense, such a market has “producers” and “consumers” of information—although that “consumption” may be largely symbolic (which, as noted below, is still significant). Connecting the knowledge creators and knowledge users are the intermediaries or “brokers” who seek to gather, process, package and promote their goods and services to both producers and consumers. Such “selling” does not necessarily entail the exchange of money or goods/services with monetary value—although it may. It can instead utilize a currency of prestige, legitimacy, affiliation, or position within a network.

For instance, on the demand side, many of our informants in the RIO study spoke of which “research evidence” they choose to use not in terms of its rigor or applicability to a policy problem. Instead, they were much more likely to refer to a well-known individual or institution—typically with which they were seeking to suggest to the interviewer some sort of affiliation—as proof of their connection to prestige and credibility. Informants would mention, for instance, that they had hosted a speaker from a well-known think tank to lecture and advise them, had learned of some findings from a popular blogger, or would cite a study out of a notable institution like Brookings, Harvard, or Stanford without really being able to discuss the quality or applicability of the report in question. Likewise, on the supply side, researchers, and representatives of organizations that produce research, are eager to show where their work has been used or cited by policymakers.

More importantly, IOs, as brokers in these markets, see it as their role to connect research information with users, be those in policy circles, public forums, or social media. And, from our data, it appears likely that brokers may operate in hierarchies just as do the buyers and sellers they seek to serve. Of course, they typically (but not always) select evidence that aligns with the agendas they represent, even as those transactions span local, national or international networks. But those selections also seem to be shaped by the prestige of their (potential) clients on both the supply and demand sides, and their ability to connect with those clients is influenced by the IO’s position in the network. A local IO in Denver, for instance, is likely to promote information from a local researcher to city and state-level policymakers—not only because of the particular relevance of that research in that context (indeed, that is not always the case), but because neither information users nor (too often) information producers have the capacity to deal in more sophisticated data and analyses, which is often left to better resourced organizations that are more prominent in the policy networks (see, for instance, A+ Denver, 2012). And IOs’ positions within the network hierarchies might be demonstrated by measures such as their citation counts, press penetration, or the prestige of the members on their boards—which are often populated by CEOs, hedge-fund managers, reformers, politicians, and sometimes professors. Thus, low-end brokers may specialize (although not always exclusively) in serving lower-end buyers and sellers, while mid- and high-end brokers may have their own market segments to serve.

In this regard, information that travels through these networks might be regarded as a positional good to some degree. Policymakers “using” evidence from more prestigious sources enjoy added credibility. On the supply side, even though researchers may like to see their work embraced by broad audiences and cited widely, researchers whose work is “used” in elite policy circles enjoy particular prestige, even if that “use” is not broadly known—for example, in a high court case, or in the formation of legislation. However, it is important to note that “use” is an intentionally vague concept. Information may be “used” substantively: for example, by helping a policymaker arrive at a particular position. Yet in the RIO project, there was virtually no evidence that information consumers “use” research evidence that way. For instance, informants are asked to offer examples of when evidence changed their position on an issue, and almost all were unable to provide examples. Instead, it is helpful to think of “use” in ways that are similar to how we can envision currency in these policy network “markets” —in more symbolic ways. Indeed, scholars have pointed to multiple ways that research can be used in contexts such as are described here. For instance, research evidence can be used conceptually in a confirmatory manner (Weiss, 1979); it might be used (or misused) tactically (Davies and Nutley, 2008); it can be utilized in “hortatory applications” where it serves a symbolic role to exhort supporters (i.e. McDonnell, 2004); or in “decision accretion,” where evidence is used gradually to limit the range of future possible policy options (Weiss, 1980).

The problem of information asymmetries inherent in these networks offers an interesting issue from which to explore the usefulness of a transactional conception of policy networks. As has been suggested above, the sophisticated and arcane nature of the production of much research evidence moving through these networks can give undue advantages to those “selling” or promoting the information. Buyers or users without the ability to observe or evaluate production processes are at a disadvantage. Inasmuch as this information is a positional good, users might not prioritize the ability to monitor quality but instead be satisfied with their association with a brand-name producer, such as a prestigious think tank or university. However, the possibility of an embarrassing fiasco where research turns out to be unreliable or simply wrong (e.g. Williams, 2014) suggests that buyers and sellers with different capacities for evaluating research need to form some level of trust.

To some extent, the institutional environment in which education policy networks operate appears to reflect monopolistic competition, where multiple sellers are generally similar but also all somewhat unique in consumers’ eyes, and there are few barriers for new providers to enter the field. Thus, for example, the Center for Education Reform promotes all forms of school choice, while the Friedman Foundation promotes research on school choice, but specializes in voucher advocacy. Each has control over a smaller segment of the market. The actual distinctions—if any exist—between products might not be clear to consumers, so promotion again often happens in terms of branding or affinity grouping.

But in instances of over-supply—as in this case where multiple producers and brokers are trying to “sell” seemingly endless supplies of information to policymakers—where there is a particular need to promote one’s product, different aspects of the good may suggest different arrangements and implications for developing and maintaining trust. If the trustworthiness of a good is readily apparent to a prospective buyer, promotion is often done based simply on direct evidence of a good’s merits relative to alternatives, as when shopping for fresh food one could sample, or clothes one could try on. In such a case in policymaking, a user would be able to evaluate the relative merits of two contradictory studies based on their methods and data, if those are apparent and understandable. But sometimes the trustworthiness of a good is something that can be assessed only after buying it—for instance, whether or not the food in the can is spoiled, or if scaling-up a program based on evaluation actually leads to better results. In other cases, a good’s relative quality can never be truly assessed. Then promotion is often based more on branding or affinity grouping. This is often the case with marketing of gasoline or supplements, for instance, or using branding and loyalty programs to encourage users to feel inclusion in a group. For the topic at hand, this might mean that one counts oneself as a supporter of a reform organization, a follower of a given think tank’s blog, or a formal member of an organized advocacy group like the American Legislative Exchange Council. ²

What this all indicates, then, is that we can conceive of education policy networks as a “marketplace of ideas,” but not in the usual sense of that phrase, where a range of ideas themselves compete for supremacy. Instead, it is a site where ideas can be seen as being bought and sold. If we can accept this market metaphor in observing the behaviors of actors within policy networks, then it may also be useful to leverage market-oriented analytical approaches in investigating these networks, and in determining the best strategies for engaging in policy discussions.

Conclusion

Many education policymakers in recent years have promoted the idea of creating more information for use by “consumers” to foster an education marketplace—for instance, through uniform metrics such as standardized test results, school ratings or league tables, or through parent information centers. Yet the degree to which such market-oriented policies themselves are based on hard evidence of their effectiveness is highly questionable, and highly contested. In an age of overabundance of both official and unofficial information being directed to sway policymakers (even when they apparently may not use it), researchers need to understand how information moves, and what drives such movement, particularly across a changing institutional landscape in education policy. Indeed, the disconnect between the advocacy and outcomes around these policies is notable, but the continued acceleration of their adoption should be a wake-up call to researchers—not simply that they need to criticize these policies, but that advocacy groups are successfully dismissing critics and their research.

Advancing from previous work on the “push-pull” factors that drive information mobility, uptake and use (Lubienski et al., 2016b), this conceptual analysis considers ways that we think about the socio-political coalitions that have been promoting market models for education through policy networks in the US. It draws on a large empirical study in order to consider ways of better understanding these issues in education policymaking. By using an economic-transactional lens to conceive of the relationships between different actors in these networks, it may help us to theorize and analyze the movement of information across networks, the drivers of such movement, and the roles of different actors in the networks.

However, this is an initial foray into this complex question. Even as the ideas herein are being further fleshed out, they are necessarily faced with challenges that may validate, expand, or alter the proposed approach. For instance, more attention will need to be paid not only to questions of currency or capital in such transactions, but how those become commonly used in a given network. Moreover, there are questions as to how not just currency, but other factors in these networks, such as roles, density, and distribution, may differ across contexts. For example, how is research evidence from a local program treated in national-level networks? Do brokers from one context bring additional (or diminished) credibility to networks in another context?

Still, while cognizant of such questions, this paper suggests the usefulness of considering alternative frameworks and perspectives to consider complex policy issues. Moreover, it highlights the fact that researchers have to gain currency in this marketplace of ideas. Simply assuming that scholarly theorizing, analysis, or credentials will lead to influence in policy discussions is shortsighted. Indeed, the academic preoccupations with theory and arcane language popular in critical research circles appear to be the wrong currency to gain advantage in these exchanges (Lubienski, 2017). Instead, less academically inclined advocacy organizations are often the most successful at producing the information that shapes the marketplace of ideas in education policy.