Design based research to develop the teaching of pupils with moderate learning difficulties (MLD): Evaluating lesson study in terms of pupil, teacher and school outcomes

Teaching and Teacher Education 34 (2013) 162e173

Contents lists available at SciVerse ScienceDirect

Teaching and Teacher Education journal homepage: www.elsevier.com/locate/tate

Design based research to develop the teaching of pupils with moderate learning difficulties (MLD): Evaluating lesson study in terms of pupil, teacher and school outcomes Brahm Norwich, Annamari Ylonen* Graduate School of Education, University of Exeter, Heavitree Road, Exeter, Devon EX1 2LU, United Kingdom

h i g h l i g h t s
Lesson study process examined in 29 secondary schools over a 2-year period.
Design-based lesson study research approach used.
Lesson study resulted for improvements in pedagogy.
Lesson study can benefit learning of pupils who have special educational needs.
Contextual factors were important for successful lesson study implementation.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 November 2012 Received in revised form 12 April 2013 Accepted 19 April 2013

This paper discusses how a design-based research approach was used to refine the use of Lesson Study in a project with novel research methods in a new area of teaching focussing on pupils with moderate learning difficulties. The study refined the Lesson Study methods in terms of the evaluation of the contexts, processes and outcomes of the Lesson Studies undertaken over two phases. The findings suggest beneficial outcomes for pupils and teachers while outcomes for schools remain more complex. The study shows that various contextual factors are critical if Lesson Study is going to become a sustained commitment in schools. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Lesson study Design-based project Outcomes Moderate learning difficulties Inclusive teaching Special educational needs

1. Introduction This paper gives an account of how Lesson Study (LS) was used and developed in a two phase project to improve the learning experiences and opportunities of pupils identified with moderate learning difficulties (MLD) at Key Stage 3 (aged 11e13 years) in English secondary schools. The aim of this paper is to show how a design-based research approach was used to refine the use of LS methods with reference to the pupil, teacher and school outcomes of undertaking these lesson studies (Brown, 1992). There were also

* Corresponding author. Tel.: þ44 1392 724942. E-mail addresses: [email protected] (B. Norwich), [email protected] (A. Ylonen). 0742-051X/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tate.2013.04.012

other aims of this project Raising Levels of Achievement through Lesson Development for Pupils with Moderate Learning Difficulties (MLD), which included the development of pedagogic knowledge and strategies, programmes and materials for improving the learning and learning dispositions of pupils with MLD (see Ylonen and Norwich, 2012 for a more detailed discussion); the evaluation of processes and outcomes of the Lesson Study in schools (see Ylonen and Norwich, 2013); and the evaluation of the usefulness of the MLD category (see Norwich, Ylonen & Gwernan-Jones, 2012). The version of LS that was used in the Raising Levels of Achievement project was based on a previous UK Lesson Study project by the National Strategies (Dudley, 2004), which drew on Japanese principles and practices (Takahashi & Yashida, 2004). In this version, an LS cycle consists of three research lessons which are jointly planned and directly observed by an LS team. Pupils’ views

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

about the lessons and their learning are also an important part of the LS cycle (see Fig. 1). Lesson planning is informed by evaluation of the previous lesson (Sims & Walsh, 2009) as well as ideas and strategies that derive from research and theory. In this way LS provides the means of bringing research, theory and practice into the planning and evaluation of specific lessons. The LS team also identify one to two pupils to focus the planning and evaluation, the case pupils, who are identified depending on the topic and aims of the Study; in this project those who had been identified as having MLD. Though LS can be used in different phases and areas of education (stage, curriculum area and pupil characteristics), we have found no references to its previous use in developing inclusive teaching of pupils with special educational needs or disabilities in ordinary schools. In Singapore, Chia and Kee (2010) have adopted a version of LS in the training of special education teachers, but not of general class teachers. Also, although there is a tradition of inquirybased approaches to inclusive developments (Ainscow, 2000; Howes, Davies, & Fox, 2009; Miles & Ainscow, 2011), these approaches have not used LS practices. The use of LS in the development of teaching of pupils identified as having MLD is also significant as MLD has been a contested area of special educational needs and neglected area of research and development in the UK (Norwich & Kelly, 2012; Desforges, 2006) (See Appendix A for the Government’s definition of MLD in England, which was used in this study). Lesson Study was chosen as a general teaching development approach because of its previous international record of promoting pedagogic knowledge and strategies. In particular the UK version of LS with its focus on case pupils’ learning (Dudley, 2004) had particular relevance to the neglected area of teaching pupils with MLD. 2. Background to project design High quality and effective lesson studies have been characterised as practices that ‘help teachers enhance their content knowledge and pedagogical content knowledge to improve instruction in classrooms, develop good “eyes” to see and analyse student learning, and ultimately to produce better student learning’ (Yoshida, 2012: 141). Existing research about the mechanisms and outcomes of the process indicates, for example, that LS

163

improves teaching by developing teachers’ knowledge, by building professional community among teachers, and by improving teaching materials (Lewis, 2009; Lewis, Perry, & Hurd, 2009). This in turn can have a positive impact on pupil learning (Lewis, Perry, & Hurd, 2009). Research has also shown that the collaborative aspects of participating in LS through the planning and review meetings can enhance teacher efficacy which can lead to improved engagement and learning of pupils (Puchner & Taylor, 2006), and that teachers can acquire new insights into teaching and how to improve their teaching strategies via participating in LS (Lee, 2008; Rock & Wilson, 2005). There is also evidence of the development of effective learning communities in schools that have engaged in a longer-term lesson study process (Lieberman, 2009). However, though the LS process has attracted much research attention in the West, much of this has not been directed towards examining the actual mechanisms that produce specific outcomes (Sibbald, 2009). Research about LS tends to focus on prescribing or describing, rather than systematically researching and developing LS. In the United States where LS has been used since the late 1990s there are still ‘only a few longitudinal and comprehensive studies of lesson study’ (Yoshida, 2012: 145). Furthermore, with particular relevance to this study, there has been no US or European research and development about LS for pupils with special educational needs (SEN) and/or disabilities in terms of either special education or inclusive practices. In a national and international context where research evidence is expected to justify educational practices (Hammersley, 2007), the question of whether LS works is an important but not a simple question. The ‘gold-standard’ of research evidence is often taken to be the randomised control trial (RCT). Though it is possible to design experimental evaluations of an LS approach compared to some other professional learning model, this would have to be in relation to some specified area of learning, phase of education and learners. Such a design may be strong on internal validity (that is, to identify causal efficacy) but rather less so on external validity (that is, the relevance of the demonstration to use in other settings, other subject areas and learners) (Cartwright, 2007). Though the RCT design has its place in research approaches, its adoption reflects more the interests of the research producers than research users who are not only interested in whether an approach like LS can be

Fig. 1. The lesson study process.

164

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

efficacious, but also whether the method is relevant to practice and what contextual factors may affect the outcomes (Cartwright & Stegenga, 2008). Lewis, Perry, & Murata (2006) describe two routes to researching school teaching improvements in terms of a general proof route and a local proof route. The general route starts by establishing causal efficacy claims and then disseminates with an emphasis on fidelity of implementation. However, controls to ensure fidelity could restrict local buy-in, and adoption may no long occur when central controls are removed. By contrast, innovation in the local proof route is local and may differ across settings, but there can be continuous adaptations, local ownership, early adoption and the organic spread of the innovation. Local conditions might also result in adaptations that undermine the potential of the innovation. As data to improve innovations are local, general efficacy claims are likely to be weaker. Lewis et al. (2006) suggest that there is a trade-off between these two proof routes, with each having advantages and disadvantages. Yet, they point out, in line with Cartwright’s analyses above, that the general proof route has disadvantages that are often overlooked. This does not imply that general route proof designs cannot be used with LS research, only that thus far these designs have not been used because LS is still mainly a development strategy rather than a specific method or intervention. In this project a local route proof approach was employed because the use of LS in professional learning and teaching development is at an early stage, and because adopting LS as a strategy still had to be defined in operational terms. There has been little LS use in the UK and none in the area of teaching pupils with SEN. A general proof route (with the use of RCTs) is more suited to interventions that are well defined and where the intervention is less likely to interact with the context of its use. The aim of this project was to refine the use of LS using a designbased research approach (Cobb, Confrey, diSessa, Lehrer, & Schauble, 2003), while providing evidence from the particular use of LS to develop teaching approaches for pupils identified as having MLD. We used a local grounded approach in examining the outcomes of the LS process on the participating pupils, teachers and schools. The aim was to contribute to the developing literature about LS based on empirical investigations in line with the local proof route. This two-year project was in two distinct phases.1 In the first phase (2010e2011) 34 teachers from 17 schools received training in LS work and support from educational consultants throughout the project. Due to unforeseen circumstances at three schools (e.g. staff illness, other work commitments), only 28 teachers from 14 secondary schools completed the programme phase and participated in its evaluation activities. Three of these were special schools that were included to act as comparator schools. Teachers in these schools specialise in teaching pupils identified with more severe MLD in smaller classes than are found in the regular classrooms of the project secondary schools, where those with MLD might be in subject based mixed attainment or attainment based class groups. Of the 28 teachers 82% were female and 18% were male, about 90% were aged under 40, and about 60% of the teachers had less than 10 years of teaching experience, while the remaining 40% had been teaching for over 10 years. The process and outcomes of the phase 1 lesson studies were evaluated and then using these findings the LS approach was adapted for phase 2 (see Fig. 2). In the second phase there were 33 teachers from 15 secondary schools (three were special schools)

1 Before the project commenced, it underwent a full university-based ethical review and received ethical clearance.

Fig. 2. Design of LS evaluation across the 2 phases.

who received training in LS all of whom completed the programme phase. However, the teachers in the second phase received less guidance and support while undertaking the LS process in order to determine how LS would operate under conditions more similar to those typical in schools. Of the 33 teachers 80% were female and 20% were male; 90% were aged under 40; while 36% had less than 10 years of teaching experience and 64% had more than 10 years of teaching experience. All of the secondary school teachers in phases 1 and 2 of the programme had previously taught pupils identified as having MLD in their classes. The secondary schools had between 750 and 2000 pupils with most having around 1000 pupils.

3. Phase 1 LS programme and evaluation 3.1. Phase 1 lesson studies Secondary schools in four local authorities (two urban and two county/rural) in the south west of England were invited to participate in the project. How participating teachers became initially involved in the project varied. In some cases teachers applied directly to participate, in others senior teacher selected teachers from those who applied internally to the school. Two teachers from each participating school were expected to represent different curriculum subjects to lead the LS work. All teachers came together on three full day conferences, at the start, after one term to review progress and after two terms to review developments and outcomes. At the initial conference teachers were introduced to the principles and practical procedures of LS, heard directly about LS from experienced teachers who used LS, were introduced to theory and research about the nature of MLDs and current knowledge about teaching and motivational approaches relevant to these pupils. The lead LS teachers were expected to recruit other colleagues for the LS teams, including the schools’ SEN coordinator (SENCO who has knowledge about SEN including MLD) and other teachers and teaching assistants. These LS teams were expected to conduct three LS cycles over the two terms (each cycle consisting of three research lessons). All schools were allocated an LS consultant who either visited and/or provided support by telephone across the two terms. The classes selected for LS were ones that had 1e2 pupils identified by the school SENCO as having MLD, who were the case pupils in these lesson studies. The sample consisted of 61 pupils with identified MLD of whom 22 were at School Action, 8 at School Action Plus, 12 Statement and 19 unknown (see Appendix B for an

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

165

explanation of the SEN classification system in England). These pupils were also assessed using reasoning, literacy and motivational measures to examine their functioning in terms of conventional definitions of MLD. This analysis showed that only a minority of pupils had difficulties in understanding concepts, many were attaining below average and might not be considered to have MLD or special educational needs (see Norwich et al., 2012 for more details). These teachers were enabled to set their own Lesson Study agendas in their own subject areas (evenly spread between English, Art, History, Geography with a few in mathematics and science) in the 11e13 age range, using a provided 10 step particular LS protocol (see DFE, 2012 for details).

focussing on one LS cycle as an example. In particular, the teachers provided the context and overall aims of the LS, a summary of each of the Research Lessons completed, the impact of the LS on pupil learning, on current and future teaching as well as on wider impact on the departments and school. Reports were received from 18 teachers from 13 schools that had completed phase 1. The reports were content analysed (Robson, 2011) for the goals of the LS undertaken and for details of the outcomes of the LS process for pupils, teachers and schools more widely.

3.2. Phase 1 lesson study evaluation

The phase 1 case reports from 18 teachers, asked the participating teachers to outline the main impact of the LS process on the case pupils with MLD. Table 1 shows the frequencies of the different themes used to analyse these qualitative accounts. The most common type of outcome reported was the beneficial impact on the pupils’ general behaviour and/or motivation, which was mentioned 12 times in the 18 reports. Examples were: ‘pupils who had been reluctant to ask or answer questions grew in confidence throughout the LS process’ and ‘marked increase in confidence of the two MLD case pupils.their attitudes to learning and participation have increased’. This table shows that outcomes focused more on behaviour/motivation than on academic/cognitive skills, and were mostly in general terms rather than in specific terms.

The evaluation methodology was informed by a realist evaluation methodological approach, based on the work of Pawson and Tilley (1997). Realistic Evaluation (RE) aims to link three distinct broad aspects of a programme: its contexts, mechanisms and outcomes (C-M-Os) by constructing a programme theory that explains what processes (mechanisms) under what conditions (contexts) result in what outcomes. What characterises this evaluative approach is its realist model of causation that recognises context as critical to the operation of processes (mechanisms) that result in various outcomes (Pawson & Tilley, 1997). It is this aspect of realist evaluation that is relevant to the local route proof approach used in this design-based evaluation of LS, as discussed above (see Section 2). Using previous LS and school improvement research literature and in consultation with LS specialists, a programme theory of LS was designed of C-M-Os at a school and a teacher level. The aim of a realist evaluation is to field test this theory to refine it in response to various data sources. In this evaluation the field test was done using a survey questionnaire and semi-structured interviews to triangulate the testing of the LS programme theory (Teddlie & Tashakkori, 2009). In the final survey questionnaire 66 statements covering teacher and school level contexts, processes and outcomes were formulated, which were rated using a four-point scale (definitely not; slightly; mostly; definitely; with a ‘can’t say’ option). After piloting the questionnaire and revisions made, it was sent to 28 participating teachers at the end of phase 1 using an online survey tool (Survey-Monkey). Responses were received from 16 teachers from 11 different schools (57% response rate). In this paper attention is focused only on examining the outcomes of the C-M-O analysis, and not the linkages that have been examined elsewhere (see Ylonen & Norwich, 2013 for more details about specific methods used). The interview schedule covered teacher and school level contexts, processes and outcomes with open-ended questions and probe questions to follow up responses. Nine semistructured interviews were carried out in five schools. The teachers were chosen to represent the range of schools that had more and less challenging experiences of using LS, as identified by the LS consultants who supported the LS teams. There was not time nor resources to interview all the teachers while the range would illuminate the key process of LS. The interviews in schools were about one hour each and were all audio recorded, transcribed and thematically analysed (Robson, 2011). In this paper, as with the survey data, only the outcomes of the C-M-O analysis are examined. In addition, participating teachers completed LS case study reports, using a provided pro-forma, at the end of phase 1 to summarise what they had done and achieved during the process by

4. Results from phase 1 4.1. Outcomes for pupils

4.2. Outcomes for teachers An analysis of the teachers’ case reports showed that they saw the impact of the LS process on themselves as positive, for example, in terms of the planning of teaching and increased confidence to take risks. By far the most common impact mentioned (in 8 of the 17 case reports) was that LS enabled the teachers to trial and develop new teaching strategies and resources. The two next most common outcomes both mentioned in six case reports were, first, a general shift in focus of attention from the teacher and teaching to pupils and their learning, and second, teachers’ increased awareness of the individual needs of pupils, including those with SEN. For example, one teacher commented that as a result of the LS process she had adopted a student-focused approach which meant that students had ‘ownership of their own learning and time to reflect on what they have done’. Other types of outcomes (mentioned by three teachers) were, first, teachers’ increased awareness of MLD Table 1 Reported outcomes of lesson study for pupils with MLD (categories derived from 18 case reports), phase 1. General or specific

Type of outcome

Frequency

General terms

Academic-cognitive (e.g. improved learning outcomes, attainment and development of thinking skills) Behavioural-motivational (e.g. increased confidence, attitudes, engagement in class and self-esteem) Academic/cognitive (e.g. improved vocabulary and speaking skills) Behavioural-motivational (e.g. more willing to listen to others; more accepting of criticism, and development and maintenance of positive friendships with peers) No outcomes mentioned

2 (9.5%)

Specific terms

Not known Total

12 (57.1%)

2 (9.5%) 2 (9.5%)

3 (14.3%) 21

166

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

and SEN more broadly, and second, impact on lesson planning and pedagogic styles. As an example of the former type of outcome, one teacher suggested that ‘Lesson Study has been useful to identify two masking strategies used by students with MLD to disguise their difficulties; copying and rushing to finish’. The following two quotes are examples of the impact on lesson planning and pedagogic styles: The process has encouraged us to take risks with our teaching in the future and to try ideas which take us outside our comfort zones. The main impact for me has been to think much more widely in my lesson planning to increase the variety of different activities with more making and doing tasks. It is clear from Table 2 above that the teachers were overall very positive about the outcomes of the Lesson Study process for themselves: in both phases the ratings were between 3.0 and 3.8 where 4 was the highest score. Most teachers agreed that taking part in the LS process had given them, for example, more confidence to make changes to their usual teaching approaches, a desire to try new and more novel approaches in teaching, the ability to be more open to learning from others and exposing their teaching to others as well as more ability to critically examine their own teaching. In addition, regarding pupils identified with MLD, the teachers felt that by taking part in the LS process their understanding of the learning needs of these pupils had increased, that they could better engage these pupils in lessons and that they could better plan and differentiate in their teaching for pupils with MLD. These findings are consistent with the case report data analysis reported above and the analysis of the selected teacher interviews, which provided an insight into how the teachers came to see teaching in a different light during the LS process. A key aspect of this was getting to know the individual pupils identified as having MLD and what may have prevented these students from getting the most out of lessons, an area which also emerged as important in the case reports. The interviews also highlighted that the teachers valued the observation process and team work aspect in LS as well as the ability to trial new teaching strategies and resources. These points are illustrated in more detail below. Two special school teachers pointed out that the observation process and the use of the video camera had been immensely helpful in making adjustments to existing practices. One of these

teachers believed that she came to know the students’ preferences better and so could better engage them, because she now had ’more empathy, more vigilance and more understanding’ towards their needs. For the other teacher LS enabled him to tailor lessons to better suit the individual students, for example by the way of asking more questions. This teacher used LS to establish a more informal manner at the start of lessons that helped students calm down and settle into the lesson. These teachers attributed their increased confidence in risk-taking in teaching to LS. It is notable that LS was even seen to have these positive effects in special school lessons with small numbers of pupils, where there is more scope for flexibility and intensified lesson planning. At one secondary school both teachers reported that LS had encouraged them to talk less while allowing the students to talk more. One teacher pointed out that the pupils with MLD prefer an environment where they can be active in speaking and listening, which would increase their engagement in lessons. Both teachers agreed that observation and collaboration in the LS team had resulted in improved understanding of the learning needs of pupils with MLD, the sharing of ideas and the development of new teaching approaches. In another secondary school one teacher talked about how observing pupils in class had made it possible to get to know the pupils and find out how they think. This teacher derived much personal satisfaction from the practical relevance of the LS process, which avoids starting from the government’s prescribed teaching approach: Teachers often start with assessment criteria and work backwards and this does not allow things to be found out about wider learning.

4.3. Outcomes for schools Although many positive outcomes were reported for the schools, these were overall not quite as clear-cut and strongly felt as the teacher and pupil outcomes. The phase 1 case reports showed that teachers at most schools (10 of 13) suggested that they were going to share their findings and what they had learnt with their colleagues. But, there were clear plans only in five schools to continue using LS either at a departmental or at a wider school level. Teachers at another five schools were hoping that they could continue using LS in some way in their schools, while teachers at the other three schools did not know. Teachers who expressed a

Table 2 Phases 1 and 2 teacher-level outcome ratings. Teacher outcomes

More confidence to try out novel teaching approaches in lessons More willingness to make changes to usual teaching approaches More theoretical and practical knowledge about lesson study More able and willing to examine own teaching to become aware of false assumptions and new possibilities More open to learning from others and exposing your teaching to others in safe settings The lesson study process has improved the quality of planning of your teaching Increased capability to engage pupils with MLD in their learning Increased capability to plan and differentiate in your teaching pupils with MLD More positive towards a dynamic concept of teaching as involving constant learning about how to deal with novel situations Increased ability to articulate aspects of practice More knowledge about how to overcome barriers to learning for pupils with MLD More personal interest in providing quality teaching to all in your lesson planning and lessons More understanding about the nature and complexity of the learning needs of pupils with MLD A more positive attitudes to pupils identified as having MLD and to their inclusion in school and teaching Deeper knowledge about your curriculum subject and subject pedagogy for pupils identified as having MLD Ratings from 1 to 4; means all in the mostly-definitely range

Phase 1

Phase 2

N

Mean

16 15 16 14 15 16 16 16 15

3.81 3.67 3.63 3.57 3.47 3.44 3.44 3.44 3.40

15 16 15 16 15 15

3.33 3.31 3.27 3.25 3.20 3.07

SD

N

Mean

SD

.40 .62 .62 .65 .91 .63 .96 .73 .99

15 15 15 15 15 15 15 15 15

3.60 3.60 3.80 3.67 3.60 3.40 3.33 3.13 3.60

.63 .74 .56 .62 .74 .74 .72 .74 .74

.72 .79 1.0 .86 1.1 1.1

15 15 15 15 15 15

3.27 3.00 3.60 3.13 3.40 3.07

.88 .76 .63 .83 .74 .80

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

167

Table 3 Phases 1 and 2 school-level outcome ratings. School outcomes

Attendance at meetings is regular and prioritised LS teachers feel supported by senior leaders in their LS work Senior teachers and those with SEN and CPD responsibilities encourage LS teachers Some teachers who are not in the LS team want join in or get involved

Phase 1

Phase 2

N

Mean

SD

Descriptor range

N

Mean

16 15 15

2.88 2.80 2.67

1.15 1.08 1.23

Slightly-mostly Slightly-mostly Slightly-mostly

14 14 14

3.07 3.21 3.29

15

2.40

1.12

Slightly-mostly

14

3.00

SD .99 .98 .91 1.2

Descriptor range Mostly-definitely Mostly-definitely Mostly-definitely Mostly-definitely

Ratings from 1 to 4

definite wish to continue using LS highlighted the reasons and likely benefits in these terms: We see lesson study as an infinitely flexible method of tracking student progress and engagement e we will recommend it as a general diagnostic tool as well as a specific way of addressing the needs of SEN pupils. We think that a long-term use of Lesson Study would result in significant long-term improvements in achievement.

Another issue with implications for longer term implementation of LS concerned senior leaders’ support for LS and getting other staff interested in LS work. This topic has been discussed in more detail elsewhere (see Ylonen & Norwich, 2013). However, it suffices to say that the project schools experienced varied levels of senior leader support. In some project schools senior leaders showed little interest towards the project while in other schools support was strong all way through the project. The following quotes by two teachers illustrate the issue:

However, a case report from another school suggested that despite many positive aspects of LS (collaboration, idea sharing ideas, insights and practical support) there were some more problematic aspects:

They [senior leaders] have ignored the LS teachers. They love it when you do these things because it looks good on paper and it’s good for the school, but they don’t want to know more and leave you to get on with it.

Putting three staff in a lesson for nine lessons [3 LS cycles] is very expensive e will school budgets be able to cover this in the future? Getting around timetable issues has been very difficult’.

Senior Management Team have been very supportive and keen to introduce LS school-wide. They have been impressed with what we [the LS teachers] have learnt from LS and how it can affect teaching within classrooms e there is concrete evidence to prove this.

These issues re-emerge below in the interview data analysis. Table 3 outlines the outcomes at the school level from questionnaire survey for phase 1 of the project. Phase 1 teachers rated the four outcome statements at a lower level (means between 2.40 and 2.88 out of 4) in the ‘slightly to mostly’ range than at the teacher level (see Section 4.2). The lowest rated statement was that ‘some teachers who are not in the LS team want join in or get involved’ suggesting that it was difficult to spread the message about the many beneficial teacher outcomes to others in school. Even one of the higher rated statements about LS teachers feeling supported by senior teacher in their LS work showed a relatively large variation in ratings implying that some teachers rated this well below the ‘slightly to mostly’ range. The interviews also highlighted some potential problem areas regarding a longer-term implementation of LS, despite all teachers reporting very positive learning outcomes for themselves and their pupils. By far the most common difficulties were about fitting the LS process in very busy school timetables: the constraints with timetabling, the difficulty of arranging for teaching cover and the subsequent fact that many of the teachers had used their own free time to undertake the LS process. Over half of the interviewed teachers suggested that if they had been unwilling to use their own time to undertake the LS activities, the process may not have worked out in practice. Many of these teachers also suggested that although they had, in theory, been able to get teaching cover, they did not want to do this because they felt that other areas of their teaching practice would suffer as a result. So, they preferred to use their own time instead or swapped a class to avoid leaving their own classes to observe a research lesson. For example, one teacher pointed out that: When project funding finished, it might be difficult to carry on with the LS. This is sad as the only way to improve teaching is through feedback, for example the review process of Lesson Study.

5. Review of phase 1 outcomes and implications for phase 2 design The fuller significance of the phase 1 outcomes of Lesson Studies will be discussed towards the end of the paper (see Section 8). In this section we focus on the pupil, teacher and school outcomes in terms of their implications for the redesign of the LS for phase 2. In phase 1 pupil outcomes were only monitored through qualitative summative accounts provided in the LS case reports. Though many positive pupil outcomes were reported across the schools these were mainly focussed on changes in learning behaviours and motivation. Although there were a few specific positive reports about academic/cognitive outcomes, these did not feature in teacher’s own evaluation of pupil outcomes. Subsequently, it was decided to introduce a more focussed way of setting learning goals for case pupils in the phase 2 design of the LS procedure. Not only would this help the LS teachers plan their research lessons in terms of more focussed learning goals, but this would also provide a way of monitoring pupil learning and outcomes for the phase 2. This system is explained in detail in the next section (see Section 6.2). Outcomes for teachers and teaching from phase 1 lesson studies were examined through qualitative analysis of LS case reports and selected individual interviews as well as by a questionnaire. There was much consistency in the positive outcomes across these data sources such as LS enabling the development and trial of new teaching strategies, increased confidence to take risks, increased awareness of individual needs of pupils with MLD, and the value of team work. Though there were no clear indications from these outcomes for the redesign of the LS procedures for phase 2, it was evident that there had been no direct observation of

168

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

the LS process in phase 1. It was therefore decided to introduce some direct observation into the monitoring and evaluation of LS for phase 2.2 Outcomes for schools were shown to be less positive and more mixed than for pupil and teacher outcomes. There were a number of problematic aspects that emerged from analysis of the case reports, individual interviews and the questionnaire survey. Despite the positive aspects of the LS process and funding to pay for teacher cover, one specific issue related to timetabling meetings, arranging teacher cover and that the process sometimes involved teachers using their free time to undertake the process. Another related issue was senior leader support in schools for LS, in some but not all participating schools. In view of these experiences, it was decided to focus more attention on the senior leaders of the phase 2 schools to persuade them to become more interested and find better ways of releasing teachers to undertake the Lesson Studies as explained below.

in 10 different schools (45% response rate). Case reports were received from 19 teachers at 13 different schools. These data sources were used to examine teacher and schools outcomes in phase 2. For pupils outcomes the GME method of monitoring the learning outcomes was used because it can be tailored to interventions like LS. It fits with the situated, variable and idiographic nature of LS goals, when no common external learner outcome measures can be used because of the diverse curriculum subjects and levels of learning involved. The GME method has been used to evaluate the outcomes of many kinds of programmes in various service contexts, initially under the name Goal Attainment Scaling (GAS) (Jones et al., 2006). Previous evaluation (Dunsmuir et al., 2009) indicates that it is most dependable when it meets these criteria: i. there is advance specification of the expected goal ii. at least three goals are used iii. there is independent review/assessment of levels attained

6. Phase 2 LS programme and evaluation 6.1. Phase 2 lesson studies Recruiting teachers and schools for phase 2 was similar to phase 1. Further schools were recruited from four local authorities in south west England in addition to a small number of schools from south-east England. In total, 33 teachers received LS training in phase 2 (2011e2012). Unlike phase 1, there were only two full day conferences, an initial briefing conference and a final conference after one term, and LS teams were only expected to undertake 2 and not 3 Lesson Studies. Identifying classes where there were pupils with MLD to act as case pupils in the LS was done similarly to phase 1. There were 31 pupils with MLD of whom 12 had a statement of SEN, 6 were School Action Plus, 10 School Action and 3 unknown.3 However, the teachers in phase 2 received less guidance and support while undertaking the LS process compared to phase 1 teachers. The aim was to examine how LS would operate under conditions more similar to typical school conditions where there was not outside support during the LS process. There were other changes to the LS programme derived from the phase 1 evaluation. The first was to introduce more specificity in the LS procedures to enable LS teachers to focus on the learning goals of the case pupils in their lesson planning. A more systematic goal setting and monitoring system based on a method known as Goal Monitoring and Evaluation (GME) was introduced into the LS procedures (Dunsmuir, Brown, Iyadurai, & Monsen, 2009; Jones et al., 2006; see details below). The second main change was to aim to engage senior leaders from the participating schools in the introduction to LS. This was done in two ways: by inviting one senior teacher from each school to attend a briefing session at the introductory LS conference and by emphasising to the LS teachers the importance of the contextual factors in their schools.

In the version of GME used the LS teams were asked to set 2e3 goals per case pupil before the LS cycle commenced. In addition, the teams were asked to specify three pupil performance levels on an 11 point progression line for each of the case pupil’s goals (ordinal scale). These levels were defined in terms of the 2e3 goals: two before the programme started (a baseline level and an expected level), and one after the programme (an achieved level). In this way, the teachers assessed pupils’ achieved levels by comparison with baseline and expected levels. They also gave descriptions for all 3 levels and evidence for those descriptions (see Appendix C for an example of a completed GME template). The LS teams were also expected to set levels so that there was some moderation of the attained levels to avoid bias that may arise from only one teacher’s assessment. The attained levels at the end of the LS cycle were then used to evaluate pupil outcomes in terms of any progress relative to baseline and the expected level (progress as expected; more than expected; less than expected; or, no progress). The degree of goal attainment could also be analysed in terms of the kinds of goals set.

7. Results from phase 2 7.1. Outcomes for pupils Completed GME data were received from 9 of the 15 schools (several LS teams gave incomplete data while a few gave none). This comprised data from 21 separate Lesson Studies with 1e2 case pupils per LS. In total, teachers set 69 pupil goals in these LSs. Table 4 shows that out of these 69 LS goals, progress was met or exceeded in just over half of the goals (54%; 37 goals) as assessed by the LS teams. Out of these 37 goals, progress was as expected in 25% of goals (17) and more than expected in 29% (20) of goals. In just under half of the goals pupils’ made some progress that did not meet the expected level (46% or 32 goals). This shows that of

6.2. Phase 2 evaluation The evaluation of phase 2 included the case report accounts and the realist questionnaire and individual interviews with selected teachers as used in phase 1. The questionnaire was sent to 33 phase 2 teachers, with responses received from 15 teachers

2

Data from the observations are not considered in this paper. These data are from the Goal Monitoring and Evaluation reports filled in by 9 of the 15 participating schools in phase 2. 3

Table 4 Goal monitoring and evaluation: 15 schools, 21 lesson studies, 69 goals set. Progress

Percentages

Progress as expected Progress more than expected Progress but not met the expected level No progress Total

25% (n ¼ 17) 29% (n ¼ 20) 46% (n ¼ 32) 0% (n ¼ 0) 100%

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

all goals the case pupils made some progress from the baseline level. Personal goals for pupils with MLD were analysed in terms of whether they were curriculum subject related e.g. ‘developing written ideas independently’ or ‘demonstrating understanding of auditory and visual information’ or referred to a learning process e.g. ‘being less disruptive and argumentative’ or ‘having more confidence in group activities’. Some goals were both subject related and about learning process, e.g. ‘independently offering contributions to class discussions and begin to record ideas in a more independent way’ or ‘being more frequently engaged in activities when emphasis is on using key vocabulary’. Overall, 49 of the 69 goals (71%) were about learning process only, while the remaining 20 goals (29%) were goals that were subject related or subject related linked to learning process. Table 5 also shows the level of attainment (scored 0e3 as shown in the table) by these three different kinds of goals. This indicates that the mean goal attainment scores for learning process and subject related goals were alike and similar to the overall attainment scores (in the range 1.79e1.83). The mean score for the relatively few combined subject related and learning process goals was slightly higher at 2.0. The mean attainment level for each pupil’s goals was calculated and shows that the modal range is 1e1.9 with those with scores above 2 the next most common range (see Table 6). The goal attainment for 13 of the 27 pupils was as expected or better than expected; for 14 pupils there was some progress but less than expected. 7.2. Outcomes for teachers Analysis of phase 2 case reports, as in phase 1 reports, showed that the most common outcome for the teachers themselves was that LS enabled them to explore new teaching strategies and resources, which was mentioned in 15 of the 19 case reports (79%). For example, one teacher commented that: Being able to plan with other subjects brought a range of ideas and strategies into my teaching which I would not have previously included. Phase 2 teachers also saw the benefits of the observation process (second most common outcome, in 6 reports) and of collaborating in the LS team (in 5 case reports). For example, one teacher suggested that: The LS process.has highlighted some very important aspects that would not have come to light if I did not have other adults in the room observing what the class was doing. Also frequently mentioned was that teachers became more aware of the needs of individual pupils with MLD and the importance of this new awareness (in 8 reports). Table 2 (see Section 4.2) shows that phase 2 teachers were also very positive about the outcomes of the Lesson Study process for themselves: in both phases the ratings were between 3.0 and 3.8 (with 4 as top of the scale).

169

Table 6 Mean goal attainment for each pupil across personal goals. Range of mean scores

Frequency

%

0e0.9 1e1.9 2e2.9 3

0 14 8 5 N ¼ 27

0 52 30 18 100%

These findings are consistent with the phase 2 case report analysis, reported above, and the analysis of the selected teacher interviews with five teachers and one deputy head teacher from three schools at the end of phase 2, discussed below. For a SEN specialist teacher in one school, the LS process had revealed new strategies with surprising effects on the pupils, many with extensive special educational needs. LS had also challenged preconceptions about what the pupils can achieve. This teacher particularly valued the observation process because it enabled new aspects to be found out about the pupils’ learning. She also pointed out that: I’d seen them all as very needy in their own right and needing my support, but I think there’s a point at which you can step back and allow them to support each other, and I hadn’t really thought about that before. Two teachers from another school had some contrasting views about LS. Although one saw some benefits in the LS process, particularly the collaborative aspects and becoming more aware of the specific needs of the pupils with MLD, she felt over-burdened by the amount of work involved. That the LS process was on top of her already heavy workload overshadowed the whole LS experience for her. The other teacher, though recognising the extra work in the LS process, was more positive and suggested many beneficial outcomes arising from the LS process. This included more analysis of her teaching, LS team collaboration and introducing some new strategies ‘that we can use to boost all students that have learning difficulties.’. She went on to suggest that as a result of the LS process she became more aware about observing the pupils in the class to see: How they are developing and what they are finding hard and what’s working or what’s not working. For the two teachers and the deputy head at the third school the LS process had been a very positive and beneficial experience. Their views reflected many of those already reported above. The process had been so enjoyable for one teacher that she felt ‘quite sad really’ when the LS process was over. For the second teacher interviewed, the LS process had: . just got us talking about MLD e I don’t think we really talked about it before e we’d talked about those students who had serious

Table 5 Level of goal attainment by kind of goal.

Subject related Learning process þ Subject related Learning process Total

None (score ¼ 0)


As expected (score ¼ 2)

>Expected (score ¼ 3)

Total

Mean

0 0 0 0

6 2 24 32/46%

5 2 10 17/25%

3 2 15 20/29%

14 6 49 69

1.79 2.00 1.82 1.83

170

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

learning difficulties, and we’d talk about the really bright ones.whereas the MLD students we just didn’t.it has made us aware of them so actually we’ve started to think about them and how to put strategies in place to help them. The deputy head of the same school highlighted how the LS process had made the LS team think more about teaching strategies: I think what it made us do was re-visit what was there and actually apply it.what it highlighted was that you don’t perhaps pay enough attention to the kids with MLD in your room e you kind of give them a piece of work and expect them to do their best with it rather than tailoring it a little bit more specifically. 7.3. Outcomes for schools The case reports from phase 2 schools show that most of the participating teachers (in 9 of the 13 schools) were planning to share their findings from the project with others in their schools. The remainder either had no intention to do so or did not know whether they would. Table 3 (see Section 4.3) shows that the phase 2 teachers rated school outcomes at a higher level (means of 3.00e3.29; all in the mostly-definitely rating range) than phase 1 teachers (2.40e2.88; all in the slightly-mostly range). It is notable that the highest rated outcome in phase 2 was about senior teachers and those with SEN and CPD responsibilities encouraging LS teachers, well above the phase 1 mean rating for this outcome. Interviews with the teachers and the deputy head from phase 2 project highlighted similar potential problems as the phase 1 teachers regarding the longer-term implementation of LS. This was despite all teachers being very positive about the outcomes for themselves and their pupils. For example, the deputy head elaborated on the issue of time and timetabling: The big.stumbling block within the school environment of Lesson Study is that it’s very time grabbing e you’ve got to have a lot of people involved in it at any one time so the cost of doing this programme, to continue it without [the project] funding, is huge. That is a big and significant financial implication for the school if we do it. He went on to consider the issue of timetabling: It would just depend on how we could release the staff from their current timetable.it is a logistical problem.it takes a lot of organisation.you’ve got to plan it out well in advance. One of the teachers at the same school pointed out that despite having cover for LS meetings and lessons, they used their own time when possible to reduce the disruption to their classes. The other teacher in that school had similar views while also noting that: Timetabling is more difficult because there’s a lot of stuff going on.sometimes me and [the other LS teacher] do our planning after school or during lunch. We enjoy doing it and that’s no problem.but I know some teachers would not [do it]. 8. Discussion and conclusions The aim of this paper was to show the use of a design-based research approach to refine the use of Lesson Study methods to

develop the teaching and learning of pupils identified as having MLD in secondary schools. LS was introduced to 17 schools in phase 1 and then another 15 schools in phase 2 of the project in the first UK LS development and research project in secondary schools with a focus on pupils with MLD. Only a few schools that started the project could not for practical reasons carry on and complete the project. For schools that completed the LS phases, the LS procedures were refined in response to the phase 1 evaluation in terms of pupil, teacher and school outcomes for further testing in phase 2. In order to ensure that the model of Lesson Study introduced to the teachers at the beginning of the process was followed as expected, the consultants visiting the schools provided some external monitoring and the case reports written by the teachers at the end of the project provided a record of the process. From this we can make a judgement that the LS model was put to practice as intended by a large majority of the LS teams. In this final section we summarise how the LS procedures were adapted and what outcomes followed these changes. We then summarise the overall pupil, teacher and school outcomes of this particular use of LS based for both phases and relate these to wider research literature about LS. But before doing this some comment is required about the gaps in some of the data collected. Questionnaire returns were lower than had been hoped. While some teachers did not manage to complete the internet questionnaires, response rates were improved when data was collected in the project conference time. There was also a gap in GME goal attainment data; completed by 60% of the phase 2 schools. However, we know from analysis of the phase 2 LS case reports where we had no GME data that five of the teams (out of six) reported positive pupil learning outcomes following the LS process. For example, at one of these schools the teachers reported that ‘during the process it became apparent that not only did the students’ behaviour improve, but that their thinking became more deep and meaningful’. These gaps in data can be related to the organisational issues in teachers finding time to undertake the LS procedures in schools as discussed previously. The LS procedures used in phase 2 involved two key changes, first, to monitor more systematically the gains in learning arising from the LS cycles and, second, about the context in which LS was to be used. This involved inviting senior teachers of phase 2 schools to attend a briefing at the initial LS introduction day conference. Though this invitation was not always taken up, there was direct communication by the project team about the importance of senior leader support. Additional attention was also given to the support and timetabling issues in the introduction conference. The evaluation of phase 2 lesson studies showed that both adaptations to LS were associated with positive outcomes. The goal setting and monitoring made it possible for goal attainment to be examined for the individual case pupils following the LS cycle. Analysis showed that there was progress from the start to the end of the LS cycle for just under half of the pupils with MLD at expected or beyond the expected levels (expected levels were set at the start of the LS cycle). The rest made progress but less than expected. However, there was no assessment of these gains independent of the LS teams. Nor could the gains be strictly attributed to the LS procedures in general terms. But, other evidence (case reports, questionnaire and interview data) suggested that LS made some difference and this can be a basis for tentative causal inferences at a local level (Maxwell, 2004). As discussed above, LS was introduced as a flexible strategy to be used in varied ways, so making experimental controls and external assessment impractical. A design-based research approach with

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

a realist evaluation methodology was used as these were suited to the local and exploratory aims of this project. Though an experimental evaluation could have in principle been used, this would have reduced the focus of this evaluation and would have been less useful to potential users of LS (Lewis et al., 2006). Participating teachers at the end of phase 2 reported much more support from senior teachers for LS than teachers in phase 1. It is unclear whether this was due to the increased attention given in the LS introduction and the written communication to the senior teachers in the participating schools. It might merely reflect a different set of schools with different commitments to the LS work. However, there was no evidence of change about timetabling and teacher cover to facilitate the running of Lesson Studies from phase 1 to 2. Evidence from both phases indicated not only variability in senior teacher support but also in how LS activities were arranged. The two can be seen as related; senior teacher support can establish a school timetable that makes time for professional learning, such as LS, to take place. For instance, in one school LS activity fitted easily into already running professional learning groups that had been established by the head teacher some years earlier. This topic has also been studied in this project in other ways (see Ylonen & Norwich, 2013). In terms of the overall teacher outcomes for both phases, teacher level evaluation data found largely similar and very positive outcomes for the teachers concerned. These were about the collaborative aspects of the LS process, improved planning of teaching and confidence to try new approaches, insights for LS observations and more awareness of the individual needs of pupils. These findings provide further evidence about the positive effect of LS on teachers and their teaching in a novel area of teaching and learning, that of secondary aged pupils with MLD (e.g. Dudley, 2012; Lee, 2008; Lewis et al., 2006; Lieberman, 2009; Puchner & Taylor, 2006; Rock & Wilson, 2005). The findings also confirm the conjecture by Lewis et al. (2006) that LS strengthens three pathways to instructional improvement: teachers’ knowledge, teachers’ commitment and community and learning resources. These findings are encouraging about LS as a strategy for secondary schools with a focus on teaching pupils with MLD. The findings indicate that the LS process enabled teachers to develop teaching approaches and a focus on the learning requirements of the pupils with MLD, who then showed some gains in their learning. Although the findings about pupil outcomes cannot be simply generalised to other settings, they act as a clear demonstration of positive pupil learning outcomes in a particular context and use of LS. Our evidence of how improved pupil learning was linked to the LS process in the project schools is consistent with Lewis et al.’s (2009) assertion that improvements in teaching through LS leads to better student learning. That phase 2 LS operated without the outside project support used in phase 1, also shows that project teachers could adopt LS procedures under more typical school conditions. Moreover, the findings show that a Goal Monitoring and Evaluation approach can be successfully used in efforts to create more robust evidence of pupil outcomes in LS contexts e an area where there are distinct gaps in research, knowledge and evidence. However, despite the positive outcomes of using LS, this research highlights the critical importance of contextual factors in the continuing use of LS in participating schools after the support and funding of the project ends. If LS observations, review and planning meetings are not timetabled in advance as part of whole school practices driven by senior teachers’

171

commitment to this kind of professional learning development, some teachers will use their own time to undertake the LS process, but others may be put off. Other international research has also shown the importance of contextual factors for effective use of the LS method in schools (e.g. Lee, 2008; Lim, Lee, Saito, & Haron, 2011; Perry & Lewis, 2009; Rock & Wilson, 2005). While demonstrating what LS can offer, this study can also be interpreted as showing that LS as a professional learning approach challenges how schools are organised and teaching managed. In addition to developing and evaluating the LS process and its outcomes for pupils with MLD, teachers and schools, this study also raises many questions for further research. A follow-up investigation with the teachers and schools from both phases of the project is being undertaken to find out if the LS approach continues to be used and, if so, what has made this possible. Further future research could focus on the development of teaching of other groups of pupils with special educational needs, not just those with MLD, as part of the movement to develop more inclusive forms of teaching. Another main line for further research could be directed at the general use of LS in schools, for example, first, to understand the school and wider contextual factors that help or hinder the LS process and its longer-term outcomes, and, second, the use of LS in pre-service teacher training. There is also the prospect of undertaking experimental evaluations (RCTs) to examine the effects of LS on the professional learning of teachers in a particular area of teaching. Despite its widespread use in the Far East and increasing use in the USA, LS has yet to be adopted more widely in Europe and in the UK. This project was a step in that direction.

Appendix A. Definition of moderate learning difficulties (MLD) Pupils with moderate learning difficulties will have attainments significantly below expected levels in most areas of the curriculum, despite appropriate interventions. Their needs will not be able to be met by normal differentiation and the flexibilities of the National Curriculum. They should only be recorded as MLD if additional educational provision is being made to help them to access the curriculum. Pupils with moderate learning difficulties have much greater difficulty than their peers in acquiring basic literacy and numeracy skills and in understanding concepts. They may also have associated speech and language delay, low selfesteem, low levels of concentration and under-developed social skills (DfES, 2005: 3).

Appendix B. Outline of the system of SEN in England. Level of need

Nature of intervention

School action

Additional to or different from the usual (differentiated) curriculum of the school. External agencies not involved. Recorded without categories in annual census. External support services are involved when the child does not make expected progress despite school interventions. Recorded with specification of categories in census. A local authority, after undertaking statutory assessment, despite interventions at previous levels, issues a statement (or record) of special educational needs and provision. Provision may be in ordinary or special schools. Recorded with specification of categories in census.

School action plus

Statement

Source: DfES, 2001, Special Educational Needs Code of Practice

172

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173

Appendix C. Example of a GME template filled in by an LS team.

References Ainscow, M. (2000). The next step for special education e supporting the development of inclusive practices. British Journal of Special Education, 27(2), 76e80. Brown, A. L. (1992). Design experiments: theoretical and methodological challenges in creating complex interventions in classroom settings. Journal of the Learning Sciences, 2(2), 141e178. Cartwright, N. C. (2007). Are RCTs the gold standard? BioSocieties, 2, 11e20. Cartwright, N. C., & Stegenga, J. (2008). A theory of evidence for evidence-based policy. In Presented to the NRC Standing Committee on social science evidence for useAvailable at http://personal.lse.ac.uk/cartwrig/Papers%20on%20Evidence. htm. Accessed on 26.10.12. Chia, N. K. W., & Kee, N. K. N. (2010). Teaching practicum for special education teachers: A modified lesson study for trainee special education teachers. Singapore: McGraw Hill East Asia. Cobb, P., Confrey, J., diSessa, A., Lehrer, R., & Schauble, L. (2003). Design experiments in educational research. Educational Researcher, 32(1), 9e13. Desforges, C. (2006). Review of literature about pupils with moderate learning difficulties. London: Esmee Fairbairn Foundation. DFE. (2012). Advance training modules for moderate learning difficulties. Available at http://www.education.gov.uk/lamb/module2/M02U09.html#. Accessed 29.10.12. DfES. (2001). Special educational needs code of practice. London: DfES. DfES. (2005). Data collection by type of special educational needs. London: DfES. Dudley, P. (2004). Lessons for learning: research lesson study, innovation, transfer and meta-pedagogy; a design experiment?. In Paper presented at the 5th Annual conference of the TLRP, 22e24 November, Cardiff, Wales. Dudley, P. (2012). Lesson study development in England: from school networks to national policy. International Journal for Lesson and Learning Studies, 1(1), 85e100. Dunsmuir, S., Brown, E., Iyadurai, S., & Monsen, J. (2009). Evidence-based practice and evaluation: from insight to impact. Educational Psychology in Practice, 25(1), 54e70.

Hammersley, M. (2007). Educational research and evidence-based practice. London: SAGE. Howes, A. J., Davies, S. M. B., & Fox, S. (2009). Improving the context for inclusion: Personalising teacher development through collaborative action research. London: Routledge. Jones, M. C., Walley, R. M., Leech, A., Paterson, M., Common, S., & Metcalf, C. (2006). Using goal attainment scaling to evaluate a needs-led exercise programme for people with severe and profound intellectual disabilities. Journal of Intellectual Disabilities, 10(4), 317e335. Lee, J. F. K. (2008). A Hong Kong case of lesson study e benefits and concerns. Teaching and Teacher Education, 24(5), 1115e1124. Lewis, C. (2009). What is the nature of knowledge development in lesson study? Educational Action Research, 17(1), 95e110. Lewis, C. C., Perry, R. R., & Hurd, J. (2009). Improving mathematics instruction through lesson study: a theoretical model and North American case. Journal of Mathematics Teacher Education, 12, 285e304. Lewis, C., Perry, R., & Murata, A. (2006). How should research contribute to instructional improvement? The case of lesson study. Educational Researcher, 35(3), 3e14. Lieberman, J. (2009). Reinventing teacher professional norms and identities: the role of lesson study and learning communities. Professional Development in Education, 35(1), 83e99. Lim, C., Lee, C., Saito, E., & Haron, S. S. (2011). Taking stock of lesson study as a platform for teacher development in Singapore. Asia-Pacific Journal of Teacher Education, 39(4), 353e365. Maxwell, J. A. (2004). Causal explanations, qualitative research and scientific enquiry in education. Educational Researcher, 33(2), 3e11. Miles, S., & Ainscow, M. (2011). Responding to diversity: An inquiry-based approach. London: Routledge. Norwich, B., Kelly, N. (2012). Moderate Learning Difficulties and the Future of Inclusion, 2005, In B. Norwich, A. Ylonen, & R. Gwernan-Jones (Eds.), Moderate

B. Norwich, A. Ylonen / Teaching and Teacher Education 34 (2013) 162e173 learning difficulties e Searching for clarity and understanding. Research Papers in Education. (published electronically). London: RoutlegeFalmer. Pawson, R., & Tilley, N. (1997). Realistic evaluation. London: SAGE. Perry, R. R., & Lewis, C. C. (2009). What is successful adaptation of lesson study in the US? Journal of Educational Change, 10, 365e391. Puchner, L. D., & Taylor, A. R. (2006). Lesson study, collaboration and teacher efficacy: stories from two school-based math lesson study groups. Teaching and Teacher Education, 22(7), 922e934. Robson, C. (2011). Real world research (3rd ed.). London: J. Wiley and Sons. Rock, T., & Wilson, C. (2005). Improving teaching through lesson study. Teacher Education Quarterly, 32(1), 77e92. Sibbald, T. (2009). The relationship between lesson study and self-efficacy. School Science and Mathematics, 109(8), 450e460. Sims, L., & Walsh, D. (2009). Lesson study with preservice teachers: lessons from lessons. Teaching and Teacher Education, 25(5), 724e733.

173

Takahashi, A., & Yashida, M. (May 2004). Ideas for establishing lesson-study communities. Teaching Children Mathematics, 436e443. Teddlie, C. B., & Tashakkori, A. (2009). Foundations of mixed methods research: Integrating quantitative and qualitative approaches in the social and behavioral sciences. London: SAGE. Ylonen, A., & Norwich, B. (2012). Using Lesson Study to develop teaching approaches for secondary school pupils with Moderate Learning Difficulties: teachers’ concepts, attitudes and pedagogic strategies. European Journal of Special Needs Education, 27(3), 301e317. Ylonen, A., & Norwich, B. (2013). Professional learning of teachers through a Lesson Study process in England: contexts, mechanisms and outcomes. International Journal of Lesson and Learning Studies, 2. Yoshida, M. (2012). Mathematics lesson study in the United States e current status and ideas for conducting high quality and effective lesson study. International Journal for Lesson and Learning Studies, 1(2), 140e152.