Projects as a policy tool: a policy ethnographic investigation in the field of education in Sweden

ABSTRACT Politics and educational policy have been increasingly projectified, and projects as an organisational form have become a symbol for solving complex educational problems and interventions. The present article is a policy ethnographic case study in the field of education in Sweden, concerning the use of projects as a policy tool in the Mountain School. The analytical focus is on using projects as a policy tool within a multi-project organisation, where the projects are aimed at promoting integration of pupils with immigrant backgrounds and improving their school achievements. Through the analysis I identified four distinct themes: Outward leadership and teacher teams, The projects have never belonged to the teachers, The lack of a common educational policy and Project as vision and reality. These four themes demonstrate the complexities and difficulties in relation to organisation, management, definition and implementation. These complexities and difficulties in the policy process are not unique to either regular or temporary organisations. However, what becomes apparent in the case of the Mountain School is that the number of projects was very high, which puts considerable pressure on the headmasters and the teachers in their decoding, interpreting, negotiating and enactment of project as a policy tool.


Introduction
We live in a projectified society, where projects can be seen as institutionalised forms of organisations, as can different kinds of development processes (Jensen, 2012;Lundin, Arvidsson, Brady, Ekstedt, Midler, & Sydow, 2015). Politics and policy have also been increasingly projectified, and projects as an organisational form have become a symbol for solving complex educational problems and interventions. The Nordic Journal of Studies in Educational Policy has a strong interest in these aspects of governance and policymaking (Lundgren, 2002;Wai-Yin Lo, 2004), as well as in policy change (Mausethagen & Mølstad, 2015). The present article, based on a policy ethnographic case study in the field of education in Sweden, relates to each of these topics and uses them to deepen the debate and analysis concerning the use of projects as a policy tool in schools.
The projectification of society is part of a shift from government (centralised control) to governance (decentralised control), which entails a change in political and administrative control of the welfare state (Godenhjelm, Lundin, & Sjöblom, 2015). Within this new governance model, policy technologies such as the policy network (Marsh & Smith, 2000) and project organisation have been established (Bakker, 2010). The field of education in Sweden illustrates a change from government to governance, 1 and the Swedish government has stressed that the multi-ethnic school was a particularly important ideological institution. For this reason, I have considered the field of education in Sweden to be particularly well suited to studies of projects as a policy tool. In this significant policy field (SOU, 1990(SOU, , p. 36, 1998, development projects have become the main government tool for educational policy change 2 (Jensen, Johansson, & Löfström, 2012) intended to promote integration of pupils with immigrant backgrounds and improve their school achievement. The present article is located here: it concerns how to use projects as a policy tool in order to promote policy change in multi-ethnic schools that have low school achievement and are located in suburban areas (Wacquant, 2008). These issues have been examined using a policy ethnographic study and interview investigations (Ball & Bowe, 1989;Ma Rhea, 2012). As in a study by Ball, Maguire, and Braun (2012), I have concentrated on one particular school (Mountain School) as a critical case 3 of using projects as a policy tool. Mountain School could be described as a multiproject school, which over time has been involved in 10 different school development projects. 4 My decision to conduct a case study of Mountain School, which had the special characteristic of being a multi-project organisation (Canonico & Söderlund, 2010;Dahlgren & Söderlund, 2010), was carefully considered, as a case study would allow me to sensitively investigate my research questions. A multiproject school is an organisation with many: '… ongoing projects and dependencies among projects in terms of knowledge, components and resources' (Dahlgren & Söderlund, 2010, p. 2). The process of policymaking concerning the use of projects as a policy tool for educational change must be investigated at an organisational level, because it is the actors (headmasters and teachers) who are decoding, interpreting, negotiating and agreeing on the actual policies. The analytical focus is on using projects as a policy tool within a multi-project organisation (Mountain School), where the projects are aimed at promoting integration of pupils with immigrant backgrounds and improving their school achievements. Ball et al. (2012) argue that it is important to understand and document the ways in which schools (headmasters and teachers) actually deal with multiple-coded policies and how they work to define and interpret policies and their interventions. As yet, little is known about using projects as a policy tool within a multi-project organisation in order to create educational change. The aim of the present study is to investigate the process of policymaking concerning the use of projects as a policy tool at Mountain School as multi-project organisation. My specific research questions are: How is the process of policymaking organised and worked out by headmasters and teachers?
How will headmasters and teachers, through their policymaking, define and interpret project policies and their interventions?
What kind of management and control strategy do headmasters use in their policymaking?
Initially, I will provide a brief background of previous research on project organisation, followed by my theoretical framework and methods. The findings related to my research interests are presented in the next section, and the article concludes with a discussion and a conclusion regarding these findings.

Background and previous research on projects
There is already a significant body of research on projects and temporary organisation in both trade, industry and the welfare sector 5 (see Bakker, 2010, Clegg & Courpasson, 2004Lundin & Söderholm, 1995;Sahlin-Andersson & Söderholm, 2002), but there are only a handful of studies on projects as a policy tool in the field of education (see Jaquith & McLaughin, 2010;Miles, 1964;Nordholm, 2015). In reviews of previous research on projects and temporary organisation, Bakker (2010) and Jacobsson, Burström, and Wilson (2013) claim that research has largely focused on concepts of task and team, and they also identify a number of knowledge gaps. For example, they identify a knowledge gap concerning whether the management of the organisation focused on the task or social relationships.
Another key issue that emerges in these reviews is whether project organisations actually have a possibility to implement policy and achieve changes. Likewise, based on previous research, we know very little about whether the knowledge that is produced by the project will remain in the temporary organisation after it has completed. This question is crucial for development projects when the knowledge and experience that were created in the project are expected to transfer and be implemented in the permanent organisation, the aim being to achieve policy change. Johansson, Löfström, and Ohlsson (2007) call this transfer process bridging and claim that this must be seen in light of the project organisation's embeddedness within the permanent organisation and that these influences will go in both directions (Blomqvist & Packendorff, 1998).
According to Jaquith and McLoughlin (2010), many school improvement projects work under unrealistic timeframes that school leaders themselves do not determine, and Miles (1964) believes that many of these project also have both unrealistic goals and conditions. Miles (1964) stated that many of these policy goals are too broadly formulated and lack clear boundaries for what is possible within the framework of the project's work conditions, and for that reason many of these schools will be balancing on a thin line between educational change and what is practically possible to implement. Crawford (2014) points out that project organisation is often used as a policy tool for implementation of new policies. One important aspect of the implementation of policy through projects is to establish and create an organisation that is in accordance with actual policy. In the context of this policy work, Crawford (2014) notes that some organisations are more focused on control, while others are more focused on change.
Organisations with a control agenda often try to establish control through management standards, project planning, budgeting and monitoring. Organisations and practitioners that focus on change are less concerned about control and more interested in the process of change and behavioural aspects of change. According to Jensen et al. (2012) the use of projects as a policy tool is a common strategy when it is difficult to formulate clear goals. However, it is not unusual for a project to be bureaucratic and to develop norms and strategies that seek to secure the organisation's survival, and this will affect the policy process. Considering the experience from previous research on projects and temporary organisations, it is interesting to note that this research has directed far too little attention and research at projects as a policy tool for educational change. The question is, what happens when projects are used as a policy tool in a multi-project organisation?

Theoretical frame: project organisation and policy
The project organisation is seen as an effective policy tool and signals both innovation and entrepreneurship, at the same time as it offers order and control . A project organisation can generally be defined as an activity carried out during a specified period, with a predetermined order, allocated resources and special working arrangements. An important aspect of project work is to establish and create an organisation that is congruent with actual policy. According to Löfström (2010), projects, as a steering strategy, tend to reduce the complexity and ambiguity contained within the policy process, thereby contributing to unpredictable policy processes involving objectives that are not clearly defined with difficult to define clear objects.
In a multi-project organisation, several projects can be performed at the same time and such projects often share some common characteristics (Macheridis & Nilsson, 2006, p. 8). Compared with a single-project organisation, a multi-project organisation is more complex regarding how, for example, it is governed, which means it is important to understand multiproject organisations. Moreover, Engvall (2001, p. 8) argues that studying multi-project organisations is an important research field, because such organisations constitute a significant feature of modern organisations. In a multi-project organisation, several projects are implemented simultaneously or so as to temporally overlap each other. Projects in a multi-project organisation often lack the permanent organisation's structure, but are designed, structured and governed in relation to it. A multi-project organisation also uses most or part of the permanent organisation's resources, such as staff (teachers), facilities, infrastructure and management (headmaster) and often uses the same administrative systems (Macheridis & Nilsson, 2006, p. 8). When policy projects are established in a multi-project school organisation, this often creates tensions between professions, subjects and knowledge, which together generate uncertainty and ambivalence owing to the presence of contradictory norms, values and definitions. According to Dahlgren and Söderlund (2010), it is important to reduce this uncertainty and ambiguity and to ask questions about what management and control mechanisms are needed within this policy work. The policy process involves a dialectic and discursive transformation between different arenas, and in line with Ball (2006), I see policies as: … representations which are encoded in complex ways (via struggles, compromises, authoritative public interpretations and reinterpretations) and decoded in complex ways (via actors' interpretations and meanings in relation to their history, experiences, skills, resources and context) (Ball, 2006, p. 44). In this process of policymaking, the organisation (Mountain School) provided a framework for both headmasters and teachers and their policymaking (Ball et al., 2012). It was in Mountain School as an organisation that the local school authority used development projects as a policy tool to promote integration of pupils with immigrant backgrounds and improve their school achievement. These development projects cannot be seen as separate entities; they are embedded in both an organisational environment and policy networks. A policy network is a political structure consisting of agents, social relationships and structures. Through its social relations, a network can be seen as a structuring resource that offers the agents within it different policy resources and positions. Policy networks involve institutionalised values, beliefs and behaviours and reflect both power and conflicts (Marsh & Smith, 2000).
The network affects the definitions and interpretations of project policies, because members act based on how they perceive their projects as well as their relations to both the permanent organisation and the policy network (Johansson et al., 2007). Therefore, it was necessary to adopt an organisational perspective and to analyse the forms of text and talk that operate through connections of knowledge and power relations that have come to be manifested in the disciplining structures of Mountain School (Chouliarki & Fairclough, 2010;Fairclough, 2005). Ball (2006) points out that there is often a struggle between different policy actors (headmasters and teachers) in this policy work (interpretations and definitions of project policies) and that the actors must often handle conflicting and sometimes capricious policies. The question is, what happens with projects as a policy tool in a multi-project organisation that articulates conflicting and capricious policy discourses? Here, this is elaborated by using a policy ethnographic investigation focused on the use of projects as a policy tool in the field of education in Sweden. It is also impossible to know which definition and interpretation will lead to expected intervention in a school. Sometimes the result of policies will simply be what Ball et al. (2012) describe as an unintended consequence. Creemers and Reezigt (2005) and Matland (1995) stress that it is particularly important that the local school and its actors are involved in the policy work, that the policies are 'owned' by the teachers, and that the teachers are intent on change. They also emphasise the importance of participant decision-making, teamwork, teacher collaboration and a trustful school culture as aspects of a successful policy work. Nordholm (2015) points out, in turn, that it is difficulty to collectively formulate a common interpretation of policy and mentions that there are relatively few collective definitions of policy texts in project organisations. Moreover, Clegg and Courpasson (2004) bring up the importance of professional control in governing the organisation and Simons (1995) identifies four management control system (levers of control) that are important for controlling and coordinating a project organisation. These four levers of control are the belief systems, the boundary systems, the diagnostic systems and the interactive systems. Belief systems could be described as core values in the organisation and are often expressed through formal documents such as mission statements. Boundary systems establish rules and norms that are important to the project organisation and that create and structure limits for cooperation and interaction. Diagnostic systems concern the organisation's performance and operate through plans and budgets, while the interactive system's goal is to establish dialogue, learning and open up for new ideas. In practice, these four levers of control are often combined and create a dynamic mix, and they can partly be seen as overlapping with Clegg and Courpasson's (2004) strategy.

Method and data production
The present study was conducted as a policy ethnographic study of Mountain School's policymaking. 6 Using observations, interviews and discourse analytical methods, I have tried to describe and analyse the processes of policymaking concerning the use of projects as a policy tool for educational change. Over the past 20 years, Mountain School has been considered one of Sweden's worst schools in terms of school achievement and grades. During this period (1995--2015), only 30% of the pupils in grade 9 received passing grades in mathematics, Swedish and English. The school is 55 years old and has its catchment area in a residential suburb in a major city in Sweden. The residential area is characterised by multidimensional poverty (Borelius, 2004), and the school's strategy to deal with this complex situation has been to participate in 10 different development projects. Through a national policy network, Mountain School has collaborated with, among others, the Swedish Migration Board, the National Agency for School Improvement, the National Centre for Swedish as a second language, the National Board of Education, the Swedish National Institute of Public Health and the Ministry of Education. Concerning the local dimension of this policy network, Mountain School collaborated with the municipal housing company, local sports clubs and parental associations.
Studies of organisations must include analyses of discourses, because discourses are part of and shape the organisational context and policymaking. According to Chouliarki and Fairclough (2010) and Fairclough (2005), critical discourse analysis is especially suitable for such studies, as it provides opportunities for detailed analyses of the relationships between (policy) discourses and social practices in organisations. It is the combination of a discursive analysis with a social analysis that allows critical discourse analysis to become a useful tool for policy studies. Critical discourse analysis considers written and spoken discourse to be a form of social practice. This implies a dialectical relationship between a particular discursive event and all the various elements of the situation(s), institution(s), and social structure-(s) that frame it (Fairclough, Mulderrig, & Wodak, 2011). Critical discourse analysis explores how texts in a broad sensewritten texts, spoken interaction, interviews and media textsconstruct representations and social relationship and how they are ideologically structured by power relations (Fairclough, 2013). From such a perspective, policymaking will be an arena of struggle over definitions and interpretations, and Ball and Bowe (1989) have labelled these kinds of studies policy ethnography. They described its distinctive purpose as being to focus on the relationship between policy texts (curriculum and projects policies) and a specific, local school and its policymaking: A policy ethnography is concerned both with exploring policy making in terms of the process of value dispute and material influence which underlie and invest the formation of policy discourses and portraying and analysing the processes of active meaning making which relate policy texts to practice (Ball & Bowe, 1989, p. 1). In concrete terms, by virtue of being a policy ethnographic study, the investigation and analysis have been carried out as a case study of using project as a policy tool in a multi-project organisation (Mountain School). Reading local and central policy texts and project policies has been an important starting point for my analysis.
The [policy] ethnographic data were obtained through observations in classrooms, corridors and the staff room. Data were also produced through interviews with two headmasters, one interview with the former headmaster and five group interviews with the teacher teams at the school (35 teachers in total). The data production was guided by the overall aim and my research questions and the focus was on aspects of policymaking, management and their organisation, asking what headmasters and teachers actually said and thought about projects and policies, but also which policy discourses were articulated in the process of policymaking. Through the observations sensitising concepts (Blumer, 1954) were developed, which then deepened through the interviews with headmasters and teachers. The aim of combining ethnographic fieldwork and [focus group] interviews (Agar & MacDonald, 1995) was to create a better understanding of the use of projects as a policy tool and institutional practice. The present thematic analysis was based on interview data with a focus on the discursive practice (Bowen, 2006;Braun & Clarke, 2006).
The starting point of the thematic analysis was to summarise the interviews and conversations through a sentence contraction (Brinkmann & Kvale, 2015). A thematic analysis in this context states that I: … search for and identification of common threads that extend throughout an entire interview or set of interviews /…/these themes are concepts indicated by the data rather than concrete entities directly described by the participants (Morse & Field, 1995, p. 139, emphasis in original). This implied that I identify keywords and common threads through repeated readings of the contracted sentences which then developed into themes within my interviews (Bowen, 2006;Braun & Clarke, 2006). These themes were grouped together and then checked in relation to my interviews and capture important narratives in relation to my research questions and can be described as a patterned response within the interviews. Through this analysis I develop four broad themes which reflect the data in relation to the use of projects as a policy tool within a multi-project organisation: Outward leadership and teacher teams, The projects have never belonged to the teachers, The lack of a common educational policy and Project as vision and reality.
These four themes aim to provide an answer to my research question about the use of projects as a policy tool at the Mountain School, here used as an example of a multi-project organisation. In the theme Outward leadership and teacher teams it emerges that headmasters have developed an outward leadership while teachers have been responsible for the implementation of projects through decentralised team organisation. This theme provide an answer to my research interest about how the process of policymaking was organised and worked out by headmasters and teachers and what management strategies headmasters use in their policymaking. The second theme, The projects have never belonged to the teachers responds to my research question about how the policy process was organised and how headmasters and teachers defined project policies and interventions. The third theme, 'The lack of a common educational policy', illustrates how both headmasters and teachers had difficulties creating a collective interpretation and definition of project policies. This theme responds to the questions of how the headmasters and teachers define and interpret project policies and their interventions and what management strategies headmasters use in their policymaking. The fourth theme concerns Projects as vision and reality and the analysis shows that they have drifted apart. This theme responds to my interest in how the process of policymaking was organised and worked out by headmasters and teachers.
The validity claim regarding the present study is tied to my ability to control and question the knowledge produced within the analysis (Cukier, Ngwenyama, Bauer, & Middleton, 2009). In terms of the legitimacy and truth aspects, I have tried to critically assess both my empirical data and the analyses that were developed. Important in this context is the credibility of my analyses and alternative interpretations. In this process, it was important to reflect on my theoretical and methodological perspectives, as these have strongly affected my analyses. As part of this questioning, I have also reflected upon my empirical material. Validity as clarity and sincerity means that the study's credibility has been tested in conversations with my informants. The conversations were conducted on the basis of my observations, interviews and thematic analysis (expressed by a research paper), and the informants respond and validate my analysis and argued that it was trustworthy (Hammersley, 1992).
Likewise, I believe it is generalisable to actors and contexts outside the case study. I have presented my analysis at lectures within teacher education programmes, postgraduate education and municipal continuing education. At these occasions, the participants felt they could relate my analyses to their own practices and, thus, experience recognition. The case study was conducted within a larger ethnographic project, and data production was carried out between 2006 and 2011. The study followed the Swedish Research Council's ethical principles (Swedish Research Council, 2011) on informed consent and anonymity. The participants were informed about the project's purpose, methods and forms of publishing. They were also told they had the right to refrain from participating. All participants were informed that the data would only be used for research purposes.

Outward leadership and teacher teams
Based on my ethnographic observations at the Mountain School as a multi-project organisation and a close reading of my interviews, it is clear that the headmasters' leadership could be described as outward, as the implementation of the projects was delegated to the five teacher team. In the interviews there were a number of the teachers who felt that the school's leadership did not work well: Harald's focus was on an outward leadership, not aimed at the internal school organisation. It was more important to profile the school outward, while others had to take care of the inner working at the organisation (Interview, Ted, teacher) Harald's leadership was outwardly orientated, and he organised the project work through the five teacher teams. He stated that each team's working methods could be very different, but the objectives of respective projects had to be completed: I do not think it's wrong that each team could decide /…/this kind of leadership is important for us and this is how we do it here. It would be possible for each team to work with very different methods / …/and they just have to fulfil their project duties. (Interview, Harald, former headmaster) Since this team organisation was introduced in the late 1990s, within the multi-project organisation, the headmaster had placed increasing responsibility for realising the projects ideas on the teacher teams. Harald mentioned that the headmaster must be very clear about their mission and that the teams have extensive freedom to work with different methods, but only as they achieve the project objectives. Many of the teachers are experiencing this decentralised project organisation as problematic as many educational issues must be resolved within respective teams. Some of these teachers work in Team 2, and say: When we were divided into the project organisation, the responsibility for everything was made part of the teams' responsibility. As soon as a question came up about projects, we were told that we must solve it in the team, next question! Yes all questions were to be resolved by the teams. (Focus Group Interview, The teachers in Team 2 felt they had been given more and more duties and responsibilities for the implementation of projects after having been organised into teacher teams. A similar experience was reported by the teachers in Team 3: the headmasters have handed over much of the workload in the projects to the teams, meaning there is a lot more responsibility placed on us / …/and the teams work very differently, and I think you would get this answer from all five teams. (Focus Group Interview, Team 3) They were also well aware that they work very differently in the various work teams and that teachers at the school were all aware of this. In this decentralised organisation, responsibility for implementation of the projects was delegated to the team level, which contributed to them working in isolation and independently. The teams also experienced that they have little opportunity to coordinate their project work between each team and do not have the energy to either listen or talk about the work with each other.

The projects have never belonged to the teachers
Over time, Mountain School's emphasis on projects has created what could be described as a multiproject organisation. Within this organisation, regular school activities and different projects have been integrated. Inga, one of the headmasters, explained the situation as follows: many of the teachers find the situation as laborious and think that they have worked in so many different projects and they've got so many different thoughts and ideas so they don't have a grip on it. (Interview, Inga, headmaster) The number of projects made it difficult for the teachers to grasp the project ideas and tasks, and the situation could be said to resemble policy interference: a clash between different policy ideas. The teachers were never involved in the formulation and establishment of the processes, and they felt that everything had been decided over their heads: '… I feel that every project and policy formulation was determined over my head. Or our heads, I would say' (Focus Group Interview, Team 2). Inga (headmasters) was of the same opinion, reporting that the projects have never belonged to the teachers: Mountain School's different projects have never been the teachers' projects; they've been political decisions and initiatives and they often came from the policy network, the school administration or headmaster (often Harald). The projects have therefore never fully been established among the teachers. (Interview, Inga, headmaster) The projects were formulated within the policy network and the teachers, as a professional group, were not invited into this policy process, beyond this formulation arena, which meant that projects were often poorly anchored at the school.

Lack of a common educational policy
Through my [policy] ethnographic fieldwork and interviews, a picture of Mountain School as a multiproject organisation emerged in which both headmasters and teachers had difficulties defining and interpreting policies and developing a common pedagogical understanding of projects. One of the headmasters (Bob) argued that: There is no common educational policy or understanding about projects /…/we have nothing to fall back on. (Interview, Bob, headmaster) Bob (headmaster) was very clear about there being no common pedagogical understanding of the projects objectives and visions. The teachers in Team 1 had the same idea as Bob, and they were very clear in stating that they also lacked a common pedagogical policy: We have no considered pedagogical policy about integration, school achievement and projects at this school; it's just talk, so to speak. More practice and less talk, so to speak. We have talked so many times about how we should work pedagogically. (Focus Group Interview, Team 1) Moreover, they felt there was too much talk and too little action linked to the educational practice. The teachers in Team 2 complain in turn that the projects from the beginning were a shared responsibility for all teams: The projects should be every team's responsibility but there was no joint discussion about projects and educational policies /…/It's our long project history that lies behind this. (Focus Group Interview,Team 2) The teachers in Team 2 say that there is no collective discussion of either policies or project implementation, and argue that this must be understood in relation to the school's long history as a multi-project organisation. Even the teachers in Team 5 argue that the lack of a common educational policy is rooted in the school's broad project participation, which has led to a high staff turnover: Through the school's participation in all projects we have had a high staff turnover /…/we has change teachers in the same extent as others change shirts. It is not possible to formulate a common educational policy or project plans under such circumstances. (Focus Group Interview,Team 5) The teachers in Team 5 state that this has resulted in a substantial turnover of teachers and that under such circumstances it is difficult to formulate a common educational policy and project planning. Another aspect of the difficulty of formulating a common educational policy in a multi-project organisation is given by teachers of Team 4: Through our extensive project commitment we just talk about the practical things, seldom, we have pedagogical discussions. The projects objectives will only be a vision. (Focus Group Interview,Team 4) As the above quotation shows, the teachers in Team 4 believe that it is the school's extensive involvement in different projects which contributes to the focus on practical problems and the limited educational discussions.

Project as vision and reality
The Mountain School has, as a multi-project organisation, carried out a large number of projects. The analysis shows that the extensive project work and organising has resulted in visions and reality drifting apart. According to the teachers in Team 3, the large number of projects makes them tired and has influenced the policy process and their commitment to the project: We have implemented a lot of different projects and during these projects we become tired and no one bothered either listen or talk about the projects' objectives and even less to implement them in practice. (Focus Group Interview, Team 3) A similar experience was expressed by the teachers in Team 2 who perceive that their project work is not progressing: I don't think we get anywhere /…/I think it is related to our workload through all projects we have at our school so even if you have good project ideas and visions, you don't have the energy or possibilities to implement them in reality. (Focus Group Interview,Team 2) The teacher argue that they often have good project ideas and visions but they do not have the strength or ability to implement these ideas and visions. Vision (theory) and reality (practice) slips apart in the project organisation as Teams 2 and 3 provide. The teachers in Team 4 in turn argue that it is frustrating that projects are based on visions and not a reality: It is a bit frustrating; the projects are based on a vision and not a reality. (Focus Group Interview, Team 4) However, it was not only the teachers who experienced the projects as problematic; one of the headmasters (Inga) also finds it difficult to combine the project vision and reality. Inga tries to avoid a split between project visions and reality through a strategy in which she: describe[s] the projects based on their reality and not on their objectives and visons. (Interview, Inga, headmaster) At the same time Inga also express difficulties in the implementation of projects and the works that make the project visions to fit the reality: It's something that I've learned; it's difficult to just walk away from a theoretical project description and to make it work in practice, in reality. Above all, in our changing reality, as it turned out, it was not as simple as I had thought … (Interview, Inga, headmaster) One more explanation according to Inga in the difficulties of getting vision and reality to function together in a multi-project organisation is that the large number of projects requires a constant reorganisation of the project: The first project came 1995 (Multilingual Education) then came the National Examples followed by the Metropolitan Development Initiative /…/finally, we have both the old and new project organisation in parallel and it is the new organisation that we need best /…/here we stand with an organisation that don't work well with our new projects and visions. (Interview, Inga, headmaster) Inga (headmaster) believes that several different projects and organisations are running in parallel, and the old project organisations do not match the new project visions and organisation. These problems contribute to the project's visions and reality (theory and practice) tending to slide apart in the framework of a decentralised multi-project organisation.

Discussion
In the present study, I have investigated the use of creating and using projects as a policy tool by conducting an ethnographic case study at the Mountain School. The aim has been to investigate the use of projects as a policy tool within a multi-project organisation and the analysis reveals a complex process of policymaking.
In my analysis I identified four distinct themes: Outward leadership and teacher teams, The projects have never belonged to the teachers, The lack of a common educational policy and Project as vision and reality. These four themes demonstrate the complexities and difficulties in relation to organisation, management, definition and implementation. These complexities and difficulties in the policy process are not unique to either regular or temporary organisations (Ball, Maguire, & Braun, 2012). However, what becomes apparent in the case of the Mountain School is that the number of projects was very high, which puts considerable pressure on the headmasters and the teachers in their decoding, interpreting, negotiating and enactment of project as a policy tool (see Ball, 2006).
The policymaking at the Mountain School was reinforced by the fact that the projects were not 'owned' by the teachers working/involved in them. Instead, these projects were formulated by key actors within the policy network, the municipal school administration and by the headmasters. Creemers and Reezigt (2005) and Matland (1995) point out that this is an important aspect of the policymaking of today, and instead stress the importance of teachers' involvement in the creation of projects and their policies. Many of the teachers also considered that the school lacks a culture of trust which is also, according to Creemers and Reezigt (2005) and Matland (1995), a central aspect of successful policy work. Furthermore Inga (headmaster) complains that many of the teachers perceive their work situation as laborious as they have worked in so many different projects that it was difficult to keep them apart. This situation was reinforced by the fact that Mountain School as a multi-project organisation using the permanent organisational resources within the project work, such as for example the teachers, which makes their work situation very complex and difficult (see Macheridis & Nilsson, 2006, p. 8).
The teachers in this study believe that the decentralised organisation, the large number of projects and the overall workload at the Mountain School makes it difficult to implement the projects as they had planned. Some even mentioned projects as often 'unrealistic'. This could be related to Miles (1964), who complains that it is important that the projects' objectives and requirements are realistic and that there are clear boundaries between the different projects. Accordingly, one wonders whether it was realistic to carry out 10 different projects over a relatively short period of time, within this kind of a decentralised organisation.
The results also illustrate some of the problems this project model created. Harald (former headmaster) describes the importance of creating a decentralised multi-project organisation where each team has exclusive freedom to developed its own working methods in its efforts to achieve the different project objectives. The teachers state in turn that the decentralised multi-project organisation was problematic because both implementation and educational discussions were delegated to each team. This decentralised organisation was also problematic in relation to the low coordination of projects combined with a weak interactive control in which dialogue and communication about policy and project was missing or delegated to the team (Simons, 1995). Dahlgren and Söderlund (2010) point out that a multi-project school such as the Mountain School requires both a strong coordination of projects and a well-developed interactive control system within the organisation. This probably, at least partly, explains some of the problems at the Mountain School, e.g. how the headmasters were more focused on tasks instead of social relations and communication.
The results illustrate how both headmasters and teachers complained that there was no common educational policy within their multi-project school. One reason for the lack of a common educational policy was that it does not function with a shared project responsibility in a decentralised multi-project organisation. A further explanation is given by the teachers in Team 5 who argue that the extensive project participation has resulted in an considerable staff turnover, which makes it difficult to create a common educational policy. The situation described above interacts with what appears as indistinct boundary and belief systems at the Mountain School, which clearly lacks a common mission, corporate statements and established rules concerning how the work should be carried out (Simons, 1995). According to Canonico and Söderlund (2010), this situation demands a clear and distinct management strategy and strong coordination of the project organisation. This was not set out at the Mountain School. Instead, the analysis shows a low level of control of the teaching profession (Clegg & Courpasson, 2004), which means that the headmasters had limited control over the teachers' interpretations of projects' and policies. The headmasters' outward oriented leadership strongly contributed to the lack of a common belief system and collective understanding of the projects' objectives (see Simon, 1995) and probably also created some of the uncertainty that the teachers felt in relation to their project work (see Dahlgren & Söderlund, 2010;Sahlin-Andersson & Söderholm, 2002).
The analysis also shows how the idea of projectbased development seems to get mixed up and drifts apart with both headmasters and teachers, i.e. in how they appear to have difficulties to distinguish between various projects and their objectives. The teachers argue that they constantly become involved in too many projects in this multi-project organisation. This extensive project engagement has made them tired and strongly contributes to their inability to orchestrate their ideas and visions into concrete action. Thus the results indicate that working in projects tends to create pressure and the need to focus on controlling goals, rather than to accomplish policy change (see Crawford, 2014;Jacobsson et al., 2013).

Conclusion
The article concerns how to use projects as model and policy tool for accomplishing educational change at the Mountain School. The analysis, however, shows that the policy making and educational change were not easy to implement.
Significantly, the results illustrate some of the difficulties in implementing policies and bringing about educational change. There was a clear absence of a strategy for the transfer of knowledge within and between the multi-project organisation and the permanent organisation. According to Johansson et al. (2007), transfer of knowledge is crucial for how a project will develop and proceed. Many of the projects at the Mountain School were developed as 'new' teaching methods, but very few among the headmasters or teachers reflected on how the experiences and knowledge gained from these projects could be implemented in the permanent organisation, i.e. there was a lack of transfer and bridging of knowledge within and between organisations (see Johansson et al., 2007). From my point of view, this lack of bridging knowledge must be understood in relation to the patterns that emerge through the analysis [the four themes] and the multi-project organisation's embeddedness in the permanent school organisation. It probably also created a general uncertainty regarding the use of policies and educational change within the Mountain School.
When these results are related to the knowledge gap that Bakker (2010), just as Jacobson, Burström and Wilson (2013), identifies in previous research, it is possible to see how the management of the organisation at the Mountain School was more focused on certain tasks and goals rather than on the social relationships. The responsibility for both social relations and the projects was instead delegated to the teams without support from the headmasters. The lack of common norms and values became problematic for the teachers, which resulted in the practices and visions of the projects drifting apart.
The analysis seems to confirm previous research on collective policy interpretations and also, like Nordholm (2015) argues, that there are very few project organisations that successfully create a collective interpretation of policy. Hence, based on previous research and the result from this study, it is justified to ask whether if it is appropriate to implement 10 different projects within a decentralised and multi-project organisation, or if it is possible to create a collective interpretation of policy. Instead, the result indicates that headmasters tended to govern by what Simons (1995) describes as a diagnostic system, while other levels of control (belief, boundary and interactive system) tended to be delegated to the teachers, which probably did not facilitate the creation of educational change (Dahlgren & Söderlund, 2010).
In relation to these result it is important to point out that earlier research (Matland, 1995;Nordholm, 2015) highlights that the formation of a collective educational policy must be done in dialogue between the schools' various actors such as headmasters and teachers. In this process of policy making, it looks as if the multi-project organisation (Mountain School), without clear objectives, strategies and definitions, together with influences from the policy network, can be described as a restricted framework and not an enabling framework in the use of projects as a policy tool.
In conclusion, and with support from Ball et al. (2012) and Miles (1964), the results presented here are not surprising. It is important to ask questions about the requirements for achieving policy and educational change within a decentralised multi-project organisation, especially for teachers to fulfil and act in accordance with policy declarations. At the same time, just as Ball et al. (2012) highlight, there are obvious difficulties in knowing what policymaking and definition of policies are aiming for, or whether they can reach educational change. The results here clearly show that policy enactment often can have unintended as well as unwanted consequences. Notes 1. The Swedish education system has changed from being one of the Western world's most centralised into one of the most decentralised and deregulated systems (Centre for Educational Research and Innovation, 2005;Lundahl, 2005). 2. A new urban policy (Government Bill, 1997/98, p. 165), which could be described as a double-coded policy discourse (see Gewirtz, Dickson & Power, 2004), aimed to create conditions for successful schooling and integration for pupils with an immigrant background. 3. The number of projects is so extensive, which places very high demands on the organisation, headmasters and teachers, and means that the Mountain School could be described as a critical case when it comes to project as a policy tool. 4. During the past 20 years, Mountain School has been involved in a variety of policy projects such as Multilingual Education (National Board of Education, 1995Education, -1996 , 2000-2005).
In light of this extensive project participation, we can reasonably describe Mountain School as a multiproject school, where all of these policy projects focused on integration and low school achievement. 5. Steinporsson (2014) argued that the concept of temporary organisation could be seen as a new research metaphor for studying projects. Instead of seeing the project as a tool for change, it can be seen as a temporary organisation consisting of individuals who are temporarily enacting common motives. 6. The fieldwork began in autumn 2006 and my contact with the school ended in the autumn of 2011.