The Swedish primary teacher education programme: at the crossroads between two education programme traditions

ABSTRACT In 2011, new teacher education programmes were introduced in Sweden. This policy case study examines the outcome of the reform, focusing on the pedagogic discourses in the examination practise of the primary teacher education programme, but also on the possible effects of dividing teacher education for primary teachers into two specialisations: for teachers of grades F-3 and 4–6. The point of departure is that the Swedish, and Nordic, teacher education programmes have been shaped by two traditions: the seminar tradition and the academic tradition. The study is based on Bernstein’s theories of pedagogic discourses and how these affect teachers’ professional knowledge base and professional identities. The results show that the primary teacher education above all prepares student teachers for everyday classroom life, but also that the examination practice and the pedagogic discourses differ to some extent between the two specialisations and that the primary teacher education (PTE) students who choose to specialise in grades 4–6 now encounter pedagogic discourses that more resemble an academic tradition than what is encountered by students who choose to specialise in F-3.

integrated education programme was something Swedish politicians had been working towards since the introduction of the 1988 primary teacher education programme. Instead, the government is now advocating a primary teacher education programme with roots in the logic of a teacher education practice from the middle of the last century, i.e. a retraditionalisation of pedagogic discourses (Beach & Bagley, 2012). In multiple articles, Beach and others have shown how this discursive rupture has affected teachers' professional knowledge base and professional identities (Beach & Bagley, 2012, 2013Nilsson Lindström & Beach, 2015), but also that the negative concern politicians have had about an integrated teacher education with a collective professional knowledge base was not justified, since the teacher education programme never really became a cohesive practice (Beach, Bagley, Eriksson, & Player Koro, 2014). Among other reasons, this is because historical traditions are so strong (Andersson 2002;Beach, 1995;Hartman, 2012;Jedemark, 2007;Linné, 1995).
This study aims to investigate the outcome of the new reform and the new primary teacher education practice. The focus of the study lies in the two specialisations of the primary teacher education programme: one for teachers of grades F-3 and one for teachers of grades 4-6 and how pedagogic discourses promoted by the new reform and educational traditions may be affecting the Swedish primary teacher education and the professional identities of the primary teachers. The study is based on the following questions: • What knowledge and competencies are legitimised in and through the primary teacher programme examination practice and accordingly construct the pedagogic discourse and shape the pedagogic/professional identities for teachers? • Has the new division of the primary teacher programme in two specialisations exhibited characteristics of historical primary education programme traditions, and if so how?
Organising of the primary teacher education programme in policy and practice Since 2011, the primary teacher education programme is one of four teacher education programmes in Sweden. 1 It is divided into three different specialisations: after-school centre education, preschool education and primary school education for grades 1-3 (F-3), and primary school education for grades 4-6. 2 Programme content is regulated by a System of Qualification, which contains intended learning outcomes that PTE students must meet to receive an exam. The System of Qualification also regulates the extent of the various subject areas taught, and for the general subject area of educational science (abbreviated UVK). The two specialisations within the primary teacher education programmethe teaching of grades F-3 and grades 4-6have learning outcomes that are basically identical. The primary teacher specialisation in teaching grades 4-6 has 28 learning outcomes, one more than the primary education specialisation for grades F-3. What differs between the two specialisations regarding the learning outcomes is that the PTE students studying to become teachers of grades 4-6 must also possess knowledge relating to grading as well as to the teaching of identity, sexuality and relationship issues. The organisation within the teacher education programme differs more, however, in that the PTE students specialising at F-3 are trained in the teaching of almost all the subjects taught, while the focus for primary education teachers who are specialising in grades 4-6 is on knowledge in four subjects. The primary education programme for primary teachers F-6, according to the reform texts (Prop 2009/10:89;SOU 2008:109) and the System of Qualification (SFS 1993:100), is thus relatively comparable but also different at the same time.
The previous teacher education programme from 2001-2010 included these two specialisations in a single programme called "early childhood education". Before that, for almost fifteen years (1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000), the education programme was divided into a programme for teachers of grades 1-7 (in which students could choose to specialise in Swedish and social studies or mathematics and science) and another one for the teachers of grades 4-9. Dividing primary education teachers into two groups (F-3 and 4-6) is accordingly a return to the practice that could be found in Sweden at the end of the 1980s. The history of the teacher education programmes and the current organisation are the motives for this study of the present primary teacher programme.

Swedish teacher education in a historical and discursive context
The teacher education programme plays a noteworthy role in Swedish education policy. There is no other programme that is so often evaluated or that has been so regulated by politicians over time (Hallsén, 2013). The reason for this, according to Hallsén, is that the teacher education programme is an extremely important instrument of control for the state. It also has a sort of interposition between school and university, which means that it has not really been viewed as an academic programme with sustainability and autonomy. Other possible reasons can be the fact that it was not that long ago that teacher education became an academic exam (1977) and that the programme did not have its own research traditions.
The abundance of policies relating to the teacher education programme implies that it is considered important for the nation. This is true not only for Sweden but for the rest of the world (Antikainen, 2010;Briant & Doherty, 2012;Cochran-Smith & Fries, 2001;Goodson, 2008;Lauder, Brown, & Halsey, 2009;Lawn & Furlong, 2009;Sarakinioti & Tsatsaroni, 2015) and is in line with the discourse of knowledge-based societies and economies (Apple, 2001;Ball, 2008;Garm & Karlsen, 2004;Young, 2009). The reform of teacher education can also be seen as part of the overarching education reform in Sweden that was made around 2010 and with the political purpose of improving the results in Swedish schools. Other reforms were new curricula a new grading system, a new upper secondary school organisation, a reform of teacher authorization etc.
The great importance of teachers, and thereby the teacher education programme, for pupils' learning and results, has also been the focus of much research, both internationally (Darling Hammond, 2000, 2006 and nationally (Gustafsson & Myrberg, 2002;Swedish National Agency of Education, 2009). Even for the EU as a creator of policies, the teacher education programme is an important factor; this is true for member states as well as for the continued economic competitiveness of the EU in a global perspective.
Historically, the Swedish teacher education programme has exhibited characteristics of two traditions: the seminar tradition and the academic tradition, whereby programmes for the teaching of younger pupils traditionally have been taught in seminars while programmes for the instruction of older pupils have been the responsibility of the universities (Hartman, 2012). This division lived on, even after the Higher Education Reform of 1977, 3 since programmes for secondary and upper secondary school education were primarily administrated by the faculties of universities, and teachers' colleges were mainly responsible for training teachers of pupils in the lower grades. The various ways this division has been realised have been manifested in the demands that teacher education programmes should approach "tested experience" ("beprövad erfarenhet") (the seminar tradition) and at the same time have strong ties to research (the academic tradition). This division has also involved different ways of organising teacher education programmes, but has also influenced the professional knowledge base and identities of educated teachers (ibid). The two teacher education reforms of 1988 and 2001 were intended to undermine the rationale behind maintaining the division, but the current teacher education policy has instead moved towards a stronger classification and specialisation of teachers (Sjöberg, 2010(Sjöberg, , 2011. Seen from a historical perspective, the programmes have not only been aimed at various parts of the school system but also targeted different social classes (Andersson, 2002). The teachers being trained to teach the lower grades were expected to prepare their pupils for more basic tasks, like teaching them about cleanliness and discipline, whereas, those that trained to teach the upper grades were educated to become cultivated instructors. Similarly, the aim of the seminar tradition was to model good behaviour, with the idea that the child was the central focus and required clearly defined social training (Hartman, 2012;Linné, 1995), while according to Erixon Arreman (2005) and Hartman (2012), the academic tradition involved striving to gain knowledge of subject matter, conducting independent studies without much external control. Since the teacher education programme, like other programmes, has not adapted the rationale behind knowledge and teaching when new reforms have been introduced, various rationales have been operating parallel to one another within the teacher education programmes, and this has led to different subcultures (Beach, 1995;Beach et al., 2014;Bernstein, 2000;Hartman, 2012). In a study of the examination practice used by the primary teacher education programme, Player Koro and Sjöberg (2018) have shown how these (sub-)cultures and pedagogic discourses come into play in the current teacher education programme. Among other things, the study shows that there are different examination cultures with different pedagogic discourses within the different parts of the programme, in the subject theory part, in the didactical part and in the educational science part (UVK). The study also shows that a subject didactics examination culture dominates examinations given on the primary teacher education programme, which is also in line with the discourse about the need for the new teacher education programme found in national policy documents (Prop 2009/10:89;SOU 2008:109).
In a Scandinavian perspective, there are both similarities and differences between the countries' teacher education practices. In general, the Swedish education programme, like those of Norway and Denmark, has deep roots in the seminar tradition. This is, however, in contrast to the Finnish programme. The Swedish, Danish and Norwegian programmes have also been more closely tied to praxis and furthermore, have had a relatively weak research base (Kvalbein, 1999;Rasmussen, 2008). According to Wågsås Afdal (2012) and Wågsås Afdal and Nerland (2014), there is still a trace of these historical organisation principles in today's Norwegian and Finnish education programmes, whereby the Finnish programme has a clearer grounding in research and the Norwegian programme has a stronger profession/context orientation. This proves to be the case not only in the content taught but also in the approach towards knowledge and the pupils. Moreover, this can be noted in the way a teacher's work is understood and in the vocabulary of teachers. One possible problem with a context oriented teacher education is that pedagogical questions run a greater risk of being individualised and thereby impossible to discuss, theorise or problematize from a more structural pedagogical perspective (Wågsås Afdal, 2014).
Accordingly, previous studies show that the teacher education reform of 2011 amounted to a clearly political and discursive reversal, by which political representatives clearly indicated their desire for the return to a teacher education programme that was characterised by strong classification. Considering the knowledge of, on the one hand, the Swedish teacher education tradition and on the other, the fact that the discourse content pervading the previous teacher education programme was so integrated, it is thus interesting to study the way the new PTE programme with the two specialisations are organised and practised, both with regard to what pedagogic discourses are related to the primary teacher's knowledge and competencies, and through this analysis, as well as their professional identities.

Methodology and empirical material
The study, which is a policy case study, was conducted at a relatively large university in Sweden and includes the primary teacher education specialisations in grades F-3 and grades 4-6. The empirical material for the study comprises all policy documents that relate to the courses on the programme: syllabi, study guides, and all examination tasks. The choice of empirics is based upon, among other things, the fact that it is examinations that most clearly control and demonstrateespecially after the Bologna reformthe knowledge and competencies that students are expected to have attained in their studies (Havnes, 2010). In total, the study includes 47 courses (syllabi, study guides) and 322 examination tasks in the two programmes (Table 1).
The empirical documents were collected during the years 2014-2015. The new teacher education programme was being taught for the fourth year in the autumn of 2014. This meant that, since the PTE programme is a 4-year programme, some of the courses included in the study had been taught up to three times before, while other courses were being taught for the first time. Even though the programme is new, it is still possible to see how the new national policy has been transformed into educational practice. Theoretical and analytical framework The theoretical points of departure for this study are grounded in a sociology of education tradition, in which the organisation of education is considered as a discursive practice that is influenced by actors at different levels, and where fields of knowledge production shape the structure and content of the programme. One way of understanding such a discursive transformation is Bernstein's theoretical concept the pedagogic device and the way pedagogic discourses operate and are construed through this transformation process (Bernstein, 2003;Singh, 2002). The concept of the pedagogic device is a way of visualising the political dimension of education and the way policy is transformed from the field where it is formulated, the official recontextualization field (ORF), to the field where it is realised, the pedagogic recontextualizing field (PRF), in this case, the teacher education institutions and teacher education practice. According to Bernstein (2000), the shaping of policy (through the distributive rules) takes place in the transformation of policy (through the recontextualizing rules) and in the implementation of policy (through the evaluative rules) via discursive struggles influenced among other things by political ideologies, research and educational traditions. The empirical documents in this study (syllabi, study guides and examination tasks) are thus part of this discursive transformation process. As Bernstein puts it, local policy functions as the result of the evaluative rules that in the end shape the pedagogical practice and the pedagogic discourses.
Pedagogic discourses operate through the content of education and how education is operated. Bernstein distinguishes between instructional discourses, which are primarily about the what-questionknowledge and skillsand regulative discourses, which are primarily about order and relationsthe how-aspect. The two discourses should, however, not be considered as separate entities; instructional discourse is embedded in and in relation to the regulative discourse. In relation to the content of pedagogic discourses, the concept of horizontal and vertical knowledge structures is also used (Bernstein, 1999). According to Beach and Bagley (2012) these knowledge structures reflect a dichotomy between common knowledge and academic knowledge, which in some way can be analogous to seminar traditions and academic traditions on the teacher education programme. Horizontal knowledge structures are characterised by knowledge that is mundane, segmentally organised, and tied to certain contexts. Vertical knowledge structures, on the other hand, are more abstract and theoretical and thereby not that tied to context. Knowledge with a vertical knowledge structure is therefore less dependent upon a given practice or a special time, and can more easily be transformed to other pedagogic contexts. Vertical knowledge structures also create greater opportunities for individuals to think outside a given framework, providing greater opportunities for pedagogical development and emancipation. In addition, researchers like Zeichner (2010) and Apple (2001) assert that vertical knowledge structures are necessary for teachers to be able to see, reflect and relate to political restructuring that is currently taking place in relation to education and consequently to the practical implication for teachers and pupils.
The concept of horizontal and vertical knowledge structures is also closely related to other theoretical concepts in Bernstein's theoretical toolbox. Bernstein (2000Bernstein ( , 2003 claims that historically, teacher education programmes have been characterised by various forms of professional knowledge that teachers are supposed to possess. The two forms of knowledge are called Trivium och Quadrivium. The first one, Trivium, refers to teaching skills of a more general type, emanating from disciplines such as pedagogics, sociology, and psychology. Trivium-knowledge, Bernstein proposes, is needed to recognise and analyse pedagogical phenomena that emanate both from inside and outside the classroom. It is also needed to understand learning and teaching in light of politics, economy and ideology (Apple, 2001;Beach et al., 2014;Brante, 2010;Ken, 2010). The other form of knowledge, Quadrivium, is related to knowledge of the subjects that the student teacher is going to teach. Historically, the weighting of the different forms of knowledge has shifted in teacher education programmes. Moreover, at times it has been thought that the two forms of knowledge need to be integrated, but at other times it was thought that they should be kept separate. Later, to bridge the gap between Trivium and Quadrivium and between theory and practice, didactics (general and subject) has been introduced as a new subject area.
According to Beach and Bagley (2012), one can see a return to both horizontal knowledge structures and towards forms of knowledge that have features of the Quadrivium in the recent teacher education programme reform, but also towards knowledge of how one should teach (know-how) rather than on what basis pedagogical questions and practices are generated (know-why) (Brante, 2010).
The last theoretical concepts used to describe the pedagogical discourse are classification and framing. Classification refers to boundary maintenance between contents or insulation of knowledge domains, i.e. how different "what" (Bernstein, 1973a), and framing refers to the degree of control teacher and pupils possess over the selection, organisation pacing and timing of the learning process (Bernstein, 1973b). The pedagogic discourse, with its related theoretical concepts, shape pedagogic identities, i.e. what a teacher and a pupil should be like (Bernstein, 2000).
As support for the analysis of the empirical material, an analytical model has been designed, produced by the research question and through an abductive interplay between the theoretical concepts as well as the empirical data. In addition to the concepts taken from Bernstein's toolbox, inspiration for the study's theoretical design has been taken from studies by Berlak and Berlak (1981) as well as Briant and Doherty (2012).
More specifically, the analysis has been initially conducted through an exploratory reading by which patterns in the data emerged. This then influenced the choice of theoretical concepts and coding that the study would be grounded in. By posing analytical questions and testing theoretical concepts, examination forms, as well as knowledge and content structures and classification and framing structures of the two programmes were brought into focus. In the next step of the analysis qualitative content analysis was conducted using the analytical focus and the coding and theoretical concepts (Table 2). This was done by examining the examination tasks for every course based on the four analytical questions (four steps). 4 The first step was to identify which forms of examinations were used on the two programmes (Question 1 in the Table 2). In the next step, the focus was to see what content structure was evidenced by the examination papers (Question 2 in Table 2). Here, the focus was on the what-questions or the instructional discourse and Bernstein's analytical concepts of Trivium and Quadrivium (subject knowledge) were employed, as well as four different aspects of didactic knowledge: if the knowledge was of the general didactic or subject didactic character, and if it was of a theoretical or practical character. Since features of the tasks meant that several content areas could be fitted into the same task, the tasks were coded, in some cases, with many different content codes. Yet another analytical question that has to do with the task's instructional discourse dealt with the knowledge structures that characterised the tasks (Question 3 in Table 2), with a point of departure in Bernstein's theoretical idea about vertical and horizontal knowledge structures (1999). According to Bernstein, horizontal knowledge structures relate to knowledge of a common nature, which are contextual, and often based on people's own experiences. Vertical knowledge structures, on the other hand, have a clearer abstract and analytical character based on established knowledge and are more unrelated to context, and thus are more sustainable in time and space (cf. also Berlak och Berlak, 1981;Briant & Doherty, 2012;Wågsås Afdal, 2012, 2014. In the analysing process, it was determined: (1) whether the tasks were of a holistic type or atomized type; (2) whether examination tasks made it possible for students to demonstrate their analytical and/or problematizing abilities or whether the tasks were more of a descriptive and/or normative type; (3) whether the tasks were set up so that the students could work together (collective) or be done individually (personal); (4) whether in doing the tasks students were supposed to use already established theories or content knowledge (public), only their own experience and knowledge (private) or whether they were supposed to combine others' and their own knowledge and experience (personal and public); and (5) whether the knowledge was tied to the classroom practice, whether it touched on questions posed by educational science that are not limited to classroom practice or whether it had a clear subject area focus.
In the last stage of coding (Question 4 in Table 2), Bernstein's concept of classification and framing was employed (regulative discourse). The classification concept was used to determine whether the content of the tasks was clearly kept apart (strong classification) or if there was an aspect of the task that had to do with integration (weak classification). With regard to the concept of framing, the possibility of regulative control (regulative discourse and the how-aspect in the pedagogic discourse), there emerged three aspects: (1) how specific the assessment criteria for the task were; (2) whether decisions about content were made by the teacher or the student; and (3) whether there was another type of regulative control in the tasks that did not primarily have to do with knowledge, such as time constraints or required formalities.
For these three aspects, three codes were created, strong, weak and middle, for when the material did not allow for completely dichotomous coding.
As an aid in the organisation of the qualitative content analysis, the qualitative analysis programme Nvivo Pro was used for the coding process. Nvivo's strength is its capacity to collect, organise and categorise qualitative data for carrying out a thematic content analysis. In this study, that type of analysis is combined with statistical calculations relating to the scope of the various coding as well as the relation between the implementation and focus of the examination, on the one hand, and between the two different primary teacher education specialisations (F-3 and 4-6), on the other. The possible correlations found in the data have been analysed statistically with the help of a cross-tabulation test (multiple group chi-square test or contingency table test) to determine whether there are any systematic differences between the two programmes in terms of examination practice (Djurfeldt, Larsson, & Stjärnhagen, 2009). The crosstabulation test, which is a non-parametric test, is used if one wishes to ensure that there is a correlation between variables when using qualitative data of nominal scale character. The significance levels will be shown in the presentation of results as the p-value *p < 0.1, **p < 0.05, ***p < 0.01, ****p < 0.001.

Results
The presentation of the results of the study will be described in relation to the four analytical questions that can be found in the table above (Table 2). For each analytical question, both general patterns will be presented regarding the pedagogic discourses that characterise the examination practice in the primary teacher education programme as well as possible differences between the two primary education teacher specialisations.
Regarding the Forms of examination (Question 1, see Table 3), the results show that the syllabus for the primary teacher education programme includes a great many different forms of examination, such as creative and multimodal presentations, workshops and teaching activities where the students use various types of presentation formats. In general, however, seminars and take-home exams dominate as forms of examination, whereby seminars comprise 39% and take-home exams 26% of all the examinations given in both specialisations.
There are also differences between the examination cultures in the two specialisations, and the cross-tabulation test shows that there is a systematic and significant difference relating to the distribution between the overall use of various forms of exams in the two specialisations****, whereby certain examination practices are significantly more common on the one specialisation compared to the other. Primary education students specialising in grades F-3 demonstrate their skills and abilities in seminars (48% as opposed to 32% 5 )** and workshops (5% as opposed to 1%)** to a much greater extent compared to primary education students specialising in grades 4-6; while takehome exams are a much more common form of examination on the primary education programme with a specialisation in grades 4-6 (15% for F-3 as opposed to 35% for 4-6) ****. Other forms of examination that are sparsely used but are nonetheless somewhat more common on the primary education programme specialising in grades 4-6 are creative and multimodal examinations (2% as opposed to 6%)* and Quizzes/web exams (0% as opposed to 2%)*. The significance of these differences is, however, not as strong.
Looking at the instructional discourses that operate in the primary education programme, focusing on content structure (Table 4), both specialisations place a very strong emphasis on didactic knowledge and competencies. 78% of examinations have a didactic content, whereby subject didactics knowledge and competencies are assessed in 54% of all examinations. Examinations that focus on the type of knowledge that Bernstein calls Trivium knowledge only make up one seventh of all examinations (14%), and focus on subject knowledge is found in a mere 8% of all examinations. There is no significant difference between the overall distribution of content structure in the examination practice. There is, however, a difference in the number of times that subject knowledge appears on the exams, where students studying to teach grades 4-6 are examined on this type of knowledge to a greater extent (5% in F-3 as opposed to 11% in 4-6) **.
The next aspect of the instructional discourse that is found in examinations has to do with the knowledge structure (Table 5). In studying the first aspect, the question was raised whether the examination tasks are of a holistic or atomistic character. In general, the results show that almost all (93%) the tasks have a holistic feature. The next aspect involved deciding whether the tasks gave students the opportunity to demonstrate and use their analytical and/or problematizing skills or if the tasks were more of the descriptive or normative sort. The results show that the tasks of a descriptive and/or normative sort dominate examination practice (73%). The ability to work together (collective) is also an important skill for teacher education students, since two thirds of the tasks require cooperation with classmates (66%). The fourth aspect was about the source of students' knowledge: is it already established knowledge and theories, the student's own experiences or a combination of the two? In one out of two tasks (50%), the students are asked to make associations and process their own experiences in terms  of theories and knowledge that they have read about, primarily in their required reading for the course. In 40% of the tasks, however, they are asked primarily to use their own experiences. The last aspect of knowledge structure shows that the tasks that have to do with classroom work dominate the examination practice (64%), while the knowledge that is required to understand the prerequisites for education, teaching and learning, is relevant for about a quarter of the examination tasks (27%). Tasks that focus on subject knowledge are actually few and only relevant for a little less than one-tenth of all examination tasks (9%). In terms of knowledge structure, the greatest difference between the two programmes again has to do with the emphasis on subject knowledge (Cf. content structure), where primary education teachers specialising in grades 4-6 are given several tasks that require them to demonstrate subject knowledge (3% for F-3 as opposed to 14% for 4-6)***. Another significant difference between the programmes has to do with cooperation, whereby the examination practice to a greater extent demands that the students specialising in grades F-3 should work together to solve the examination tasks (72% as opposed to 61%)**. Otherwise, there are no other significant differences. Possibly there is a tendency for the 4-6 programme students to receive more tasks that have an atomistic character (5% as opposed to 9%). The level of significance is, however, low in this case.
Last, the regulative discourse in the examination tasks are presented, which is in the form of its classification and framing (Bernstein, 2000) (Table 6). A review of classification in the common examination practice involves deciding whether the tasks that are given student teachers focus on limited content knowledge or if they have more of a focus on integrated knowledge. Most of the examination tasks have a weak or mediumweak classification, which means that they integrate various skills and forms of knowledge (93%). Here, a slightly differentiation can be found between the programmes, whereby the 4-6 programme contains slightly more tasks with strong classification. The difference was, however, not significant. The framing of the examination practice has to do with the regulative control that takes place between the teacher educators who formulate and assess the tasks and the PTE students. Since the introduction of the Bologna-reform, the demand for clearly defined intended learning outcomes and assessment criteria has increased in European universities. The purpose has been to clarify for the students what knowledge and skills they need to show to pass a course, but also to make it easier for presumptive employers to see what a student should know upon graduation. Formulating clear assessment criteria has thus become more and more important for teacher education programmes, just as it has become for other academic programmes. While they function as a support for students' learning, according to formative theories of learning (Black & Wiliam, 1998;Hattie, 2009), assessment criteria also function as a regulative control between educator and student. Clear assessment criteria are, however, not particularly common in this teacher education practice (39%), and for many tasks there are none, or very vaguely formulated ones (53%).
The next aspect relates to the control of content in the tasks. In a little more than half of the tasks, the formulation of what it is the student needs to present in terms of content is clear or relatively clear (55%). A large proportion of the tasks (45%) are, however, formulated in such a way as to allow the students themselves the possibility to determine the content, and thereby what knowledge and competencies they want to present for assessment. The last framing aspect relates to another form of control present in the examination tasks. This control is aimed at the way the task is carried out and can affect time constraints, length of the task and other formalities (pacing). Approximately half of the tasks contain no other controls other than those affecting content and criteria (48%). One fourth of the tasks (24%), however, did contain formulations that involved students adapting their presentation of knowledge to aspects of formalities and time constraints, in addition to when and how the examination task was to be handed in or presented. There were no significant differences between the two programmes with regard to classification and framing and the regulative discourse. There are weak tendencies toward the 4-6 programme using clearer assessment criteria (35% as opposed to 42%) and more frequently some form of formality control (criteria of communication/ pacing) (46% as opposed to 56%), but these possible correlations are somewhat too weak to be viewed as significant.

Discussion
A few years ago, the Swedish teacher education programme was changed and became a more distinctly classified practice, whereby every category of teacher (preschool teacher, after-school centre teacher, F-3 teacher, 4-6 teacher, secondary school teacher, uppersecondary school teacher and vocational education teacher) now takes a programme of courses tailored to their specific target group. The reform was an ideological criticism of the two previous teacher education programmes which both had striven towards a more consolidated and integrated teacher education practice to create a common professional identity and a consolidated educational practice for children and teens. The reform texts of the current teacher education programme indicated that both the special conditions present in schools and the differing levels of maturity and development of children and teens made it necessary to differentiate between teacher education programme specialisations (Prop 2009/10:89;Sjöberg, 2010Sjöberg, , 2011SOU2008:109). Therefore, the primary teacher education programme related to primary school was divided into two different specialisations, F-3 and 4-6, which in practice function as two education programmes. Intended learning outcomes for the two specialisations are mainly the same, but the organisation of them differs; F-3 teachers are trained to teach basically all that is to be taught to children in that age group, but the syllabus for the 4-6 teachers might be called a hybrid of F-3 and the secondary school education major, since 4-6 teachers are trained to teach a certain number of subjectsusually four. The F-3 specialisation is thus structured as a more weakly classified practice, where teachers' work has a comprehensive character. This differs from the 4-6 specialisation, where teachers are trained for a more classified practice, and where pupils encounter teachers with different specialisations (Bernstein, 2000). The Swedish teacher education programme, like other Scandinavian teacher education programmes, also carries a long, entrenched tradition of determining how different teacher categories have been and should be trained (Andersson, 2002;Erixon Arreman, 2005;Hartman, 2012: Jedemark, 2007Kvalbein, 1999;Linné, 1995;Rasmussen, 2008;Wågsås Afdal, 2012, 2014Wågsås Afdal & Nerland, 2014). The focus of this study has been to see how the intentions that brought about reform are implemented in the teacher education practice, based upon the fact that previous research has shown that reforms have not made much of an impression on teaching practice (Beach, 1995;Beach et al., 2014;Bernstein, 2000).
The results of the study show that the instructional discourse, at work in both the syllabi for the primary education specialisations, first and foremost prepares primary school teachers for a specialty in classroom teaching and subject didactics. This is particularly evident in the choice of content (content structures) made by the examiners of the teacher education, where the didactic content in examination tasks has a strong dominance, but it is also evident in the aspect of knowledge structure and the focus on classroom specific content. Furthermore PTE students have not so many opportunities to demonstrate Trivium-knowledge, knowledge of theories and phenomena that influence the practice where teachers work through their examinations. Nor are future primary teachers given examinations that provide an opportunity to practice and show problematizing and analytic competencies and skills. This result shows that the instructional discourse operative in the primary teacher education programme embraces to quite a high degree a horizontal knowledge structure, with knowledge, skills and competencies that are dependent upon context and that are geared more toward knowing how instruction should be carried out than toward knowing why education practice is the way it is (Beach & Bagley, 2012;Brante, 2010;Wågsås Afdal, 2014).
In light of the descriptions covering Swedish teacher education offered by Hartman (2012), Andersson (2002) and others, the results of this study furthermore show that the current primary teacher education programme has features of both the seminar as well as the academic tradition. The pedagogic discourse, with a strong focus on classroom management skills and competencies that are operative in the seminar tradition, is hence powerful, even if, for 40 years, the education programme has formally been an academic programme. When the empirical data for this study were collected, the first primary education students studying on the new programme had not yet graduated. Even though the programme is relatively new, certain cultures have been constructed in the two primary teacher education specialisations. The most salient differences between the two specialisations have to do with the relationship between the forms of examination that are used, whereby the syllabus for the F-3 specialisation is dominated by seminar-based examinations, but the students studying to be teachers for grades 4-6 mostly write take-home exams. This means that teachers teaching the youngest school children develop their professional identity in an oral culture while the written culture dominates for teachers for grades 4-6. At the same time, there is a growing trend that the F-3 teachers' learning practice takes place in a collective context while the 4-6 teachers are offered a learning practice where knowledge is acquired individually. Another difference between the two specialisations has to do with the 4-6 specialisation placing a stronger emphasis on specific subjects that will be taught. In view of Hartman's (2012) definition of historical traditions, and results from, among others, Wågsås Afdal (2012), the results of this study also show that there are differences between programme specialisations that bear traces of various teacher education traditions, whereby the F-3 specialisation has a somewhat stronger seminar tradition compared to the 4-6 specialisation, which in this respect is more of a hybrid. These differences may also influence, first and foremost, the primary teachers' professional identities and their knowledge base, but may also affect the pedagogic discourses the future primary teachers will construct in their own pedagogic practices.
According to Singh (2002) reforms have further implications than just organisational changes. She claims, as do Beach andBagley (2012, 2013) and Nilsson Lindström and Beach (2015) that reforms generate discourses of how teachers/students should work, and individuals subsequently become positioned, taking on specific professional identities, and that these identities today are related to neoliberal discourses. Several educational science scholars, such as Zeichner (2010) and Apple (2001), also show that knowledge and skills with a vertical knowledge structure are necessary, on the one hand, to prevent pedagogical questions from being individualised, and on the other, to ensure that teachers can think critically about their own workplace as well as political intentions, restructuring practices and reforms. This study shows that in all likelihood we are on a path towards a school system with teachers who have partially different pedagogic/professional identities, just as was intended by the reformed teacher education, but we are also on the way to a school system with teachers who are better at doing classroom work than at thinking critically about what is happing inside the classroom. Notes 1. The other three teacher education programmes are: the preschool teacher programme, the secondary and upper secondary school teacher programme and the vocational education teacher programme. 2. The specialisation of after-school centre education is not included in this study. 3. With the Higher Education Reform of 1977, parallel post-secondary school programmes such as the teaching and healthcare programmes were incorporated into institutions of higher education. 4. The examination tasks could be found in the syllabus, in the study guide or as a special document. 5. For each figure showing size/percentage, the primary education programme for teachers of F-3 is presented first, followed by the programme for teachers of grades 4-6.

Notes on contributor
Lena Sjöberg is a lecturer at University West, Department of Social and Behavioural Studies. Her main research interests concern education policy and pedagogic discourses. She is currently working on a project concerning professional knowledge in primary teacher education and a project concerning policy discourses in higher education.

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