Decision-making and control: perceived autonomy of teachers in Germany and Sweden

ABSTRACT Teacher autonomy has become an increasingly popular research topic over the past decade, reflecting wider national and global education trends. In this light, this article investigates and compares the perceptions of German and Swedish teachers concerning their professional autonomy. We analyse teachers’ perceptions using a grid, and view teacher autonomy as a multidimensional phenomenon taking place in different domains (educational, social, developmental and administrative) and at different levels (classroom, school, profession). The findings show that the teachers interviewed in Germany and Sweden value autonomy in various domains and dimensions differently, even if there also are many similarities. In instruction, that is, the educational autonomy domain, they perceive themselves to be very autonomous, in particular in relation to choices of content and method. Autonomous work in the classroom arena is also seen as the very core of the teaching profession. Overall, German teachers perceive themselves to be significantly involved in more areas of their work, and they refer much more to decisions which are to be made, whereas their Swedish colleagues are more concerned about control. Finally, we discuss the findings in relation to different nation-specific forms of extended or restricted autonomy teacher autonomy.


Introduction
Teacher autonomy has become an increasingly popular research topic over the past decade, and this reflects wider national and global education trends. The ways in which teacher autonomy (among other things) has been offered as a key ingredient for Finland's success in PISA (Sahlberg, 2011), and the teachers matter movement which promotes the importance of the individual teacher in student performance (Hattie, 2011) are two cases in point. Indeed, two international research reviews (Parker, 2015;Wilches, 2007) have been published quite recently, and both reviews summarize studies that clearly illustrate the positive impact of teacher autonomy on different aspects of teachers' work. These studies show that teachers' perceived autonomy can be seen to correlate positively with perceived selfefficacy, work-satisfaction, empowerment and a positive work climate. On the other hand, teacher autonomy seems to correlate negatively with staff turnover and risk of burn out (ibid.). Many authors in the field would argue that teacher autonomy is an important, almost magical, ingredient for a successful school and professional development. The literature grees that teacher autonomy should be taken to mean teachers' freedom, authority, scope of action or discretion in different areas of their professional lives (Parker, 2015).
Another key issue raised in the teacher autonomy literature concerns the multidimensional nature and contextual embeddedness of the phenomenon (e.g. Frostensson, 2015;Ingersoll, 2003;Wermke & Hösfält, 2014). This body of literature suggests that the teaching profession should be understood as a layered phenomenon, including the individual teacher as well as different groupings of teachers at the school-level and more widely. Also, due to the complex nature of teachers' work, including, for example, pedagogical, social and administrative tasks, teachers should be understood as operating in multiple areas. Put simply, teachers may be more autonomous in certain areas or dimensions of their work than in others, which should be reflected in research. Furthermore, research should remain sensitive to the possible differences between the individual teachers, as opposed to professional groups engaged with teacher autonomy. Furthermore, it has been shown that teacher autonomy depends on different contextual factors such as regimes of governance or educational traditions (for overviews see e.g. Wermke & Höstfält, 2014;or Mølstad, 2015; in this journal, also Wills & Sandholtz, 2009;Wermke & Forsberg, 2017).
This article presents an empirical interview study that analyses how teachers in Sweden and Germany perceive autonomy in their daily work, thereby investigating the multidimensionality and different nature of teacher autonomy in various national contexts. To this end, we will build our empirical foundations on the considerations concerning the control of decision-making in schools made by Richard M. Ingersoll. We argue that a German-Swedish comparison is particularly interesting because it enables us to relate the phenomenon of teacher autonomy to contextual factors such as different governance regimes: Both countries present western K-12 school systems, based on democratic and meritocratic values. However, since the beginning of the 1990s, the school governance regime used in each of the countries has been almost the converse of the other. Strategies that were implemented in Sweden at the beginning of the 1990s are on the one hand, 'standards-based reforms', envisaged to control the schools' output through the setting of standards and the evaluation of schools. Competition strategies, or so-called choice policies, have been central. This means mainly that quasi-markets are established within the education system; for example, by shifting from supply to demand financing (Englund, 2012). Although there have been many attempts at reform, with the new model of school governance, with governing by goals and results as its aim ('neue Steuerung'), Germany is-even on the state level-an example of centralized regulation of inputs and economic resources, detailed curricula or the professionalization of teachers are still emphasized (see for an overview of this Altrichter & Maag Merki, 2016;Wermke, 2013). The role of teachers as state employed and certified civil servants remained untouched (Terhart, 2011; see for a more detailed on this discussion, Wermke, 2013). These configurations provide an interesting variety of national contexts (Jahn, 2007) that might assist us to further understand the impact of context on teacher autonomy. This article is delineated in the following way. We commence by presenting an empirical understanding of the teacher autonomy question as presented in the seminal work of Richard Ingersoll, who proposed to render teacher autonomy empirically investigable by understanding the issue in terms of what are the most important decisions to be made in schools, who makes them and who controls whether the decisions have been made appropriately. From this perspective, we present an analytical device that will guide the interviews conducted in both countries. Finally, after the analyses we will discuss our findings.

Understanding teacher autonomy as decision-making and control
Teacher autonomy is often seen in a normative way, as something professional teachers must be. It has also often been confused with freedom or teacher empowerment. This lack of conceptual clarity has consequences, not only for the conceptual understanding but also for empirical investigations of the phenomenon (Wilches, 2007).
Yet, for a variety of reasons and despite its widespread use, the meaning of teacher autonomy and its implications for schooling and school stakeholders remain opaque. First of all, the lack of correlation among theorists within and across subject areas has resulted in a notable inconsistency in the use of the concept [. . .]. Second, although teacher autonomy has been connected to a number of theories including professional development, teacher decision-making, teacher efficiency, and empowerment, this relationship still remains unclear [. . .]. (ibid. 2007, p. 246) The life in schools and of teachers is very complex, which complicates in international comparisons of these phenomena. Here, we employ a theory of teacher autonomy that reduces this complexity in order to make it empirically investigable. In this study, we facilitate an understanding of autonomy as combining (only) aspects of decision-making and control in teachers professional work.
The starting point of this comparative study is Ingersoll's (1996Ingersoll's ( , 2003) theory on power distribution in schools and who controls teachers' work. This rationale has also been put forward by Cribb and Gewirtz (2007). However, while the latter's work was merely theoretical, did Ingersoll (1996Ingersoll ( , 2003) build and test his ideas on thick empirical data. That is why we think, his approach is an appropriate vantage point for our study. Despite the nation-specific differences, his reasoning is not dependent on national but rather organizational context. The focal point is the key decisions that affect the content and conditions of teachers' work and how teachers' abilities in decisionmaking relate to a well-functioning school. Teacher's decision-making abilities are seen in relation to the degree of conflict between teachers, students and school administration.
Autonomy is here seen as having a high degree of control over issues concerning daily activities. Ingersoll (1996) states that teachers' capacity to exercise their autonomy has a positive effect on school function, but that the effect is dependent on which activities they can influence. His results indicate that teachers' scope of action within the school's social sphere is of greatest importance to how well schools function.
In his work 'Who controls teachers' work? Power and accountability in America's schools ', Ingersoll (2003) takes his thesis a step further and defines which questions should be asked to determine who holds the power within an organization: what are the most important decisions, what standards and criteria should be used to evaluate the distribution of power between members of an organization and how should we accurately measure who has the power in regard to the relevant questions and decisions? Power here refers to the extent to which an individual or group influences or controls certain questions, decisions, behaviours and individuals (Ingersoll, 2003, p.18).

The school organization
Raising children and youths to become citizens (whatever their eventual characteristics) is seen as one of the school system's most important functions. The teacher is responsible for the transference of societal norms and values to the students. According to Ingersoll (2003), rationalized bureaucracy is an obvious way to implement this mass education. Organizations that are hard to bureaucratize are called loosely coupled systems, the school being the archetype. In a complex environment like the school, there can never be rules to cover all eventualities and thus the organization has to have certain flexibility. A loosely coupled organization does not equal a high degree of teacher autonomy or a decentralized organization; a lack of standardized rules can instead be a source of administrative power and centralization. Non-bureaucratic environments can in actuality be more centralized than bureaucratic ones. In such an organization, the employees have considerable control and responsibility over their work, but unclear liability. School power structures can be visible, invisible, formal or informal. Ingersoll (2003) makes an important distinction between the delegation of responsibility and the delegation of power, 'the division of labour is a distribution of responsibility and control' (Ingersoll, 2003, p.143), this being one of the administration's most important mandates. Delegating decision-making on less important issues is often used as a subtle centralization of power where the employees experience participation in organizational control that in reality does not exist. Normally teachers lack control over key decisions concerning their work but have a great deal of responsibility (Ingersoll, 1996). Ingersoll (2003) asks the question of whether embedded structures control, influence or limit teachers' work. Guidelines, rules and policies that become part of everyday norms are less likely to be questioned and can thus be used as effective means of control (p.137). This also means that a lack of obvious steering methods can indicate the presence of effective hidden methods of control.
An individual focus, where more responsibility and commitment leads to a higher salary, leads to competition between teachers to excel at work. Horizontal conflicts between employees instead of vertical ones between teachers and administrations are advantageous from a management perspective. Deciding what constitutes meritorious performance is also a form of control (ibid, 2003). Administration is another mechanism of hidden control, since teachers' obligations to document their teaching and other tasks indirectly creates the opportunity for management to follow up and ensure that teachers do what they should (Ingersoll, 2003, p.116). External actors can also exert a form of hidden control, since teachers have a large responsibility for their students both socially and academically and society and parents often see teachers as more accountable than they really are, which in turn leads to teachers believing that they will be held accountable. The expectation of a negative reaction is a control mechanism in itself (Ingersoll, 2003).
Control over resources means power, this not only includes money but also scheduling, classroom allocation, etc. This is a potential control factor under which teachers compete between themselves to get their wishes catered for, and there is also a risk of developing a culture of favours and favours in kind. Teachers automatically end up in dependency relationships not just because of resource allocation but because teachers are dependent on the school management for support in their decisions and in conflict situations.

Control mechanisms
The conventional view of the division of power in schools sees the teacher as largely free to do as he or she wishes in the classroom. Ingersoll (2003) states that teachers' work is often described using the 'egg crate model' where teachers are controlled by the school administration and separated and isolated in their classrooms. Even in schools with a great deal of collegial collaboration, the majority of work is done individually and each teacher is responsible for his or her own work, something Ingersoll (2003) calls 'individual solutions for organizational problems'. However, this can also foster commitment and loyalty from the employees (p.171). This responsibility and scope of action can easily be mistaken for autonomy, but in reality teachers have considerable responsibility to implement decisions but little influence over the decisions. A large workload leads teachers to find ways to simplify their work by standardizing their teaching, for example, using preplanned lessons (Ingersoll, 2003, p.152). Simplifying and standardizing work requires a less skilled work force, and this method of separating complex work into simple parts is called 'deskilling'. From a management perspective, a workforce where everyone can perform all tasks makes it easier and less costly to organize work. However, while the academic work in schools can be simplified, the social work cannot, since humans are complex by nature (Ingersoll, 2003, p.154-158). Ingersoll (2003) states that the degree of control and the question of who exercises it depends on where you look, but generally speaking, administrators take strategic decisions outside of the classroom and teachers take operational decisions within the classroom (Ingersoll, 1996).
In conclusion, an advantage with Ingersoll's work is its empirical usefulness. His questions are universal; they can actually be employed in different contexts, which is very important in a comparative study such as ours. We will develop this in the following method section. However, we want to remind here also, firstly, that the perspective of Ingersoll is of organization theoretical nature. That might limited its explanation value for curriculum and schooling questions at a macro school system level. We will still use this because we handle here the perceptions of teachers regarding autonomy questions at different levels, and not the decisionmaking power they actually have. Regarding this assumption, teachers are still part of the organization of a school, even when they consider the situation of the teaching profession in a school system. Secondly, the value of his ideas for empirical and comparative investigations comes at the price of complexity reduction in our material. A focus on teacher autonomy as decision-making and control excludes other themes related to the concept in question, such as teacher empowerment, the structure of teacher agency and also issues of teacher self-determination, as put forward by research on motivation.

Method and material
Asking teachers about autonomy The operationalization of the interview questions originates from Ingersoll's perspective on how organization and control of teachers' work affects schooling. As in Ingersoll's (1996Ingersoll's ( , 2003 work, the term control here refers to steering. The aim was to capture teachers' perspectives on autonomy through the use of semi-structured qualitative interviews. The purpose of the interview guide was to expose organizational power structures through teachers' perceptions of autonomy and decision-making.
The interview questions are based on those of Ingersoll: (1) What do teachers consider to be the most important decisions to be made regarding school function? (2) Who do teachers consider makes these decisions (the principal, teachers, private or municipality actors, parents, etc.)? (3) How are decisions monitored and followed up and who ensures that decisions are implemented 'correctly'?

Sample
Asking teachers within an interview study on autonomy with a limited sample of 10 Swedish and 15 German teachers embeds this study within interpretative research, which has consequences for which findings can be generalized in which way. We will discuss this below. In his classic Handbook on research on teaching chapter, 'Qualitative Research on teaching', Erickson (1986) presents a thorough elaboration of interpretative research on teachers and teaching. The present study views itself as continuing this tradition. We aim to expand the understanding of the multidimensionality of teacher autonomy as it takes form in relation to different national school systems. In Erickson's words, we are interested in 'social topology rather than social geometry.' (ibid., p. 133). We focus not on how a certain population actually is, but instead show patterns and relations of different phenomena (Larsson, 2009). This principle guides our sample construction. Speaking about German teachers and Swedish teachers means first of all an easily understandable distinction between two national groups. Properly, it means to speak about teachers working under circumstances as they exist in Germany or Sweden. However, even this needs further qualification. Germany is a federal republic, consisting of 16 so called Bundesländer states, that have sovereignty over public education. This would mean that the German system actually consists of 16 school systems. This can become more complicated when we look at the decentralized nature of schooling in Sweden, in which the municipalities and charter school owners have even more autonomy in shaping schooling than the German states. From this point of view, we might envisage more than 290 Swedish school systems (the same as the number of Swedish municipalities).
The reference to German and Swedish leads to the issue of what is seen as relevant in a school system and what teachers do even in small-scale samples, and this can be seen as representative of the respective national teaching profession. This may be particular political conditions, similar structures in teacher education or similar employment conditions: For example, are teachers employed by the state, following a fixed salary scheme for all teachers; or are they employed in a school or company and able to bargain their salary themselves? It is consequently concerned with finding the conditions that are alike for the teachers in a country and examining the possible impact of these on the phenomenon in question. For example, even if, in Germany, the different states are responsible for teacher education, all states harmonize their school system in an institution: The Standing Conference of the Ministers of Education and Cultural Affairs of the Länder in the Federal Republic of Germany. The Standing Conference of the Ministers of Education and Cultural Affairs is the oldest conference of ministers in Germany and plays a significant role as an instrument for the coordination and development of education in the country. It is a consortium of ministers responsible for education and schooling, institutes of higher education and research and cultural affairs, and in this capacity formulates the joint interests and objectives of all 16 federal states. The tracking of school systems, the tenure-track civil servantship of teachers, and a teacher education that is separated into two periods are the same in all lander. Moreover, more recent trends, as referred to above, are rather similar (Tenorth 2008). 1 Nevertheless, we aimed to generate a particular variance in our sample in order to deal with what Erickson (1986, p. 144) states as a risk in interpretative research: The researcher's tendency to leap to conclusions inductively early in the research process can be called the problem of premature typification. This problem makes it necessary to conduct (while in the field, and in subsequent reflection after leaving the field) deliberate searches for disconfirming evidence in the form of discrepant cases-instances of the phenomena of interest whose organization does not fit the terms of one's emerging theory.
Our sample, as described in the following, aimed to generate a variance that allowed us to test our ideas even against distinct cases, in other words, having a certain variance of different school contexts and indeed also different individuals in our sample.
School culture can differ widely between schools. In the Swedish case, two schools with different types of steering were examined to make it possible to compare contexts instead of individuals. In Sweden, both municipality and independent schools have adhered to the marketization mindset, which views students and parents as customers (Frostensson, 2015). Choosing an independent school that has a more unusual form of steering makes for more interesting comparisons than choosing two schools that, despite different mandates, have similar steering. Five individual interviews were held at the municipal Sjösala School and five at St. Isidors School which is run by a foundation and has a Christian heritage. They are both located in the city centres of mid-sized to large Swedish towns.
The German schools in our sample also represent a variety of possible schools in the German context. We focus on the voices of 15 teachers who work at 3 different schools in a larger German city. The first represents a traditional secondary and upper secondary school (a so-called Gymnasium) with a population of students from high socio-economic backgrounds, called the Goethe School. Most pupils at the school have ambitions to continue their education at university level. The second school, Town Hall School, represents a school with students from a lower socioeconomic status, with many students at risk. It is also a rather small school with a faculty of 30 staff. Finally, school number 3, Ghandi School, represents a school with heterogeneous population of students. It is a very large secondary school with around 120 staff in faculty.
Much of the volume and quality of the interviews depends on the field access you have. If we were to employ an ecological terminology, we might say that schools in many countries are 'overfished' today, in both countries in focus. Not only is there a tremendous amount of empirical research today which focuses on schools, but also surveys by state departments and municipalities, as part of different but interrelated processes such as massification, marketization and managerialism, have exhausted teachers in a way that it becomes increasingly difficult to be given access to schools. In particular, for our study, this led to a sense that there was greater sensitivity, concerning, for instance, teacher autonomy in relation to principals, etc., challenges in recruiting enough teachers, and this limited our opportunities to construct deliberate samples. We employed gatekeepers who helped us to gain entrance into the schools and helped to persuade individuals to commit to the study. We recruited the gatekeepers in our professional networks, in Masters courses, through colleagues and so forth. The teachers required that we followed the two most important ethical guidelines in qualitative research, as again Erickson (1986) describes. Those studied, especially those studied as focal research subjects, needed to be (a) as informed as possible about the purposes and activities of the research that will occur, and of any burdens (additional work load) or risks that may be incurred by them while being studied. Focal research subjects also needed to be (b) protected as much as possible from risks. The nature of this study is not one of physical risk, as in some medical experiments. Psychological and social risks (embarrassment and/or vulnerability to administrative sanction) are the most common in our fieldwork research.
Still, the risks of psychological and social harm can be substantial when fieldwork is done by an institutionally naive researcher who has not adequately anticipated the range of different kinds of harm to which persons of varying social position in the setting are potentially liable. Liability to risk is often greatest between members of differing interest groups in the local setting. (Ibid., p.141).
For the sake of transparency, the teachers we were able to convince to participate in our study included the following: The teachers in the Swedish sub-sample had a working experience of between 7 and 36 years (average approx. 20 years). We interviewed six women and four men, a majority of whom had a background in Social science and language subjects. The German subsample comprised teachers with a working experience of between 1 and 31 years (average: 11 years). We interviewed nine women and six men. The majority had a natural sciences background. 2

Analysing the interviews
In the interpretative approach to our interviews, we have developed an analytic device in order to be able to compare our nation-specific findings. In the following, we present an analytic device depicting teacher autonomy as a multidimensional empirical phenomenon. This means that teachers can obtain and/or lack different forms of autonomy (individual, collegial and professional autonomy) at different levels of autonomy in different domains of the profession. The device shows these different domains and levels in terms of where teachers do or do not have the ability to exercise their autonomy.
Drawing from conceptual work that distinguishes different dimensions of teacher autonomy, we distinguish teacher autonomy on three horizontal dimensions: classroom-related, school-related and profession.
(1) With the classroom dimension, we refer to the scope of action related to classroom work. Studies with such a focus examine the micro context in which a teacher operates, and the scope of action they have as individual professionals. This relates to Frostensson's (2015) conception of teachers' individual autonomy and Wermke's and Höstfält's (2014) idea of service autonomy. Examples of such studies would include, for example, a teacher's choice of taught content and methods.
(2) The second dimension is the school as the arena in which teacher autonomy is formed. Frostensson (2015) refers to this as the staff or faculty dimension, whereas we call it the school dimension. We argue that at this level, relations with parents can also be of interest for the autonomy researcher. In other words, studies with a focus of this kind tend to pay attention to the local contexts within which the teachers operate. This refers to studies that explain schools as complex social systems with multiple actors, that is, systems in which teachers not only operate in different roles but also in which the autonomy of one party may promote or inhibit the autonomy of others. This has been shown in research on teacher cooperation and collegiality (see overview by Kelchtermans, 2006). Here, we will warn against confusing teacher autonomy, which emerges at the school level, with increased local autonomy, or school autonomy, witnessed, for example, in relation to the Swedish Friskola movement and the English Academy movement, since the latter does not automatically equip teachers with increased autonomy (Salokangas & Chapman, 2014).
(3) Finally, there is a professional dimension of teacher autonomy-as proposed by Frostensson (2015). This refers to the overall group of teachers. Autonomy is discussed as an asset of the teaching profession in relation to others and includes teachers' academic knowledge base, status, education and so forth. Here the relations of teachers to other actors in a school system, such as textbook publishers, unions, school inspectorate and so on, are central.
This horizontal dimension is complemented by a vertical dimension, which refers to the different domains in which teacher autonomy can be formed. In relation in particular to the work of Wilches (2007), Ingersoll (2003), Friedman (1999) and also the classic study of Rosenholtz (1989) concerning the workplaces of teachers, we propose four different domains. Our definition of the domains as presented below may be open to conjecture, but above all our grid attempts to capture the multidimensionality of teacher autonomy. The discriminatory power of the domains must also be investigated in further empirical research.
(a) Firstly, with educational domain we refer to matters related to activities and responsibilities related to teaching and learning, including, but not limited to: planning, instruction/delivery and assessment/evaluation. (b) From a sociological point of view, education and schools more specifically play a crucial role in the socialization of students. We call this the social domain. Examples of such processes would include, for example, grouping students, either randomly or based on their gender, ability or developmental stage. Another example of socialization would be disciplining of students, and the extent to which teachers contribute in the actual act of disciplining, as well as in developing school-level discipline policies. A third example would be the treatment of students with special needs.
(c) Developmental domain refers to decisions that relate to identifying and steering the school towards a 'vision' or a plan of action. To what extent are teachers involved in developing the school, and in steering the direction of the school in matters such as professional development of staff decisions related, for example, to developing overall school subject specialization, or other strategic functions. Here, we propose to include only formal activities. Teachers' professional development is often also characterized by informal professional development (see Wermke, 2013). However, this happens privately and is privately funded and therefore sees teachers rather as private individuals and not as professionals.
(d) Finally, by administrative domain, we refer to the administrative work of schools that facilitates learning, and other possible activities in schools, distinguishing educational and administrative duties of teachers. Decision-making concerning, for example, timetabling, use of resources or teachers' pay or office space are examples of administrative functions in which teachers may or may not be involved.
Perceived decision-making and described control were coded following the analytical device. We used similar colours to display how teachers perceived decision-making and control in the two countries (see Figure 1): domain: education/level: classroom; domain education/level: school; domain education/level profession; domain social/level classroom; domain social/level school; domain social/ level profession and so forth. The coding work was processed in ATLAS TI. We coded each interview transcript following the device, and then compared between the schools and between teachers with more or less work experience and different subjects. Variances in the national sub-samples can be seen regarding which dimension/cell had been discussed more, or perceived as more or less important, not regarding patterns of decision-making and control. However, the device and the questions asked might have limited the degrees of freedom in the question in focus. Regarding a description of decision-making and control, the teachers in each country answered in quite similar ways, which illustrates the impact of the different governance models as described in the introduction. In other words, from a governance perspective on teacher autonomy, operationalized as decision-making and control, the differences between the subsamples are greater than within the two subsamples.
Employing the described coding strategy, we have chosen citations from the material that we believe are both representative and also illustrative of the national teachers' reasoning on autonomy in order to report them to the audience. We decided to present citation from the different schools evenly within a national context. In order to present the citations in a more lively manner we employ fictive names.

Decision-making and control of Swedish teachers
As we can see in Figure 2, Swedish teachers in our study describe an extended decision-making power in the classroom. In the classroom, results are what matters to the administration and the parents, not methods. The grades are under scrutiny from both the parents and the school administration and it is important for teachers not to deviate too much from the statistics since that means opening oneself to questioning.
Parents also serve as a controlling factor when it comes to student well-being in the classroom. Partly, the findings of the Swedish teacher study provide a positive answer to Ingersoll's (2003) question of whether embedded structures control, influence or limit teachers' work. When asked about who controls teachers in their everyday decision-making, the participants mention that the grapevine, or word of mouth (djungeltelefon), serves as a control function to disclose colleague behaviour or decisions. Eva-Lena from Sjösala school says that, '[. . .] if there's a teacher who doesn't manage to perform the teaching or gives weird grades or something, then it's somehow revealed through the grapevine.' Jörgen from St. Isidors school makes a similar statement; 'It can be students or colleagues that notice that this wasn't good [. . .]' The collegium also serves as an important factor here in that the application and agreement of social rules and control functions are largely decided upon collegially.
A great deal of the responsibility (regarding social questions) is on the teacher teams but also on the collegium, where you agree on the rules, how to maintain them and what the control function is [. . .]. (Britta at Sjösala School) Regarding Ingersoll (2003) are informal codes of conduct that are applicable at schools. A lack of formal consequences in the teaching profession does not mean there is no control; this social contract serves as an important control function, since breaking it would mean being ostracized.
At the school level, teachers also experience a high intensity of control in the administrative domain, since the principal and administration decide on matters regarding resource allocation. There is a great difference between different types of decisions, and when it comes to school-wide issues like scheduling and resource distribution, teachers can make suggestions and participate in discussions, but this does not mean they have any say in practice.
We get to voice our opinions, but they do as they wish anyway. . .then it doesn't always turn out as they decided either, for some strange reason (laughs). (Eva-Lena, Sjösala School) However, teachers do have some say at a classroom level, since they can decide how to distribute the economic resources allocated to their subjects. When it comes to smaller issues, teachers sometimes feel like they participate too much in what they experience as unimportant decision-making. Jesper, a teacher at St. Isidors School had a telling answer when asked about the most important decisions regarding professional development; 'What did you say it was called? [laughter]'. As shown in the matrix (Figure 3), Swedish teachers do not experience control in the developmental domain, which is no surprise, since professional development is a rare occurrence for the study participants. Teachers are expected to take part in current research and improve their skills; however, since they are not given the resources, there is no real incentive to do so except their own sense of professional responsibility.
There is a heavy focus on assessment in the Swedish school system, both at a national and at a school level. Due to the decline in PISA statistics over the recent decades, the results have become a societal issue and are debated among politicians, scholars and the general public alike. As a result of the marketization, grades are also one of the most important competitive advantages for the schools to attract students. This has resulted in increased standardized testing, a new and more specified curriculum (Lgr11) and a new grading system.
If you look at PISA results they have more impact on the societal debate than before, when there was sort of a focus on creating a democratic citizen. (Heiki, St. Isidors School).
Several of the participants in the study think that the recent curriculum has limited teachers' autonomy and is a form of soft governance from the state; however, some of them feel that this is a good thing since it has made it easier to plan their teaching. By simplifying and clarifying teachers' work, the complexity of the profession decreases. A consequence of this governing is the escalation in administrative work for teachers. Lars at St. Isidors School thinks that; 'There's too little teaching, it's more about collecting assessment bases, that's what I feel sometimes.' The teachers feel that they need to document and motivate all grading, since there is always a risk of parents, and sometimes the administration, questioning grades. In addition to increased control from the state, these administrative duties also give the school administration the possibility to indirectly control whether teachers perform satisfactorily (Ingersoll, 2003).
It is a paradox that Finland, and also Germany, with its strong but conservative teaching profession, are so strongly emphasized as positive role models when in recent years educational reforms have focused on increasing school autonomy instead of strengthening the professions' institutional autonomy (Wermke & Forsberg, 2017). Instead, management and political governance has increased even though a reform argument held that extended local or school autonomy would increase teacher autonomy. The introduction of individual salaries also shifted the focus from the collective to the individual, which serves as a means of steering for school administrations, since they decide what counts as meritorious performance. Ingersoll (2003) claims that the solution to raising the quality at school is to raise the quality of the teaching profession (p. 249). This is something Swedish politics has embraced with the introduction of teaching permits, career jobs and higher salaries among other things. The question is whether these measures are enough. Frostensson (2015) states that even though the teacher certification reform was supported by the unions, it was organized in such a way that the content was decided upon by the state, meaning that in reality it weakened the professional institution.
Several participants pointed out that beyond supporting the students' academic development and ensuring their success, parents expect them to know about their whole social background and personal issues, something that can easily create a conflict between answering to all standards and demands while at the same time being autonomous. At the same time, the customer perspective of New Public Management has resulted in a greater emphasis on students' wellbeing and personal growth at the expense of the profession's traditional knowledge values which further undermined teachers' professional and institutional autonomy (Frostensson, 2015). All of the participating teachers stated that all these developments have resulted in a focus shift towards assessment in their teaching.
It's changed and it's also the fact that you have to be on your toes due to the fact that students and parent that can contest grades and things like that so you need to know what you're doing and that's a change I think compared to when I started, ok sure, maybe they had the opportunity but it wasn't done with the same frequency, yes maybe you had to discuss it but usually you could warrant it and that was like enough and they bought it. . . (Mattias, St. Isidors School) In particular, the new curriculum can thus be said to be a success insofar as it has managed to focus teachers' practice on the results, which was one of the goals. The latest PISA results show some improvements, but whether this remains a lasting trend has yet to be seen. While steering and standardization serve as a means of deskilling that has restricted teachers' autonomy concerning assessment and teaching content, it has at the same time increased the focus on academic knowledge. However, a greater knowledge emphasis demands highly knowledgeable teachers, so the question is then to what extent can teaching be deprofessionalized.
Yes, I think that the most striking difference is the new curriculum, that it limits teachers' autonomy. The last curriculum was based on teacher's autonomy, meaning the teachers themselves were supposed to make these curriculums and knowledge criteria and so on, meanwhile this new one has removed the decision-making from the teacher since the national testing has increased and includes more subjects, you can see them as guidelines but it's also a form of control. . . There's also a change in the fact that direct results are what counts. (Maria, Sjösala School).

Decision-making and control of German teachers
In Figure 4, we show how teachers perceive decision-making in Germany. We can see that teachers view most decisions at a classroom level as being in their hands. The significant school level decisions are made by teachers collectively or together with the principal. The principal in the German system is a teacher, and traditionally seen as primus inter pares (first among equals). A school is apparently owned by its teachers, and led by a principal who is also a teacher. That teachers experience the classroom to be sacred and that they share the responsibility in the school with colleagues might be seen in the light of what Lortie (1975Lortie ( /2009) classically described as a parity-autonomy principle of schools. Teachers are equal and autonomous from each other in the classroom (a phenomenon also described by our Swedish teachers). I: What must happen, what must they take away from you, that you would say: now my professional integrity is so harmed, that I will throw in the towel? Petra: If I were forced to permanently report my content and how I teach to someone. . .like every day. I: That means, if they took your freedom in teaching? Petra: Yes exactly, I would go immediately.
Eva: If I weren't allowed to decide any more how I teach different stuff to the pupils. If they prescribed the way. This would be a standardization I could not live with. I: Do you mean here the methods, or rather a standardization of your personality as a teacher? Andreas: Both, a) if they standardized the teachers, and b) if they prescribed how teachers have to methodologically teach certain things. I want to find my way on my own.

(Teachers at Ghandi School)
It is indeed true that the most important decisions are discussed and decided together in a subject department. However, one significant actor is the subject head. In other words, there is also a hierarchical structure in German schools, which we will discuss below in the control section. The professional level was not discussed so much, which is itself an interesting result.
Iris: And then there is also the syllabi developed in our school (school curriculum), which means a tremendous control, for example, in the subject German, which we both teach. The syllabi expect so many subjects to be taught, that it is actually impossible to manage, at least if you have any ambition of being thorough. Maike: Of course.
Iris: . . .you are, however, not allowed to say this. Otherwise you get in trouble. Deep red: Individual teachers; light red: teachers with colleagues; orange principal; white: other actors (e.g. at municipality or state level) I: With whom? Iris: With the head of the subject department.
I: So, the head of the subject department is an institution of control? Maike: At this school it might differ between different subjects.

(Teachers at the Goethe School)
German teachers perceive control to be less intensive than their Swedish colleagues ( Figure 5), even if the control patterns are quite similar, in particular with respect to the educational domain, which is the most controlled domain, which is not surprising. So, according to parents, the importance of teachers differs not nationally, but between schools.
Final examinations! The future! They are panicked, afraid. Parents go to the principal directly. (Sina, Ghandi School) However, control is exercised by the profession itself and located within the schools. A very important control institution in German schools are subject groups with their subject heads. Today they actually do have a high impact instrument in their hands to facilitate collegial decision-making, the school internal curriculum, which means syllabi developed in the school for each year and each subject. In particular, whether such documents exist in a coherent way is controlled by the school inspectorate. Another aspect is the final examinations that are organized centrally in each German state, which influences and controls teachers decision-making [central examinations Abitur (after 13 years of schooling) or mittlere Reife (after 10 years)]: The 10th grade graduation examinations (Mittlerer Schulabschluss, MSA) influence the whole school year for the 10th graders. It is like a sword of Damocles over their heads. The examinations, the examinations, everything is only directed toward the examinations. We do everything for the examinations, we train for the examinations, we do not do anything else. (Linda at Town Hall School) The social component our teachers perceive as mostly controlled by their pupils' parents. Here, only the classroom level is relevant, and the control is felt to be rather less intensive. It seems that teachers do not feel that the social domain of schools and the teaching profession as such are so controlled at all. This challenges Ingersoll's (2003) argument on the primacy of social questions when it comes to teacher autonomy. Since the classroom is sacred for teachers, the social component in education and teachers coping with it is experienced as a natural part of teachers' duties which is somehow a conditio sine qua non of teacher autonomy and goes largely unquestioned by the German teachers: I: Okay, last question: What is teacher autonomy? Anna: I am allowed to have an impact on sometimes 15, sometimes 25 young people. Who can say this about themselves? Lisa: To guide someone in life. In particular, from grade 7, when the important developmental period of puberty is going on, in which attitudes and beliefs about life are shaped, you can be a role model for pupils. Peter: For me it is about my very own values. Presenting these to the developing young people, and maybe sometimes have some positive impact on one or another.

(Teachers at Ghandi School)
The most significant finding regarding our control issue is that it is actually processed locally, in the subject departments, and here the control is felt by our German teachers to be highly intensive, a phenomenon that recalls Bird's and Warren-Little's (1986) early statement that the significant units for understanding schools might actually be rather subject departments.
I: So, you as subject head, are the link between teachers and school management? Linda: Yes, this is true for the subject German, my subject department. If there are any problems, you would go to the department head and cope with the problem within the department.

(Teachers at Town hall school)
Nevertheless, teacher education is an issue where teachers feel the control of actors other than their subject heads or parents. The no-control picture at the professional level relates to the picture described above, in which teachers do not have any interest in the decisions made at this highest, the professional level. Formal professional development is up to teachers. The control happens only in an administrative way. Principals only control whether any formal professional development has taken place. The administrative domain of teachers' professional work is not an issue at all.
Here the centralized nature of the German system that for a long time has allocated resources to teachers is apparently still valid.
I: How much influence do you have on resource and organizational questions? Linda: Not much. Thomas: Yes, not much (laughs).
I: Even as subject head? Linda: Yes

(Teachers at Town Hall School)
Finally, what becomes apparent is that although teachers experience control, there is apparently a lack of consequences, even when everybody knows about a teacher's poor performance. That means that control becomes relative for teachers if there are no consequences even for negative professional performances. We will discuss this in the discussion section. The arguments of Ingersoll (2003) on control are valid. However, here we obviously see nation-specific particularities and the fertility of comparative research designs.
There are unfortunately no consequences. [. . .] I feel that with teachers with tenure status (Beamte) in particular nothing happens. There are some who never conduct written texts during the term. This is unacceptable. I mean grades are not everything, but this a significant feedback for students.

Discussion
In this last section, we perform a comparative analysis across both cases and provide a possible explanation of our findings. German teachers enjoy decision-making power at the classroom level to a great extent. They also have the possibility to exert collective autonomy at the school level, the exception being the administrative domain, but just as in the Swedish case they have no autonomy at the professional level.
As in Germany, the Swedish teachers rule the classroom in the educational and social domains, they also have a collective say in these domains at the school level. The principal as an individual has more power even when decisions are made collegially and collectively, since in reality the principal has the final say. The teachers have no decision-making power in the developmental and administrative domain, they can make their wishes heard but have no sure fire way to make them come true. In municipality schools, teachers' professional development is often decided upon at a municipal level thus it is also out of the principals' hands.
German teachers mostly experience control at the classroom level and in the educational domain. Student results, but not methods, are heavily controlled by the school subject heads and also by the parents. The school inspectorate controls whether there are means of control, such as school curricula, or school internal standard tests. Central examinations at the end of the 10th or 13th grade are important as means of control, also hold by the state. However, overall the German teachers are subject to a control of much lower intensity than the Swedish teachers, probably because there are fewer stakeholders involved. Swedish teachers feel heavily controlled in the educational and social domains at both the classroom and the school level, and in the educational domain also at a professional level. There is also a large number of stakeholders, the principal, the parents, the state, school inspectorate and so on. This coheres with the view of Frostensson (2015) that Swedish teachers have become restricted in their autonomy.

Similarities between Germany and Sweden
In both the German and the Swedish cases, the classroom can be seen as a sacred space for teachers' decision-making. This is where teachers can develop and express their own professional identities. Although educational planning is related to steering documents, teachers see the autonomous meeting with the students in the classroom as the core of the teaching profession. Parents are attributed a significant control function in instances such as grades and behavioural patterns by the teachers in both countries, but there are school related differences, in particular in relation to the socio-demographic background of the students. In the Swedish data, parents from higher social backgrounds voice more opinions regarding their children's education and tend to question teachers' decisions more often.
The results, but not the methods, are controlled, since it is the results that count. One issue to note is the fact that the dimensions in which teachers exercise their own decision-making must not be the ones that are most intensively controlled. For example, Swedish teachers are highly controlled in the educational domain at both a school and a professional level as well as in the social domain at a school level, meanwhile they have very little individual autonomy in these areas.
The relation of perceived control and real consequences is also very ambivalent, since there are practically no real consequences. For a teacher to forfeit their permit to teach or to lose their job requires extreme causes.
In both instances, the developmental domain is controlled very little or not at all, though German teachers have a higher degree of autonomy in this dimension and also take part in more professional development, something that for Swedish teachers can be practically nonexistent.

Differences between Germany and Sweden
In Germany, the school level is the focal level of autonomy from a decision-making perspective.
Here collegial decision-making is of very great importance, since all decision-making is based on so called conferences (from the Latin 'to bear together'). For example, there are school conferences, subject conferences and marking conferences. There are many control technologies but no formal consequences; however, there can certainly be informal ones if a teacher fails to perform his or her work to an acceptable standard. The Swedish school system is quite heavily controlled on the professional level through standard testing. This is seen as a yardstick that student results should not deviate from. Teachers also experience pressure from the media, the research community and different policies, since, due to Sweden's decline in PISA results, there is pressure to deliver good results. The marketization of the school system, resulting in a wide variety of schools, has led to significant differences in intra-school control regimes between private and municipality schools in terms of more or less collegial control.

Different governance regimes and teachers' perceived autonomy
When discussing decision-making and control in schooling, Swedish teachers are more likely to mention control issues, which is not surprising since control is also perceived as being more intense. There are intense accountability issues. German teachers reason more about decisionmaking than about control. There is a greater wish for control from outside. This picture confirms to some extent the reasoning of comparative governance research.
A plethora of comparative education research (cf. Wermke & Höstfält, 2014;in this journal;Wössmann, 2007;Benner, 2009;Hopmann, 2003) has put forward two opposing government regimes for the teaching profession, which can be seen in western school systems (See Table 1). More outcome-controlled systems open up possibilities for a greater diversity in service, which means different forms of schooling and instruction. The price to pay is that control is rather rigidly processed from outside the teaching profession, such as by school administrators and also scientific researchers. In contrast, in input-controlled regimes, teachers are civil servants, and the path into the profession is rather difficult. Resources for schools and teachers are standardized. These, as well as rather strict syllabi, do not allow for much diversity of teaching and schooling. However, in relation to the strict input regulation, there has been little control of the teachers throughout their career. This does not mean that there is no control, but that the control is contained within the profession, exerted by hierarchical structures within schools in which principals as head teachers, subject heads and teacher educators in the school play significant roles. Since control is from within, it might not be so visible. Swedish teachers work in an outcome-controlled system, in which external control is the primary issue. Their German colleagues are active in an input-controlled system (grey in the table). Here the issues of schooling draw more on collegial decision-making. Table 1 relates our findings to the different regimes of governance.
Finally, regarding our material, in both systems parents' satisfaction plays a major role for teachers as a reference of control. As described in the introduction, the Swedish system is characterized by marketization with school funding related to parental choice policies. Parents have greater actual influence in Sweden. However, German teachers nevertheless perceive highly intense control by parents as well. This finding cannot be related to the dichotomous model of governance. Other models, perhaps, focus on public education in a field of tension between civil society and state government. However, we argue that regarding the role of parents and teacher autonomy further empirical comparative investigation is needed.
In conclusion, this study of two countries' teachers' perceptions on autonomy has drawn on an empirical understanding of teacher autonomy that focuses on the processes of decision-making and control. This approach we have complimented with an analytical grid that enables to unpack and show the complex relations in schools' and teachers' life. We hope that this approach might inspire future advanced comparative research in education. The presented national cases have illustrated the relations between nation-specific particularities and teachers' perceptions of their autonomy. This might contribute to a further understanding on the complex and contingent relations of context-specific governing traditions and the life in schools. Notes 1. Another argument for a variance within the German teaching profession has been the East-West divide between 1949 and 1989. The differences are however quite little. Already, 1999 (Gehrmann, 2003) states drawing on a large study of West and Berlin teachers on professional values no more significant differences, as that East Berlin teachers were more conservative regarding certain values. In our study, we take east west differences not into consideration. 2. For a more detailed description of the individual in our sample, see e-complement of this article.

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

Funding
This work was supported by the Vetenskapsrådet. -More externally (i.e. not by teachers) regulated standards and measurement -Teacher accountability through external control and incentive regimes -Legitimation of good practice through efficiency in standards achievement in relation to the curriculum -Rather little external standards testing (or testing with no consequences) -Strong boundaries for the profession, administrators/teacher educators have teaching background -Legitimation through teacher owned professional educational system of reason The practical dimension of the profession (service dimension) -More externally (i.e. not by teachers) and centrally regulated applications of resources and content of schooling and instruction, and teacher education -More liberal local (or decentralized) choice over content of instruction, application of resources Perceived autonomy in terms of decisionmaking and control -Extended discretion in educational decisions in classroom teaching, but intense control of results and grades through the principal and different stakeholders at a professional (governmental organizations etc.); Strong control through parents (as powerful actors in a marketized school system -Little collegial decision-making and control -Extended discretion in educational decisions in classroom teaching, but intense control of results and grades through department heads, principal and, first of all parents -Decision-making is often seen as collegially embedded in arenas of decision-making at school level -Little importance is given to the professional level regarding decision-making and control

Notes on contributor
Wieland Wermke is associate professor at Uppsala University, Department of Education. His research focusses on teachers' work and teacher professionalism from a comparative perspective. He has lead a research project on teacher autonomy in Sweden, Finland, Germany and Ireland.