Contradictory policy demands – perspectives on educational leaders’ shared responsibility for educational improvement

In recent times, educational research has raised particular interest in leaders at the intermediate level between the state and local schools (Paulsen et al., 2014; Prøitz et al., 2021). This is related to changes in approaches to educational improvement (Fullan, 2009), visible in recurrent policy demands that stress the need for actors at all levels to promote student learning and results (Darling-Hammond, 2004; Datnow et al., 2020; Sahlberg, 2007). In Sweden, which is the context of this study, the intermediate level is of special interest, since the chain of command assigns shared responsibility for educational improvement to principals in local schools and to the school organizers, i.e. the local education authorities (LEAs) ^[1] (Skolverket, 2015; Svensk författningssamling [SFS] 2010, p. 800). Swedish national policies also stipulate that principals decide on their school’s internal organization and are responsible for distributing resources according to contextual conditions and student needs. Consequently, there are inherent hierarchical tensions in the policies that those responsible for implementation must deal with. In times of declining results and increasing differences in and between Swedish schools (Organization for Economic Co-operation and Development [OECD], 2023), the importance of shared responsibilities for educational improvement has been further emphasized.

During the early twenty-first century, approaches to educational improvement have changed from bottom-up initiatives based on professional freedom and autonomy to top-down initiatives based on goals and performance targets (Fullan, 2009). Since standardized and test-based strategies have negatively influenced educational leaders to “chase numbers” in fear of sanctions rather than in the hope of improving education (e.g. Datnow et al., 2020), tri-level strategies combining top-down initiatives with bottom-up professional involvement have been promoted. Tri-level strategies have paved the way for educational improvement based on respect for professional knowledge through partnerships between educational leaders at the national and local (school and district) levels (Fullan, 2009; Hargreaves and Shirley, 2009). To improve cultures of learning across organizational levels, LEA managers and principals are recommended to work in partnership and become resources for mutual learning (e.g. Honig, 2012; Hooge et al., 2019; Paulsen and Hjertø, 2019; Rorrer et al., 2008). However, research has also shown that it is difficult for LEA managers and principals to embrace a nonhierarchical partnership since it challenges previous experiences and beliefs (King Smith et al., 2020). This raises interest in how shared responsibility for educational improvement is played out at the intermediate level when LEA managers and principals meet and interact.

Even though Swedish policies stress shared responsibility for educational improvement, there are signs that the increased focus on results and regulations has pushed LEA managers into a controlling governing style toward principals, causing tensions between the two groups (e.g. Adolfsson and Alvunger, 2017, 2020; Ståhlkrantz and Rapp, 2022). This contrasts not only with policies but also with a tradition of a high degree of autonomy for Swedish principals (Nordholm et al., 2023). Educational improvement might be addressed in various ways and principals’ needs might differ, yet Swedish LEAs have tended to offer general support (Arora-Jonsson et al., 2024; Rojas and Wermke, 2025). Research and national reports have shown that in turn this approach has led to difficulties in fulfilling LEA responsibilities, particularly in meeting the needs of principals and schools with severe challenges (Liljenberg et al., 2023; Skolinspektionen, 2021). Since research about relations between principals and managers at the intermediate level is limited (Håkansson and Adolfsson, 2022), we still do not know much about how shared policy responsibility between these educational leaders is understood and enacted in practice.

This study aims to contribute to knowledge about how contradictory policy demands play out in practice in the intersection where educational leaders meet and interact. It uses sensemaking (Weick, 1995) and framing theory (Benford and Snow, 2000) to explore how LEA managers make sense of and frame problems related to national policy of shared responsibility for educational improvement and what this means for policy enactment in local contexts.

In the Swedish goal- and result-oriented school system, the state sets educational goals which locally elected politicians and professionals in the 290 municipalities and various independent school providers are responsible for achieving. With its strong state involvement, municipal autonomy and far-reaching marketization, the Swedish decentralized school system, moving towards recentralization (Wermke et al., 2022), is an example of a complex context for educational leaders and their responsibilities at different hierarchical levels.

Nordholm, Wermke and colleagues (e.g. Nordholm et al., 2023; Wermke et al., 2022) have deepened our understanding of principals’ perceptions of autonomy and control in the Swedish context. Taking a historical perspective, they state that Swedish principals have experienced a high degree of autonomy in their work and, to some extent, still do. However, in relation to the economy, administration and quality work, principals' experienced autonomy has decreased due to various LEA strategies. However, Nordholm et al. (2024) also show that principals in some respects find ways to get around LEA strategies to increase their autonomy. Of particular interest in relation to the focus of this study is their results stating that some principals experience a high degree of autonomy in educational improvement issues and even stress the need for intensified mutual engagement and collaboration between managers at different levels. The need for mutual engagement and shared responsibility in educational improvement is particularly prominent for principals leading schools with severe challenges (Liljenberg et al., 2023).

In contemporary Sweden, principals’ work is characterized by increased workload, external pressure and high turnover, making it challenging, especially for less-experienced principals (Arvidsson et al., 2021; Jerdborg, 2023). In this perspective, the relationship between autonomy and control, as well as the widespread normative presumption that more autonomy is always to the better, needs to be considered carefully (Cribb and Gewirtz, 2007). Although multiple policy efforts have aimed at providing principals greater autonomy, studies have shown mixed effects on organizational outcomes and student results (Kim and Weiner, 2022; Nordholm et al., 2025).

Altogether, this brief presentation of educational leaders’ relations and responsibilities further highlights problems related to the national policy of shared responsibility for educational improvement, which the study takes on.

To analyze how educational leaders make sense of and frame problems related to national demands of shared responsibility for educational improvement, we draw on two bodies of literature: sensemaking (Weick, 1995; Weick et al., 2005) and framing theory (Benford and Snow, 2000). By combining these perspectives, we align with influential researchers’ work on policy implementation and educational leaders’ new work situations (Coburn, 2006; Spillane and Anderson, 2014). In what follows, we present how these perspectives are informing this study.

The contradictory policy demands of shared responsibility for educational improvement and autonomous principals deciding about their local schools can be understood as an ambiguous or confusing situation (Maitlis and Christianson, 2014) for LEA managers to make sense of and handle. According to Weick (1995) and Weick et al. (2005), sensemaking is an ongoing retrospective process in which people make sense of situations through their prior knowledge, experiences, beliefs and values. Sensemaking is linked to identity and embedded in social contexts. People tend to work on autopilot under normal conditions; however, when ambiguous situations such as organizational changes arise (Maitlis and Christianson, 2014), they become conscious and try to create meaning of what is happening. Hence, in this study, the LEA managers construct the meaning of policy demands based on their prior and situated knowledge and beliefs. Since LEA managers can be regarded as key actors in mediating the policy message of shared responsibility being forwarded to the principals, they contribute to shaping the two-way interaction of enactment and hence what the message may imply for principals’ responsibility (cf. Coburn, 2005). When creating meaning from previous experiences, the idea is that people, i.e., in this study, LEA managers, rationalize what has occurred and consider what should be done next. Therefore, a perceived change in the environment is followed by enactment to reduce the ambiguity of events, which in turn opens up for certain kinds of justifications by LEA managers in the study. According to Weick (1988, p. 306), “The term ‘enactment’ is used to preserve the central point that when people act, they bring events and structures into existence and set them in motion.” Thus enactment, the process of making ideas real by acting on them, is seminal in organizational theory and opposed to the idea of policy implementation. Weick (1995) also argued that unless common sensemaking is fostered in the organization, people tend to rely on their previous assumptions and act as they “used to do it”.

By adding framing theory (Benford and Snow, 2000) to a sensemaking perspective, we get a complementary perspective for understanding how LEA managers prioritize and construct the policy problem. It goes beyond the traditional contextual and cognitive emphasis of sensemaking and offers an analytical lens for examining how power relations among various actors influence the construction of policy meaning and outcomes within local contexts. Framing also constitutes a crucial aspect of the meaning-making process, in which certain elements of reality are emphasized to mobilize others toward specific forms of action. In the process, problematic instances in need of change are identified. In our analysis, we use diagnostic and prognostic framing where diagnostic frames define problems, including location and attribution of blame, whereas prognostic frames propose possible solutions and actions (Benford and Snow, 2000; Coburn, 2006).

Therefore, consistent with Coburn (2006, p. 345), “conceptual tools from two distinct yet complementary theoretical traditions: sensemaking theory and frame analysis” are brought together. Adapting Coburn’s ideas about how people frame problems during policy implementation, this study regards LEA managers and principals as policymakers. Policy problems are regarded as social constructions identified by policymakers as problematic, highlighting some aspects while ignoring others (Coburn, 2006). The way policy ideas are enacted in practice varies due to local interpretations. Consequently, how LEA managers diagnostically frame policy problems of shared responsibility for educational improvement (i.e. define and attribute blame) is crucial since it embraces some actions and closes others. However, framing always depends on the response given by the addressed party, which, in this paper, are the principals. According to Coburn (2006), framing activities shape implementation and respondents’ sensemaking and practices, which in this paper relates to LEA managers’ framing in relation to principals’ sensemaking and work. Benford and Snow (2000) defined this part of the framing process as resonance. Alternative understandings of the situation might challenge current framing, meaning that frame disputes will occur and frames will be reconstituted and reframed, creating counter-frames in negotiation and interaction over time.

This qualitative multiple-case study (Yin, 2009) is part of a larger project on how six Swedish municipalities organize and manage quality and improvement work. The six municipalities (A–F) were strategically selected (Flyvbjerg, 2011) based on national data on municipality classification, number of inhabitants and geographic location (Sveriges Kommuner och Regioner [SKR], 2022). Therefore, the selection not only allowed national representativeness regarding municipal school organizers but also variation in organizational structures, such as personal and financial resources. Although the cases are embedded in distinct contexts, the study does not aim to conduct a contextual comparison. Instead, it examines potential differences and similarities across cases to gain a deeper understanding of the policy enactment processes and what it means for practice. In order to protect anonymity, the exact number of inhabitants and geographic location within the country are omitted (Table 1).

The larger project data consists of observations of meetings with LEA managers, LEA administrators and principals (about 100 h). It also includes semi-structured individual interviews with LEA managers (n = 17), LEA administrators (n = 9) and principals (n = 55) responsible for compulsory schools (Table 1). This study draws on the semi-structured interviews. All principals were interviewed in the smaller municipalities (C, F), whereas a selection of principals were interviewed in the larger municipalities (A, B, D, E). All respondents gave written informed consent to participate in the study. Their names are omitted to further protect their anonymity. The interviews addressed responsibility, mandates, previous and present work procedures and organizational relationships. In addition, questions were added based on reflections from reading documentation, making observations and conducting earlier interviews. These questions were incorporated into the interviews to improve their relevance. The interviews lasted 60–75 min and were audio-recorded and transcribed verbatim.

The entire dataset was read several times to ensure an adequate understanding of the respondents’ statements. Next, an explorative analysis using open questions (what, how, who and why?) was conducted to identify specific segments in the data concerning the policy demands of shared responsibilities for educational improvement. These segments were then categorized according to the specific analytical question they addressed for example: What constitutes the shared responsibility for educational improvement? How is it enacted? Who is involved? Why is it pursued in this way? From this point in the analysis, the theoretical constructs of sensemaking and framing (identity, retrospect, enactment, diagnostic and prognostic framing) were explored (Benford and Snow, 2000; Weick, 1995). The sensemaking analysis was inspired by Spillane and Anderson (2014), focusing on (1) roles through “I am” statements describing self, beliefs and motivation; (2) improvement and relations in retrospect in terms of statements describing previous experiences and times; and (3) enactment by capturing their actions. Then, LEA managers’ diagnostic framing and attributing blame were analyzed through the questions “What has happened/what is happening here?” and “Who/what is to blame?” followed by their prognostic framing of policy problems and relations through the question “What is to be done about it?” In this analytical step, the codes emerging from data were clustered based on similarities and differences in how LEA managers framed problems related to contradictory policy demands of shared responsibility (i.e. related to roles, mandate and relations). This generated two broad categories on LEA managers’ prognostic framing labeled (1) responsibility for educational improvement as intertwined across hierarchical levels and (2) responsibility for educational improvement as divided between hierarchical levels. After this, resonance and implications for policy enactment in local contexts were identified by analyzing principals’ responses also revealing counter-frames. Finally, cross-case analysis was conducted (e.g. Miles and Huberman, 1994). Therefore, the coding and analysis can be characterized as both data- and concept-driven (Kvale and Brinkmann, 2009).

This section presents LEA managers’ sensemaking and framing of problems related to the national policy of shared responsibility for educational improvement. Since the policy has inherent hierarchical tensions in relation to principal autonomy, we analyzed LEA managers’ understanding of their roles and mandate in relation to principals’ roles and mandate. From the analysis of roles and relations in retrospect, we identified two overarching ways of framing the problems related to national policy of shared responsibility: either principal autonomy as subordinated to shared responsibility, or principal autonomy as superior to shared responsibility. The following two subsections present a retrospective perspective and how the problem framing is played out in practice. It ends with a description of how frames change due to external pressure. The section concludes with a summary table.

In Municipalities A and D, the LEA level had previously decided on how to enact policies of educational improvement. In retrospect, LEA managers and principals perceived this as a period of short-term and ad hoc initiatives, with one principal describing it as consisting of “day lions and grasshoppers” ^[2]. The LEA managers framed the previous problem with shared responsibility as a dominant but distanced LEA level, lacking in their analysis of both needs and outcomes. Therefore, LEA managers attributed the problem to their own level and renounced the previous hierarchical, uncritical and short-term approach to shared responsibility of educational improvement. Related to the idea that principal autonomy is subordinated to shared responsibility and that LEA managers and principals mutually depend on each other, LEA managers prognostically framed the importance of principals’ perspectives and competencies in the organization. For example, one LEA manager in Municipality A stated that her role was to invite principals by asking them what they could do together and not to overrule them, and also to take advantage of principals’ different perspectives. In Municipality D, one LEA manager emphasized that the only way to establish shared responsibility was to engage in principals’ learning. Following this, the LEA managers framed the policy demands of shared responsibility for educational improvement as intertwined across hierarchical levels. They regarded themselves as both decision-makers and co-creators of educational improvement. This approach clearly contrasted with statements on how shared responsibility had been dealt with in the past.

Hence, LEA managers’ prognostic framing was to pave new ways of working across hierarchical levels. In close collaboration with principals at local schools, they introduced research-based work methods. However, to respect the principals’ local responsibility and mandate, the LEA managers articulated that the principals were not required to adhere to any particular method:

The principals responded by highlighting the strength of mutual collaboration while also appreciating their local responsibilities and mandates:

In both municipalities, the principals perceived the LEA managers’ enactment as supportive, which was especially appreciated by the less-experienced principals:

Framing and enacting shared responsibility in this way increased communication across levels, which for the principals meant more participation, fewer surprises, better adaptation to local needs and predictability. Overall, LEA managers’ enactment of shared responsibility as intertwined across hierarchical levels was framed differently than before, taking previous experiences into account. Even though principals’ autonomy was diagnostically framed as subordinated, when played out in practice, it was still high, which the principals appreciated.

However, the findings also revealed difficulties for LEA managers in adhering to this framing when pressure for improved results increased impatience in the organization. This issue was particularly evident when improvements did not pay off quickly enough in visible results:

Therefore, the pressure of visible results reframed LEA managers’ initial understandings, changing the relations. In contrast to LEA managers’ prognostic framing of shared responsibility as intertwined, they used their hierarchical position to intensify control. The principals understood this counter-frame as a disappearance of the previous coherence in the organization. They responded negatively by perceiving themselves as being excluded:

When pressured, LEA managers demonstrated power and expressed that they needed to know more about the principals’ data analysis. They exemplified this with examples of questions to the principals, such as “What is missing?” “What are the reasons? “Do you have any planned efforts at your schools?” The principals responded to such intensified control by reluctantly playing along:

Therefore, institutional pressure to increase numbers challenged shared responsibility as intertwined between hierarchical levels. This change was a setback to mutual collaboration and the principals’ local responsibility and mandate, in sharp contrast to the framing of LEA managers as co-creators of educational improvement. Hence, even though common sensemaking across hierarchical levels was seemingly fostered through increased communication, institutional pressure pushed LEA managers into old habits and assumptions and, in turn, they pushed principals to make renewed sense of their mandate and responsibility.

Also, in Municipalities B, C, E and F, the previously enacted policy on shared responsibility had been decided on by the LEA level, leaving principals behind. Their decisions on “shared” responsibility were often based on general initiatives, in terms of national grants and local policy goals. In the interviews, both LEA managers and principals rejected this way of enactment, since short-term initiatives did not pay off in the long term:

In retrospect, the LEA managers stressed that when adhering to national initiatives or local policy initiatives, the principals withdrew from their responsibilities of delving into the analysis of the local needs and how to address them, and rather expected the LEA managers should tell them “which direction to run” and “what to prioritize”.

Also in these municipalities, the LEA managers framed the problem with previous policy enactment of shared responsibility for educational improvement as depending on a dominant LEA level accepting insufficient analysis and leaning on general initiatives. In addition, they framed the principals in dichotomous ways, either as confident or unconfident enough to take responsibility at the local level. Therefore, unlike the previously described LEA managers, these managers attributed blame to both the LEA level and the principals. Related to the framing that principal autonomy is superior to shared responsibility, these LEA managers expressed that principals should be given charge of the improvement in their local schools, while LEA managers should take responsibility for the overall issues. For principals to do a good job, the LEA managers framed that they had to provide them with prerequisites without interfering with their work. The LEA managers communicated that it was important not to influence principals’ decision-making but to stay neutral:

Furthermore, the LEA managers communicated to the principals that they had to make their own wise decisions about improvement. To do so, they stressed that the principals must collaboratively learn and exchange experiences both in the management team and in the provided collegial arenas for continuing professional development. Overall, this shows that the LEA managers in Municipalities B, C, E and F regarded themselves as providers of prerequisites for educational improvement and withdraw to strengthen principals’ autonomy, framing the policy demands of shared responsibility for educational improvement as divided by hierarchical levels.

Some principals responded to this way of framing and enacting shared responsibility by perceiving it as very beneficial, while others stressed that it meant a lot of work that could have been reduced if coordinated more collaboratively:

Novice principals and principals responsible for schools with more extensive needs responded by articulating loneliness. While the LEA managers communicated that they were ready to support principals when needed, this was not always consistent with how the principals perceived it:

The LEA managers initially framed principal autonomy as superior to the policy demand of shared responsibility. However, when pressured, LEA managers in Municipalities E and F sidestepped principals’ local responsibility by making strong recommendations for actions in the local schools. In Municipality F, detailed management increased and in Municipality E, LEA managers prompted that results should be improved:

Thus, these LEA managers reframed their initial understandings when pressured. In contrast to their prognostic framing of shared responsibility as divided between hierarchical levels, they became more dominant and intensified their recommendations. From the principals’ perspective, the LEA managers demonstrated their power to increase control, sending signals of impatience rather than trust in principals. The LEA managers’ desire to meet external expectations made them return to their previously rejected decision-making despite the principals’ response by expressing the need for something else:

Hence, when pressured, LEA managers created counter-frames that relinquished responsibility as divided across hierarchical levels, showing that this way of framing was also difficult to uphold in practice.

In conclusion, the findings illustrate how educational leaders framed contradictory policy demands of shared responsibility in relation to ideas of principal autonomy and how it played out in practice. Two broad categories of shared responsibility for educational improvement were identified (Table 2). These, in turn, have implications for how principals understood their responsibility and carried out their work, which in turn was related to both experience and workload. Notably, external pressure made LEA managers create counter-frames revealing additional differentiation within both categories.

Our study aimed to contribute to knowledge about how contradictory policy demands play out in practice in the intersection where educational leaders meet and interact. It focused on how LEA managers made sense of and framed problems related to the national policy of shared responsibility for educational improvement and what this meant for policy enactment in local contexts. The analysis led to two broad categories of educational leaders’ framing and enactment of shared responsibility for educational improvement. However, the analyses also revealed that some LEA managers tended to relinquish their initial framing when pressured to increase numbers and student learning (cf. Datnow et al., 2020). Below, our findings will be discussed in relation to previous research, policy and practice.

Previous research stresses the need for partnership between educational leaders at different levels (Adolfsson and Alvunger, 2017; Fullan, 2009; Hargreaves and Shirley, 2009). It also highlights the importance of trust and learning between educational leaders (cf. Hooge et al., 2019; Paulsen and Hjertø, 2019; Rorrer et al., 2008). Such ideas are visible in our findings in terms of LEA-level awareness of policy demands of shared responsibility. However, the findings also show differences in LEA managers’ framing of how to share responsibility. This framing sometimes contradicts research stressing the need for partnership and raises questions about capacity and relations. In a decentralized educational system such as the one in Sweden, LEA managers are powerful actors in determining how policy demands should be enacted in practice. While LEA managers initially wanted to dissociate themselves from dominating, when pressured, shared responsibility turned out as cosmetic and rather controlling. This led to blaming or guilt up or down the system, limiting partnership and trust, risking educational improvement (cf. Hargreaves and Shirley, 2009). By ignoring context and local needs, LEA managers distanced themselves from the shared responsibility, which in turn raises questions about capacity in terms of fostering common sensemaking, of particular importance for powerful actors (cf. Liljenberg et al., 2023; Weick, 1995).

Another contribution of our study is demonstrating how LEA managers’ sensemaking and framing mattered to principals’ work, which is known to be lonely and challenging, with high turnover and exhaustion (Arvidsson et al., 2021; Jerdborg, 2023). When LEA managers acted as co-creators of educational improvement (cf. King Smith et al., 2020), the principals’ sense of belonging to a collegial community increased. Even so, they still experienced a high degree of autonomy. However, when LEA managers, to uphold principals’ autonomy, provided prerequisites from a distance, principals’ sense of loneliness increased. Experienced principals with stable organizations could handle both situations without problems, whereas novice principals and those in charge of schools with severe challenges expressed that collaboration and a sense of belonging could be conceived as a work-life changer. Rather than idealizing high autonomy as “one-size fits all” solutions (cf. Arora-Jonsson et al., 2024; Nordholm et al., 2025), this study highlights the necessity for LEA managers to balance the needs of experienced and novice principals responsible for schools in different contexts. Thus, context matters and as Arora-Jonsson et al. (2024) conclude, school turnaround can be a result of bottom-up initiatives, but when principal leadership and/or teacher collaboration is lacking, alternative solutions, most often including the LEA level, are needed. In line with this, we argue that autonomy needs to be combined with a culture of mutual trust and collaboration across hierarchical levels. However, our findings clearly highlight that, regardless of context, it is difficult to hold on to ideals and framings in a time when external pressure demands quick results, with the risk of neglecting time-consuming trust-building activities.

Finally, despite increasing interest in shared responsibility for educational improvement at the intermediate level (cf. Paulsen et al., 2014), our findings reveal that local decision-making was sometimes broken by national directives. Initiatives from the state regarding national funding challenged local and contextual decision-making. General financial constraints in education make it difficult or impossible for the intermediate level to refrain from applying for national funding (cf. Rojas and Wermke, 2025). Therefore, by using national funding, the state challenges the shared responsibility at the local level and instead increases control and sets the direction for how to solve declining results. Thereby, it can be understood as a rejection of LEA managers’ and principals’ capacity for fulfilling their shared responsibility for educational improvement. In our study, national funding overruled both LEAs and principals’ autonomy. It also increased the usage of “day lions and grasshoppers” at the cost of contextual knowledge, local needs and long-term goals. This issue is of interest as previous research highlights the need to combine top-down governance with bottom-up professional involvement (Fullan, 2009; Hargreaves and Shirley, 2009). However, our findings illustrate how these strategies are challenging and difficult to fulfill in practice. If problems related to contradictory policy demands are to be addressed, we argue that it is necessary to be aware of how power relations influence framing and enactment in local contexts. Furthermore, educational leaders at all levels need to upgrade the importance and knowledge of local needs. This approach requires patience and collaboration across national and local levels. It also requires trust, and that educational leaders are encouraged to challenge prevailing cultures instead of taking them for granted.

How LEA managers make sense of and frame problems related to national demands for shared responsibility influences principals’ work and possibilities for enacting educational improvement. Therefore, one implication for practice is the need to explore and understand how shared responsibility for educational improvement is enacted in an organization. It is also important for practice to remember that dealing with contradictory policy demands requires conscious and persistent actions. In a decentralized school system, LEA managers are powerful actors and how they make sense of and frame their responsibilities significantly impacts contemporary and future education. We argue that it is vital for educational leaders at different levels to develop a shared understanding of how local responsibility for educational improvement should be addressed and sustained. One question for policy is to consider what LEA-level capacity is needed to meet contemporary complex school contexts and how it can be achieved and maintained.

Our study focused on six Swedish municipalities, limiting its ability to generalize conventionally to a larger population of LEA managers and principals. However, using a multiple case study approach, strategically selecting cases representative of national variation, allows for a more robust analytical generalization (cf. Yin, 2009). By providing transparency in methodology and theoretical implications, we encourage both researchers and practitioners to make informed judgments on how our results can be applied in their practices. Considering our findings and the limited research focusing on the intermediate level, the need for future multiple case studies is paramount, particularly how local policy enactment affects student outcomes. Challenges with increased differences within and between schools must be taken seriously, but assigning blame to those who serve within the system is not productive. Instead, attention must be redirected to those with both power and the possibility of changing the preconditions of the system. Long-term prerequisites must be provided with the local context and pupils in mind.

Finally, it is crucial for LEA managers as public servants to pay continuous attention to their assignments, what they must master and who and what context they are to serve. Educational leaders at all levels must also avoid automatically switching to increased control in times of external pressure, as it can undermine autonomy and trust. A decentralized system with several levels of responsible leaders cannot rely on individual capacity but must be ensured by organizational structures and work processes consciously built on scientific knowledge. We argue that this approach is particularly important for reducing differences within and between schools and for retaining principals.