Between ideal teaching and ‘what works’: The transmission and transformation of a content area from university to school placements within physical education teacher education

The purpose of this study is to explore the recontextualisation of Assessment for Learning (AfL) as a particular content area in the transition between a university course and a school placement course within Swedish physical education teacher education (PETE). By combining Basil Bernstein’s pedagogic device and Stephen Ball’s performativity perspective, we alternately ask how AfL is constructed as a pedagogic discourse and what AfL becomes in different contexts within PETE. Nine students attending a Swedish PETE programme participated in the study. The empirical material was collected through one seminar and two group interviews at the university, as well as through nine individual interviews based on lesson observations at different school placements. Our findings highlight five recontextualising rules, which indicate that: (1) the task of integrating assessment into teaching enables the use of AfL; (2) an exclusive focus on summative assessment and grading constrains the use of AfL; (3) a lack of critical engagement with physical education teaching traditions constrains the use of AfL; (4) knowing the pupils is crucial for the use of AfL; and (5) the framing of the school placements determines how AfL can be used. As a consequence of these rules, AfL was transformed into three different fabrications: (1) AfL as ideal teaching; (2) AfL as correction of shortcomings; and (3) AfL as ‘what works’. One conclusion from this study is that increased collaboration between teacher educators and cooperating teachers in schools can help strengthen PETE’s influence on school physical education.


Introduction
Can teacher education make a difference? This question was the point of departure for Brouwer and Korthagen's (2005) longitudinal study, which indicated that integration of theoretical studies and practical experiences in teacher education programmes had an impact on student teachers' future teaching. Narrowing in on physical education teacher education (PETE), Ní Chróinín and Coulter (2012: 221) also suggest 'that teacher education in physical education can have an impact on pre-service teachers' knowledge and understanding of physical education which, in turn, should impact on teaching and learning in their classrooms'. Nevertheless, Martin and Dismuke (2018: 22) claim that not enough 'research has investigated the linkages between teacher education and classroom practice'. Thus, our intention is to contribute to the debate on whether teacher education in general, and PETE in particular, matters for teaching practice in schools.
This particular study is part of a larger research project, funded by the Swedish Research Council and the Swedish Research Council for Sports Sciences during a period from 2019 to 2023, exploring the transition from Swedish PETE to the school subject physical education. Our choice of 'what' to study in this transition is Assessment for Learning (AfL) as a particular content area, transmitted and transformed through different recontextualisations (Bernstein, 1996) within teacher education, from university courses to school placement courses. According to Black et al. (2002, preface): Assessment for learning is any assessment for which the first priority in its design and practice is to serve the purpose of promoting pupils' learning. It thus differs from assessment designed primarily to serve the purposes of accountability, or of ranking, or of certifying competence. Wiliam (2011) explains that AfL is based on five key strategies: (1) clarifying and sharing learning intentions with pupils; (2) engineering effective classroom discussions, tasks and activities that elicit evidence of learning; (3) providing feedback that moves the learner forward; (4) activating pupils as learning resources for one another; and (5) activating pupils as owners of their own learning. These strategies are 'tight but loose' (Thompson and Wiliam, 2007: 2), which means that they can be realised in different ways, as long as they are used to decide 'where the learners are in their learning, where they need to go and how best to get there' (Wiliam, 2011: 16). Thus, progression across a series of lessons is a common feature of AfL. This is also how AfL was introduced to the PETE students in the current study during the assessment course at the university. Accordingly, in the context of the university course, AfL is regarded as a content area, while in the context of school placements, AfL is regarded as a pedagogic approach. The reason why we have chosen to focus on AfL is that it is unlikely PETE students will have come across this approach to assessment prior to their teacher education. AfL is, thus, different from other content areas, such as ball games or physical training, which makes it an appropriate object of research in a project exploring whether teacher education matters.
Historically, teacher education in Sweden, as well as in other countries, has been criticised for being non-academic and related to a weak and unclear research base. At times, it has been regarded as vocational training (Alvunger and Wahlström, 2018;Larsson, 2009). Today, there are different views on how to organise teacher education in the most appropriate way. On the one hand, Loewenberg Ball and Forzani (2009: 497) 'argue for making practice the core of teachers' professional preparation'. On the other hand, Beach and Bagley (2012) are critical of the tendency for teacher education to become re-vocationalised. In their view, the movement towards a narrower focus on professional training and measurable skills means that practical 'know-how' gains ground at the expense of scientific 'know-why'. In order to bridge the gap between university-based and school-based teacher education, Zeichner (2010: 89) advocates 'that the old paradigm of university-based teacher education where academic knowledge is viewed as the authoritative source of knowledge about teaching needs to change to one where there is a non-hierarchical interplay between academic, practitioner, and community expertise'. In the same spirit, Darling-Hammond (2006: 300) highlights some of the essential components of teacher education programmes, including 'tight coherence and integration among courses and between course work and clinical work in schools'.
Our research interest in this article is the interplay between a university course and a school placement course within PETE and the pedagogic discourse produced in the transition between these educational settings. Accordingly, the focus is on how AfL is perceived and expressed by PETE students in the context of a course on assessment at university and, shortly afterwards, realised and reflected upon in the context of a school placement course. By combining Basil Bernstein's (1996) pedagogic device and Stephen Ball's (2000Ball's ( , 2003 performativity perspective, we alternately ask how AfL is constructed as a pedagogic discourse and what AfL becomes in different contexts within PETE. Consequently, the purpose of this study is to explore the recontextualisation of AfL as a particular content area in the transition between a university course and a school placement course within Swedish PETE. Based on the presentation of AfL in influential research literature, the research questions are: How is the pedagogic discourse of AfL recontextualised in university and school placement courses within a PETE programme in Sweden? What becomes of AfL in this process of recontextualisation?

'PETE on repeat'
A recurring theme in international research on PETE is that the teaching practice of physical education seems to be unaffected, no matter what happens in society or in teacher education (see, for example, Larsson et al., 2018;Matanin and Collier, 2003). For instance, PETE students' beliefs, based on previous experiences of physical education and sports, seem to have a strong impact on their views of what is valued knowledge in school physical education. Mordal-Moen and Green (2014: 806) have explored student beliefs with the aim of finding out 'what it is that makes them so resistant to change'. One reason seems to be that PETE students find the development of sporting skills and teaching methods far more interesting than any theoretical content, which is expected to be less applicable in school. All of these scholars thus highlight the need for sociocritical approaches that challenge PETE students' views of what is important to master as a teacher in physical education. However, Larsson et al. (2018) have shown that many assessment components in PETE tend to reproduce and verify beliefs that emphasise the importance of skills proficiency in different sport techniques. As practical skills are often assessed in isolation from pedagogical contexts, Backman and Pearson (2016) argue that PETE students' ability to teach movement activities should be part of their overall assessment. This idea resonates with an initiative by MacPhail et al. (2013), who have developed instructional strategies based on the notion of constructive alignment employed through 'rich tasks'. These tasks allow PETE students 'to be productively engaged in their own learning, and learning to teach, thus reinforcing their understanding' (MacPhail et al., 2013: 111). This way of engaging students in their own learning is also one of the key strategies of AfL (Wiliam, 2011).
According to Lorente-Catalán and Kirk (2016), there is a widespread consensus on the need for AfL in research literature on PETE and physical education. Nevertheless, they point out that the transmission and transformation of AfL from PETE to physical education may be a complex process, so their 'question to guide future research is, what happens to well-supported concepts and good intentions once the student teachers start to teach?' (Lorente-Catalán and Kirk, 2016: 78). Traditional approaches to assessment have proved to be resistant to change (Penney et al., 2009) and different implications of AfL have been identified in teaching practices of physical education in schools (Tolgfors and Ö hman, 2016). Thus, Lorente-Catalán and Kirk (2016: 78) claim that 'we need to know more about the structures that support and reproduce [ . . . ] the prevalent and dominant practices of teaching and learning in school physical education that work against the use of AfL', if this alternative form of assessment is to become an embedded routine in the teaching practices of physical education. In regards to school placement studies, the same scholars argue: 'If PE teachers are to be competent users of alternative assessment in schools, then it is of crucial importance that they are able to experience such practices as part of their professional preparation' (Lorente and Kirk, 2013: 79).
Based on this argument, AfL should not only be introduced in courses on assessment at university. In order to challenge the notion of 'PETE on repeat', student teachers should also get opportunities to use AfL during their school placements. Macken et al. (2020) have recently studied PETE students' use of AfL during placements in primary school. The scholars found that PETE students had opportunities to develop a certain assessment literacy in this context, but there was not enough time to implement AfL techniques such as peer-and self-assessment. From our point of view, these findings raise new questions, such as: what becomes of AfL in the context of school placements?

School placements as key worksites
According to Chambers and Armour (2012), school placements can be seen as key worksites for PETE students' learning in the collaboration process between universities and schools. In line with this suggestion, both Mordal-Moen and Green (2014) and Amaral-da-Cunha et al. (2020) highlight that the PETE students in their studies viewed the school placement settings and the cooperating teachers as the two most important elements of their PETE programmes. Behets and Vergauwen (2006: 407) explain that school placement teaching is a laboratory experience, which allows PETE students to 'try out different ways of teaching, reflect on the outcomes and make modifications accordingly'. This explorative working method requires conscious supervision. Thus, Amaral-da- Cunha et al. (2020) highlight the importance of the cooperating teachers' own development as supervisors, involving role performance, pedagogical strategies and relational issues. According to Belton et al. (2010), a cooperating teacher programme that gives cooperating teachers an opportunity to offer feedback in line with the PETE students' curriculum can facilitate such a development.

Theoretical framework
In the choice of theoretical framework for this study, we have been inspired by Macdonald et al. (1999: 32), who have investigated 'the ways in which knowledge in the physical activity field is produced and reproduced in a cyclical fashion at the interface between senior school and higher education'. Drawing on Bernstein, among others, Macdonald et al. (1999) explored how educational discourse was constructed, transmitted and adapted in the transitions between some key sites of physical education practice. By adopting a similar methodological approach, we focus on AfL in the transition between a university course and a school placement course within PETE. Accordingly, we have turned to Bernstein's (1996) theory of how pedagogic discourses are constructed, recontextualised and realised through the pedagogic device and to Ball's (2000Ball's ( , 2003 theory of performativity. According to Bernstein (1996), a pedagogic discourse arises when regulative and instructional discourses meet at different levels in the education system. For instance, at the distributive level, the limits for thinkable and unthinkable interpretations of AfL are regulated through influential research literature and curricular documents. When this distributive level version of AfL meets conceptions of AfL inherent in school physical education (for example, among cooperating teachers at school placements), a process of recontextualisation takes place. According to Singh (2002), teacher education and teacher educators have important roles in the system of rules structuring the construction of pedagogic discourses in schools. At the evaluative level (Bernstein, 1996), different frame factors also determine how a particular content area or pedagogical approach, such as AfL, is realised in teaching practice. In this study, frame factors for the use of AfL in physical education school teaching are understood as enablers and constraints (Braun et al., 2011). Thus, Bernstein's (1996) pedagogic device will help us identify the recontextualising rules regulating the pedagogic discourse of AfL in different PETE contexts.
When exploring what AfL becomes in this process of recontextualisation from a performativity perspective, the focus is on how current educational policy regarding AfL determines PETE students' possible action spectrums and the consequences thereof (cf. Ball 2003). In this article, current educational policy regarding AfL is to be understood as synonymous with the normative version of AfL at the distributive level (Bernstein, 1996) -that is, the version of AfL's five key strategies that is stipulated in the course literature (Wiliam, 2011). According to Ball (2000: 4), '[p]erformativity works from the outside in and from the inside out'. This means that the pedagogic discourse of AfL, constructed in the context of the university course, is likely to be contested, negotiated and transformed in the context of the school placement course. Thus, the performativity perspective helps us identify what becomes of AfL in the transition between different educational contexts within PETE. The product of what is performed in the interactions between PETE students, teacher educators, cooperating teachers, pupils and researchers can be seen as a fabrication (Ball, 2000). In our case, fabrications of AfL are to be understood as cultural patterns within the PETE settings (see also Tolgfors, 2018).
To sum up, by combining the two perspectives of this study, we are able to move from and between questions of how AfL is constructed as a pedagogic discourse in different contexts within PETE (Bernstein, 1996) to what AfL becomes in terms of different fabrications (Ball, 2000).

Participants and collection of empirical material
Participants in this study were a group of student teachers in the final year of their PETE programme at one of eight PETE universities in Sweden. All nine students in the class, seven male and two female, agreed to participate in the study. The syllabus for the investigated PETE programme stipulates that student teachers shall know how to 'use documentation, assessment and evaluation as grounds for development and planning of teaching' and be able to 'handle, compare and value different approaches to pedagogical assessment'. In line with these goals, Swedish teacher education involves both university courses and school placement courses, allowing student teachers to develop subject content knowledge and address practical issues of teaching, learning and assessment. During the collection of empirical material, the students were involved in two courses, one assessment course at the university (covering assessment for and of learning) and one school placement course. These courses were interlinked, which meant that a specific learning task was to use AfL during the school placement course. However, how to realise the five key strategies was up to the individual student. One of the researchers was responsible for teaching the course on assessment while another researcher handled the data collection. By avoiding dual roles, we prevented power relations from affecting the quality of teaching and research. The school placements lasted for five weeks (24 days) at different upper secondary schools in the city and some bordering municipalities. During this period, the cooperating physical education teachers at the schools acted as supervisors. They were also responsible for assessing the PETE students according to criteria provided by the teacher education department at the university. One of these criteria focused on the PETE students' ability to use and reflect on AfL.
The empirical material analysed in this paper consists of: (1) curriculum documents, study manuals, lectures and course literature from both courses (providing contextual information); (2) a sound-recording from one two-hour seminar on the PETE students' interpretations of and reflections on AfL, as presented in the assessment course; (3) nine individual, semi-structured interviews (40-50 min) with the PETE students in connection to the observations made at their school placements; and (4) two group interviews (90 min each) at the university after the school placement course had ended. At the seminar, the researcher who was teaching the course on assessment acted as a moderator for the PETE students' discussion about the lectures and course literature on AfL. During the fieldwork, the observations focused on how the PETE students integrated the five key strategies of AfL into their teaching. Field notes were kept, to be followed by individual interviews. The interviews concerned the participants' views of AfL in general, and in relation to the observed lessons in particular. The questions focused on whether the PETE students had used any of AfL's key strategies and in that case how, why and what this had led to. During the group interviews, the PETE students discussed questions based on the researcher's interpretations of the observations and individual interviews. The researcher introduced different areas of interest by summarising his impressions of the prior fieldwork and asked the participants to comment on them. This prompted further reflections and explanations from the PETE students' perspective regarding their ways of using AfL (or not) during school placements in physical education. Altogether, these methods provided rich data on the transmission and transformation of AfL as a particular content area from university to school placements within PETE.

Analysis
The recordings of the seminar and the interviews (group and individual) were transcribed verbatim. Excerpts of particular interest were translated from Swedish to English. All four researchers in the project were involved in the analysis of the collected data (Braun et al., 2017;Patton, 2002). As a first step of our analysis, we searched for not only explicit ideas stated in the material but also latent and implicit meanings in order to get to know and comprehensively structure the data (Burr, 2003). This familiarisation was intertwined with a process of coding -that is, identifying and labelling certain things of interest for our research questions. For instance, the PETE students' ways of relating to the five key strategies of AfL were highlighted in the transcripts. As a second step, we operationalised the theories of Bernstein (1996) and Ball (2000Ball ( , 2003Braun et al., 2011) in order to identify patterns in the empirical material, by posing the following theoretically grounded questions: 1. How is the pedagogic discourse of AfL constructed by PETE students in the context of the university course? 2. How is the pedagogic discourse of AfL constructed by PETE students in the context of the school placement course? 3. What recontextualising rules, in terms of enablers and constraints, regulate the pedagogic discourse of AfL in the different educational contexts within PETE? 4. What fabrications of AfL are produced in the recontextualisation process?
Initially, the four researchers carried out their familiarisation and coding individually. Then followed a collaborative analysis process when themes were identified in line with what Goodyear et al. (2019) describe as a deliberative strategy. During this deliberative process, the researchers discussed their individual readings in relation to the theoretically grounded questions. Finally, the categories answering the two overarching research questions were agreed upon. As Goodyear et al. (2019: 217) argue: the deliberative strategy is accordingly not a process to come closer to an essential truth in the study, but as a procedure to reach as high quality research as we can. The goal of the deliberation is thus a form of collective agreement where all co-authors are given the possibility to make judgements in relation to different alternatives, views and arguments.
Our analysis led to the following collective agreement.

Findings
The findings section is in two parts. First, we present how the pedagogic discourse of AfL is recontextualised in the university and school placement courses within PETE. Second, we present what AfL becomes in the process of recontextualisation.

The pedagogic discourse of AfL within PETE
In our analysis, we identified five categories that regulate the pedagogic discourse of AfL. These categories are: (1) the task of integrating assessment into teaching practice; (2) the assessment culture of school physical education; (3) teaching traditions in physical education; (4) knowing the pupils; and (5) the framing of school placements. In the discussion of each category, we describe how AfL is constructed through contextual enablers and constraints, encompassing an empirically generated recontextualising rule (Bernstein, 1996).
(1) The task of integrating assessment into teaching practice. An open-ended assignment was designed for the school placement course that tested students' ability to integrate assessment into their teaching practice of physical education. The PETE students knew in advance that their ways of solving this task would be part of the assessment of their school placement course. This task prompted the students to use in practice what they had covered in their assessment course, which entailed various attempts to embed AfL in their teaching. The following statement indicates that the first key strategy of AfL (clarifying and sharing learning intentions with the pupils) was considered very important by some PETE students: 'The pupils should understand the purpose and goal of a lesson and how they will be assessed. That is a basic principle!' (Student B, individual interview).
In line with this basic principle, some PETE students often communicated their learning goals by writing them on the whiteboard during lessons. Other students focused on the fifth key strategy of AfL (activating the pupils as owners of their own learning). Two of them introduced 'training logs' to the pupils, in order to give them the opportunity to assess their own developmental needs regarding physical ability and health: 'I want the pupils themselves to understand how they can develop and take responsibility for their own learning' (Student F, individual interview).
In addition, the third key strategy of AfL (providing feedback that moves the learner forward) was used in order to support pupils' self-regulatory processes. Finally, students frequently combined the second and fourth key strategy of AfL (engineering effective classroom discussions and activating pupils as learning resources for one another), by organising group discussions at the end of their physical education lessons. Consequently, the first recontextualising rule is that the task of integrating assessment into teaching enables the use of AfL in physical education.
(2) The assessment culture of school physical education. In Sweden, the strong focus on grades in an education system based on high-stakes assessment tends to influence the assessment culture of physical education, which also affects the PETE students' perceptions of the possibility of using AfL: 'One of my supervisors assesses all the time. But not for the purpose of AfL or providing feedback. The pupils think it is all about summative assessment: "Was that an A, or what grade did I get?"' (Student G, group interview).
The PETE students seem to have discovered that the assessment culture of school physical education differs from how assessment for and of learning is presented at the university. In the teaching practice of physical education: 'AfL is more used to predict a particular grade than promoting learning' (Student G, group interview).
Given that the cooperating teachers often inspire their PETE students, this predominant assessment culture constrains the use of AfL and transforms its idea of promoting learning into an issue of accountability. Our analysis indicates that the PETE students' conception of AfL, as it was introduced in the assessment course, often clashes with the assessment culture of school physical education. This struggle between two competing assessment discourses hampers students in their ambitions to realise AfL. Consequently, the second recontextualising rule is that an exclusive focus on summative assessment and grading constrains the use of AfL in physical education.
(3) Teaching traditions in physical education. According to the PETE students, some cooperating teachers planned and organised their physical education classes in such a way that only the potential benefits of variation and joy of movement were realised. This tradition of exposing children to a wide range of physical activities is resistant to change, and was noticeable in some PETE students' reflections on the physical education practices they had observed while on placement: 'It was only about starting up the lesson, letting the pupils activate themselves and then finishing it off. It was nothing like what we have learnt about providing opportunities for reflection, clarifying intended learning outcomes and so on' (Student A, group interview).
Some of the PETE students were encouraged by their cooperating teachers to try teaching as many different activities as possible rather than working with knowledge areas involving some kind of progression. They were also encouraged to prioritise action and physical activities, rather than spending too much time on talking. On the one hand, then, the physical education traditions at these school placements constrained the PETE students' use of AfL. On the other hand, the practical nature of the subject meant that the pupils' performances were often visible, and it was possible to give them feedback anyhow, even if served in a somewhat 'smorgasbord' fashion: 'There are always certain techniques to give feedback on' (Student I, individual interview).
However, feedback on sports techniques was only common in some activities, such as swimming and gymnastics. Otherwise, feedback was more of a collective and encouraging kind -such as 'Good job!' (Student A, individual interview) -rather than related to any particular learning or progression point.
In some instances, the PETE students' own experiences as pupils, within a traditional discourse of physical education, also constrained the use of AfL in their teaching practice. One of the students explained that s/he often reproduced the same exercises and activities s/he was used to from school physical education: 'I am pretty much like the physical education teachers I had myself' (Student A, individual interview). Thus, the third recontextualising rule is that a lack of critical engagement with physical education teaching traditions constrains the use of AfL.
(4) Knowing the pupils. Before the school placement course, some PETE students were worried that the version of AfL described in research literature would be hard to realise in practice: Since the PETE students did not know the pupils' names, interests or capabilities, they found it challenging to adapt the teaching to their needs and prerequisites. Some students also doubted whether their own subject knowledge was sufficient for them to provide relevant feedback to the pupils.
Being, as Student H (individual interview) put it, 'thrown into a class' without knowing the pupils constrained the use of AfL. Consequently, most PETE students highlighted the importance of building relationships with pupils if AfL was to become meaningful. Student G (individual interview) explained that they tried to find out what the pupils already knew, in order to adapt the teaching to their previous experiences and understandings. Student D (individual interview) expressed the ambition to challenge those who were high-performing while moving back down the progression ladder for those who needed some more help. Yet another student (Student C, individual interview) had been inspired by a cooperating teacher to always pose the generic question, 'What is needed in this class?' in order to create a safe learning environment. Accordingly, by approaching the individual pupil 'to show that you see him or her', Student C cared about 'soft values', such as self-esteem, well-being and belonging, in physical education. Thus, feedback was mainly used for building trustful relationships. However, the PETE students' lack of continuity with their classes affected their conceptions of the pupils' expectations. On the one hand, the PETE students reported some resistance from pupils to the idea of engaging with the same physical activity several lessons in a row in order to work on their progression: 'I do not want them to be bored -they are supposed to be entertained!' (Student E, individual interview). On the other hand, they also reported meeting pupils who seemed to accept whatever they had planned for the physical education lessons: 'Some pupils are willing to adapt' (Student F, individual interview). The latter provided the conditions for the PETE students to work according to AfL's key strategies. Consequently, the fourth recontextualising rule is that knowing the pupils is crucial for the use of AfL in physical education.
(5) The framing of school placements. The fact that the school placement course only lasted for five weeks was generally seen as the principal constraint for the use of AfL. Student H (individual interview) regarded AfL as a 'dream picture' at the university, rarely transferable to school practice. Other PETE students also doubted that they would be able to promote any development of pupils' abilities when they were on placement for such a short time: 'The pupils only have one PE lesson per week and we are only here for five weeks. Then we want to try many different things instead of striving for progression' (Student E, individual interview).
According to the students, some of the cooperating teachers' colleagues also tended to constrain the use of AfL: 'Honestly, when we sat in the staff room discussing something, some teachers at our school expressed their opinion that 'AfL is an insight that comes and goes' (Student B, individual interview).' This suggests that experienced teachers do not necessarily regard AfL as a unique pedagogical approach. They have likely encountered some aspects of this concept before, but under other labels and in an unsustained way, which is why many of them find AfL contestable. According to the PETE students, cooperating teachers felt that 'AfL is something we talk aboutnothing we do' (Student G, individual interview).
The PETE students were also sometimes advised to adjust their teaching to suit the facilities. If there was a floorball rim in the gym the cooperating teachers asked, 'Why not play some floorball?', even if that interfered with the students' original lesson plans and ideas of learning and progression. Thus, 'inconsistent planning, due to sudden changes in the educational setting' (Student A, individual interview) constrained the use of AfL. Accordingly, the fifth recontextualising rule is that the framing of the school placements determines how AfL can be used.

What AfL becomes
Depending on the recontextualising rules regulating the pedagogic discourse of AfL, different fabrications (Ball, 2000) of AfL are produced in the transition between the PETE settings. We have identified three fabrications, named after their most prominent features: (1) AfL as ideal teaching; (2) AfL as correction of shortcomings; and (3) AfL as 'what works'.

AfL as ideal teaching
The pedagogic discourse of AfL, collectively produced by the PETE students while studying their assessment course at university, corresponded quite well to the normative description of distributive-level AfL in the research literature (e.g. Wiliam, 2011). For instance, the PETE students highlighted the importance of adapting teaching to the prerequisites of the learners: 'The first key strategy is to share the learning intentions. And if pupils need any kind of adaptation, the goals may be reached in many different ways' (Student C, seminar). This line of reasoning highlights that the PETE students had embraced the idea of AfL, which is to adapt the teaching to the needs of the pupils in order to promote learning (Wiliam, 2011). The PETE students' interpretation of the five key strategies was that they belonged together and built on each other: 'As I understand the five key strategies, they are interdependent and non-hierarchical in relation to one another' (Student C, seminar). Thus, the students regarded the key strategies as a holistic pedagogic approach. To some extent, this conception of AfL survived the transition from university to school placements. In comments made at the end of the study, the PETE students accentuated a physical education teaching where 'AfL should work as an integrated part of teaching' (Student H, group interview). Hence, the students found it important to plan their teaching in themes or knowledge areas involving a series of lessons, thus facilitating a process of feedback based on the intended learning outcomes. With the aim of promoting life-long learning, the fabrication of AfL as ideal teaching emphasises that the teaching should be adapted to the pupils' needs and prerequisites through a holistic pedagogic approach to the five key strategies.

AfL as correction of shortcomings
It was not obvious to all PETE students what kind of feedback would move high-performing learners forward. The fixed norm of how to move or act in physical education meant that some PETE students focused their feedback on the limitations of low-performing pupils. Thus, the third key strategy of AfL (providing feedback that moves the learner forward) was often reduced to correction of pupils' shortcomings. This fabrication was mainly produced in the context of the school placement course.
However, since feedback in physical education is so closely related to pupils' bodies and identities, the PETE students found troubleshooting pupils' shortcomings a sensitive area. As a consequence, they hesitated to provide feedback at all -for instance, by temporarily stopping a learning activity in order to improve some aspect of it: 'If one interrupts, let us say, a game of floorball in order to provide individual feedback to someone, that might feel a bit distinctive, I think' (Student A, group interview).
Due to the pupils' different physical capabilities and previous experiences of sports, the PETE students did not want to intervene when they saw pupils performing exercises that could have been performed with better movement qualities. Instead, they 'lowered their expectations' (Student H, group interview). This meant that individual feedback was often neglected: 'At the school placement, we just lead the lesson and then there is nothing more to it. So, in regards to speaking to individual students -what are we supposed to talk about?' (Student C, group interview). The uncertainty of what to talk about means that not all PETE students were completely familiar with or guided by the intended learning outcomes of physical education. Collective feedback was also accompanied by some doubts: 'I want to spend some time on reflection at the end of the lesson, for what it is worth. [ . . . ] But the pupils mostly say what they think I want to hear' (Student G, individual interview).
The three quotations above indicate that the PETE students had not yet considered how their own subject knowledge from PETE could be transferred to school physical education. The fact that the PETE students did not know the pupils so well also hindered conversations with the aim of promoting learning. Consequently, the fabrication of AfL as correction of shortcomings can be derived from the contextual limitations of school placements.

AfL as 'what works'
The different enablers and constraints for using AfL in the teaching practice of physical education imply that making things work in the classroom is significant and that it is important to keep activities going without disruption: 'One has to have a pragmatic approach to it all. [ . . . ] At the school placement, you sometimes forget about the goals' (Student F, individual interview).
Rather than sharing what the pupils were supposed to 'learn', the PETE students often explained what the pupils were supposed to 'do'. So, in order to connect the lessons in terms of progression, the PETE students often reviewed with their pupils what they had done before: 'I start the lesson by reminding the pupils about what we did last time, in order to tie the lessons together' (Student C, individual interview, emphasis added).
After the introduction, most lessons followed a traditional physical education plan, including a final gathering meant for evaluation and group reflection. During the lessons, the PETE students circulated and gave feedback to the pupils. In the fabrication of AfL as 'what works', the third key strategy of AfL (providing feedback that moves the learner forward) survived the transition from university course to school placement course, but not without transformation. According to one PETE student: 'Feedback is a way of building relationships with the pupils. It is not always goal related, striving for proficiency. The feedback is more "encouraging" than "achievementoriented"' (Student C, group interview).
A common opinion among the PETE students was that it was impossible to watch activity and give comments to all pupils in each lesson, due to the limited time. Thus, encouraging feedback to the whole group replaced individual feedback based on the learning intentions. In general, the PETE students also preferred to save all feedback until the end of the lessons. At these gatherings, they gave the pupils the opportunity to reflect on what they had experienced during the lessons. Thus, the fabrication of AfL as 'what works' determines the formulation of a time efficient teaching routine of ongoing activity, interlinked lessons with encouraging feedback and group reflections.

Discussion
Inspired by Macdonald et al. (1999), we have explored AfL as a particular content area in the transition from university to school placements within PETE. In these educational contexts, the PETE students encountered different contextual enablers and constraints (Braun et al., 2011), which regulated the pedagogic discourse of AfL. Our findings highlight five recontextualising rules, which indicate that: (1) the task of integrating assessment into teaching enables the use of AfL; (2) an exclusive focus on summative assessment and grading constrains the use of AfL; (3) a lack of critical engagement with physical education teaching traditions constrains the use of AfL; (4) knowing the pupils is crucial for the use of AfL; and (5) the framing of the school placements determines how AfL can be used. As a consequence of these rules, AfL was transformed into three different fabrications (Ball, 2000): (1) AfL as ideal teaching; (2) AfL as correction of shortcomings; and (3) AfL as 'what works'. Together, these fabrications provide a complex picture of the transitions between university courses and school placement courses within PETE. Interestingly, the fabrication of AfL as ideal teaching was mainly produced in the context of the university course. However, the normative 'dream picture' of AfL was hard to realise in the context of school placements. When the PETE students met 'real' pupils in 'real' schools, with physical education teaching traditions resistant to change (Larsson et al., 2018;Mordal-Moen and Green, 2014), some of them lost track of AfL, introduced as it had been in the more theoretical context of the university. Using Bernstein's (1996) pedagogic device, the encounter between the regulative and instructional discourses of AfL entailed a process of recontextualisation, which implied a poor alignment between goals, teaching and assessment (cf. Matanin and Collier, 2003). Instead of, for example, providing feedback based on specific learning intentions, the PETE students either corrected pupils' shortcomings, gave collective encouraging feedback or neglected providing feedback at all. These findings correspond with Macken et al. (2020), who have shown that AfL techniques such as self-and peer-assessment are rarely used during school placements. Due to limited lesson time and a lack of trustful relationships with pupils, the PETE students concentrated on 'what worked' under the particular circumstances of their placements rather than adopting the holistic pedagogic approach of AfL that they had been taught.
According to previous research, ideas about what knowledge and skills to develop, what teaching methods to use and how to assess pupils' knowledge and capabilities are strongly affected by the PETE students' beliefs, based on their own previous experiences of physical education and sports (Matanin and Collier, 2003;Mordal-Moen and Green, 2014). The students in our study were also affected by their own beliefs in relation to AfL, which most likely originated from the assessment course they took at the university. The five key strategies (Wiliam, 2011) were also integrated in the interlinked university and school placement courses, in that: (1) the learning intentions and assessment criteria were shared with the students in advance; (2) several seminars facilitated discussions on AfL; (3) feedback was provided by both teacher educators and cooperating teachers; (4) two or three students were placed together at each school, which gave them opportunities to give and get feedback and act as learning resources for one another; but still (5) the individual student was activated as owner of his or her own learning, as the task of integrating AfL in the teaching practice could be solved in different ways. Thus, other researchers' claims about physical education practice being unaffected by PETE are not fully supported in our study (cf. Larsson et al., 2018). However, the PETE students' conceptions of the pupils' preferences (that physical education should mainly be about fun and enjoyment) tended to prevent them from planning lessons that aimed at progression. Further, if the PETE students' experiences are taken as evidence, the predominant assessment culture of physical education and teaching traditions mediated through the cooperating teachers, such as introducing a wide range of physical activities for the benefit of joy and variation, seems to be resistant to change (Larsson et al., 2018).
Amaral-da-Cunha et al. (2020) note that PETE students find school placement settings and cooperating teachers very important for their development as teachers. Although we agree with the great potential of school placement courses within PETE, the recontextualising rules we have identified in our study need to be addressed if AfL, as presented in the research literature (or other subject content areas in teacher education), is to transition into the teaching practice of physical education. Hence, drawing on Belton et al. (2010), Darling-Hammond (2006) and Zeichner (2010), we argue that cooperating teacher programmes should ensure more active collaboration between universities and schools. For instance, AfL, as taught in an assessment course, is more likely to survive the transition to school physical education if its ideas are shared and discussed among teacher educators and cooperating teachers involved in PETE. Based on our findings, these discussions should focus on: 1. how to develop frames for school placement courses that provide opportunities for PETE students to get to know pupils; 2. how to handle the intertwined process of assessment for and of learning in physical education; 3. how to share the responsibility for assignments and learning outcomes in school placement courses so that PETE becomes more of a joint project between university and schools; and 4. how to provide worked examples of how AfL can be implemented in practice, based on a problematisation of AfL in relation to teaching and learning in physical education. Lorente-Catalán and Kirk (2016) warn that the transition of AfL from PETE to school physical education might be problematic. It seems that although the PETE students in this study understood the meaning of the university version of AfL, they were not ready to use this version of AfL when they entered the school placement course. To expect student teachers or early career teachers to apply a fully developed version of AfL in their initial teaching plans seems to be a goal too high to reach. One solution could be 'rich tasks', as suggested by MacPhail et al. (2013), planned by teacher educators and cooperating teachers working together, and involving different aspects of teaching in authentic school settings. This study shows that a 'rich task' cannot only be defined in relation to educational models described in the literature at university (such as AfL), but must also be rooted in the problems that are identified in pedagogic practice -for example, by cooperating teachers in schools. If these tasks are created collaboratively, school placements can become the key worksites (Chambers and Armour, 2012) for laboratory experiences (Behets and Vergauwen, 2006), with PETE students supervised by cooperating teachers and coordinated by teacher educators at the university (Amaral-da-Cunha et al., 2020).
To sum up, if the doxa of PETE, as well as of school physical education, is to be overcome, it is not only the students' beliefs that must be challenged (Larsson et al., 2018). Rather than pressing the 'PETE on repeat' button, all actors involved in PETE 'need to proceed'. Through increased collaboration between teacher educators and cooperating teachers, teacher education could, in line with Brouwer and Korthagen (2005), make an important difference.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors were funded by the Swedish Research Council and the Swedish Research Council for Sports Sciences.