Organizational Routines and Theorizing in Large-Scale School Improvement—Exploring a Swedish Design

ABSTRACT

In many countries, we can see a movement to more large-scale and system-wide approaches to school improvement. The aim of this article is to study and theorize on conditions for school actors’ knowledge building within such large-scale school improvement reforms. Using a Swedish large-scale school improvement program, “Collaboration for the Best School Possible” (CBSP), as a case and the concepts of organizational routines and theorizing as theoretical lenses, we explore questions linked to school improvement design and conditions for school actors’ knowledge building. Internal policy documents from the Swedish National Agency of Education and interviews with responsible officials constitute the empirical material for the study. The results point to a great variation concerning the degree of standardization and formalization of the CBSP design. This implies that some parts of the CBSP process are to a large degree conditional on individuals’ competences, experiences, and attitudes. An important conclusion from the study is that an efficient large-scale school improvement design should contain organizational routines that can reinforce the interplay between the different levels of the school systems, but also create conditions for integrating research to nurture different types of theorizing. Such design will generate conditions for professional knowledge building and educational change.

Introduction

In school systems around the world, there has been an increasing focus on pupils’ academic achievement and school quality. In light of this movement, much attention has been drawn to the issues of effective school reform and school improvement. Substantial resources, not least time and money, are invested in the hope of improving schools and pupil achievement (Desimone, 2009; Kennedy, 2016). That is, school improvement and effectiveness have become highly political and have currently become an important issue at all levels in the school system, from policymakers to school actors at the individual schools as well as for the school improvement research field. Due to the experience of decades of school reform efforts, in combination with insights from an expanding school improvement research field, some researchers have emphasized that when it comes to changing and improving the educational system, working only with one school at a time does not seem to be effective (Harris & Dinham, 2011). Accordingly, a more systemic approach to school improvement is desirable, which implies that all organizational levels within the school system (i.e., the national, district, and school levels) should be taken into account with the aim of making them work and move in the same direction. From this perspective, it is through the interplay between these levels that the conditions for sustainable and long-term educational change will be generated (Fullan, 2009; Hopkins et al., 2014).

However, previous research also centers on the recurring challenges of large-scale school improvement (LSSI) and whether system-wide approaches are actually a “panacea” for school improvement. For instance, by examining US school reforms, Cohen and Mehta (2017) identified some key factors for (un)successful reforms. For example, in their exploration, they found that failing reforms tend to offer solutions to problems that educators do not think they have in their daily practice. Other reforms fail because they do not provide a profound infrastructure comprising tools, materials, and practical guidance that educators need to implement reform ideas. Departing from the British and Canadian contexts, Levin’s (2010) work also reveals that large-scale change involves the true challenge of changing a large number of schools and classrooms on a sustained basis, which, for example, addresses important issues regarding resources, competency, and professional development. It also becomes important to consider the bureaucratic challenge of improving connections among different areas of social policy to improve students’ results. These and other examples show that educational reforming and LSSI have been, and still are, a complex quest for educational professionals (cf. Cuban, 1990, 2013).

Without overlooking these and other results, this article argues that there is a need to further explore the concept and components of LSSI in various contexts. In particular, former research has revealed that we have limited knowledge, empirically and theoretically, of how organizers of LSSI programs can plan for and design programs to enhance individual and organizational knowledge-building processes (Brix, 2019; Easterby-Smith & Lyles, 2011; Ertås & Irgens, 2021). The current article aims to contribute to this question.

This article draws attention to an LSSI program in Sweden, “Collaboration for the Best School Possible” (CBSP) (Samverkan för bästa skola in Swedish), which is subsequently detailed. The CBSP program started in 2015 and stands out as a relevant example of an LSSI reform from both national and international perspectives. In short, the idea behind the program is that Local Education Authorities (LEAs) and single schools that have received serious critique from the Swedish Schools’ Inspectorate are offered support from the Swedish National Agency for Education (SNAE) and participating universities to improve their schools’ results. Moreover, and what makes this program of particular interest, besides its systemic school improvement approach, is the program’s research-based approach in terms of strengthening the relationship between academic (theoretical) knowledge and local professional (practical) knowledge. By analyzing documents from the SNAE detailing the operationalization of the CBSP program and interviews conducted with directors of education from the same National Agency, and using the concepts of organizational routines and theorizing as theoretical lenses, we explored questions linked to school improvement design and the conditions for individual and organizational knowledge building. The aim of this article is to explore the design of the CBSP program and its conditions for different forms of knowledge-building processes. Two research questions (RQs) directed the analytical work:

RQ1:

How is the CBSP program designed, and what conditions for knowledge building does it create?

RQ2:

What possibilities and limitations in terms of LSSI can be distinguished in the design of the CBSP program?

The article is structured as follows. First, the background section sets the focus of the article from a historical perspective and introduces the Swedish school system. This section also presents the CBSP program. The theoretical starting points are then detailed. Thereafter, the materials and methods are presented, followed by a presentation of the results. The article ends with a discussion and some conclusions, including directives for further research.

Background

From a historical point of view, large-scale school reforming is not a new phenomenon (Fullan, 2000). In a North American reform context, large-scale improvement reforms were first attempted in the 1960s. However, these reform efforts had, according to Fullan (2000), a modest impact, mainly because the reform advocates often ignored issues of implementation in addition to not considering local institutions and cultures (see also Fullan, 2001). In the 1990s, large-scale improvement reforming re-occurred, with greater attention being paid to implementation strategies and with a growing sense of the importance of current reform work. Fullan (2009) has also pointed out that since then, large-scale or whole-system reform policies and strategies have become increasingly evident.

The Swedish Policy Context and the Initiation of the CBSP Program

Former research also details that national, district, and local contexts are essential for understanding large-scale reforming and successful school improvement (Hargreaves & Shirley, 2009; Seashore Louis et al., 2008). In this sense, Sweden is often described as a notable example of educational reform and governance in terms of decentralization and marketization (Blossing et al., 2014; Lundahl, 2002). During the 1990s, the municipalities in Sweden were given increased responsibility as LEAs, becoming more directly responsible and accountable for their schools, which was a significant break from the former school system in which LEAs and principals received detailed directives from state authorities. Akin to parallel international movements, for example, in the United States, in which ideas of greater school autonomy, site-based management and distributed leadership gained increased attention (see e.g., Ortiz & Ogawa, 2000; J. P. Spillane & Healey, 2010), the LEAs and local school pThe working process of the CBSP programrincipals received increased autonomy. Analogously, and essentially, they also became more directly responsible for school results and school improvement work.

However, years of declining student achievement and deficient equality between schools have spurred a re-centralization trend in Sweden, where the national level has gradually tried to take greater control over school outcomes (Adolfsson, 2018; Adolfsson & Alvunger, 2020). In addition, this increased national re-regulation has also been legitimized by several Organisation for Economic Co-operation and Development (OECD) reports (i.e., Organisation for Economic Co-operation and Development [OECD], 2014) that identified major weaknesses in the existing Swedish school system concerning, for example, the division of roles and responsibilities. For example, the unclear division of roles and responsibilities between the state, municipal, and local school levels has evoked questions regarding school system governance and municipal and school leadership. Another strand of criticism has addressed the fact that the market-oriented school system disfavors students from socioeconomically disadvantaged areas and/or without academic traditions at home. Consequently, in light of this policy movement of re-centralization, the Swedish government implemented a number of reforms and policy initiatives, including i) the establishment of the Swedish Schools Inspectorate in 2008 with the aim of strengthening the national audit and monitoring of schools; ii) the reformation of the Education Act in 2010, emphasizing how local authorities and principals shared responsibility for educational quality, student achievement, and equality; iii) the initiation of a number of national professional development programs; and iv) a national curriculum for compulsory and upper secondary schooling with strengthened national knowledge standards, assessment criteria, and a predefined knowledge corpus (Author, 2018; Ronnberg, 2011; Sundberg & Wahlstrom, 2012). Aside from these reforms, it is also essential to highlight that there is a considerable amount of state funding for which LEAs can apply. In 2021, the total amount was SEK 6.3 billion (about 63 million US dollars) (SNAE, 2021b). Arguably, state funding of this amount becomes a strategy for governing LEAs and schools in a desirable direction.

Accordingly, it is within this policy context that the initiation of the national LSSI CBSP program should be understood. The CBSP program started in 2015 (U2015/3357/S) and stands out as a somewhat unique example of an LSSI program from both national and international perspectives (cf. Harris & Hargreaves, 2012). As noted, LEAs and individual schools that have received serious critique from the Swedish Schools Inspectorate are invited to join a multiannual national school improvement program in collaboration with the SNAE and a university. It is voluntary for the school organizers and the schools to participate in the program. However, in light of the critique from the national inspectorate in combination with the extensive resources that come with the program, it is very rare for school organizers to dismiss the effort. The general policy idea behind the CBSP can be described as a national effort to level out the unequal conditions that exist in the Swedish decentralized school system regarding differences between schools’ and school organizers’ resources, educational quality, and capacity. At the beginning of 2022, the program involved school actors from approximately 250 schools, 50 process leaders at the National Agency for Education, and 130 educators from Swedish universities. The financial resources were close to SEK 100 million (about 10 million US dollars) annually (Swedish National Agency for Education [SNAE], 2019).

The working process of the CBSP program can be divided into several steps or phases. These phases are the focus of this study and will be detailed further below. However, before that, the theoretical starting points of the analysis will be discussed.

Theoretical Points of Departure

The focus of the study is on the design of the CBSP program and what conditions of knowledge building become possible or not through this program design. One way to analytically understand and study the design of this school improvement program is through the concept of organizational routines (Author, 2018; Becker, 2004; Feldman & Pentland, 2003; J. Spillane et al., 2011). Using Feldman and Pentland’s words (2003), an organizational routine includes an organization’s structures, regularities, and actions in terms of “a repetitive, recognizable pattern of interdependent actions, involving multiple actors” (p. 311). In the research literature, two different aspects of an organizational routine can be distinguished. With the ostensive aspect of a routine, the focus is on the formal and schematic part of an organization – that is, the ideal or intended way in which an organization should operate. The aim of ostensive routines is to generate regularity and predictability within an organization and create the conditions for desirable processes and activities. The performative aspect of an organizational routine refers to the actions, interactions, and activities that occur among people within the current organization and in relation to the existing ostensive routines: “specific actions, by specific people, in specific places and at specific times. It is the routine in practice” (Feldman & Pentland, 2003, p. 101). In other words, the performative aspect of a routine can be described as comprising the “living part” of an organization. Consequently, the concept of organizational routines enables an analysis of the patterns of formal structures, regularities, and intended activities that the CBSP program is composed of. Based on the aim of the article, the analytical focus is primarily directed at the CBSP program’s ostensive routines regarding the program’s formal structures and actions related to the school actors’ knowledge-building processes.

A way to analytically understand and explore different forms of the knowledge-building process is through the concept of “theorizing.” Inspired by the work of Ertås and Irgens (2021), the concept of theorizing is defined in the article as the process through which new theories and professional experiences are transformed into new knowledge by educational professionals on an individual or collective level. Using the concept of “theorizing” instead of “theory,” the focus is directed toward the knowledge-building process rather than toward understanding theory as something fixed and static. From such an understanding, three degrees or aspects of theorizing can be distinguished, here termed “T1,” “T2,” and “T3.”

With inspiration from Ertås and Irgens (2021), the T3 degree refers to what people often associate with theory, something abstract and less context-specific that scholars and researchers produce, and which is often found aggregated at a generic level in, for example, textbooks and scientific articles. In this article, T3 is therefore understood as and termed “research-based theorizing.” These theories are normally formulated and presented more explicitly than the T1 and T2 theories, which are more embedded in practice (cf. Weick, 1989). That is, from this perspective, attention is directed to how scientific and less context-dependent knowledge is operationalized in school improvement practices, including if and how school improvement programs are conducted using scientific-based methods and whether the work is systematized and properly evaluated.

Regarding the T2 level, this level of theorizing is associated more with individual and collective proven beliefs and knowledge within a specific organization. These are often formulated in one way or another. Individual and collective theorizing in T2 can, for example, result in espoused theories in the form of plans, charts, rules, procedures, standard value documents, and programs, as well as less formalized ways to describe “how we do things here.” In this article, this aspect of theorizing is termed and understood as “professional practice-based theorizing” – that is, knowledge that is produced, shared, and communicated among local school professionals. This comprises the development and formulation of local concepts and strategies for effective school improvement, the evaluation and assessment of quality, various types of exchanges of experience, professional dialogs, and so forth. Consequently, this type of theorizing is linked to local conditions and to knowledge established within a specific municipal and/or school context. Lastly, the T1 level is, according to Ertås and Irgens (2021), more hidden in what people actually do in daily practice. Accordingly, this kind of personal experience and non-reflected action is characterized as being individual and tacit. In this study, it is termed “context- and experience-based theorizing” and is understood as involving more individual and personal-based experiences compared to the other degrees of theorizing. The knowledge developed in this type of theorizing process, for instance, among principals and teachers, is therefore less formalized and tacit compared to T2 and T3.

Accordingly, the combination of the concepts of organizational routines and theorizing constitutes the theoretical framework of the study. With this framework, the analytical focus is placed on the relationship between the design of the LSSI program and the conditions for different forms of knowledge building this design enables. In the next section, the empirical material of the study will be discussed along with how the analyzes were conducted within this theoretical framework.

Materials and Methods

As noted above, the CBSP program and its working processes can be divided into four different phases or steps. Together, these four phases provide important insights into the design of the CBSP program in terms of theorizing. Going into detail, the first step centers on participating municipalities/independent school organizers and schools analyzing their improvement capacity together with directors of education from the National Agency for Education and universities to identify strengths and shortcomings (hereafter termed “the analysis phase”). The second phase in focus is the step in which participating municipalities/independent school owners and the schools, based on the analysis that has been undertaken, decide on and formulate which improvement interventions and activities will be carried out (hereafter termed “the formulation phase”). The third phase is the so-called improvement phase. This is the phase in which the improvement interventions and activities that have been decided upon are implemented and undertaken in collaboration with one or several universities. The fourth and final phase is the evaluation phase, in which the outcomes of the improvement efforts that were undertaken are summarized, evaluated, and formulated in a number of reports (hereafter termed “the evaluation phase”). The analysis below is structured according to these four phases, focusing on how the CBSP program is designed to generate the conditions for theorizing in terms of T1, T2, and T3.

To gain a deeper understanding of the CBSP program and its four phases, a request was sent to the National Agency to gain access to documents detailing how the CBSP process is formalized and operationalized. After a dialogue with the agency, 16 documents (n  = 16) were sent to the research group. These documents were of several different types. The first type of document (n  = 12) had the character of internal working material (drafts, meeting notes, plans, figures, images, etc.). The documents in this category together offered background facts on how the CBSP program and its organizational routines were established. They also gave the research group valuable input for formulating suitable interview questions. The second type of document (n  = 4) was characterized as being more formalized and public. This implies that this type of document was, to a larger degree, available for and used by the directors of education at the Agency, the participating school organizers, and the individual schools in the CBSP process. In this sense, the documents in this category could be described as the formalized results of the internal work conducted by the Agency. The second type of document that was explicitly integrated into the analysis is detailed in Table 1.

Table 1. Analyzed documents.

Name:	Referred to as:	Character:
“National agreement between Swedish National Agency for Education and X university”	Swedish National Agency for Education (2018)	A document/an example of an agreement between the National Agency for Education and one university.
“Working logic for the government assignment ‘Collaboration for the Best School Possible’”	Swedish National Agency for Education (2021a)	Descriptions and details on the work processes between the National Agency for Education, participating school organizers, and universities.
“Process description: To carry out targeted interventions”	Swedish National Agency for Education (2021b)	Descriptions and details of how the improvement work should be carried out.
“Action plan”	Swedish National Agency for Education (2022)	The basis for the Swedish National Agency for Education’s decision on which initiatives are carried out. A support mechanism for systematically planning, following up, and developing the planned work.

In addition, a sample of five (n  = 5) directors of education at the SNAE were interviewed. They were all leaders of the four working teams that organized and conducted the CBSP program. The interviewed directors, who were strategically sampled, had extensive experience of working with the CBSP program in designing the collaborations and working processes between the Agency, participating municipalities or independent school organizers, schools, and universities. Based on their key position at the Agency, they also had many years of experience working directly under the ministry level, interpreting the political intentions of the program and translating these intentions into concrete school improvement programs and strategies (see also Nordholm & Adolfsson, 2024). Thus, having access to these informants who, independently and together, all had a quite autonomous role in the CBSP program, offered unique data on the interface between two levels of the education system that hitherto has received limited attention in Sweden and internationally.

It is also worth noting that the two authors of the article have worked on the CBSP program in different municipalities and, therefore, have had discussions with several interviewees over the years. This probably had a positive impact in terms of gaining access to the field and collecting data. It is also worth noting, however, that the researchers have never been employed by the Agency or involved in CBSP project design as such, or in project governance, which is the main focus of this analysis.

The Analysis

The sampled documents and transcribed interviews were imported into software for qualitative data analysis (InVivo 20). In terms of organizational routines, the policy documents enabled an analysis of the formal structures, routines, and activities that have been developed by the SNAE with the aim of framing and regulating the CBSP process. In addition, the interviews gave us a deepened understanding of these routines and important insights into the activities and actions that occur within them, as well as how the officials at the SNAE perceive the current design in terms of benefits and problems.

In the second step, and based on the concept of theorizing, an analysis instrument was developed. The three aspects of theorizing were applied as deductive categories throughout the analysis process. Consequently, a number of inclusion criteria for each aspect had to be defined. The T3 degree of theorizing, in this article termed “research-based theorizing,” includes how the CBSP design integrates with and capitalizes on more generic (academic) results and strategies throughout the four steps. This category also comprises whether the joint work undertaken is systematized and properly evaluated and problematized, as well as to what degree researchers with adequate knowledge and experience are involved and facilitate dialogs with local professionals. Regarding the inclusion criteria of the second category of theorizing, here termed “professional practice-based theorizing,” these comprise the locally produced and shared experiences and knowledge among local school professionals, often materialized in the form of school development plans, charts, rules, procedures, and standards. Regarding the inclusion criteria for the T1 degree, here termed “context- and experience-based theorizing,” these focus on how the CBSP design considers context-dependent and tacit-oriented knowledge developed by local professionals in their daily work, often individually. Compared to T2, this form of theorizing is often not formalized and verbalized and, therefore, not collectively shared in the same way.

The results presentation below is structured according to the two steps of the analytical work. In the first section, the design of the four CBSP phases is explored and presented through the lens of organizational routines. In the second section, the focus is on the conditions for knowledge building in the CBSP process. Accordingly, in this section, the results from the first step are analyzed in light of the three aspects of theorizing (T1, T2, and T3).

The Design of the CBSP: Prominent Routines and Processes

The Analysis Phase

The study of the internal policy documents linked to the CBSP process in combination with the interviews revealed clarifying details on how the work in the CBSP program is designed and how the ostensive aspects of the routines are formulated in this initial phase. For instance, the “operational logic” on which the CBSP program should rest and what the different actors should do in the different phases are described (Swedish National Agency for Education [SNAE], 2021). The documents detail how the first operating logic was decided in 2016. Since then, the assignment has changed at some point. Consequently, the need for a more specific and clear operational logic has gradually emerged to better structure and support the employees at the SNAE in the initial analysis phase, as well as for the employees who are responsible for the evaluation of the CBSP program: “The purpose of the operational logic is that it should provide guidance in the work of providing school organizers with support within the CBSP” (SNAE, 2021a, p. 3).

Looking further into the ostensive routines and ideas formulated during the analysis phase, the policy documents describe how the SNAE shall initiate a dialogue with participating school organizers about the school’s results and development prospects (SNAE, 2021b). In the analysis, concepts such as “process support” and “dialogue-based working methods” reappear and constitute central routines linked to the analysis process. Specifically, the document describes how the SNAE shall support participating school organizers in analyzing their schools’ results with the aim of identifying significant improvement areas and, in the next step, formulating these areas and needs into concrete improvement activities. The routines linked to the analysis phase further reveal that the school organizer and the Agency’s officers shall carry out this analysis process of the schools’ results and the formulation of the school improvement plans together:

Targeted efforts are a way of working based on dialogue and collaboration. The work includes a selection of school organizers and units, declarations of intent, agreements, decisions on interventions based on the school organizers’ current situation, analyses, and improvement plans. Improvement plans are formed in dialogue and cooperation between the SNAE and the school organizers and participants. (SNAE, 2021)

The formal documents also describe the roles and functions of the various actors in the analysis phase. The “process supporters” from the SNAE are, for instance, described as central persons in this phase. The policy documents state that the work they conduct, together with the school actors and linked to the analysis phase, must be based on research and proven experience.

The documents also describe the expectations and demands directed toward the participating municipalities, principals, and other school actors. For example, they are all required to undergo the obligatory analysis phase and take an active role and responsibility in subsequent improvement interventions. All of these commitments are formalized in an agreement between the National Agency and the participating school organizers and schools. However, it is worth noting that this agreement does not describe the roles of the participating universities. Consequently, due to the way in which the CBPS is designed, the initial analysis phase seems to be an issue only for the SNAE to handle, together with the municipalities and the schools, with little or no involvement of the universities.

Regarding the actions and activities that occur within the existing ostensive routines – in this case, the analysis phase – the results generally point to an accordance between the directives in the documents and how the CBPS program is realized. For instance, Alex, from the SNAE, explained it as follows:

The government directive is very clearly formulated, and it is written that we must work in dialogue, and that has been our main strength, and we really do that. We are not some kind of control (agency) that comes out because it is really a form of dialogue throughout, and we have devoted a lot of time to developing our skills within that.

The interviewed employees at the SNAE described how, at the same time, they had a great amount of autonomy when it came to designing and implementing this LSSI program, including the above-described analysis phase. However, this autonomy implied that they often had to “lay the rails while simultaneously driving the train.” One of the representatives of the National Agency emphasized that this implied that the routines and processes within the CBPS program had been constantly developed since 2015, when the CBPS was initiated. In this process, the formal documents and directives have contributed to pointing out the direction and also creating common sense within the National Agency. At the same time, the SNAE officials stated that these formalized documents can be used quite pragmatically in relation to specific contexts.

To sum up, the analysis of this first phase of the CBSP design points to how, in the way in which the routines were formalized, the officials at the SNAE had a high degree of autonomy linked to the way the analysis phase was carried out. However, dialogue seemed to be the dominant way of working in this phase among the SNAE officials, which was also in accordance with the CBSP directives. Furthermore, although the formal directives emphasized that the CBSP should be based on research, the universities only had a limited role in this part of the CBSP process.

The Formulation Phase

In the formulation phase of the CBSP process, the outcomes from the analyzes conducted in the first phase should be formulated into concrete school improvement strategies. In the same way as in the first phase, there were clear formulations regarding working processes, tasks, and expectations of the school organizers and the SNAE. For instance, the routines concerning the formulation of concrete improvement plans and strategies are described in detail:

The improvement plan is the basis for the Swedish National Agency for Education’s decision on which initiatives are carried out within the framework of the CBSP task. The improvement plan is also a support for the systematic planning, follow up, and development of efforts throughout the collaboration period. … It is the school organizer who is responsible for developing improvement plans based on the current analyses.(Working material, SNAE, 2022)

Accordingly, the ostensive aspect of the program design clarifies that an “improvement plan” has to be formulated where the current improvement program’s strategies, objectives, responsibilities, participants, and resources are specified. The improvement plan should also contain a timetable for each effort and note when various costs will be charged to appropriations (see e.g., SNAE, 2021b). In a similar way as in the first phase, dialogue between the SNAE, the school organizers, and the schools is strongly emphasized: “The sub-process clarifies the Swedish National Agency for Education’s responsibilities and commitments, and the working method is based on dialogue and cooperation” (SNAE, 2021).

Even if the universities have no formal roles or obligations in the formulation phase, the documents now start to address participating universities in their forthcoming contributions. For instance, in the national agreement (see SNAE, 2018), it is stated that the universities, in addition to fulfilling the joint commitments, are expected to “act in accordance with what is stated in the improvement plan” (p. 2) and to “ensure that the necessary resources and competences are available at the current time” (p. 2) Accordingly, the improvement plan constitutes the basis for decisions concerning suitable improvement interventions (SNAE, 2021b, process description).

However, when it comes to the performative aspects of the routines linked to the formulation phase, all four respondents underlined the importance of a dialogic approach, but regarding a direct question on who has the main responsibility and the task of formulating the improvement plan, Alex, one of the respondents, answered, “It must be the school organizer. But my experience is that we want them to feel secure in how they select and prioritize. Then, we need to support and challenge them.”

Even when it comes to the roles and functions of the participating universities, there is a correspondence between what was expressed in the Agency’s documents (ostensive aspect) and how this, according to SNAE officials, is then carried out. Blake, for example, described it as follows:

The university should present itself and then join in and listen, and create its own map based on its own perspectives. And, thus, not being active at all in the first stage. It is a win for the university to be able to participate and listen and to see more and more, and not start its efforts when we are in the analysis phase. So that, before the improvement plan begins, they should participate and reflect but not contribute with any research yet.

To sum up this second phase of the CBSP process – the analysis of the policy documents and the interviews – there is a rather clear and detailed design even in this second phase. In line with the ostensive routines linked to the formulation phase, universities are not expected – nor is it desirable for them – to contribute to the formulation phase in an active way. However, the representatives from the National Agency stressed the importance of participating in this phase because it is supposed to give the universities important insights for the ongoing process.

The Improvement Phase

The third phase of the CBSP process is called the improvement phase. It is in this phase of the program that the school organizers and the schools, with support from the universities, work for almost two years to meet the identified development needs in line with the action plans. The results from the analysis of the policy documents indicate a less detailed design and fewer formalized routines regarding how the current improvement work should be organized and carried out compared to the previous two phases. The SNAE officials continue to facilitate the improvement process, but the documents state that the Agency should now take a step back in this phase. At the same time, the university, together with the school organizer, gains more responsibility over the process and the content. That is, the universities are expected to undertake their work with a starting point in the improvement plan and involving the analysis conducted in the first phase. However, the National Agency, in this phase, acts more in the background in terms of supporting the process, but it also controls the improvement work so that it is in line with the formulized improvement plan: “The agency ensures that the efforts respond to the needs of the target groups and that the agency’s priorities are well-founded, while the work is resource efficient” (SNAE, 2021b).

Regarding the role of the universities, the documents point to the importance of good and well-functioning collaboration between the universities and the National Agency (SNAE, 2018). Yet no specific routines are mentioned. However, when it comes to collaboration between the universities and the participating school organizers, the routines are more explicit. One of the university’s main tasks, according to the routines, is to provide research-based support linked to the school organizers’ and the school’s improvement work. It is also made clear that universities may not use subcontractors without the approval of the SNAE – this is because their research-based competencies must be assessed by the National Education Agency. Besides this, the universities are also obligated to participate in different learning seminars with the National Agency, the participating school organizers, principals, and teachers. With a focus on the current ongoing program, the aim of these seminars is to illuminate changes and results linked to the improvement process and draw conclusions about important knowledge and experiences.

Moving on to how the ostensive aspects of the routines condition practice, all respondents described a shift where the National Agency took a step back and the universities, to a greater degree, took on more responsibility for the process. Some also described this shift as a central but also somewhat challenging stage in the CBSP process, here described by Alex:

Well, and this is both the good and the difficult part of it, because here we lose control. We buy services from you [the university] through an agreement, yes, based on an estimation and agreement. … But then we can’t know what happens.

Even if the National Agency generally took a step back, the directors of education still explained how they had a responsibility to act if problems arose in the process or in collaboration between the university and the school organizer. Blake described this in terms of the following:

I put 100 percent confidence in the university that they are doing what they are supposed to do. And if it doesn’t work, I will be informed, and then I have to do something.

Summing up this third phase, the results show that the third improvement phase is characterized as much less formalized and with routines that are less detailed compared to the previous two phases. However, the results point to coherence between the design intentions and how the program plays out in practice. Roughly speaking, the SNAE takes a step back, and the universities take on more of a key role when it comes to working with the improvement plan together with the school organizers, principals, and teachers. In addition, even if the policy documents reveal what skills and knowledge universities are expected to support school organizers with, few details are provided regarding how this collaboration is to be organized.

The Evaluation Phase

In the fourth and final phase of the CBSP process, the document analysis details how different types of evaluation are expected from the different partners involved in the CBSP program. The overall purpose of this phase is expressed as follows:

The process aims to deal with lessons learned, questions, and improvement needs that have arisen in order to improve working methods and support development work both internally and externally. The process is initiated by the need to follow up on the working method for continued development of the Swedish National Agency for Education’s common working methods and processes. (SNAE, 2021b, p. 6)

After finalizing the task, participating universities must deliver a final report (see e.g., SNAE, 2018). The report must contain a description of how the research and development support has contributed to the improvement work undertaken, how activities have been supported by current research, and how the interventions have been research-based. The final report is also expected to describe what activities and strategies have been established to follow up on, systematize, and share the knowledge that the participating universities, school organizers, and the SNAE have acquired through the phases of the CBSP program. School organizers and the SNAE are also required to write reports in which the goals and results of the improvement efforts undertaken within the CBSP program are evaluated. In this step, the National Agency and the school organizers work closer together again, generally without formal dialogs with the universities.

The strategies and routines formulated in the documents and the directors at the National Agency described the work process in the evaluation phase in a similar way. They, for instance, described a rather structured and systematized way in which the work in this phase should be carried out:

There are given steps in the process with regard to the fact that we have a closing seminar. They (the school organizers) must write a final report and they must have feedback. … What I can see in our team is that we have developed—to think together how to do this. We try different ways of evaluating or using dialogue models, and then they get to decide how they want to do it. [I think] we think we should pay more attention to this phase. (Jordan)

As the quote above indicates, the directors from the SNAE described how they tried different approaches and, once again, laid the rails while the train was running. At the same time, they stressed that this evaluation phase needed to be developed further. For example, some of the respondents emphasized that this phase could have more of a learning focus rather than a strict evaluation focus.

Summing up the final phase of the CBSP process – the analysis of the documents and the interviews – points to the quite clear routines and expectations of the different actors involved. These routines and expectations, expressed in the ostensive aspect of the design, were also performed by the different actors. In addition, the analysis showed that the SNAE occupies a more prominent role by again working in dialogue with school organizers, while the universities evaluated their actions and interventions quite separately, which is highlighted as a potential shortcoming by some of the directors.

Theorizing for Individual and Organizational Knowledge Building

Based on the identified organizational routines linked to the CBSP process’s different phases, the next step of the analysis aims to explore how the prevailing design conditions enable different forms of knowledge building in the CBSP program. The results presentation of this step is structured based on the three degrees of theorizing while still addressing the four phases of the work process of the CBSP program.

Context- and Experience-Based Theorizing

As discussed in the introduction to this article, this type of knowledge building focuses on how the CBSP design considers context-dependent and often tacit knowledge developed by local professionals in their everyday work.

As shown in the previous results section, a key idea in all phases of the CBSP program is that collaboration should be based on dialogue and reciprocity between the program’s different participating actors. In accordance with this, the Agency strives to create different platforms and methods to involve LEA officials, principals, and individual teachers. Another key idea is that participating LEA officials, principals, and teachers must “own” their own improvement processes and that they are the ones who must take the initiative and drive the school improvement process forward:

We constantly need to put them in the driver’s seat. It is so different, depending on who we meet. From having very little competence and no control over the results at all to giant organizations where there is plenty of competence. … We need to constantly have a dialogue with, for example, principals and managers, so that we don’t take over the hosting. (Jordan)

Given this quote, the overall design aims to create the conditions and back up for T1 theorizing. To encourage dialogue and to be able to make use of the knowledge that local professionals possess, the SNAE has developed an “arsenal” of different dialogue models. Due to government directives, all respondents explained that this way of working was not negotiable. Instead, it constitutes the basis of the program:

Yes, we try our hand at different dialogue models, but it is very clear that we cannot work without dialogue models. We do not have talk meetings as we say; they are not rocking chair meetings, but really workshops where they work. We give them different models to get everyone involved in the dialogue. (Charlie)

However, even if dialogue and the importance of meeting on an “equal level” reappeared in the interviews, the analysis also revealed inherent challenges in creating conditions for this type of theorizing. Given the fact that individual school organizers and schools have previously received substantial criticism from the Swedish Schools Inspectorate, it is worth considering the conditions for initiating and implementing these dialogs. That is, the school organizers and schools are often under strong external pressure from the national level, which in itself comprises an inequality in relation to the SNAE. Alex’s statement demonstrates these challenges:

[They are] selected because they have poor results and they have had them for a long time. … And often they realize themselves or discover along the way that they need help at all levels and they get the systematics in this with a follow-up and they get help with what is an analysis.

In addition, the results indicate that this type of theorizing primarily appears in the third phase of the CBSP process – that is, the improvement phase. It is worth noting here that the SNAE, in this phase, takes a step back and that the dialogue is conducted between the school organizer and the participating universities, which are obliged to work together in line with the so-called improvement plan. Overall, the directors of the National Agency stated that the work in this phase mostly turned out well. At the same time, several examples were given that this was not always the case. For instance, the respondents emphasized that an important condition for obtaining an effective improvement process in phase three was that the representatives of the universities not only had profound academic knowledge but also knowledge about school improvement in practice:

I think that the quality of those who come in and support the improvement work in this phase from the universities varies. … You need to have a combination of a practitioner and research background when you come in and support. Because I believe that, yes, someone who problematizes and thinks, and asks the right questions. (Charlie)

Accordingly, the CBSP design with its organizational routines seems to characterize both the strengths and challenges concerning creating conditions for the T1 level of theorizing. However, a key factor seems to be the dual competencies of the educators from the universities in terms of their academic research competence in combination with practical experiences and knowledge of the school practice’s conditions and work. However, as will be discussed in the next section, these conditions for theorizing seem to have an intimate connection with the subsequent T2 level.

Professional Practice-Based Theorizing

Regarding the T2 degree of theorizing, the study focused explicitly on knowledge developed in the mutual and formalized work conducted at the local level before and during the CBSP program. As noted, such extracts detail how the work process intends to generate concrete plans, charts, rules, procedures, and standards that could support schools’ improvement processes. In that sense, this category focuses on collective and mutual work processes and formalized results through the four phases of the program.

Arguably, the analysis becomes extra important when it comes to this type of theorizing. In this phase, the school actors develop insights and a deepened understanding of the school’s results, educational quality, and improvement needs. In addition, the “improvement plan,” which then structures and guides the subsequent improvement work, is written based on this new understanding and insights. Consequently, it becomes important to see how common and/or collective knowledge, manifested in formulated plans, templates, schedules, and so forth, is created and plays out in the CBSP program. The four directors of education answered somewhat differently and sometimes vaguely regarding the question of how the knowledge that was already formalized in the local context was used and integrated into the formalized work in the CBSP program. However, once again, dialogue and critical-reflective conversation were emphasized as a fundament of the improvement work, as exemplified by Blake:

My role is to help the school organizer capture school activities in a different way than they are used to. … We work a lot with documents. Is this a basis that we can use in the analysis process? Is it possible to use this? To analyze? That, I think, is my mission—not to go in and audit their business, but to support them in doing so, with their own business.

In addition, the interviewees indicated some differences regarding the approach to formalized knowledge that had been developed at the local level. One position is exemplified by Blake above, which is characterized more by a problematizing and “critical friend” view. Another somewhat different position that emerged indicated that the directors, instead, took a more operational and leading role, here explained by Alex describing a school organizer with whom they had just started collaborating:

So, we thought that we need to “poke holes” in them [their quality reports] quite quickly, because here we got a quality report that talked about how difficult everything is, how difficult the conditions are. … So, we will, in a humble way, work our way away from them. … And if they use models and matrices, sure, if they fit like that, but no, if they ask us to look at them, we do it, but they rarely do that.

Accordingly, the CBSP design seems to be open to differences in how to handle formalized context-specific knowledge. Another challenge in the current design raised by the interviewees concerns the link between the locally formalized knowledge, for example, the plans formulated by the school organizers (with support from the SNAE) and the more theoretical knowledge provided by the universities. The analyzes of the interviews point to how culture clashes and friction could sometimes arise in this meeting, for example, when it came to questions concerning how the improvement plan and its efforts should be formulated and implemented:

We [the Agency for Education and school organizers] are more generalists, and you are specialists. The challenge I can sometimes experience is that, depending on which person represents the university, the big challenge is to make the [theoretical] knowledge available. To actually make it accessible so that it actually leads to actions as well. … Some are skilled at it, but others are not. (Jordan)

That is, the couplings between the research-based knowledge (T3) and the professional-based knowledge (T2) tended to be quite weak sometimes, which implied that the desirable connection between the two knowledge forms did not appear. However, several of the interviewees also emphasized that the collaboration between the three parts had improved over the years and that conflicts occurred less frequently. A contributing factor to this improvement, highlighted in the interviews, was that the universities (sometimes) now entered the CBSP process at an earlier stage and, in that way, gained a greater pre-understanding of the schools’ development needs and conditions and thereby also provided input for the improvement plan.

Research-Based Theorizing

To recapitulate the T3 aspect of theorizing, in this article termed “research-based theorizing,” the analysis focuses on how the CBSP design conditions the possibilities for integrating and capitalizing from more generic (academic) results and strategies throughout the four steps of the process. This category also focuses on whether the joint work undertaken is systematized and properly evaluated and problematized, whether researchers with adequate knowledge and experience are involved in facilitating dialogs with local professionals in each step, and so forth.

The respondents returned several times to some overall and vitally important questions regarding research-based theorizing. To start with, they all reflected on the fact that the universities, in a more active way, first became involved in the process in phase three and whether this should be seen as a strength or a potential disadvantage. A director of the SNAE pointed to some potential problems when the universities did not participate more extensively earlier in the CBSP process. However, some changes can be noted linked to this:

There hasn’t been much [collaboration with the universities] in this [analysis] phase. Now there has been more cooperation with the new improvement agreement … my colleagues and I have been able to link the universities earlier than before and you feel yes, this school or this school organizer needs more support in their analysis. … So there is every chance to start working together earlier. (Charlie)

At the same time, other directors from the National Agency had a different way of reasoning in that they emphasized that all parts of the CBSP program had their specific roles. This was an essential idea behind the program per se. From this point of view, this means that each part might have different roles throughout the CBSP program, which does not necessarily mean that they cannot participate in the process, but that they have different degrees of participation and collaboration, as explained by Blake:

Firstly, it is important that we work together. We need to talk about the roles that we should have. … It is a win for the university to be able to participate and listen and to see more and more, and not start its efforts when we are in the analysis phase … they should participate and reflect but not contribute with any research yet.

That is, despite a quite formalized program design, even here it is possible to notice different approaches among the directors, in this case, regarding the researchers’ role within the CBSP process. Another crucial question related to the topic above concerns the conditions for T3 theorizing. All directors reflected on this topic and described their efforts to ensure that the CBSP process should rest on a research basis. According to the directors from the National Agency, an overall strategy to ensure that the CBSP program is research-based is that the modules that the SNAE have developed are based on high-quality research. These modules constitute a structure and a guide for the CBSP process as well as for the participating universities.

Regarding universities and their researchers, the analysis showed that there were no formal directives detailing the knowledge they were expected to have and expected to contribute. Rather, this knowledge is described as rather implicit in the agreements and contracts between the universities and the Agency. The directors also described how the universities were not always able to provide expertise with exactly the right skills for the specific task, as explained by Alex:

No, there is no template. … I would like to say that it is probably a little more embedded that it is researchers who work or it is like universities that carry out the efforts, and then that in itself is a certainty that there is a scientific basis behind it. … It really depends on which university and which competence this university delivers.

Summing up this final level of theorizing, the analysis of the interviews revealed strengths and potential contradictions in the current design and its organizational routines. Even in this analysis, it was possible to distinguish different approaches and attitudes among the SNAE officials. That is, despite a quite formalized program design, there were, in some cases, different opinions and ideas among the SNAE officials concerning what role the researchers should have in the initial analytical and formulation phases. In other words, to what degree should these phases be informed by more research-based knowledge (T3)? However, when it came to the so-called improvement phase, the universities, in accordance with the program design, took on a more active role, while the National Agency took a big step back. Nevertheless, what seems to be a central condition for the successful integration of research-based knowledge in relation to local contextual knowledge is the competence and experience of the educators from the university. However, the degree of formalization of what knowledge the representatives from the university are expected to have and expected to contribute is very low in the program design.

Based on the results above concerning what different forms of theorizing processes the CBSP design seemed to enable or hinder, in the final section we will discuss these findings in relation to previous research with the aim of deepening our understanding of LSSI and the conditions for individual and organizational knowledge building.

Discussion and Conclusions

Using the CBSP in Sweden as a strategic case and the concepts of organizational routines and theorizing as theoretical lenses, we explored questions linked to school improvement design and the conditions for individual and organizational knowledge building. We argue that these questions are of high relevance not only for this specific case, but for all who are involved in designing and working with LSSI programs. In the following discussion, we will summarize the results from this study and point to some crucial insights.

Considering the large number of participating schools and school organizers and the extensive financial resources, the CBSP program stands out as an interesting example of LSSI. A general insight from the study, given the considerable size of the CBSP program, relates to the noticeable fairly low degree of central standardization and formalization of the program design. Put another way, the ostensive aspects offered few details to the SNAE, to participating school organizers, and to participating universities navigating through the processes. In the policy documents formulated during the initiation of the CBSP, some general directives and an overall structure linked to the CBSP process can be found, but thereafter, the officials from the SNAE had to develop and improve the more specific routines and processes within the CBSP.

However, one formalized aspect was the idea concerning the working process, where “dialogue” was emphasized as a central approach that should pervade the whole CBSP process. That is, dialogue was considered a crucial force for learning and development with the aim of what Harris and Dinham (2011) described as collectively building professional capacity at the schools and with the school organizers. The focus for the dialogue, as well as the improvement work, should be the schools’ and the school organizers’ problem and development needs. Understanding this in the light of the theoretical framework of the study, this implies that T2, as “professional practice-based theorizing,” seemed to be the form of theorizing that the CBSP is primarily designed around. However, an obstacle to realizing this was that the so-called receiving capacity at the current schools and of the school organizers was often very low. In other words, the school organizers and schools that participated in the CBSP were often characterized by a lack of shared and communicated knowledge and competencies, and there was often an absence of necessary infrastructure at the local level. Instead, the knowledge and experience were often more individual, personal-based, and uncommunicative, including an organization that did not enable exchanges of experience and professional knowledge. Consequently, in this kind of school organization, school actors are often not used to, and are sometimes even resistant to, working with dialogical-based forms of work. Accordingly, schools and school organizers with such low “receiving capacity” results in challenges in designing and running large-scale improvement programs in an effective way.

Based on the findings of this article, and in light of the systemic school improvement approach that was discussed initially (Fullan, 2009; Hopkins et al., 2014), it therefore becomes clear that it is not only the interplay between the different levels of school systems that generates the conditions for knowledge building and long-term educational change. It also becomes relevant to consider how this interplay integrates universities to nurture different types of theorizing in and between the different levels of school systems. According to us, the design of the CBSP evokes questions on infrastructure, as raised by Cohen and Mehta (2017), and whether expertise and skills are used in a wise manner to actually support local professionals. More specifically, and as shown above, the universities enter at a rather late stage of the process – in fact, after the analysis is completed and the concrete work is about to start. Consequently, the degree of the quality of this interplay and thus the quality of the improvement phase per se depended on and were mainly conditional on individual factors, for example, to what degree the researcher from the university had the capacity to “translate” and convert the research-based knowledge (T3) in a way that was perceived as relevant and useful in relation to the schools’ needs and current improvement work.

Consequently, there are, according to us, reasons for problematizing this path from several perspectives. One could, for instance, reason about what type of expertise and skills a university does (not) possess. Arguably, most researchers can support LSSI work with analysis and systematization and they can probably also contribute to the design work based on the analyzes that are undertaken. Linked to the CBSP, universities could also contribute important knowledge in the final evaluation phase. Then, and as also shown above, researchers may not necessarily be the most skilled to work with LEAs, principals, and teachers “on the ground.” Thus, this might be a better task for officials and their colleagues at the SNAE, who often have extensive experience of leading school improvement work – that is, fostering T2 skills. A more efficient program design could possibly decrease such vulnerability. For instance, instead of taking quite a passive role in the improvement phase, the officials from the SNAE could be a part of the process in a more active way and work as “knowledge brokers” and facilitators with the aim of strengthening the couplings between T3 and T1. Accordingly, one important finding regarding future LSSI in Sweden and internationally is that if we involve one or several universities in LSSI programs, we must build an infrastructure that carefully considers what researchers should (not) contribute and at what stage of the work process.

From a broader standpoint, it also becomes relevant to set the results in a wider context in terms of educational governance, school leadership, and LSSI. For instance, and as noted above, ideas on school-based management and distributed leadership (see e.g., Ortiz & Ogawa, 2000; J. P. Spillane & Healey, 2010) have influenced school governance both in the Scandinavian and US contexts, which, on central points, assume that local professionals possess the skills and expertise to operate complex issues of school improvement, both in locally initiated projects (i.e. bottom-up) to regionally and/or nationally initiated projects (i.e. top-down) (see also Pont et al., 2008). Given that ideas on decentralization and marketization were introduced in the 1990s in the Swedish context and that we can now see an evident re-centralization of the school system, of which the CBSP is a striking example, we believe that there are important lessons to be learned here as well. More precisely, regardless of the relationship between centralization and decentralization, we believe that the empirical findings and three types of theorizing (T1–T3) offer important tools and perspectives on the knowledge that local professionals must possess or develop, with or without external support from the state level, to foster profound improvement capacity at the local level. This, as well, becomes a key issue to consider when designing LSSI projects in various contexts.

Regarding limitations, the findings of the article are mostly presented from the perspective of SNAE managers, and the sample of documents and interviewees is restricted. Thus, even though the interviewees possessed essential knowledge of the CBSP program and the link to the state level offered a unique perspective, further studies using, for example, questionnaires and/or focus-group interviews could include a larger number of officials involved in the program and achieve a deeper understanding of the state agency level’s role and function in interpreting and operationalizing LSSI. Another option linked to inherent limitations could be to focus more on LEAs, principals, and teachers to obtain important details on how the three levels of theorizing actually play out in practice and whether they are seen as relevant knowledge (cf. Cohen & Mehta, 2017). Lastly, even if this article resonates and extends the theoretical work of Ertås and Irgens (2021), there is still an evident need to refine the theorizing concept and delve deeper into the relationship between the three types of theorizing and how these relationships create the conditions for school improvement.

Acknowledgments

The article is the result of equal collaboration between the authors, who worked on and contributed to the different sections of the article in equal degrees.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Swedish Research Council [grant number 2022-03017].