Organizational characteristics of successful and failing schools: a theoretical framework for explaining variation in student achievement

ABSTRACT Why do some schools bring forth high achievement levels, whereas others do not? To understand the mechanisms behind school success, one must examine schools as organizations and the social relations and interactions among their members. By integrating the literature on educational effectiveness and sociological institutionalism, this article presents a study of four successful and four failing schools in Sweden regarding the impact of schools’ internal organization on student achievement. The results demonstrate that the two types of schools have distinct organizational characteristics, which show significant stability over time. Linking organizational characteristics and student outcomes, we argue that (a) the varying social reality of teachers and principals among the two types of schools fosters different types of actions and beliefs among teachers and principals, which (b) causes different learning environments for students that subsequently lead to variations in their learning trajectories.


Introduction
Why are some schools associated with high achievement levels among students, whereas other schools are not? Why are there systematic differences in student outcomes between schools? The composition of students and students' individual backgrounds, such as their home conditions and their parents' educational level and income, are important answers, but they do not tell the whole story. Consequently, an increasing body of literature is focusing on the individual school level in an attempt to understand the factors and processes involved in school success (Hopkins et al., 2014;Kyriakides, 2007;Reynolds et al., 2014;Teddlie & Reynolds, 2000). In this article, we present a study comparing four successful and four failing schools in Sweden. The argument is that in order to understand the mechanisms behind school success, one must examine schools as organizations and the social relations and interactions among their members. School performance is the result of human behavior and relations. Therefore, it can be affected and changed by the actions of individual teachers and principals as well as by policymakers. This study's underlying assumption is that schools' internal organization differences may explain some of the variation in student outcomes between schools.
There are two major conclusions. First, for a school to become successful in terms of continuously improving student outcomes, it must develop an organizational nature that both encourages and reproduces cooperation and support among colleagues as well as a shared vision of teachers and principals acting on behalf of the learning of all students. Such a culture creates social bonds and a genuine sharing of professional experiences that foster a learning environment beneficial for students' learning. Second, such organizational cultures grow spontaneously due to teachers' and principals' social interactions. The growth of such cultures requires (a) an agent, or a group of agents, who carry (b) a shared vision of student learning being at the very heart of the school's teaching mission, which (c) becomes rooted and maintained in the school's day-to-day activities.
The study's overarching aim is explanatory. Consequently, case selection was based on the student outcome dependent variable. We agree with Scheerens (2015), who, in contrast to the main tradition in the educational effectiveness literature, argues for the necessity of including aspects of ineffectiveness in effectiveness studies. Because structural changes in Swedish public education in the early 1990s increased school professionals' discretion and autonomy, we expect to find differences in the independent variable (i.e., the organization of schools' inner life and work).
Drawing on sociological institutionalism (DiMaggio & Powell, 1991;March & Olsen, 1989), this article shows that different sets of institutions emerge at successful and failing schools and form distinct organizational characteristics. Linking organizational characteristics and student outcomes, we argue that (a) the varying social reality of teachers and principals among the two types of schools fosters different types of actions and beliefs among teachers and principals, which (b) causes different learning environments for students and subsequently results in their learning trajectories to vary.
First, we turn to the educational effectiveness literature, arguing for the importance of including "negative" examples in empirical studies and challenging the dominance of the rationality paradigm. Second, we argue that sociological institutionalism answers the call to strengthen the theoretical base of educational effectiveness research by viewing schools not only as formal systems but also as social and cultural systems. Third, we introduce a comparative case study. Fourth, we present the schools' organization of leadership functions, teachers' work, and teaching, outlining the organizational characteristics of successful and failing schools. Finally, we conclude and discuss the implications of our findings in the context of public education.

Background
The educational effectiveness research shows systematic differences between schools, despite the fact that they face similar conditions in terms of student composition (Hopkins et al., 2014;Hoy, 2012;Reynolds et al., 2014;Teddlie & Reynolds, 2000). Following a quantitative tradition, the literature has identified various factors associated with school success, such as committed leadership, a positive school climate, efficient use of time, and professional development for teachers (Harris, 2009;Hopkins et al., 2014;Kyriakides, 2007;Scheerens & Bosker, 1997). The school improvement research has deepened the knowledge of improvement processes at the local level. Mainly based on case studies and indepth descriptions of schools' improvement processes, knowledge has expanded regarding the context of the effectiveness-enhancing factors demonstrated by the effectiveness research (Bogotch et al., 2007;Reynolds et al., 2014;Teddlie & Sammons, 2010).
However, there is significant demand for the theoretical foundations of educational effectiveness to be broadened (Bogotch et al., 2007;Creemers et al., 2000Creemers et al., , 2010Kyriakides et al., 2010;Scheerens, 2016). Through systematically reviewing 109 educational effectiveness studies, Scheerens (2013) showed that only a minority (11 studies) used theory to "shape the research process" (p. 4). To deepen the knowledge of the mechanisms behind school success, Scheerens (2013) argues for the importance of connecting educational effectiveness studies "to more established theory" in order "to explain findings and construct hypotheses on the basis of more established theoretical principles" (p. 3).

The importance of including "negative" examples
To expand the theoretical base of educational effectiveness research, it is of the utmost importance to challenge the tradition of studying mainly positive cases. With some exceptions (Sammons et al., 1998;Stringfield & Teddlie, 1988), the literature is dominated by positive examples, which can be traced back to the origin of the field (Reynolds et al., 2014). It can also be associated with the rationality paradigm, which assumes that individual behavior is governed by self-interest. Relying on this paradigm, most studies include mainly positive cases because they take for granted that "the antipodes of effectivenessenhancing factors are associated with school failure" (Scheerens, 2015, p. 19). In order to generate a theoretical understanding of the mechanisms behind school success, it is essential to make comparisons between successful and failing schools (Jarl et al., 2017).
Yet, the reliance on the rationality paradigm gives rise to another problem: Important aspects of the school organization are left out (Reynolds et al., 2014;Scheerens, 2015). Effective schools might not only, we assume, be the result of rational planning and professionals making well-informed choices of different ways to act to reach preferred goals. We must make explicit the informal aspects of schools as local organizations: human relations, cultural aspects, and informal norms and routines. To answer the call for a broader understanding of schools as organizations, including both formal and informal aspects of organizing, we turn to sociological institutionalism.

Sociological institutionalism and public education
New institutionalism has contributed to a deeper understanding of the dynamics of complex organizations (Scott, 2008). The interpretation of new institutionalism differs between social science disciplines (Powell & Bromley, 2015). This article adopts the view of sociological institutionalism, which challenges the traditional view that organizations are mainly formal systems where rational decisions are made regarding how to structure the daily activities in order to reach the preferred goals (DiMaggio & Powell, 1991). Rooted in rational or economic models of human behavior, organizations, in the traditional view, are means to reach certain goals; the core activities of the organization are easily changed by the decisions of its leader or by changes in its formal structure (Scott, 2008).
According to sociological institutionalism, organizations are also social and cultural systems, consisting of an inner life that could not be rationally scheduled or planned. Seen as social and cultural systems, organizations are established based on human interaction; they develop gradually and as a result of the actions and social interplay of their members (DiMaggio & Powell, 1991). Organizational members give meaning to an organization's inner life. The organization also shapes the members' activities and preferences: Rules, or institutions, govern human behavior (March & Olsen, 1989). Institutions include regulative, normative, and cultural-cognitive elements (Scott, 2008). Regulative elements refer to formal rules such as organizational charts and role descriptions. The normative and cultural-cognitive elements refer to informal routines and practices that dominate the organization's daily activities, although they are not easily captured in formal documents or scripts. They capture what teachers and principals view as important and will act on, and also provide deeper perceptions of what it means to work as a professional at a specific school. Thus, over time, and to varying degrees, organizations develop their own sets of norms and values. These institutions give the organization a distinct nature (Scott, 2008), which we refer to as unique organizational characteristics of formal and informal institutions.
Institutions constrain individual choice through the logic of appropriateness: Individuals consider the appropriate ways to act given the present situation or context (March & Olsen, 1989). Preferences are formed within the organization in relation to other members. The individual takes the institutions of the organization into account when considering different ways to act.
When organizations are viewed as social or cultural systems, organizational change is likely to take place gradually in small steps. Processes of change are incremental; they grow spontaneously from within the organization, often following specific paths framed by the organization's history and values (Bennich-Björkman, 1997; Morgan, 1986). Leaders may be agents of change through their way of being or acting as leaders.
Sociological institutionalism also covers factors external to the organization that may explain why organizations within a system tend to show similar characteristics over time (DiMaggio & Powell, 1991). An important delimitation of this article is that it does not take exogenous aspects into account.

Methodology
Due to structural reforms of Swedish education in the early 1990s, organizing public education and hiring and paying school professionals went from being responsibilities of the national government to being handled by the municipalities (Jarl et al., 2012). Structural changes also included market reforms (Lundahl, 2002). Furthermore, management by objectives and results replaced the traditional model of public governance in terms of detailed state regulation. As a result, the role of the state is limited to setting educational goals in the national curriculum and to different control mechanisms, such as national tests and the establishment of a national school inspectorate. Moreover, the reforms included increased school principal autonomy. Principals have a mandate to decide on schools' internal organization (SFS 2010:800).
In parallel, variation between schools' student outcomes has grown substantially (Gustafsson & Yang Hansen, 2011; Organisation for Economic Co-operation and Development [OECD], 2015). Programme for International Student Achievement (PISA) results show that the between-school variation increased from 9% in 20039% in to 16% in 20159% in (OECD, 2015. This variation in Sweden in 2015 was, admittedly, lower than the average of the countries included in the OECD (30%), but substantially higher than in other Nordic countries with similar educational models (OECD, 2015).

Case selectionsample of schools
Case selection was based on statistical analysis, making use of the fact that in Sweden, the registry data are stored with a personal identification number (for a detailed description of the case selection process, see Jarl et al., 2017). The database used includes information about Swedish primary schools, such as student numbers, grade levels, and geographic location.
First, using ordinary least squares (OLS) regression, we investigated whether student results (mean merit rating based on each student's 16 best grades in the ninth grade, with a maximum of 320) at the level of individual schools increased or decreased during 1998-2011. Student migration background (the student and/or both of their parents born outside Sweden categorized as immigrants) and educational background of students' parents (university, secondary, or primary schooling) were included as control variables. A complementary test to avoid selection biases based on grade inflation was conducted, showing no substantial differences.
Second, we reduced the sample of schools using constraints to ensure the inclusion of schools with Grade Levels 1-9 that had been active during the entire period (1998-2011) with a minimum number of students in Grade Level 1 and in total. All primary schools in Sweden, public and independent, were included at this stage.
Third, schools were sorted based on the results from the OLS regression from most successful (continuous improvement in student outcomes over 14 years) to most failing (continuous negative outcomes over 14 years). Most schools had a low mean merit change per year. However, at "the top" and "the bottom," there were schools with considerable changes (i.e., more than +1.5 and less than −1.5 mean merit change per year, respectively). Finally, we studied information on 20 schools (10 schools at "the top" and 10 schools at "the bottom") provided by the Swedish School Inspectorate. The aim was to ensure that no school faced external circumstances that could disrupt the trustworthiness of the empirical analysis. To select eight cases, we matched candidates in the top and bottom groups of schools based on geographic location and size (number of students). Case selection is summarized in Table 1. Successful schools are labeled Schools S1-S4 and failing schools Schools F1-F4.

Data collection and analysis
Schools were contacted, and agreed to participate, in the fall of 2013. Initially, respondents were informed that the study included two groups of schools. Information on which of the two groups the respondent belonged to was presented at the end of each interview. We interviewed teachers and principals working at the schools both during the time of the study and prior to that. Interviews were carried out in two phases: late fall of 2013 to early spring of 2014, and late spring to summer of 2014. At two schools, some second phase interviews were conducted early in the spring of 2015. In total, 119 individuals were interviewed separately or in focus groups: 28 principals or deputy principals, 82 teachers, and 9 superintendents. Interviews (conducted by the authors) were recorded and transcribed.
The interviews followed a semistructured guide (Brinkmann & Kvale, 2015) with questions categorized into four themes: the organization of (a) leadership functions, (b) teachers' work, (c) teaching, and (d) student outcomes. Regarding a-c, respondents were asked to describe the present situation and changes over time, and to reflect on rules, routines, and underlying ideas of the school's work. The analysis was done in four steps. First, we wrote individual case reports which summarized respondents' statements concerning a school's organization of leadership functions, teachers' work, and teaching, as well as its development over time. We later provided each school with its report to support schools' improvement processes. Second, we produced schoolspecific within-case matrices with separate columns for leadership functions, teachers' work, teaching, student achievement, and organizational change (Miles et al., 2014). For each school, we registered key concepts that summarized the themes of the different columns and captured the inner life of the organization as a whole. Third, we produced cross-case matrices to make comparisons between the two school types. We used the same column headers and looked for similarities and differences in key concepts of the two types. We registered key concepts that summarized the columns' themes.
Fourth, we entered a more analytical phase that aimed to detect patterns of the internal organization of the two types of schools separately. In parallel, we analyzed the case reports and the within-and cross-case matrices, attempting to make sense of the patterns that emerged. We identified recurring themes, which we argue are typical of the successful schools on the one hand and of the failing schools on the other hand. We label these recurring themes "qualities." On the basis of these qualities, we conclude that the two types of schools have developed different organizational characteristics that are composed of regulative, normative, and cultural-cognitive elements. We acknowledge that it is difficult to separate the three elements in the Small city Note: Student numbers are approximate. Schools are paired as follows: S1 and F1 are located in one large city (over 1 million citizens); S2 and F2 are located in one large city; S3 and F3 are located in rural parts; S4 and F4 are located in smaller cities (approximately 50,000 citizens). *F2 was reorganized during the study, explaining the lower student numbers.
empirical data. Therefore, we treat these characteristics as composite organizational profiles (Scott, 2008).

Results
Next, we present the qualities identified among the two school groups, arguing that these qualities signal distinct organizational characteristics.
Organizational characteristics of successful schools

Organization of leadership functions
We found three qualities regarding the organization of leadership functions. First, leadership is goal oriented. Principals at the successful schools have a clear vision of the work of the school that is centered around student outcomes. All students can learn, and the schools expect nothing else. Principals are described as standard-bearers of these visions. As described by School S4's principal, "My role is to safeguard the goals, to make trouble when anyone tries to cut corners." A similar picture emerges from School S2, where the principal has a leadership philosophy focused on enabling "teachers to do the best job they can," which includes constantly reminding them of the curriculum goals and providing guidance when required. There is also a continuity in leadership functions that facilitates goal orientation. In most cases, the principals and former principals we interviewed had worked at the schools for a long time, in other positions such as teachers or deputy principals. For example, a former principal of School S1 had worked at the school for 30 years in different positions, and the present principal of School S4 has worked at the school for 25 years, first as a teacher and then as the principal. These visions of putting students' learning at the center have permeated schools' work for a number of years. However, respondents refer to a certain point in time when the vision was formulated and, thus, started to become rooted in day-to-day activities. For example, at School S4, the person who took on the principal position in the late 1990s introduced structures for teacher collaboration and teaching that were still present at the time of the study. A similar story was told at School S2. The present structure for teacher collaboration and teaching goes back to the 1990s, when a teacher, who later became deputy principal, started at the school. According to teachers and principals at School S1, the school's work has, as far back as they can recall, centered on the vision of putting students' learning first. Further, there is a tradition of principals being pedagogical leaders who work in close dialogue with the teachers. At School S3, goal orientation was formulated and, at times, upheld by a group of teachers.
Second, leadership is organized collectively. Principals are not the sole bearers of the vision. There are structures and role descriptions which ensure that team leaders have a central position in leading teachers' work in line with the vision. There are groups of teachers, in some cases teachers and deputy principals, who are assigned specific improvement tasks. Formal structures distributing responsibility for high standards are combined with a shared responsibility. Schools have an open discussion climate where everybody is expected to openly share both positive and more demanding teaching experiences. "It is very natural to talk about what you do in different situations, how you planned a certain exercise in class, for example," said a teacher at School S1. Or, as stated by a teacher at School S2, "If the principal says something that I think is wrong, I can say that … . It is an open climate, you can discuss decisions." At School S4, one deputy principal stated that "the social climate among teachers is extremely open, without prestige; they are open to sharing thoughts and ideas and to learning from each other." Third, there is a leadership community that includes the central administration. Successful schools' principals experience both demands and support from the central administration. They unanimously describe the dialogue with the central administration as well functioning. The superintendents also describe well-functioning relations with the schools. At School S1, the chair of the board expects frequent reports on student outcomes from the principal. The chair and the principal are engaged in what they both describe as a constructive dialogue regarding how to further improve the school. Since the early 2000s, School S2 principals have taken part in management groups with other principals and the central administration in the municipality. An improvement agenda for all schools in the municipality is the topic of discussion. The relations between School S4 and the central administration were described as strained at the beginning of the period. However, both principals and the superintendent reported improvements since then.

Organization of teachers' work
In relation to the organization of teachers' work, we identified one qualityteacher collaboration focused on student learning. There are structures in place at the schools to facilitate collaboration. Teachers and principals also seem to value collaboration and the sharing of experiences and have engaged in these activities for a long time.
The specific structures for collaboration vary between the schools. At School S1, the principal has organized and scheduled time for planning to maintain and support collaboration across subjects and grade levels. Once a week, on a certain weekday, all teachers meet to discuss and plan the teaching. Collaboration at School S2 is content driven. The school has a history of teaching in themes that span over several subjects. The themes bring teachers together to follow up on progress in their colleagues' classrooms and to plan and adjust their own teaching in relation to that of others. At School S4, structures are deliberately changed during the year, although the baseline is that all teachers are expected to report their teaching to their colleagues. At one point, some teachers may have meeting time scheduled every morning, whereas others are expected to coordinate teaching throughout the school day.
One might expect cooperation to be considered time consuming and to restrict professional autonomy. Yet, teachers show no such signs. Colleagues are considered to be assets that enhance one's own work. "It could be the case," one teacher at School S2 argued when reflecting on the purpose of planning together, "but in my teacher team, I do not see it like that; instead, we cooperate … . We go in and out of each other's classrooms to watch and to learn." Further, as illustrated in the below quote from a School S2 teacher, norms of sharing ideas, asking for help, and engaging in critical dialogue with other teachers are recurrent signs of successful schools: if you notice something and want to comment on it, you can do that without problems and even in discussions outside [the classroom], like in the teacher teams or with colleagues in the staff room. Yes, in every informal meeting, I think it is very openyou can share what you are doing, and you can give tips to others and so on.

Organization of teaching
Regarding the organization of teaching, four qualities were identified. First, there are high expectations of the students. Students know what is expected of them in relation to cognitive outcomes. At School S2, teachers constantly remind the students of the curriculum objectives. Teachers report on a tradition of encouraging students to attempt course components for which they might feel unready. Teachers and principals at School S4 sometimes worry that the school puts too much pressure on the students. According to teachers and principals at School S1, there is no doubt that students know what is expected of them. "They get to hear the same message from the whole staff," explained the deputy director, adding that students graduating in Year 9 without having passed all subjects is about the worst thing that can happen. At School S3, the scheduled slots for doing homework at the school are always crowded, illustrating that high expectations may also be canalized via the parents.
Second, schools map and follow up on students' achievements. There are structures to map student achievement systematically. The structures vary between the schools, but structures are in place that are considered to be important tools to benefit the learning of all students and are used in order to organize teaching. School S4 maps students' reading ability in Grade Levels 1 and 2. They use a screening system with tests for different grade levels recommended by the municipal administration. School S2 similarly maps students' learning to read in Grade Levels 1 and 2. Based on the results, students may be recommended different tools for reading skills. At School S3, initial tests of reading and writing ability are carried out in preschool, followed by tests in Grade Level 1 and, in recent years, in Grade Levels 2 and 3. The focus of screening activities at School S1 is not only reading and writing, but also mathematics and the English language.
Third, schools adjust teaching to students' past achievements and needs. Teachers and principals report on structures for making deliberate adjustments to teaching in relation to what is known about students' past achievements and individual needs. Teaching is planned in a deliberate and systematic manner with the aim of enhancing students' learning. It is consistently viewed as important that teaching is both planned and carried out in relation to syllabus objectives. Teachers and principals are well aware of students' past achievements and any specific needs. Student results are the focus of daily talks at the schools, in teacher teams, in the staff room, as well as in more informal settings. As illustrated by a teacher at School S2, the school never blames students for not reaching the objectives: "it is the other way around; we think about how we can improve teaching, how we can lift it and change it." Or, as put by a deputy principal at School S4, "We who work at the school need to take responsibility for what the student needs. That is a core focus for us." Fourth, teacher leadership characterizes successful schools. Common to the schools is the importance of teacher-structured teaching and teacher leadership, not only in the classroom but also socially. Teachers are seen to have an important task in being in charge of classroom activities and learning situations. Teacher leadership has been the focus of professional development activities at School S4, noted the principal. Teachers at the school confirmed that teaching is led by the teachers. At School S1, there is a culture among teachers that expects students to meet teachers with respect, which is an approach that the students seem to appreciate. "As teachers, we expect to be met with respect and we are. Students accept it when they get used to it," said a teacher at School S1. At School S3, teachers have agreed on using similar routines for teaching in class, resembling what teachers view as a traditional teacher rolethe teacher teaching from the lectern.

In sum
The qualities of the organization of leadership functions, teachers' work, and teaching at the successful schools signal organizational characteristics of consensus and coordination. There are common norms of cooperation, support among colleagues, and the value of sharing experiences. Norms are upheld by not only teachers but also principals and school administration. These norms circle common perceptions of the deeper purpose of schools' workthat students' opportunities to learn and targeting high performance of all students are at the heart of the common task. There are formal structures laying the foundation for collaboration and support. The shape of these structures, however, varies between the schools. Thus, we conclude that it is the normative and cultural-cognitive institutional elements that give distinct characteristics to the organization of successful schools. These norms of cooperation, support, and trust between colleagues show significant stability over time, as do views of expecting no less than high performance from all students. The voices of teachers and principals do not hold a trace of things being different. To generate cultures of school success, it is essential to have a common goal or vision, with one actor, or a group of actors, acting as "goal defenders."

Organization of leadership functions
Regarding the organization of leadership functions, we identified two qualities. First, there are frequent leadership changes. Changes include a high degree of principal turnover as well as constant reorganization of leadership functions. Teachers describe a tradition of principals not staying particularly long in their positions. One teacher at School F3 recalled having four principals in her first year. Similar but less extreme situations were reported in Schools F1 and F2. For instance, a teacher who had worked at School F2 for more than 15 years argued, "I do not know how many times, but we have changed principals many times over the years I've been here … . I didn't have very much contact with any of them." In parallel, failing schools constantly reorganize leadership functions. On several occasions, Schools F1 and F2, for example, have changed their internal structure in order to deal with conflicts between different sections of the schools. This resulted in changes in the responsibilities or tasks of principals and deputy principals. At School F3, there were reports of mistrust between teachers and principals, making it difficult for principals to succeed in their aim of developing their school's internal work. This is an issue that, in the long term, makes principals want to leave the school.
Second, there is a lack of support from the central administration. Principals tend to be left alone to deal with their work-related problems. There is no structure of support from the administration or other schools. Rather, principals feel that the central administration interferes negatively with schools' internal work. For example, a former principal of School F3 recalled an organizational change in the early 2000s, when 14 school districts in the municipality were merged into four larger districts. School F3 was located in a small rural district, and the principal was appointed to the school even though he did not want the change: "One can say that it was not a positive development. I would have preferred to stay at my former school." The superintendent summarized the continuing turbulence at the municipal school administration by arguing that "It is not a very stable organization," and "unfortunately, I think it is characterized by a lack of trust." Yet another example was School F2, where teachers observed that the principal, due to directives from the school board, was continually occupied with the budget at the expense of pedagogical matters. As one teacher put it, "I do not think there have ever been meetings where we have discussed pedagogic issues."

Organization of teachers' work
Teachers' work is organized based on the quality of teachers working alone, taking responsibility for their own classrooms. There are no structures for teachers working together, and teachers do not express a need to share experiences with their colleagues.
Structures for teacher collaboration are glaringly absent. There are no examples of teacher teams covering the entire personnel; instead, there are loosely organized teacher teams or working groups of teachers at a particular grade level or section. Often, these groups work side by side without obligations to collaborate with each other, not seeing the need to either. For example, according to a former principal, School F4 has a "conservative culture" with strong boundaries between groups of teachers as well as between grade levels. Teachers and principals found it difficult to describe the internal work organization of the schools. Work routines at the schools cannot be captured in formal documents or organization plans; they just seem to happen.
Frequently, teachers are reluctant to agree to principals' initiatives attempting to establish teacher teams or scheduled time for planning. Teachers view the initiatives as attempts to restrict their professional autonomy, whereas principals see them as urgent in terms of creating some order in a messy everyday situation. The principal at School F3, newly appointed to the position, was taken by surprise by the school's lack of internal structure. The principal tried to set up formal meetings, but was met by resistance from the teachers, who believed that they were being too tightly controlled. "They [leaders] do not have any confidence in their employees, in their professionalism, and therefore no issues can be freely handled. It's a very short leash, all the time," as one teacher put it. A teacher at School F2 explained that teachers guard their planning time, while another confirmed that "you meet occasionally; if it's something urgent you try to meet up, but it's not that it's organized." The culture among teachers is "prestigious," with everyone being "closest to oneself," noted a principal at School F1. The principal at School F4 described what they believe is teachers' lack of interest in individual students' learningteachers focus on handling their classrooms the same way they have always done. Principals at Schools F3 and F2 stated that the culture among teachers has always circled around "going to their classroom and shutting the door." However, teachers at the two schools criticized principals for not working hard enough to instill a sense of belonging and purpose. One teacher said that for a long time, they had missed "some sort of belonging … these pieces are important, saying 'these are our priorities,' 'this is our vision.'" Organization of teaching Two qualities address the organization of teaching. First, there is a lack of focus on student results. Collegial discussions on students' progress are mostly missing. Principals stated almost unanimously that student results have not been the focus of the schools' dayto-day operations and discussions. At School F4, there is a tradition of documenting students' performances. The data have not, however, been analyzed in a systematic manner, noted the principal. According to most principals, things have improved in recent years due to changes in the national curriculum, which have increased schools' responsibilities in systematically following up on internal work and making adjustments to teaching practices accordingly. However, this was not a typical state in the period studied. Instead, respondents provided vague answers when the interviews circled around student outcomes. "One simply did not talk about students' results," a former principal at School F1 concluded. At School F2, the principal expressed that there is a "lack of culture among teachers where they believe that their teaching can make a difference". At School F3, teachers vaguely described that "student progress was just not in focus" and that they "simply did not think about it." Second, there is a tradition of blaming low achievement on the students. Respondents were asked to reflect on the student outcomes of their schools. Even though one quality, as described above, was that student outcomes were not the focus of systematic discussions, most respondents were familiar with the downward trend. There were statements that they "had a feeling" that the students did not perform particularly well at the failing schools. However, teachers' and principals' reflections on the reasons for this trend circled around the students. School F1 teachers see the entrance of new pupils in Grade Level 4 as a contributing factor. Because these students are new, the school does not know them particularly well, meaning that they have not been able to foster them in the school's way of teaching and learning. The superintendent of School F3 pointed to increasing social problems in the school's neighborhood, saying that the community is "without prospects." Similar pictures emerged in teachers' and principals' descriptions. Teachers argued that the neighborhood culture "had never prioritized schooling." At School F4, teachers and principals explained that traditions of prioritizing schooling and learning are absent in the community. At School F2, similar stories were associated with the students. Despite the fact that principals and teachers largely attribute the negative outcomes of the schools to the composition of students, there were no reports of how this fed back to the organizing of teaching. There were examples where the absence of joint structures for evaluating students' results and progress made teachers anxious. As described by a School F2 teacher: Perhaps everybody was afraid that any teacher would take it personally. Everyone knew that there were poor results, and then they wanted to … . I mean, maybe someone could feel offended when you do your best, and still the results are not getting better.

In sum
The qualities of the organization of leadership functions, teachers' work, and teaching signal organizational characteristics of fragmentation and conflicts of interest. There are common norms at the failing schools. These are norms of primarily working alone, taking responsibility for one's own classroom, and not expecting support or guidance from colleagues or principals. Also, there are norms of students being responsible for their own learning. Common perceptions of a school's deeper purpose are missing; instead, different views of the essence of the school's work are present. Different grade levels, sections, or teacher groups tend to follow separate regulations. The norms of working alone, doing what is believed to be best in one's own classroom, and of students being responsible for their own learning have dominated for a long period. Consequently, parallel views of the essence of schoolwork exist within the schools. Attempts to challenge these norms through the establishment of common structures for teacher cooperation have met resistance in organizational structure.

Conclusions
This study has, as assumed, demonstrated that successful and failing schools have distinct organizational characteristics. We argue that these characteristics explain some of the systematic differences in student achievement between schools. The organizational characteristics impose and value a structure for professional work that differs between the two school types. In accordance with the logic of appropriateness (March & Olsen, 1989), the distinct nature of successful schools tells teachers and principals what the appropriate behavior in specific situations would be. When challenged, they can fall back on colleagues who act in a supportive manner and invite them to collegial discussions. Failing schools have unique characteristics that are not merely due to the absence of the norms of successful schools. There are norms of working alone and of viewing the students as responsible for their own learning. These norms not only signal that collaboration is difficult or does not work, but also that working alone is believed to be the best way of dealing with day-to-day interactions. When presented with challenges, teachers fall back on doing things as they always have (i.e., taking responsibility for their own classrooms). Also, norms signal that there is little point in adjusting teaching in relation to students' various opportunities to learn.
At successful schools, teachers and principals have high expectations of students and signal what is expected of themin the classroom and socially. Teachers adjust teaching according to students' prior achievements and needs. Teaching in different grade levels and subjects is similarly structured, providing stability in the learning environment. Teachers and principals are aware of the challenges and opportunities of each student. Successful schools have created stability and continuity in students' learning environments that enable learning to prosper.
At failing schools, teachers and principals doubt students' abilities to learn. Teachers are aware of the students' challenges and opportunities in the subjects they teach. However, teachers do not share knowledge on students' specific learning challenges. Because students are viewed as responsible for their own learning, the teaching practice is not regularly adjusted to students' prior achievements. Teaching and learning are structured differently in various grade levels and subjects. Failing schools have organized fragmented learning environments which place significant pressure on individual students' trajectory in education.

Discussion
Grounded in sociological institutionalism, and based on a comparative, longitudinal design, this study makes an important contribution to the school effectiveness literature. Effective schools are the result of not only rational planning but also human relations, culture, and informal norms and routines that evolve gradually within the school (Reynolds et al., 2014;Scheerens, 2015). The results validate that effectiveness cultures evolve over time and that teachers must regularly adjust their pedagogical strategies according to students' needs (Kyriakides & Creemers, 2008).
Considering schools as social and cultural systems means that teachers and principals alone cannot be credited, nor blamed, for their students' level of success. The essence of sociological institutionalism is that the preferences and actions of teachers and principals both form and are formed by the structures, norms, and values that permeate schools.
The literature on professional learning communities has indicated that teacher collaboration has a positive effect on teaching and learning (Stoll et al., 2006). However, there is a lack of knowledge on the potential of establishing "enduringly effective" professional learning communities due to the shortage of longitudinal data (Stoll et al., 2006, p. 247). This study makes a significant contribution: It shows a persistent culture of teacher collaboration at successful schools, which is most likely supported by stability in the leadership functions, and that such a culture is missing at failing schools. Additionally, this study answers the call for longitudinal studies to better understand the importance of principals' leadership practices over time (Hallinger & Heck, 1998). This study shows that principals are important agents of change through the way they act as leaders (Leithwood & Louis, 2011). Successful school principals signal a clear vision of the school's work, which, when embraced by the teachers, becomes rooted in the school's activities.
The results have important implications for educational policy. Organizational change must happen naturally through schools' activities and social interactions. Teachers and principals at failing schools must look deeply at the norms and values that are rooted and structured in the social interactions of the school and initiate processes of change and improvement that specifically target these norms and values (Hallinger & Heck, 2011). Policymakers must create and maintain institutional arrangements of educational systems that promote professional collaboration. The Swedish government recently launched national school development programs that support schools with low achievement levels, as identified by the Swedish School Inspectorate. This policy agrees with the Swedish tradition of equity and a "school for all." It may be seen as an alternative to policy responses that include strengthened accountability systems in, for example, England and the US.
One important limitation of this work is that teacher competence was not taken into account when selecting cases. We recommend further research to elaborate on the use of comparative case studies, and to operationalize the dependent variable in numerous ways. Questions to be answered in future research include whether traces of the organizational characteristics of successful schools can be identified in schools that go from low to high achievement levels or in schools with typical profiles. Such studies may also shed light on the question of why schools develop different sets of institutions to begin with; this is a question of the utmost importance for improving students' opportunities to learn at all schools.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the Swedish Research Council under Grant B0474901.

Notes on contributors
Maria Jarl is an assistant professor at the Department of Education and Special Education, University of Gothenburg. Jarl has a PhD in political science, and her research interests include education policy and governance. She has studied the origins of the decentralization and marketization reforms of Swedish education in the beginning of the 1990s, as well as their implications for schools and professionals. Jarl's theoretical expertise includes organizational and neo-institutional theory. Recently, she was the leader of the research program To Organize for School Success, financed by the Swedish Research Council. She has also been an adviser to the Swedish Minister of Education.
Klas Andersson is an assistant professor at the Department of Education and Special Education, University of Gothenburg. His research interests are the relationship between school organization and student outcomes, and issues concerning school leadership and teaching. His work also includes studies on the effects of teaching and education, especially how different approaches to teaching civics affect democratic virtues among students. Currently, he is conducting research as part of two projects targeting civic education, one funded by the Swedish Institute for Educational Research and the other funded by The Institute for Evaluation of Labour Market and Education Policy.
Ulf Blossing is a professor in education at the Department of Education and Special Education, University of Gothenburg. His research interest is school improvement and school leadership. Blossing has led and participated in several research projects where school culture, change agents, leadership, and improvement work have been in focus. He is a member of several research councils and is the chair of the council for school leader education at the University of Gothenburg.