A professional development practice to enhance principals ’ instructional leadership – enabling and constraining arrangements

Purpose – Thepurposeofthispaperistoincreasetheknowledgeofprincipals ’ professionaldevelopment(PD) by focusing on the arrangements that shape a PD practice initiated to enhance principals ’ instructional leadership. Design/methodology/approach – Thepaperdrawsonfindingsfromathree-yearPDinitiativeinaSwedish school district. The data consist of field notes and semi-structured interviews with principals and managers. Theoretically, the paper takes its starting point in the theory of practice architectures and the cultural-discursive, material-economic and social-political arrangements that shape practices. Findings – Practicesfor principals ’ PD requireafine balanceto prevent the perceptionofhierarchicalcontrol. Designated time, common artifacts and external expertise are arrangements of practice that enable principals ’ PD. High expectations and relations between principals and managers both enable and constrain principals ’ PD. As trustful relations are of particular importance, the implication for managers and others organizing for principals ’ PD is to make sure that collaborative work settings also become a natural way of working for principals. As building relations takes time, a longer time perspective is also recommended. Practical implications – The findings have practical implications for educational leaders responsible for organizing PD practices for principals in any context. Originality/value – This paper adopts a practice theory approach to its study of principals ’ PD and provide an elaborated illustration of arrangements that enable and constrain principals ’ PD in collegial settings.


Introduction
Based on the knowledge of the importance of school leadership for student learning, professional development (PD) for principals has gained increased attention (Grissom and Harrington, 2010;Gurr and Drysdale, 2012).Unfortunately, PD for principals has, as PD for teachers, faced challenges when it comes to school improvement.Over the years, researchers have stated that difficulties in making an impact in school practices depend on both the content and the arrangement of the PD practices (Harris, 2014;Timperley, 2011).For PD to promote learning that leads to improvement in schools, it needs to have a clear focus on instruction and student results.Further, it needs to be continuous, provide opportunities for collegial inquiry and systematic learning (Newmann et al., 2000).In a review of the field of PD for principals, Goldring et al. (2012) identify a number of aspects that emerge as essential to defining high-quality PD.They state that PD for principals should be job-embedded and address real issues in principals' context, have a long-time perspective and offer multiple learning opportunities in various formats.Campbell et al. (2017) confirm these results but also stress the importance of external support and mentoring for leaders at different stages in their career trajectory.As principals in their position are lonely, an important element of highquality PD is networking with other principals who could stimulate critical reflection and help identify gaps in knowledge and skills (Service and Thornton, 2021).
Research on principals' PD has also shed light on the role of school districts.Burch and Spillane (2004), for example, found that central office administrators showed difficulties in relinquishing control and providing opportunities for principals to take ownership of their own learning.Although the intention was the opposite, the administrators had a top-down approach, monitoring the principals rather than working collaboratively with them.Also, Honig (2012) and Honig andRainey (2014, 2019) have shown similar results as they, from a social learning perspective, followed administrators in their roles as teachers of instructional leadership for principals.Hence, as not all principals take an active lead in their own PD, external and internal leaders can play a crucial role.Similar results have also been found in the work by Lechasseur et al. (2019), who studied professional learning networks for principals implemented in an urban district.They concluded that implementing new strategies for principals' PD challenges well-established district structures and cultures.Changes in the bureaucracy of schooling from top-down to more peer-led models of PD have become possible through negotiation with support from "principal learning managers" and external consultants.Consequently, how to arrange practices for PD that meet the needs of principals as well as districts has been identified as a challenge in need of further research.This study intends to make a contribution by focusing on a PD practice for principals, initiated and organized by the principals' managers (i.e. the school area mangers) in a Swedish school district [1].The research questions considered in the study are: how can a professional development practice to enhance principals' instructional leadership be organized?What arrangements enable and constrain principals' professional development in this practice?Principals' instructional leadership from the perspective of PD Researchers increasingly agree that instructional leadership is significant for a successful school organization but also for the quality of teaching and learning (e.g.Grissom et al., 2013;Leithwood et al., 2020).For Swedish principals who work in a highly decentralized school system, increased administration together with greater focus on effectiveness have made it difficult to provide time to instructional leadership [2] (Liljenberg, 2015).However, Leo (2015), who has addressed the norms guiding principals' leadership, has argued that principals also must ask themselves about their priorities as limited time might not be the only explanation for their lack of focus on instructional leadership.Lack of competence and uncertainty in the relationship with teachers might be other reasons.
Consequently, principals' instructional leadership can also be addressed from the perspective of PD.Likewise, Huber (2011Huber ( , 2013) ) argued that the importance of principals' PD for enhancing the quality of schools has become a central concern of educational actors in most countries.However, as most of the research has been conducted within Anglo-Saxion countries, more knowledge of principals' PD practices is needed, especially from outside of North America (Huber, 2011).Thus far, the results have shown that there is no best method or strategy.Rather, a mix of strategies and methods, embedded in practice and adjusted to local needs, is recommended.In the Swedish context, research focusing on principals' instructional leadership from the perspective of PD is limited, although some does exist (e.g.Aas and Blom, 2017;Forssten Seiser, 2019;Nehez and Blossing, 2020;Salo et al., 2015).Forssten Seiser (2019) arranged a university-school partnership with 12 principals who regularly met over one and a half years to explore their leadership.The trustworthy relationship that developed overtime Principals' professional development among the principals turned out to be necessary for the principals' possibility to share experiences and failures, and thus for learning to take place.Contributing to PD was also the cyclic process of action research that prevented the principals from uncritically adopting new leadership ideas.Also, Nehez and Blossing (2020) addressed the relationship among principals and concluded that competition among principals together with vague ideas, centralized control and external pressure are aspects that can prevent improvement and learning, although other promoting arrangements (e.g.time, supervision and action research method) are put in place.

Theoretical framework
Theoretically, this study takes its starting point in practice theory and specifically in the theory of practice architectures (Kemmis and Grootenboer, 2008).This theory has gained increased attention as our interest in understanding collaborative processes such as practices in schools has increased.The theory has been used, for example, in studies focusing on middle leading and teachers and principals' PD (e.g.Edward-Groves et al., 2019;Forssten Seiser, 2019;Langelotz, 2017).
To study practice, Kemmis and Grootenboer (2008) and Kemmis et al. (2014) argued that both actors and structures need to be taken into consideration as they mutually contribute to practice.They have built on Schatzki (2002Schatzki ( , 2005)), who asserted that social life is conducted in practices and that practices are interconnected with and take place in social sites.To understand PD in an educational setting as a phenomenon of social interaction, the site in which it occurs must be examined.According to the theory of practice architectures, people taking part in practices do so, not only based on their own intentions but also based on historically developed arrangements or practice architectures that prefigure, but not predetermine, individuals' practices.Taking part in practices means being involved in sayings, doings and relatings.Kemmis et al. (2014, p. 30) claimed that it is the culturaldiscursive, material-economic and social-political arrangements that enable and constrain what can be said and done, and how people can relate to each other in a particular practice.The cultural-discursive arrangements consist of discourses, ideas and norms; the materialeconomic arrangements, of time, space and material artefacts, and the social-political arrangements, of power, solidarity, roles and relationships.Accordingly, it is these arrangements that give content to the "sayings" (language), "doings" (activities) and "relatings" (social relations), which together constitute a practice of one kind or another.Although sayings, doings and relatings are analytically separated in theory, in reality, they come together in projects (Kemmis and Groothenboer, 2008;Kemmis et al., 2014).The project is the very purpose of the practice and constitutes how people in the practice talk about, act and relate to each other in order to reach the aim of the project.Schatzki (2002) defined projects as expressions of agents' goal-oriented actions.The planned project in this study is PD to enhance principals' instructional leadership.However, although the project of a practice is defined, tensions and contradictions among different interests can be present.People taking part in a practice can have different ideas about the goal or project of the practice, and sometimes this may contribute to the common project not being achieved (Mahon et al., 2017;R€ onnerman and Kemmis, 2016).In this study sample, it was the managers who took the initiative and organized the principals' PD.The principals who took part in the initiative may have had other ideas about the planned project.

The Swedish case
The Swedish school system is considered to be highly decentralized and market adopted (e.g.Lundahl and Alexiadou, 2016).Although curriculums (i.e. the consensually constructed political document which identifies the steering values in education) are centralized at the national level, principals and school districts have the authority to locally decide how to organize schools and preschools to best accomplish the goals and guidelines stated.Moreover, Swedish School Law (SFS 2010:800) states that school districts should systematically and continuously plan, follow up and develop schools and preschools and ensure that principals take part in PD based on the needs identified.In addition, Swedish principals have a long history of working independently (Jarl et al., 2012) at the same time as Sweden, in line with the other Nordic countries, has a long history of a well-established "employeeship," characterized by less managerial control and shared responsibility between employees and managers.Consequently, tensions among different interests may occur and also influence principals' PD practices.

Method
Empirically this study has a qualitative case study design with a purposive sampling.Yin (2013) recommends a case study design when complex social phenomena are to be understood in depth and in their real-life settings.In addition, the case study "can "close-in" on real-life situations and test views directly in relation to phenomena as they unfold in practice" (Flyvbjerg, 2011, p. 309).In this case study, a PD practice for principals is in focus.The study is linked to a PD initiative taken by the managers in a Swedish school district with three school areas.Before starting it, the managers contacted the author of this paper (the researcher) to guide them in the process.To discuss and make plans for the PD initiative, the researcher and the managers met on three occasions.In these meetings, the researcher presented relevant literature about pedagogical leadership and introduced different ways of working to promote collegial learning.When the initiative started in September 2017, altogether about 70 principals responsible for all preschools and compulsory schools (preschool class to school year nine) in the district were included.During the three years that the initiative continued, ending in June 2020, two phases were distinguished.In the first phase (the first one and a half years), the researcher continued to meet the managers to plan for the initiative but also gave shorter lectures to the principals as they met altogether once a month.At these meetings, the principals were divided into geographically mixed groups in which they processed the lectures, discussed common literature and followed up on leadership actions that they had conducted in their schools between the meetings.During the second phase, the remaining time of the second year and third year of the initiative, the principals continued to meet once a month, not altogether but in geographically discrete groups in the three school areas.During this time, the researcher did not participate in practice but met the managers on a few occasions to follow up on the initiative.

Data
The paper is based on data collected over three years.Data from the first and the second year of the initiative consist of field notes taken by the researcher when participating in practice.Data from the third year of the initiative consist of notes taken by the researcher in the meetings with the managers, together with transcripts from interviews with the managers (n 5 3) and a selection of volunteer principals (n 5 15) conducted during the last two months of the initiative.Managers and principals' statements about practice form the basis for the analysis.To get volunteers for the interviews, an email was sent to all principals who had taken part in the initiative.The principals who responded positively to participate were interviewed.These principals, both men and women, had experience of their profession ranging from a few years to over 10 years and represented all school forms.In order to preserve anonymity no further details are given.The managers were interviewed together and the principals individually.The interviews were conducted digitally on Microsoft Teams Principals' professional development (due to the COVID-19 pandemic) and audio recorded.In the interviews, the interviewees were asked open-ended questions, enabling them to give their perspective of the PD initiative in which they had taken part.To help the interviewees concentrate on the practice, they were initially given five areas to elaborate their thoughts around: (1) the structure of the initiative, (2) the work in the groups, (3) the literature they read, (4) their prerequisites for participating and ( 5) their relationships with other principals.These areas are connected to the culturaldiscursive, material-economic and social-political arrangements in the theoretical framework.The interviews lasted about 40-80 min.The length depended on whether the principal had been part of the initiative since the start or entered later due to being newly employed in the school district.Most parts of the interviews were transcribed verbatim.The relationship among the researcher, the managers and the principals, of course, had an influence on the interviews.As Cohen et al. (2018) emphasized, researchers are never completely objective as they already take part in the world that they research.Being reflexive and openly elaborating on how the researcher has taken part in the research is preferable as trying to eliminate the researcher's effect is impossible.Based on the outcome of the interviews, there is no indication other than that the participants sincerely contributed with their opinions.

Analysis
In the first step of the analysis, a detailed examination of data was conducted by foregrounding the PD practice, and based on indictive reasoning, working to identify emerging themes from data to find the answer to the first research question, i.e.How can a professional development practice to enhance principals' instructional leadership be organized?In this step, qualitative content analysis was used (Miles and Huberman, 1994) and open questions such as: how do principals describe the PD practice that they have been part of?What happened in practice?How can it be understood?where asked in connection to the field notes and the interview transcripts.The themes that emerged included features such as control, relations, artifacts, common concepts and expectations.In the second step of the analysis, the themes from the first phase were related to the theory of practice architectures to examine how the cultural-discursive, material-economic and social-political arrangements enabled and constrained principals' PD.This step of the analysis was conducted to find the answer to the second research question, i.e.What arrangements enable and constrain principals' professional development in this practice?Coding and analysis can thus be characterized as both data-driven and concept-driven (Kvale and Brinkmann, 2009).The identified arrangements are presented in the result section and summarized in the discussion.Due to the limitation of space, this paper focuses exclusively on PD practice and its arrangements.However, practices are connected to one another, and aspects of one practice create practice architectures for other practices (Kemmis et al., 2014, p. 44).

Results
The results are presented in two sections following the two phases of the initiative.Each section starts with an illustration, written based on the notes taken by the researcher, to set the scene of the PD practice that the principals took part in, followed by a description of the arrangement that gave premises for the project, that is, PD to enhance principals' instructional leadership in that phase.

Arrangement shaping the PD practice in the first phase
It's Friday morning.Principals from all schools and preschools in the school district arrive to the assembly hall for a first gathering.The managers have arranged the room with groups of tables and chairs and meet up to welcome the principals at the door.It's not the first time that they have met all JPCC 6,4 together, but it's the first time that they will work in geographically mixed groups.As the clock strikes 8:30, the managers take the stage.Some principals have already taken their seats at the assigned tables.Others are unsure of which group they belong to.Most principals speak freely around the tables, new employees greet and introduce themselves.Some principals pick up their computers and the book that they all have been instructed to read some chapters of.Others do not pick up anything.Then, the managers start to talk (Field notes, September 2017).
The identified material-economic arrangements that shaped the practice in the first phase were the "collective use of time and space," "scientific expertise" and "common artifacts."The managers had set aside time in everyone's calendar well in advance to make it possible for all principals to meet together half a day a month.To further support collegial learning, they had divided the principals into geographically mixed groups and arranged the tables in the room, making it possible for the principals to work in these groups.At several of the meetings, the managers had invited the researcher to give a short lecture on a specific theme connected to instructional leadership.The principals were also instructed to read some chapters in a book selected and distributed by the managers.As a complement to the chapters, the managers had created so-called "reading keys," which contained questions and matrices to encourage the principals to reflect on their own leadership in relation to what they had read.In the interviews, some principals noted that they thought about the reading keys as a way of controlling their dialogues, while others felt that the reading keys provided them with structure for their reading, but foremost contributed to deepening the dialogue in the groups.One principal expressed approval of this: I like structures because they give you an entrance to the discussion.If it had been completely free, it would be easier to just talk, and then there would be no deeper discussions.It would not have given me anything.I want to discuss, and I think the reading keys contributed to that.They were valuable and created structure in our group.We stuck to them all the time.(Principal 13) The identified cultural-discursive arrangements that shaped the practice in the first phase were the managers' recurring talk about the initiative as an "investment in instructional leadership" that they all were part of.At the beginning of each meeting, the managers returned to their purpose and goals in a way that marked the context for what they all were there to work toward.They linked previous meetings to the present one and connected the initiative to what was going on in the local context.Some principals perceived this as supportive; others perceived it as somewhat overambitious.One noted the latter view: I think it matters what attitude the managers show because it spreads further down in the organization. . . .The managers made many context remarks.They were supportive and tried to challenge us.They wanted to achieve a lot, and it certainly became a lot, but from my perspective, it did not have to be so ambitious.(Principal 6) In addition, the literature introduced the principals to new concepts such as "the school's infrastructure" and "goal-claimer."These concepts became part of their language as they discussed their leadership actions in the groups.For a newly employed principal who had not been part of the initiative from the start, this became particularly prominent as other principals talked about themselves as "goal-claimers" helping teachers focus on the goals in the curriculum.
"High expectations" and "new relations" were identified as social-political arrangements that further shaped the practice in the first phase.Upon reflection, the principals emphasized the high expectations expressed by the managers on their being present, cooperative and dedicated.Although most principals expressed that the managers communicated their expectations positively, there were still some uncertainties among the principals, especially related to the focus on instructional leadership.Some principals perceived the focus as criticism of their leadership.For these principals, the initiative was an expression of Principals' professional development hierarchal control, which contributed to their more skeptical attitude.Others perceived the opportunity to meet principals from different school areas and jointly develop as instructional leaders a privilege.These principals were more willing to discuss their leadership in the dialogues: For me, it was like a great welcoming embrace into instructional leadership.But I also heard that there were mixed feelings among colleagues, thoughts about why we should go into this.Is it because they are not pleased with what we do?A feeling that we have not done it right, and now we'll be reprimanded.(Principal 5) Regardless of attitude, all principals had to make "new relations" in their groups, an arrangement that had a significant impact on what happened in the practice.Exchange among principals in the three school areas in the school district had been rather limited to that point.The arrangement with geographically mixed groups was an attempt to change this.The principals noted that working in mixed groups with principals from different school areas and different school levels was initially challenging; however, it became easier as time went on.One principal described this situation: There was a bit of uncertainty at first.We came from different areas and we had different lengths of our careers, but we found common ground for how we would make it work. . . .It was still enriching.What was obvious to me was not always as obvious to someone else.And something that was not obvious to me was more obvious to them.(Principal 2) Based on principals' evaluations of the initiative after each term, the managers came to the conclusion that the reading common literature and working collaboratively had contributed to an understanding of what instructional leadership could be.However, the managers' ambition for the principals to design, test and document their leadership actions and bring their reflections to the groups was less fulfilled.This resulted in frustration among the managers, although they knew that they could not force things to happen.The principals believed that this could be improved if they were to work in geographically specific groups (instead of mixed) as that would make the dialogue more locally-based, and thus more relevant.The managers, on the other hand, were hesitant to do this as it conflicted with their desire to create unity across the school district.However, they were also eager to give principals autonomy in determining their own PD.Accordingly, after one and a half years, they decided to put together new, geographically discrete groups.They also decided to continue to meet for half a day, once a month, from that point on in the three school areas.
Arrangement shaping the PD practice in the second phase of the initiative It's Thursday at lunch time.The manager responsible for the preschools and schools in the West school area of the school district has just finished the agenda.The principals in the area are now free to continue with their own work in their groups.In the first group, the principals are about to continue with the chapters in the book that they have chosen together.However, at first, they have to finish up some daily work on student transitions.In the second group, the principals "check in" by briefly telling what leadership actions they are involved in.One of the principals, chosen to be moderator, guides the others through the meeting.The third group has had a tough start.Mistrust and frictions among the principals ended up in the need to take a break.After consultation, relations have now been improved, and the principals are about to start up the intended work again.(Notes, September 2019).
The identified material-economic arrangements that shaped the practice in the second phase were "distributed use of time and space" and "less use of artifacts."When the principals reflected on the second phase of the PD initiative, most were slightly disappointed over how it turned out.The managers had taken a step back and given them the time and opportunity to set the premises for their practices.Unfortunately, the principals concluded that they had not taken this opportunity.As the PD practice in this phase turned out to be less structured due to lack of joint agreements, the conversations also turned out to be less challenging and more confirming.
I do not know.There might have been some who requested less control, but in my opinion, it [the meeting] gave me more before, even if it also required more of me.(Principal 12) In this phase, it was also up to each group to decide how to use the artifacts they had been introduced to in the first phase.Few groups continued to use them, and they did not create their own either.Yet, groups that continued to use an agreed structure for their dialogues, jointly delimited the content and appointed a leader for the group found this to be an arrangement that enabled deeper dialogues.
The identified cultural-discursive arrangements that shaped the practice in the second phase were the less structured dialogues that opened up the session for talk about "practical issues."For most principals, less structure made it difficult to maintain focus and give priority to instructional leadership: In the dialogue, it becomes a lot of "I did this and I did that."It is just fluffy talk.Not really challenging, or not challenging at all. . . .The conversation does not give me anything.(Principal 7) However, those who found common ground for their work in this phase talked about "leadership actions" based on "local needs" rather that needs identified by others.One stated it this way: I think, we in this group have managed to have an even better focus on pedagogical actions; the group is perhaps a little smaller, and we are a little more familiar with each other . . .or that we are simply mature for this now in the process.(Principal 1) "Individual priorities" and "previously established relations" were identified as socialpolitical arrangements that shaped the practice in the second phase.Although it was the principals who had asked for more influence and new groups, they were critical of the result.First, fewer common directives had opened up for individual priorities and deletions.Second, new groups meant that they had to start from the beginning in their collaborative process.Most of the principals stated that the geographically discrete groups had turned out to be less favorable for collaboratively enhancing their instructional leadership.Having practical issues in common, related to the school area, made it more difficult for them to keep their focus on instructional leadership.It also became evident that some relations among principals working in the same area were rather tense, and it was not suitable for them to critically review each other's leadership actions.One participant offered this insight: But in my group, disaster.It did not work well at all.I think it has to do with the fact that we are so close to each other, and it becomes so difficult to expose ourselves.It's about prestige.(Principal 14)

Discussion
The first research question of this study initiated an exploration of how a PD practice to enhance principals' instructional leadership could be organized.The results presented above show that the managers used different strategies as they organized PDs for principals.In the first phase, the managers used a top-down strategy with a significant impact on the arrangements that shaped the practice.In the second phase, the managers changed and utilized a bottom-up strategy.Top-down strategies have repeatedly been shown to present difficulties when used by school districts to support principals' PD (e.g.Burch and Spillane, 2004;Nehez and Blossing, 2020).However, studies have also shown that district leaders can Principals' professional development be of great importance for principals' PD (Honig andRainey, 2014, 2019).In a way, highquality PD for principals in parts requires a top-down approach, for example to keep away less thoughtful initiatives and counteract loneliness in the principal position by providing networks and opportunities to build professional learning communities (Goldring et al., 2012;Service and Thornton, 2021).In this study, the top-down strategy that the managers used in the first phase was appreciated by many of the principals as it provided them with structure, external expertise and new collegial input that gave them favorable conditions for PD, thus enhancing their instructional leadership.However, the top-down strategy was perceived as hierarchical control and a restriction of professional autonomy by other principals.To counteract this, the managers decided to step back and be open for a more co-constructed practice (cf.Lechasseur et al., 2019;Newmann et al., 2000) in the second phase.This decision opened up for principals to address the needs of their local organization but also for adaption to principals' different stages in their career trajectory, aspects that are significant of highquality PD (Campbell et al., 2017).However, as arrangements of practice intermesh and mutually influence each other (Kemmis et al., 2014), their decision had a significant impact on the practice architectures, not only to the advantage of principals' PD.
The second research question asked what arrangements enable and constrain principals' PD in this practice.Analyzing the two phases of the initiative revealed enabling and constraining arrangements for principals' PD in both phases.The identified arrangements are summarized in Table 1.
In the first phase, the material-economic arrangements with regulated time, supported by scientific experience, and common artifacts enabled principals' PD, although adaption to local needs was lacking.In addition, the mixed groups opened up new perspectives and relations spanning the different school areas, although some principals were more skeptical.This is in line with Honig andRainey (2014, 2019), who state that work in smaller groups provide principals with beneficial conditions for learning collaboratively.In the second phase, time was likewise set aside, but less structure was provided.This opened up space for individual priorities, but also for the ever-present practical issues to become the main focus, a priority that can be explained by principals' intense work situation (Leo, 2015;Liljenberg and Andersson, 2019).Moreover, few groups continued to use reading keys and they did not create their own either.Yet, the groups that continued to use some sort of agreed upon First phase Second phase

Material-economic arrangements
The managers organized the collective use of time and space, and they provided scientific expertise and artifacts, such as literature, reading keys, and ground rules for dialogues The principals decided how to use time and space.Few groups continued to use artifacts as reading keys and ground rules for dialogues

Cultural-discursive arrangements
The managers repeated the purpose and goals and talked about the initiative as an investment in instructional leadership The principals used new concepts from the literature in the dialogues Practical issues related to the school area became the focus in many dialogues Leadership actions based on local needs became the focus in some dialogues

Social-political arrangements
The high expectations from the managers made some principals become skeptical while others were motivated by them New relations among principals initially limited the dialogues, but also opened up new perspectives on instructional leadership Individual priorities prevailed Tense relations among principals inhibited collaboration and a focus on instructional leadership Table 1.Identified arrangements in principals' PD practice JPCC 6,4 structure for their dialogues found this to be significant for their PD.This is in accordance with previous research showing that artifacts can bring structure to dialogues and contribute to principals' PD (Liljenberg and Andersson, 2020;Burch and Spillane, 2004).However, artifacts can initially be perceived as constraining as it takes time to get comfortable working with them.They may also challenge established power relations.When the principals were given more influence, letting go of the artifacts can be understood as an attempt to take control, but also as an attempt to get rid of what was perceived as challenging.Without the artifacts and without scientific input, the dialogues ended up being more simple conversations which constrained the principals' PD.However, the groups that, in the second phase, shaped the material-economic and the cultural-discursive arrangements, agreeing on their own structure, and directing the focus toward the local needs of their organizations rather than needs identified by others, also enabled the principals' PD.
Likewise, the social-political arrangements both enabled and constrained principals' PD.Principals' attitudes toward the initiative and how they related the initiative to themselves as school leaders directed how they took part in practice.Some principals perceived the initiative as an expression of hierarchal control, while others perceived the initiative as a privilege, giving them the opportunity to meet other principals, and jointly develop as instructional leaders.However, regardless of their attitude toward the initiative, all principals had to work in groups and build relations with other principals.In some groups this, together with differences in years of experience and knowledge, created tensions in relations.This is a result that corroborates with previous research (Nehez and Blossing, 2020).Moreover, it highlights the important balance between top-down and bottom-up strategies when managers organize for principals' PD (Honig, 2012;Lechasseur et al., 2019;Thessin, 2019).Trying to improve this balance the managers in this study changed their strategy, from topdown to bottom-up, in the middle of the initiative.Although this was done based on requests from the principals, a hybrid model integrating top-down and bottom-up could perhaps hold more promise as both strategies have their strengths and drawbacks (cf.Lechasseur et al., 2019;Newmann et al., 2000).

Conclusion
To conclude, viewing PD for principals as a practice provides some important insights about principals' PD in collegial settings.In this study, arrangements of practice that promoted sense-made ideas about why, together with trustful relations and the mutual influences on how, were found to be essential for principals' PD, and thus for enhancing their instructional leadership.However, the study, of course, has limitations.First, the case was not independently selected.The researcher was invited by the managers to bring expertise and scientific grounds to the initiative.In the first phase, the researcher also participated in practice.This can be seen as a limitation of the study (not for the PD practice), but also as a strength as taking part in practice gave the researcher first-hand information.Second, the researcher's involvement in the initiative might have contributed to principals with more critical viewpoints choosing not to participate in the interviews.Despite its limitations, the results of the study contribute to the ongoing discussion of how to arrange PD practices for principals and the role of district managers in this (Honig and Rainey, 2014).Although linked to the national and local contexts, the findings have implications for leaders at all levels facing the challenges of organizing similar initiatives.Previous research stresses the importance of addressing real issues in the principals' context (e.g.Goldring et al., 2012).However, real issues do not have to be addressed individually.Rather, in high-quality PD principals learn from the experiences of others (Campbell et al., 2017).In this study, the relations among the principals were of great importance for the PD practice.A primary implication for district managers and others working with principals is to make sure that collaborative work settings become a natural way of working preferably to make it possible for principals to share experiences and failures.This also applies to principals themselves in how they manage their respective work settings.Working with colleagues to improve practice has been advocated for teachers for several years (Harris, 2014;Timperley, 2011).Therefore, the same collegial approach can be successfully applied to the continuing PD of principals.
Notes 1.In this paper, I use the term "school district."In Sweden, a school district can be a municipality, as in this case, or a private school organization.
2. In the Swedish context as well as in the other Nordic countries, the concept of "pedagogical leadership" is most often used in the literature when these aspects of instructional leadership are referred to.