Introduction

This chapter takes as its point of departure the ascendant position of the evidence discourse within education. In stark contrast to its genesis within the medical field, the evidence discourse has been launched into the field of education by external stakeholders in mostly top-down moves that have largely bypassed professionals within the field (Hammersley 2007; Krejsler 2013). Danish examples will serve to illustrate the argument; the trend, however, can be seen across the Nordic countries (e.g. Bergmark and Lundström 2006; Oscarsson 2006; Telhaug et al. 2006; Utdanningsforbundet 2008) and beyond, globally (Furlong et al. 2009a, b; Henry et al. 2001; Hopmann 2008; Meyer and Benavot 2013; OECD 2007; Rizvi and Lingard 2010; Wells 2007; Krejsler 2017 (forthcoming)).

I shall argue that the evidence discourse mirrors a cultural struggle that currently rages about how key areas within modern societies are to be defined (Biesta 2007, 2010; Gibbons 1994; Hammersley 2013). How we perceive evidence for what works has significant implications regarding how a hospital or a school may conceive of their visions, targets and practices, and what kinds of research and research paradigms are considered legitimate in the production of knowledge. A too strict focus on Randomized Controlled Trials (RCT tests) in the health services and education will tend to marginalize other and ‘softer’ professional practices of validation. In other words, discourses about the population’s health and education proliferate in close reciprocity with the criteria for verification that such knowledge is subjected to. Health and educational issues must be conceived of in ways that somehow satisfy the criteria for producing evidence which mandate powers funding those activities demand.

On this background, I shall argue that evidence may be conceived of as a floating signifier with great advantage, and support this claim by introducing three distinctions to analytically qualify critique of, and challenge the current dominant regime of, evidence: By introducing (1) the distinction between evidence-based vs evidence-informed knowledge, I wish to exploit an already fruitful approach to distinguishing between issues that merit so-called ‘hard’ and others that merit ‘soft’ evidence approaches to evaluating what works. By introducing (2) the distinction between global vs local forms of evidence I wish to highlight that some issues merit knowledge that is valid regardless of context, whereas intervention in other contexts requires forms of knowledge that are highly responsive to the particular context of intervention. By introducing (3) the distinction between external vs internal evidence I want to problematize the question about who has the right – or more precisely the executive powers – to make decisions about which forms of evidence count as knowledge that works. This points to the tensions between the considerable production of knowledge and documentation for what works by professionals and educational researchers, which function as supplements or contesting knowledge to the forms of evidence that powerful external stakeholders currently impose on education.

In total, this chapter is an argument that serving the public good in a democratic society requires tools for producing ‘evidence’ and knowledge about ‘what works’ that reflect the diversity of values and purposes that members of this society strive to accomplish. And education in particular cannot be dissociated from the visions of the good life and the good society that give education direction, unless you accept an instrumental and objectivist vision of knowledge and education. Consequently, the term of the floating signifier is not a license that any ‘evidence’ goes, but a reflection that truths about what works in education can seldom be dissociated from the perspectives, values and purposes that give direction to educational activities (e.g. Biesta 2007, 2010).

Discourses , Genealogies and Floating Signifiers

A Foucauldian approach allows us to make evidence discourse problematic as a particular truth regime that makes some ways of speaking and acting possible while excluding others (Foucault 1971). Like any other discourse, it is constituted as a pattern of interconnected statements which reciprocally refer to one another, thereby continually reinforcing the totality of the discourse. The immanent logic thus construed forms a strategic space wherein a number of different subject positions emerge to be occupied by willing individuals. Obviously, one must subject one’s self to the discursive regime in question in order to be included as a legitimate subject within this regime.

Foucault argued that the power-knowledge effects of a given discourse must be measured by the extent to which it matches and mirrors the dominant configuration of dominant and less dominant discourses that set the boundaries for how individuals can think and act at a given time and space in history (Foucault 1993, 1997). Foucault considered it his task to chart, via a genealogical method, the topological contours of the battlefields with which subjects within different discursive fields struggle to come to terms (e.g. the fields of madness, reason, imprisonment, subjectivity, sexuality and so forth).

Drawing on insights from Foucauldian genealogy, I shall sketch major threads that appear to have coalesced, making evidence discourse an increasingly dominant voice which cannot be ignored when considering what works in education. Or, formulated as a question in a Foucauldian genealogical vein: How has it come about that researchers, policy-makers and practitioners today make education problematic in terms of ‘evidence’ and ‘what works’?

In order to explore potentials for expanding what evidence and what works may mean, I shall introduce ‘floating signifiers’ as a useful thinking tool. This entails an analytical concept coined by Ernesto Laclau (1993) in order to capture a category of open empirical concepts that have become increasingly prevalent in modern societies, producing meaning and consensus at a sufficiently general level in order to cover the diversity of lifeforms at play when various subject positions negotiate what shall count as legitimate truths within different policy areas. The current political climate abounds with dominant floating signifiers such as ‘quality’, ‘efficiency ’ and ‘excellence’. They set new agendas and dislocate established truths by expressing flexibly the interests of dominant configurations of stakeholders.

A particular consensus-producing feature of the floating signifier is that it makes it difficult to disagree until you require specifics about how it is intended to be operationalized. You cannot disagree with ‘quality’ as such. Similarly, it is hard to disagree that evidence and what works approaches to dealing with reality are desirable. As documented in this chapter, the dominant version of evidence performs exactly that operation. It appears open to interpretation at a surface level while being simultaneously already coded with meaning and woven into powerful genealogies of possible meanings by virtue of lengthy negotiations among powerful stakeholders. In medicine as well as education, evidence for what works is thus established by linking particular perceptions of science to the pragmatic powers and agendas of dominant stakeholders within policy, science and market.

Observing evidence as a floating signifier thus assists us in mapping dominant policy agendas and, hopefully, challenges us to experiment, looking for new interpretations in the ruptures and inconsistencies that appear from the maelstrom of highly charged political contexts (Deleuze and Guattari 1994; Krejsler 2016). Overall, this chapter represents a thinking technology for researchers and professional groups within education in their struggle to produce knowledge and practices that gain legitimacy by exploiting the interstices and possible loopholes in current dominant policy configurations to which they are subjected. As a hybrid technology, it represents potential pitfalls as well, as engaging in the dangerous struggles around dominant policy practices entails the risk of being co-opted into negotiations that may well narrow the room for manoeuver, giving legitimacy to an evidence consensus that may disregard useful knowledge and practices which educational researchers and professionals have built up over decades. That is the risk taken when entering this game. However, the costs of staying out of it may well be considerably higher, as it leaves the space entirely to others. As such, this chapter claims to do service to the public good by supplying a conceptual apparatus that may expand what counts as ‘evidence’ and ‘what works’ knowledge , and thus may assist more groups in entering the struggle about education and how it may be thought and organized.

The Genealogy of ‘Evidence’ and ‘What Works’ Discourse in Education

In this section I shall clarify how the evidence discourse became a dominant regime of knowledge in education by mapping how major sources from medicine coalesced with agendas of school effectiveness research and transnational agencies like the OECD and its focus upon optimizing human capital. This would eventually transform conditions for producing knowledge about education (Bhatti et al. 2006; Hammersley 2007; OECD 1996, 2007; Rieper and Hansen 2007).

Historically, the concept of evidence has many sources, ranging from everyday common sense discourses to judicial and economy discourses. From the late 1980s onwards, however, the concept of evidence has mainly – albeit not exclusively – been associated with a resurging neo-positivist paradigm and its procedures for producing knowledge about what works in relation to particular interventions (Alvesson and Skjöldberg 2000; Hammersley 2007; Krejsler 2013; Pawson 2006; Rieper and Hansen 2007; Sackett et al. 1996).

The evidence discourse that has gained pre-eminence in relation to producing knowledge about what works is rooted within the medical field (Bhatti et al. 2006; Browman 1999; Rieper and Hansen 2007; Sackett et al. 1996). In 1972, Archie Cochrane (1909–1988) published “Effectiveness and Efficiency  – Random reflections on health services” (Cochrane 1972), the groundbreaking book that was destined to achieve an almost mythical position. Cochrane argued that public resources are scarce, and therefore it is important that they are spent on practices with a proven record of effectiveness . He argued that a systematic base of scientifically tested knowledge of what works should be accumulated, preferably based on Randomized Controlled Trials (RCTs) or similar experimental designs that aim at testing well-defined causal relations . He was convinced that this approach would supply reliable knowledge to the health services with greater probability than other approaches to evidence. A substantial element to Cochrane’s argument was that the universality and objectivity of evidence produced by RCT would contribute to continually ensuring equal access for all to effective treatment. Cochrane’s simple suggestion, and the fact that RCTs were becoming mainstream in medical research, contributed to a breakthrough within the health services. By 1990, The Cochrane Collaboration (www.cochrane.org) was established, which has since then been under continuous expansion and is today an international and highly influential collaboration of professionals, with considerable impact on policy. These events gave momentum to the efforts to elaborate so-called systematic reviews of international research about what works in relation to particular medical substances and treatments. The evidence ladder here serves as a methodological device to rank the quality of studies according to an ideal of objectivity , validity and reliability that celebrates the Randomized Controlled Trial as its golden standard (e.g. http://www.controlled-trials.com/). This development has been facilitated by the fact that in the 1980s, meta-analysis developed into an independent field of research. Meta-analysis is the methodological basis of this review form designed to statistically calculate and synthesize what primary studies say about what works according to rigorous standards privileging the evidence ladder. The emergence of the evidence discourse , Cochrane’s initiative and its repercussions mainly arrived ‘bottom-up’, i.e. from the profession of medical doctors themselves. This truth regime signifies a neo-positivist, quantitative and experiment-oriented approach to doing science.

Inspired by the work of the Cochrane Collaboration , the Campbell Collaboration was established in 2000 with the ambition to make social welfare, crime prevention and education as evidence-based as medicine (Bhatti et al. 2006; Petrosine et al. 2001). The Campbell Collaboration upholds basically the same review procedures as the Cochrane Collaboration , including the evidence ladder that places at its top systematic reviews conducted as meta-analyses of primary studies that place RCTs as the golden standard.

In relation to social welfare, crime and education, the evidence discourse is considerably more directed by the agendas of policy-makers and administrators. In part, this can be explained by the fact that these fields differ from the medical field in crucial aspects (e.g. Bhatti et al. 2006; Hammersley 2007; Pawson 2006). Among other issues, these fields are more diverse in terms of professionals’ levels of education, lack of unified professional identity and questioned public status and state authorization. Further, their knowledge base refers to a number of competing scientific paradigms, largely social sciences and humanities paradigms that are often more difficult to convert into regimes that produce so-called ‘certain knowledge about what works’ which is measurable and can be standardized.

One could argue that education constitutes the field that has been impacted most severely by the evidence discourse understood as a pressure that has largely come from external stakeholders representing discursive regimes which resonate poorly with most existing discourses among professionals and researchers in education (Ball 2007; Bhatti et al. 2006; Borgnakke et al. 2006; Krejsler 2013). It should be added, though, that established traditions among researchers and professionals within education have since long adopted evidence discourse. The latter often engage in close collaboration with policy-makers and administrators in an effort to develop and implement evidence-based or evidence-informed policy , with many researchers gathering in or referring to the well-established International Congress of School Effectiveness and Improvement (e.g. http://www.icsei.net/).

This intensified focus on education is hardly surprising in a period when national governments and transnational bodies like the OECD , EU and the Bologna Process are increasingly occupied with discourses of Knowledge Economies and Life-Long Learning (Henry et al. 2001; Meyer and Benavot 2013; Nóvoa and Lawn 2002; Prewitt 2012; Rizvi and Lingard 2010). The OECD has increased attention to improve the role and efficiency of educational research as a background for decisions made by policy-makers, in addition to the ongoing efforts of practitioners to make educational institutions more result and efficiency oriented and more evidence-based (Burns and Schuller 2007; OECD 2007). An increasing number of comparative surveys, statistics and country reports have thus been developed as aids to policy and practice advice.

In 2004, one of these country reports was accomplished at the request of the Danish government (OECD /CERI 2004). It came to the main conclusions that Danish educational research (R&D) was characterized by too little focused research on key areas, and that links were too weak between educational research and the needs of practitioners and policy-makers. It was within this context that the OECD made the recommendation that a Clearinghouse for Educational Research should be established, drawing on the accomplishments of the British Evidence for Policy and Practice Information and Co-Ordinating Centre (EPPI) as well as the American What Works Clearinghouse (WWC) (http://ies.ed.gov/ncee/wwc/). WWC was established in close collaboration with the Campbell Collaboration (Boruch and Herman 2007). This was done in continuation of the No Child Left Behind School Act (U.S. Department of Education 2002), with the explicit purpose of making school practice more evidence-based (Coalition for Evidence-Based Policy 2003). EPPI and – even more so – WWC have drawn extensively upon inspiration from, and collaboration with, the Campbell Collaboration . To illustrate the impact of these developments in terms of what counts as evidence and knowledge about what works, it should be noted that legislation following the NCLB act requires schools to rely on scientifically based research for programs and teaching methods. The act defines this as “research that involves the application of rigorous , systematic, and objective procedures to obtain reliable and valid knowledge relevant to education activities and programs” (Hamilton et al. 2008; Zucker 2004). This means scientifically based research results in “replicable and applicable findings” from research that uses appropriate methods to generate persuasive, empirical conclusions. Non-scientific methods – according to this discourse – include following tradition, personal preferences , and what is claimed to be non-scientific research such as research based on case studies, ethnographies, personal interviews, discourse analysis, grounded theory, action research and other forms of qualitative research . The latter are no longer seen as an acceptable basis for making decisions about teaching children under the act, which makes them ineligible for federal funding.

The Danish Clearinghouse for Educational Research applies a broader definition of evidence than the one represented by the Campbell Collaboration , WWC and the evidence ladder. It explicitly mentions EPPI as a key inspiration. In addition, it should be mentioned, however, that very few RCT-primary studies have been carried out in Danish or Scandinavian contexts, which for a considerable time to come would make systematic reviews relying on such studies dubiously reliant on American studies in particular. And, judging from existing reviews, the Danish Clearinghouse approach appears to be closer to the more inclusive and narrative approaches to synthesizing reviews of what works that thrive among scholars like Hilbert Meyer (2004) and Brophy and Good, who for decades have represented alternative approaches to synthesizing evidence for what works in education (Brophy and Good 1986; Good and Brophy 2003). Currently, the approach to conceiving of evidence represented by John Hattie and his best-seller Visible Learning is gaining considerable ground (Hattie 2009). This approach mimics the Campbell approach in some aspects by carrying out quantitatively based reviews from thousands of primary studies – albeit not adopting the evidence ladder as such – in order to identify correlations that enable statements about what works in relation to facilitating learning.

It will thus be interesting to observe in the years to come which strategies for determining what counts as evidence will be adopted by the Danish Clearinghouse for Educational Research . Will this clearinghouse succeed in bridging the gaps between policy-makers, practitioners and the research community, in terms of ideas about what works, and thus gain broad legitimacy? Or will it rather be looked upon as an unbearable attempt to reduce the multi-paradigmatic features of existing educational research in order to elevate more mono-paradigmatic quantifiable measuring to the legitimate standard for producing truths to teachers and pre-school teachers (e.g. Borgnakke et al. 2006; Christensen and Krejsler 2013; Lihme 2005; MandagMorgen 2004; Moos et al. 2005; Thorslund 2005; Utdanningsforbundet 2008)? Or will it be ignored?

In summary, we observe that the evidence discourse is subtly transformed as it travels from a medical context to an education context. The educational research community and professionals appear to have less of a say than policy-makers and administrators. Human capital and knowledge economy discourses visibly intervene, carried forth by strong transnational players, the OECD in particular. A truth regime that often resonates poorly with existing discourses and regimes for producing truths about what works among the bulk of professionals has to be accommodated (Ball 2007; Biesta 2007, 2010).

‘Evidence’, ‘What Works’ and the Reconfiguration of Dominant Policy

The expansion of the evidence discourse from the medical field to education signifies the establishment of a dominant discursive formation that cannot be ignored. Thus, in a Danish context, Nordic Cochrane , Nordic Campbell Centre and the Danish Clearinghouse for Educational Research were all established in alliance with dominant players in Danish society such as the National Board of Health, the Ministry of Health, the Ministry of Social Welfare, the Ministry of Education and the Ministry of Science, Technology & Innovation (Bhatti et al. 2006; Moos et al. 2005). Nordic Cochrane and Nordic Campbell Centre received substantial supplementary funding from the Ministry of Health and the Ministry of Social Welfare, which, however, was not the case with the Danish Clearinghouse for Educational Research . In England, the Evidence for Policy and Practice Information & Co-Ordination Centre (EPPI), the Campbell Collaboration and other key evidence-producing institutions succeeded in elevating this truth regime into a dominant discourse by joining forces with New Labour and its interest in legitimizing policy choices with reference to evidence-based or, at least, evidence-informed scientific backing (e.g. Furlong et al. 2009a, b; Oakley 2007; Wells 2007).

As already argued, the evidence discourse is furthermore closely aligned with dominant players on the global stage such as the OECD and EU, as knowledge economy discourse surges (e.g. Henry et al. 2001; Krejsler et al. 2014; Pawson 2006). It is part of a global regime of knowledge that standardizes knowledge production across national boundaries and academic disciplines (Larner and Walters 2004; Nóvoa and Yariv-Mashal 2003). Comparisons and rankings of countries gain ground as education and the competitiveness of nations are increasingly linked discursively (Henry et al. 2001; Hopmann 2008; Meyer and Benavot 2013; Nóvoa and Lawn 2002).

New Public Management strategies are employed to implement the idea of a market of public services , which presupposes a market of competing suppliers of services that are comparable and transparent to consumers (Hood 1995; Sahlin-Andersson 2001). Here, the evidence discourse expediently offers a methodology that makes services measurable and comparable with the explicit purpose of exposing what works (Hammersley 2013; OECD 2007; Pawson 2006; Prewitt 2012; Rieper and Hansen 2007). A seductive imagery of standards and transparency is produced, which comes in handy for policy-makers and practitioners who need simplified criteria for prioritizing choices in a complex world abounding with information and possible choices. Floating signifiers like ‘freedom’, ‘quality’, ‘choice’ and ‘evidence’ profoundly influence health services and education as consumers are allotted tax-financed vouchers to choose among public services made comparable in order to optimize what they believe will provide the best treatment or learning environment.

In light of the battles with positivism in the late 1960s and the rise of social constructivism since the 1980s, it is noteworthy that the evidence discourse drawing, as it does so explicitly, on neo-positivist ideals, succeeds in gaining such momentum within the hitherto largely humanities and social sciences dominated fields of education (Alvesson and Skjöldberg 2000; Gibbons 1994; Habermas 1971; Hammersley 2013; Lyotard 1999/1984).

Thinking ‘Evidence’ and ‘What Works’ Differently: Exploring the Potentials of a Floating Signifier

The rise of the evidence discourse as a dominant truth regime draws attention to the fact that over time, different academic and professional fields have developed different criteria for producing truths, in basic research as well as applied research (Clarke 2006; Gibbons 1994; Habermas 1971; Hammersley 2007). They exercise different functions, serve different stakeholders, and produce, consequently, different knowledge bases that represent different approaches to what works. This, however, is no argument for not continuously scrutinizing whether already established criteria for scientific practice are unduly taken for granted. Nonetheless, it brings into question whether the RCT-based or similar criteria for evidence may be ascribed universal validity across academic and professional boundaries. Among many educational researchers and practitioners, contestation is rising against an evidence discourse that is experienced as largely disregarding the ‘nature’ and particularities of the educational field as they understand it (Ball 2007; Biesta 2007, 2010; Hammersley 2013).

Nevertheless, it is recognized that the field of education is under increasing pressure from external stakeholders to document what works according to particular formats, as public debate increasingly takes direction from what counts as evidence in large-scale international quantitative comparisons. This is largely driven by a pressure to produce so-called certain knowledge which, allegedly, makes it easier to prioritize interventions with a proven track record in education and thus make more efficient use of limited tax revenue. The OECD PISA surveys have been particularly influential in shaping national agendas in that direction (Hopmann 2008; Meyer and Benavot 2013). It is still highly disputed, however, among educational researchers and professionals whether the breakthrough of the evidence discourse in the form of quantitative measuring, statistics, rankings, benchmarking and so forth actually increases the overall quality of education (Biesta 2010; Hammersley 2013). A strong critical voice, Stephen J. Ball (2007), warns that educational studies should not be reduced to a mere technician’s approach to finding tools that work, thereby potentially excluding the role of the intellectual’s theoretically informed approach to research.

In order to manoeuver within the opportunities and pitfalls of this new – and at times treacherous – discursive landscape, I suggest that we enter into a struggle to expand the meanings of evidence and ideas of what works in ways that link constructively to the long genealogies of producing truths within different professional languages and academic disciplines in education. In claiming to serve the public good , I do this to emphasize that evidence and what works should be dealt with as means to an end: better education. Education – as well as other professional fields – first and foremost needs professionals who can exercise professional judgment according to the educational situation to be dealt with. Obviously they need knowledge about what works, but the question about which knowledge that works, many would argue, can seldom be decided in advance and independently of the educational situation in question. Consequently, it becomes evident – in my mind – that it would be a tactical blunder of considerable dimensions to voluntarily surrender the right to define the floating signifiers of evidence and what works too quickly and conclusively to the dominant evidence discourse and its particular staging of how one produces evidence about education. What counts as evidence about what works in relation to a given issue thus appears to merit – more often than not – a rigorous appraisal of the ‘nature’ of the professional field and the influence of the particular context involved.

Working with evidence as a floating signifier requires careful appreciation of the actual strategic spaces available in current policy and scientific discourse and their associated contexts. Nonetheless, by engaging in constantly challenging the limits of the actual possible one may find spaces and opportunities to expand and – possibly – loosen bits of meaning of evidence from the currently dominant evidence discourse (Deleuze and Guattari 1994; Krejsler 2006, 2016).

In this light, I suggest that we introduce three analytical distinctions into the evidence discourse to sharpen the gaze when the educational researcher or professional manoeuvers within current evidence and what works discourse:

  1. 1.

    A distinction between evidence-based and evidence-informed knowledge allows for distinction between ‘hard’ and ‘soft’ evidence issues;

  2. 2.

    A distinction between global and local forms of evidence enables distinction between a global form of evidence that is valid among all within a well-defined group of intervention across contextual particularities and a local form of evidence that primarily makes sense with reference to the particularities of the context involved;

  3. 3.

    A distinction between external and internal forms of evidence makes visible the distribution of power relations concerning who owns the right to produce legitimate knowledge about what works.

Evidence-Based or Evidence-Informed Knowledge and Professional Discretion

Following the genealogy of evidence discourse closely will reveal that as it moves from medicine towards ‘softer’ areas such as education, the need for a distinction between evidence-based and evidence-informed knowledge about what works gradually develops (e.g. Hammersley 2013; Hammersley 2007). In medicine , it often – but not always – makes perfect sense to talk about evidence-based knowledge about what works, i.e. causal or quasi-causal knowledge. In education, however, contextual factors often – but not always – become so important that it makes sense to talk more humbly about evidence-informed knowledge. In relation to a pill for headaches, it usually makes sense to talk about evidence-based knowledge; however, when doing counseling in relation to a child at school whose parents are in the process of being divorced, it is mostly difficult to speak of what works in a strong sense. This does not mean, however, that choice of interventions cannot be informed by knowledge about what works although decisions will often require close and careful reference to the particular context, the ones involved as well as appreciation of and dialogue about the issue as it evolves.

Even strong proponents of dominant evidence approaches, like Ann Oakley, previous director of EPPI (2007) and David H. Hargreaves (2007) concede that in softer, multi-paradigmatic policy-fields it makes sense to speak more tentatively about evidence-informed policy and practice rather than employing a more rigorous evidence-based approach like the one within the medical field. Further, as mentioned, the similar but broader approaches to looking for correlations based on large quantitative systematic reviews , like those carried out by researchers such as John Hattie, are gaining considerable momentum (Hattie 2009). Already existing approaches to synthesizing knowledge about what works offer potential alliance partners in challenging the dominant evidence discourse in ways that resonate better with understandings among professionals and researchers of various paradigms. For instance, the strong German tradition represented by Hilbert Meyer and Andreas Helmke among others has attempted to make more inclusive and narrative syntheses of what research claims works in education (Helmke et al. 2008; Meyer 2004). For instance, in Was ist guter Unterricht? (What is good teaching?) Hilbert Meyer concludes that research shows considerable agreement that there are traits which characterize good teaching across contexts, including: it is well structured; the teachers know their subject; methods cannot be chosen independently of context; and teaching that works takes into consideration the individual differences and learning needs of students. Syntheses of the latter kind have been criticized for being so general that they do not offer much guidance to teachers in relation to concrete instruction. They do not capture the causal or quasi-causal relations that the dominant evidence regime is looking for, and they do not isolate particular methods, ways of organizing class and so forth in manners that are valid across contexts as global evidence . The counterargument is that this level of generality reflects the contextually dependent ‘nature’ of most educational issues.

Global Versus Local Forms of Evidence: Which Kind of Knowledge Is Relevant?

Here, focus is on the features of the knowledge base that is applied to deal with professional issues (e.g. Moos et al. 2005). Global evidence represents knowledge that is valid with large probability for all within a well-defined group of intervention, for example taking a particular medicine against a certain type of symptoms, or the application of a particular reading method to deal with particular pronunciation problems. Most dominant evidence-discourse aspires for knowledge about what works that represents global evidence. Local evidence points to the importance of contextual knowledge for deciding whether a given intervention is likely to work.

Local evidence should not be confused with what is called internal evidence in the following section, which rather refers to the aspect of who demands that a particular kind of knowledge shall be applied, be it external or internal stakeholders. Local evidence refers to a kind of knowledge production that is sensitive to local contexts and may or may not be relevant in other contexts. Production of this kind of – often qualitative – knowledge often employs other research paradigms, such as action research that directly involves students, teachers and other stakeholders in producing knowledge about what works, or ethnographic qualitative approaches that are sensitive to the particularities of a given school and classroom culture located within a particular local, social, ethnic and national culture. What counts as evidence and what works as best practice for a nurse in their care for a particular patient is often hard to settle without reference to the particular context, i.e. local evidence : is it a cancerous patient overwhelmed by fear of death? Is the patient a child, an elderly spouse or a single and lonely patient without relatives? Is the patient religious or an atheist? Here, it hardly makes sense to apply methods that fit all. The same would often apply to the teacher dealing with a child in crisis, and so forth.

Since the 1960s and 1970s, individualizing discourse has developed into a particularly strong regime for producing concepts and technologies that resonate with local evidence and context sensitivity. Over the decades, this discourse has highlighted a plethora of terminologies about the academic, social and personal competences to be developed by students in each their individual way, be it the self-realization, project-oriented, unique individuality or the lifelong learner version (Gardner 2011; Gartenschlaeger and Hinzen 2001; Rogers 1969). Here, demands for evidence that particular interventions work refer more to professionals’ and students’ process evaluation within the institution, i.e. local forms of evidence. The social technologies employed to produce evidence of student achievement would be project work, log books, portfolios, social contracts, formative evaluation and so forth. Attention is directed at dialogue, supervision and other process technologies that reflect interaction between teacher and students. At an organisational level, appraisal interviews with teachers and teams of teachers, and dialogue with parents are employed (Krejsler 2006, 2007). When society increasingly demands that individuals learn to document their particular academic, social and personal competences, obviously school will be required to produce evidence that students acquire such competences (Undervisningsministeriet (Danish Ministry of Education) 1996, 2003). It may be reasonably assumed that a host of knowledge and practices developed and practiced by the teaching profession for decades qualifies as evidence for best practice, be it project work, log books, portfolios, self-appraisal or social contracts. Concerning educational research , Stephen Kemmis argues that participatory action research, which involves teachers and students, is an indispensable approach to producing knowledge and practices, i.e. an alternative understanding of evidence for what works (Kemmis 2007). Obviously, even local forms of evidence should be scrutinised for their potentials and pitfalls, as has been done extensively by, among others, Foucault-inspired educational research (e.g. Krejsler 2006; Popkewitz 1998; Rose 1999/1989).

Most educational situations, however, tend to call for interventions that make use of both global and local forms of evidence. Even within contextually very particular situations there are plenty of partial aspects that may benefit from reference to global evidence . For the nurse, this could be knowledge about and access to pain-relieving medicine . For the teacher, it could be knowledge about and access to methods for learning links between sound and letters, knowledge about physiological development for children of particular ages and ensuing learning barriers, and so forth. Exercise of professional discretion here demands that professionals have significant mastery of various theories and methods representing both global and local forms of evidence that can be brought into play when a given practice must be dealt with. Global and local forms of evidence are here exposed as ‘just’ the means – or the repertoire of knowledge – that professionals draw on when performing the art of exercising professional judgment in situations of application. Blind evidence refers to situations where global or local forms of evidence are applied blindly by virtue of habit, dogmatism or tradition, i.e. where professional judgment is suspended when it comes to determining whether one or the other form of evidence should be applied to a concrete situation.

Even strong proponents of the dominant evidence and what works truth regime, such as Professor Robert E. Slavin , director of the Center for Research and Reform in Education at Johns Hopkins University, shows considerable understanding for employing the distinction between global and local evidence , albeit from a narrower epistemological perspective than argued in this chapter:

However, there is a big distinction between two kinds of good evidence that I think it is useful to make … One kind of good evidence relates to proven programs … A hallmark of proven programs is that they are designed for replication … The other type of good evidence, local evidence , is derived internally to a given school, district, city, or state. Such evidence helps policymakers and educators understand their own situation, opportunities, and problems, and to evaluate policies or practices already underway or being considered. Such data may be particularly valued by the local leadership, because it addresses problems they care about, but it is not intended to produce answers to universal problems, except perhaps as a byproduct … Of course, proven programs and local evidence can overlap, as when a given district or state implements and evaluates a replicable program that responds to its own needs … Because the local leadership was involved all along, they may have greater commitment to obtaining good data and then acting on it. Local evaluations exist in a particular context, which may make the findings of interest in that context and in other places with similar contexts … As we build up stronger and broader evidence of both kinds (i.e. proven programs and local evidence (JBK)), it will be important to learn how each contributes to learning about optimal practice in education. (Slavin 2016)

External or Internal Evidence : Who Has the Decision-Making Powers to Require Which Knowledge Should Be Applied?

Whereas the above distinction between global and local forms of evidence focuses on differences in the forms of knowledge that the professional may employ, the distinction between external or internal evidence asks who has the right to make the decisions about which forms of evidence should be applied. External evidence refers to requirements that external stakeholders such as politicians, officials, administrators or external experts impose upon a given professional area, its organizations and professionals. Internal evidence refers to knowledge and evidence for what works that professionals and associated researchers choose to develop and apply based on research, development and practice within their field, be it global or local forms of evidence.

External evidence typically refers to situations where policy-makers or national and municipal funding bodies require that professionals use special methods or approaches when it comes to documenting that there is evidence showing that applied interventions have effects that work better than other available interventions. When such external evidence demands are linked to funding or specific municipal or national quality assurance measures, they usually have the effect that professional practices are aligned to satisfy such requirements. Here, one may find that external evidence tends to impact on educational practice by enhancing attention towards fulfilling the demands that are required. This may produce positive effects in that students know which well-defined demands to achieve. It may even ensure that most students achieve some basic literacy and numeracy skills. Further, it may encourage professionals to become more systematic and stringent in conceptualizing and implementing interventions that may previously have been conducted with less reflection, relying more on tradition and habits. This may ensure that systematic methods are applied, and that obligations are honored to measure and compare individuals’ behavior before and after a given intervention in quantitative terms, across organizations, municipalities and nations. This may lead to development of a reservoir of knowledge to provide overview of the multitude of possible interventions, and potentially develop global evidence that may qualify professional judgment . When professionals know that they are expected to abide by relatively well-defined requirements, they may actually enhance performance due to having tangible demands to live up to. This may provide some certainty in working procedures to ensure that all clients/users meet some basic requirements in relation to social skills, job readiness or other demands.

Inversely, however, it may produce negative effects in terms of undermining trust-based relationships between external stakeholders and large sections of professionals and the educational research community. The latter may experience a loss of ownership and professional autonomy in relation to the practice they are required to exercise professional judgment within or produce knowledge about. When professionals experience that, first and foremost, they are expected to abide by a given manual, they may – as a consequence – lose attention to all the other potential solutions to the problem which their professional repertoire might otherwise have sensitized them to activate. This could mean that interventions that do not fit into an experimental and quantifiable design are likely to be excluded; not because they do not work, but because their effects cannot be made comparable or measured in Randomized Controlled Trials or similar formats. Interventions based on qualitative methods that seek to adapt approaches to the particular context will not be chosen – e.g. an approach that seeks to involve, by means of explorative dialogue, the clients whose resources are assumed pivotal to engage in order to make real and long-lasting changes in habits and lifestyle. The imposition of external evidence measures may result in substantial washback effects on interpretation of curricula and on teaching (e.g. Graham 2006). If demands are primarily directed at literacy, numeracy and science subjects – as is currently the case in Denmark and in the PISA surveys – it may be at the expense of other subject areas such as history, music, arts and physical education. It may turn much social and educational work unnecessarily instrumental. There is a risk that the cultural heritage may lose some of its features of having a value in its own right – in contributing to developing ‘the good life’ and ‘the good society’ – if students experience that they come to school mainly to get good grades and do well in tests (Biesta 2007, 2010; Hammersley 2013).

In contrast, internal evidence points to knowledge and practice which professionals and/or associated researchers choose to develop and apply within their organizations, and which simultaneously works to qualify their knowledge base and professional practice. Such practice involves various mixtures of global and local forms of evidence. In order to qualify such a concept of internal evidence, it must be linked with procedures regarding how local practice may be continuously qualified on a scientific basis. If one blindly exalts local traditions and knowledge of what we usually do to ‘knowledge about what works’, one’s knowledge base is in imminent danger of becoming provincialized. As an example, the Norwegian educational philosopher Erling Lars Dale spent much of his career developing a scientific and systematic approach to conceptualizing how internal evidence about what works can be qualified – an approach aimed at integrating global forms of evidence with local forms of evidence. Dale developed the notion of ‘didactic rationality’, built on pedagogy as a scientific discipline and as a criterion for adequate scientific justification of educational practice (e.g. Dale 2008). The main idea is that an educational practice can only count as rational if there is a systematic link between three levels of competence and the corresponding role expectations: K1 (to carry out teaching/educational practice ), K2 (to construct teaching/educational programs) and K3 (to communicate on the basis of and construct didactic theory). At the K1 level, the professional teacher exercises the role of teaching and is subject to immediate demands to act. At K2 level, the professional exercises the role of being a team colleague, and plans instruction and class activities for the immediate future. Here, they are under suspended demands to act within the context of a compelling local school culture, a particular mix of students and given local, municipal and national frameworks. At competence level K3, the professional exercises the role of researcher in their own professional practice. Here, the professional is freed from the demand to act which everyday school life and its context usually impose upon them. And the practice levels K1 and K2 can be questioned and discussed in relation to the larger societal contexts which frame them. To the extent that the professional can build up systematic coherence between the three levels of competence on a scientific base , their practice acquires the potential to qualify and develop evidence, i.e. ‘knowledge about what works’. Other approaches that resonate with such an understanding of developing internal evidence would draw on concepts such as the reflective practitioner, the transition from novice to expert, tacit knowledge, communities of practice and so forth (Dreyfus and Dreyfus 1986; Polanyi 1967; Schön 1983; Wenger 1998).

Join the Struggle to Expand ‘Evidence’ and ‘What Works’!

As education and other public service areas are linked to global knowledge economy discourse (e.g. Henry et al. 2001; Larner and Walters 2004; OECD 2007; Rizvi and Lingard 2010), a growing need is produced for instruments to prioritize an increasing number of possible interventions in education to optimize learning within a framework of limited tax resources (Prewitt 2012; Sahlin-Andersson 2001). This state of affairs has facilitated the advent of a particular dominant evidence discourse for producing knowledge about what works which tends to reduce the kinds of knowledge and approaches that will be deemed legitimate and funded by states under pressure. This is done to reduce the public services expenditure and to better their perceived rankings among competing knowledge economies (Hammersley 2013; OECD 2007; Wells 2007). Consequently, the need to enable evidence-based or evidence-informed choices about what works is growing (Prewitt 2012). This chapter has had the double purpose of (1) mapping the genealogy of this discourse in order to (2) enable the reader to constructively join the cultural struggle about defining and, hopefully, expanding how knowledge about what works can be defined.

Establishing the regime of truth regarding what counts as evidence about education and other public services is obviously a high-stakes endeavour that includes dominant players in society and has serious implications for what may count as public good . Recognizing this pressure, this chapter argues that a fruitful strategy for educational researchers and professionals may be to adopt the concept of evidence as a floating signifier (Laclau 1993). It is argued that this may prove helpful in their struggle to expand the meanings of evidence to also cover substantial parts of their professional knowledge and experience. In the long run, this may prove beneficial also to students, citizens, consumers, policy-makers and others, as a considerably broader knowledge and experience base will come to inform how education – but also health and social welfare – practices are carried out. And thus public good will be served!

Viewing and dealing with evidence as part of a cultural struggle is, as argued in this chapter, no easy endeavour. If one aspires to be taken seriously, it requires up-to-date knowledge about, and proficiency in, how to play the complex constellations of subject positions among policy, market and professional players who compose the game to be mastered.

From the point of departure of understanding evidence as a floating signifier whose meanings can be expanded, the reader is incited to venture into the struggle with aid and inspiration from three analytical distinctions:

  • Between evidence-based and evidence-informed knowledge;

  • Between global and local forms of evidence; and

  • Between external and internal evidence .

This serves the purpose of enabling professionals and others to think differently as they struggle to come to terms with the potentials and pitfalls of the pressure from the evidence discourse (Krejsler 2016). This chapter thus constitutes an encouragement to engage in the battle about what public good may be and become in terms of defining what works in public services , recognizing the particular challenges posed by the advent of the evidence discourse .