Education-line Home Page

How does CPD affect teaching and learning? Issues in systematic reviewing from a practitioner perspective

P. Cordingley, M. Bell, & B. Rundell
Centre for the Use of Research and Evidence in Education (CUREE), Coventry, email: Philippa.Cordingley@curee.org.uk

Paper presented at the British Educational Research Association Annual Conference, Heriot-Watt University, Edinburgh,
11-13 September 2003

This paper is in draft form. Please do not quote without prior permission of the authors.

Section 1

1.1 Introduction

Recent initiatives to encourage teacher engagement with research and evidence in their professional practice have now been taken up in policy circles to the extent that several large-scale projects have been developed and funded at government level. One example is the Networked Learning Communities by means of which teachers and school leaders are supported in collaborative inquiry and knowledge sharing. Another, different, example is the Research Evidence in Education Library (REEL) which is the home site of the Centre for Evidence-Informed Policy and Practice in Education, (EPPI Centre) funded by the Department for Education and Skills. The Centre's vision is "to be a centralised resource for people wishing to undertake systematic reviews of research in education and those wishing to use reviews to inform policy and practice". (REEL) Practitioner involvement in conducting reviews is an essential feature of the EPPI review process. Although the reviews were intended to be of use to other educational 'users' as well as policy-makers, it was our experience of working with and for practitioners during this process which has prompted this paper.

The systematic review process developed by the Centre grew out of EPPI's work in the health field and the belief that there is "much that researchers in education and users of educational research can learn from work in these other areas, although some of the challenges of research synthesis in education are particular to that setting". (REEL) Our aim in what follows is to:

  • explore the nature of these "challenges";
  • highlight the findings of one such review on the impact of collaborative CPD; and, through this example,
  • assess to what extent the health-derived review methods have the potential to support education practitioners in developing an evidence-informed approach to their practice.

1.2 Evidence-based practice or evidence-informed practice?

There is no quarrel here with the requirement for rigour in all spheres of educational research - on the contrary, the National Teacher Research Panel (Cordingley, Philippa NTRP 2000) has made it clear that practitioners are not interested in research which has not been rigorously designed and carried out. Hammersley (2002) emphasises that "teachers need to know that research findings are valid and might provide answers to their own needs and concerns." Reviews of research, he suggests "are increasingly finding their way into schools and organisations as a means of delivering messages to inform practice".

The reality of teaching is that what works in one classroom is unlikely to work in quite the same way in another. Research into teachers' professional development and studies of teachers' and leaders' use of research and evidence in practice (Desforges (2000), Guskey (1986), Huberman (1993), Mitchell (1999), Wikeley (1998), Cordingley & Bell, (2002), and Williams & Coles ESRC (2003) points consistently to the need for education practitioners to interpret and adapt information from research for use in their own contexts. Because of this context specificity practitioners do not expect evidence from educational contexts to provide readily transferable, 'safe' knowledge, and, unlike health practitioners, they would regard with suspicion any such claims, no matter how extensive and rigorous the research. They are reminded minute by minute of the multiple and dynamically interactive variables involved in learning in school classrooms.

This, of course, has significant implications for the way in which research reviews in education are presented. What does it mean for how they are carried out? Hammersley (2002) identifies major differences between reviews aimed at fellow researchers and those aimed at lay audiences ("polar opposites"), particularly in the type of language used and the type and amount of information needed. Our own experiences provide evidence of what is involved in seeking to resolve or at least address these tensions.

1.3 Background and context: the CPD Review Group

Because this paper claims to be representing a practitioner oriented perspective on the systematic review process, we believe it to be important to describe the composition of the review group and to offer a brief rationale for its establishment.

The review was initiated and substantially sponsored by the NUT who were also the principal funders. Because of the potential interest to teachers, additional resources were provided by the GTC. DfES funding came through registration with the EPPI-Centre. A systematic approach to research in CPD was thought to be timely, both because of the many national and international initiatives dependent upon significant advances in teacher learning and because of the NUT's own initiatives in professional development. The government's CPD strategy aimed at enabling teachers to take more control of their own professional development and give schools much more direct control of the funding for CPD. The NUT believed that teachers and schools needed and wanted to know more about how professional development might help them develop professional knowledge, skills and careers at the same time as enhancing pupil learning.

The membership of the review group and its advisory group included teachers, NUT and GTC officers, academic CPD research specialists, DfES CPD specialists and members of the Centre for the Use of Research and Evidence in Education (CUREE). The hands on review group comprised a group of (mainly network) teachers, CUREE colleagues and one HEI colleague. At the time, it was the only EPPI review group which was not based in a higher education institution. The expectation was that registered review groups would undertake a series of systematic reviews, over time, to explore aspects of their specialist field.

1.4 Issues in Systematic Reviewing

In the next sections of the paper we:

We have first described the aims and findings of the review in sections 2 and 3 in order to set the context for our analysis. We wanted to illustrate what we were looking for and what we were able to find out as the context for discussing the processes we followed and the challenges we encountered in doing so. In sections 4 and 5 we outline the EPPI review process which we followed and show how the practitioner and researcher inputs generated tensions which were often healthy and creative. However we found the balance in the review methodology to be significantly weighted towards the researcher perspective and we have identified a number of areas where the EPPI framework and the medical research tradition was at odds with the needs of education practitioners as creators and users of reviews.

Section 2

2.1 Aims of the Review

Our review was focused on CPD for teachers of the 5-16 age group which was both sustained and collaborative. We wanted to know how CPD affected teaching and learning, so information about the nature of the CPD, its context and processes were important aspects of our enquiry and were a strict condition for the inclusion of studies. We will describe this process in more detail in section 4. The review was initiated in the context of an earlier, interpretative review of teachers' acquisition and use of knowledge (Cordingley and Bell, 2002) which drew extensively on evidence about the importance of teacher experimentation and coaching (e.g. Joyce and Showers, 1988). The review also drew on the work of various authors about the stages of teacher development, such as Hargreaves' (1993) modelling of the way in which teachers are able cumulatively to extend aspects of practice.

By collaborative CPD we meant teachers working together or teachers working with LEA, HEI or other professional colleagues on a sustained basis. In fact 12 of the studies finally included involved teachers working together with teacher colleagues. Whilst the core purpose of CPD is enhancing student learning, it embraces teacher learning and teacher beliefs, knowledge, attitudes and behaviours as a means to that end. The review was therefore conducted with a strong focus on the expressed needs and interests of teachers in relation to their students' learning.

As we explained in our review report, the decision to pursue studies that attempted to relate teacher learning and pupil learning was a radical one given the number of intervening variables and the apparent paucity of studies in this area. This goal and the focus on sustained and collaborative CPD were fuelled by teacher interest. Early trial searches informed by the work on CPD outcomes of Harland and Kinder (1997), Joyce and Showers (1988) and Day (1999) gave us confidence that the question would generate studies likely to produce positive findings of interest to teachers. In particular, we wanted to be able to attend to teachers' interest in the nature of the CPD and the different ways in which it affected teachers and students.

Section 3

3.1 The Review Findings: a brief outline

It was clear from the reporting of the studies that it was not a straightforward process to follow often complex interventions through to their effects first on teaching and secondly on students. Whereas researchers in other disciplines, such as medical research, are often able to justify their claims about impact by reference to relatively easily measurable differences in outcomes and through comparisons with control groups, research into CPD is not always able to track inputs or measure outcomes quite so rigorously. Measurement of the effects of CPD not only had to address pupil outcomes, but to embrace the fundamental changes in much less easily evidenced factors such as attitudes, beliefs, knowledge and behaviours of teachers and their dynamic relationship between these factors and the responses of their students. In doing so, it also had to pursue similarly complex, hard-to-observe factors for the students taught by those teachers. Changes following CPD interventions, while very real to the teachers and pupils involved, were difficult, time-consuming and costly to record and quantify in terms of research data. Some academic colleagues in our Review Group were deeply sceptical about the possibility of unearthing studies that attended to both pupil and teacher outcomes and advocated focusing predominately upon the latter. However, the teachers participating in various aspects of our review were adamant that the review should explore links between CPD and both teaching and student learning. We therefore held to this aim.

3.2 How did the CPD interventions affect teachers and teaching?

The reports from which we drew our findings cited changes in terms of teachers' attitudes and beliefs, knowledge and understanding as well as their classroom practice.

Changes in attitudes and understanding

All but one of the studies reported changes in teachers' attitudes. Evidence from observations, interviews, questionnaires or teacher diaries indicated that participation in the collaborative CPD programmes was linked to enhanced teacher confidence. Six of the studies in the review also indicated that teachers shared a stronger belief in their own power to make a difference (self efficacy). There were reports in seven of the studies of increased teacher enthusiasm for professional development through collaborative working. Positive outcomes of the impact of collaborative CPD often emerged only after periods of relative discomfort in trying out new approaches; things usually got worse before they got better. In the words of one teacher "I think at first everyone had a lot of reservations, a lot of trepidation. I think now we're all in a learning mode". A further eight studies reported an increase in teachers' willingness to take risks including trying things that they had previously thought to be too difficult. Collaboration was important in sustaining change.

Reviewers were keen to explore any changes in teachers' understanding of the subject or in their knowledge of teaching methods although the main focus of the review was on exploring how this led to improved classroom practice. Evidence from the Saxe study indicated that when teacher CPD developed both their own understanding of mathematics and that of their students, this led to greater gains in their pupils' conceptual understanding compared to pupils in other groups. Collaborative discussion in a climate of 'critical openness' in examples such as the Kirkwood study enabled teachers to "get beneath the surface of issues" leading to greater competence and understanding. There were also examples of collaborative development of new curriculum units. Other specific examples of increased pedagogic knowledge included: greater insight into students' thinking, understanding of new teaching strategies such as advance organisers, or decoding skills in reading.

Nine studies reported the use of strategies for supporting and encouraging more active learning, such as making stronger connections between ideas, developing co-operative learning strategies between students, enhancing problem solving and involving students in designing learning activities. Development of teachers' ability to support student self-evaluation was cited in three studies.

Pedagogical change

Teachers made changes either to the content of lessons through specific teacher activities, or in generic learning processes. Changes to the content of lessons tended to be related to the aims of individual studies and included: greater use of computers for teaching and problem solving, more effective planning for pupils with special needs, or the use of specific student support strategies. Several studies reported more effective teaching and learning after teachers had increased their own knowledge in science or mathematics. As teachers benefited themselves from more active learning opportunities, so this became manifest in their practice, with greater focus on active student-learning. Individual studies reported for example that teachers involved in active learning through collaborative CPD were "trying to teach with less telling " and using student problems as a focus for learning, or that teachers provided more feedback to students and teaching became "learning rather than task oriented".

There was one study where the collaborative and sustained CPD did not lead to the targeted improvements. This CPD simultaneously targeted changing the learning environment and increases in teachers' use of ICT. Student views that their learning environment had not changed led the teachers in this study to commit themselves to an additional, more specifically focused year of action research. Sustained and collaborative CPD was also less effective where one of two groups focusing on the most challenging pupils were novices and much less able to benefit from the programme than experienced colleagues. Other noteworthy findings included differences in outcomes when groups were or were not involved in direct classroom observation, or when there was no subject input into an intervention intended to achieve subject specific changes. Time for discussion, planning and feedback, and access to suitable resources were a common concern in many of the studies.

3.3 How did the CPD interventions affect pupils?

Pupil outcomes were reported in terms of changes in pupils' attitudes and behaviours, or in their learning, based on a range of evidence, including questionnaires, interviews, observation and teacher report. All but two of the studies reported observable improvements in attitudes to learning and included increased active participation in lessons and enhanced motivation and enthusiasm. Evidence of pupils' increased confidence also emerged: for example "students enjoyed co-operative work leading to greater confidence and increased satisfaction with their work" (Britt). Another study reported that "the majority of students involved in the programme enjoyed a very positive learning experience and were motivated by the new units" (Kirkwood). The Ross study reported that "students believed the new process of self evaluation was fairer and they appreciated being given a voice".

In general, researchers were cautious in claiming causal links with performance outcomes for students. Nonetheless, evidence from seven of the studies indicated that participation in collaborative CPD by teachers was linked to measured increases in students' achievement. Such changes included: increased performance in mathematics, science, English or economics as measured by pre and post testing, greater ability to explain mathematical thinking, improved decoding and comprehension skills for struggling readers, or development of technology skills. Often such outcomes became evident only after the programme had been underway for some time. The Bryant study focussed on supporting less able readers with the demands of subject specific texts and reported that "teachers were beginning to see the effects of the strategies.....In the maths classes students were begging for multisyllabic words to decode". As reported previously, in one study teachers failed to note changes in pupils' perceptions of specific programme objectives but teachers did find student feedback sufficiently powerful to motivate them to enter a second round of action research to address their goals and bring about changes that would be noticed by their students.

Teachers' collaboration in the CPD process was linked with greater pupil-teacher collaboration in the classroom. Evidence also indicated that students had begun to question each other, evaluate each others' work and show an interest in the process of their own learning. This led to a review hypothesis that teachers' modelling and engagement in collaborative learning generated an enthusiasm for creating similar opportunities for their students.

3.4 What were the main features of the CPD interventions? What explains what worked?

To help understand the implications of the review findings for practitioners, the reviewers wanted to try to identify common features of the collaborative CPD reported in the studies. In several studies the emphasis on reporting was so firmly upon outcomes that intervention details were described only in outline or woven haphazardly into the exploration of outcomes. The methodological bias of the data extraction process (see below) also created problems in teasing out intervention data. Thus reviewers spent considerable time examining individual studies to unpick the sort of details that were needed to inform our review question. We found the most consistently reported elements of the CPD interventions to be:

Outside Expertise

All of the studies reviewed involved the input of external 'experts'. This was not simply a story of outsiders riding to the rescue of ignorant teachers. Sensitivity and flexibility were needed to ensure that such inputs took place within a framework of partnership with teachers. External specialist input and teacher peer support took the form of collaboration between equals, where each of the partners brought "separate but complementary bodies of knowledge" (Ross). External consultants were typically from neighbouring university research departments sometimes supported by district (local authority) advisors. Their support included providing access to relevant existing research, helping teachers to refine their development aims to make new work both useful and manageable and (for those involving enquiry) in the processes of data collection and analysis. Experts were seen as useful in providing a focus for debate, encouraging professional reflection on existing teaching practice and offering a menu of possible options which could then be modified to teachers' own contexts. Consultants also modelled practice and supported teachers through mentoring or coaching.

Observation

Researchers or teachers were involved in observing classroom practice in nine of the studies. Sometimes this was an informal arrangement between teachers, sometimes a more formal part of the CPD process involving exposure of lessons to outside 'experts' Tracking the benefits of observation and feedback to improvements in teaching and learning was complex and difficult. However the review provides evidence of cases where comparisons were made between teachers who received such coaching and those who did not. One study (Da Costa) reported that collaborative consultation between teachers without direct classroom observation was the least effective method, of the four studied, in promoting pupil achievement and personal teacher efficacy.

Peer Support

In twelve studies teachers were reported as offering crucial support to each other, often through coaching. For example in at least eight of the studies the teachers were encouraged to undertake development of new lessons collaboratively within workshop or coaching sessions. Thus cross fertilisation of ideas and shared development helped to reduce the load on individuals and create a meaningful learning environment while simultaneously enhancing the productivity of the group. Teachers in a number of studies also valued the opportunity to develop a team spirit amongst professionals who shared the same work experiences and goals. For example one study reported that 'engaging and developing teachers' interest, expertise and energy may be enhanced by the collaborative nature of the CPD' (Kirkwood). Working collaboratively was also reported as being important in keeping the projects moving when enthusiasm might have waned, supported by the presence of individuals who could drive ideas forward.

Ownership

Seven studies explicitly reported that teachers had been given the opportunity to select their own focus for the CPD. Thus projects often resulted from teachers' genuine interests in exploring a 'burning issue' or developing specific expertise. In these and three other studies, teachers and consultants examined existing research outcomes or specific programmes which were then adapted to suit the needs of both teachers and their pupils. In other examples, sensitivity to teachers' needs took the form of allowing teachers control over the intervention timetable or professional development sessions they attended. Reported benefits resulting from 'giving teachers a voice in their CPD' included the creation of informal networks which enabled teachers to determine their own priorities. They were "architects" of the new curricula and "designers" of challenging classroom environments. (Parke & Coble) A sense of ownership of the focus of the CPD programme appeared to be a strong motivator. Teachers were able to focus on strategies that they believed could really benefit their students.

Differentiation

There was explicit evidence in five studies of the need to differentiate between teachers' individual starting points. The majority of CPD interventions involved were designed to match needs to processes. The process of observation was reported as important in enabling colleagues or consultants to understand 'where teachers were coming from'. For example one study reported that observation feedback was a useful tool in promoting discussion on the effects of the current intervention before teachers moved on to try something else. (Britt)

Action research programmes were also reported as helpful in enabling teachers to start at a level and pace they felt comfortable with. Ross found that 'Participation in collaborative research helped teachers to add an item to their agendas for professional renewal and to determine when they would deal with that item.' By establishing more than one learning cycle this study also created an opportunity for teachers to identify and refine what they wanted to research before implementing the action research.

Although six studies provided some degree of baseline needs assessment, it was not clear whether the results were used to diagnose teacher needs and thus inform the programme design. Such information could inform the design of the CPD to ensure that it was appropriately differentiated.

There is evidence from two of the studies of the importance of paying particular attention to starting points for beginning teachers. Gersten showed how the beginner teachers in his study needed more support and possibly a different type of CPD, as they were still learning the basic craft of teaching. O'Sullivan reported that an established model of CPD had to be considerably modified for teachers who had limited qualifications in Namibia.

3.5 The EPPI Review process: a brief outline

The review protocol set out in details the aims and scope of the review, the review question and the methods by which the review would be undertaken.

For practical reasons, the review focused on studies carried out since 1988, across the 5-16 age range that were reported in English, although there were no geographical limits.

Methods of identifying studies for the systematic map and in-depth review comprised:

4. The Challenges Of The Review Process From A Practitioner Perspective

4.1 ' Healthy tensions': debates on methodology

It is certainly not our intention to give the impression that the systematic review process was an unproductive experience from a practitioner perspective. It was, on the contrary, a formative experience, requiring a good deal of flexibility, adaptation and scaffolding to embrace the diverse starting points of the review group members. In particular, we learned a great deal about good and bad practice in reporting research and we are still feeding some of the lessons learned into our ongoing projects. The latter are all aimed, in different ways, at bringing sound research evidence to the attention of practitioners in accessible and useable formats.

Admittedly there were some tricky practical problems as a result of involving 'users' (practitioners and policy-makers) in all aspects of the review process, especially the complex and time-consuming data-extraction process. It called for a considerable amount of costly training and support. Most practitioners were also (understandably) unused to the statistical terms and techniques and had to be supported in these areas as well. However the mix between the practitioner perspective and that of experienced academic researchers, including the EPPI reviewers, did lead to some productive tensions. For example, studies which might have been judged by EPPI as of high quality on purely methodological grounds were valued more cautiously by users because of lack of information about the school/teacher contexts and about the processes and strategies used in the intervention. Teachers were simply not interested in knowing whether an intervention was effective without also knowing the details of the design, delivery and implementation of the intervention. They needed to get a grip on the scale and nature of the work if they were going to be able to relate it to their own professional contexts.

Similarly, studies which practitioners might have found particularly relevant and helpful in exploring the practical implications of the review may have failed to meet the more exacting methodological assessments of the EPPI approach to methodology. These related particularly to the nature of the sample and the sampling process and to the presence or absence of control groups - always a tricky ethical issue in education. Probably as a result of this, a high number of studies were judged to be of overall 'medium' weight of evidence in relation to the review specific question. Some of the differences in perspective were less productive and we outline the problems we experienced in the following sections.

4.2 Setting the review question: what practitioners want to know

The review question sets the scope for every stage of the review process. Inclusion and exclusion criteria are dictated by the question, as is the search strategy. Not surprisingly, review groups are advised to keep their questions 'manageable' and tightly focused. In the CPD review group it soon became evident that a focus on one particular sort of CPD intervention in one particular area of the curriculum was not proving attractive to practitioners -and wasn't particularly exciting for policy-makers either. The same lack of enthusiasm was evident when the group turned to the types of outcomes it wanted to look for. Changes in teacher behaviours were not enough; the teachers we consulted wanted to know whether such changes had an impact on the pupils involved: did their learning improve during or following the CPD? They were also opposed to limiting the focus either by curriculum area or by age range or phase which would have made the review more manageable. They argued that if we wanted teachers to take any notice of the review findings we had to make sure they had the widest possible reach.

Because of this, it proved impossible to create a narrow and easily manageable focus for the review. The teachers were particularly interested in forms of collaborative and sustained CPD, so we were able to eliminate all one-off or one-day courses and inputs and all forms of CPD where teachers worked in isolation. However collaborative CPD encompassed a variety of types of intervention in a wide range of settings which meant that the review group had to make ensure that the search strategy was broad enough to find what was out there.

The result of practitioner influence at this critical early stage of the review therefore was that we had to devise a much broader initial search strategy than we had initially envisaged. We believe this to be an important lesson. If reviews are to be genuinely geared towards practitioners' (and even policy makers') interests and if such folk are genuinely to be involved in making crucial decisions about the review then they are unlikely to be satisfied with the kind of tight and narrow focus implicit in the EPPI systematic review design. This also has implications for the number of titles and studies likely to be identified and subsequently for the data extraction process in the final stages of the review. We will deal with the latter in more detail below.

4.3 Keywording: creating meaningful maps for education practitioners

Another early indication of the gap between the review methodology designed by the EPPI team and the potential value of the review for teachers was the choice of keywords, where the legacy of the health approach was much in evidence. Ours was part of the second wave of education reviews, yet we found the keywording instrument still appeared to be unsympathetic to the needs and interests of an education audience. Teachers, like health practitioners, tend to specialise, especially at secondary level. Any research map needs to be designed to reflect those specialist interests if potential users are to make connections between the research and their own practice.

Some of the keywords were either too broad (teaching and learning; teaching staff) to be helpful to practitioners with specific concerns - or too remote from the practice of education (lawyer, treatment) to warrant inclusion as core keywords. Practitioners do not readily turn to research for its own sake. It has to relate to specific issues or problems which concern them. We wanted to make sure that the core keywords (which would fundamentally shape the searchability of our review) reflected this as far as possible to encourage and enable practitioner searching between reviews whilst we were keen to devise a complementary set of review specific keywords. It seemed to us that these should relate closely to CPD rather than to core aspects of the education system. For example, there was only the word 'teacher' in the subsection dealing with the population focus of the research. he CPD group felt strongly that irrespective of the main focus of individual reviews, the core keywords should at a minimum distinguish specialist teachers such as SENCOs and subject co-ordinators and between teachers and learning assistants.

Other examples in the key words for the Topic Focus included:

Of course part of EPPI's reasoning in structuring the core keywords in this way is to enable read-across between education reviews and those in other fields. But the reality of keywording multiple articles is that only a certain number of keywords can be coded reliably in a single pass. In our experience it was important to be able to contain all key words in a two sided paper document since few reviewers found it sensible or practical to work on line because of the time needed for reading, the response rate of the software and problems of access to computers with modems. Including words from other domains is problematic because it risks giving, operational and symbolic priority to the needs and interests of researchers rather than those of educational practitioners and policy makers.

There were many other examples in the detailed comments the CPD Group prepared as we worked through the keywording process and, along the way some keywords were changed and much for the better. Much of the medical referencing disappeared, although references to teachers remained sadly undifferentiated. Again, our main purpose in rehearsing these experiences is to reinforce our earlier contention that research reviews which are genuinely directed at least as much at practitioners as at the research community need to be able to produce maps which are readily accessible by and meaningful to teachers' own concerns and interests.

4.4 Data Extraction: what data?

By the time we were ready to review our included studies in-depth, we had sifted systematically 13,479 titles and abstracts, reviewed 266 full studies, identified 72 studies as relevant and so keyworded their content to create a map of the literature. (The review process is described in section 3.5 above).

We had seventeen studies in total which had been through the process and met all of our review criteria, including the critical last fence which required that they be able to describe the impact of the CPD on pupils as well as teachers. After an introduction to the data extraction methods described in section three above, the group established its review pairs, distributed two copies of each of the seventeen included studies amongst them and went to work.

Reviewers had to answer over 100 questions for each study, many of which had accompanying text boxes where summaries of the relevant text from the studies could be pasted, before comparing notes with their co-reviewer. Each reviewer spent between four to seven hours on each study - and more in some complex cases. At this point we would like to remind readers what we said we were looking for, as described in the review protocol (Appendix 1) and in sections 1.1 and 4.1 above. We wanted to know what processes were involved and how the CPD affected teaching and learning, not just whether it did or whether it didn't. We hoped the review findings would help us understand more about how CPD could lead to improvements in teaching and in pupil learning and we approached the task of data extraction with these questions very much in mind. However the reality of the data extraction was very different. Of the 100 or so questions only one dealt with the substance of what we wanted to know: "Please describe in more detail the specific phenomena, factors, services or interventions with which the study is concerned." In contrast, for example, twenty-nine questions dealt with aspects of the sampling methodology, including allocation into groups.

The variation in the reporting of the studies themselves was also problematic. Some studies went into detail about the CPD interventions whilst others reported very little. In the event the group tackled the issue by assessing the 'whether' and 'how' questions separately in terms of the specific review question.

On completion of the review we discovered that we could have devised an additional set of review-specific data extraction questions targeted at what we wanted to know about the actual processes, design, inputs and outcomes which we had specified in the protocol. However we are all human and the degree of methodological interrogation to which each study is subjected under the current system is already extremely-complex and time-consuming. It was only after data extraction was completed that we realised the significant gaps in the resulting material and so had to conduct a further round of analysis. To find the information we wanted about the CPD processes we had to revisit each study after the data extraction process, putting yet more time into the review. Many of the studies were comparatively small-scale investigations. Our findings, useful and widely welcomed as they have been, should not have taken us so long to identify. We believe that the data extraction process would be more appropriately balanced if questions or a substantive kind were given more prominence by EPPI and if some of the more detailed questions about, for example samples were refined or at least organized into main and sub questions, avoiding the impression that methodological minutiae rather than research questions drive the process.

This is not to say that teachers don't want to know that the research was sound. A reliable and transparent quality appraisal mechanism is clearly necessary to establish whether the findings and conclusions from the research are trustworthy. But this is different from an exhaustive commitment to the capture of data relating to the quality of the research as an end in itself. Data about the processes involved in the intervention and the nature of the outcomes are of greater interest and use as far as teachers are concerned and the EPPI framework needs to model and support this. The data-extraction tool has the potential to be a 'gold standard' for systematic literature reviews. However a more practitioner-oriented approach to data extraction needs to be devised to ensure that the tool has the capacity to yield the type of data which can ultimately help to inform practice. None of the above should detract from the fact that reviewers learned a great deal from the data extraction process - not least about variations in reporting research and the challenges these pose for systematic reviewing.

4.5 Synthesis: creating meaning from the data

The final stage of the systematic review process involves synthesizing the data extracted from the included studies. What, if anything, does all this tell us about what works in CPD? Although most of the EPPI review processes are tightly specified, the process for synthesis is not. Yet it is in many ways the most challenging aspect of the review and one that is central to both the usability of the outcomes and to the rigour of the findings. After data extraction the only further generic structuring takes the form of "guidelines" about the structure of the report

The EPPI guidelines for reporting are attached as appendix 2 and broadly require for section 4:

From this it is clear that the synthesis process must start by being descriptive: the pattern of studies, their design, the issues they address and their reliability in answering the question. Such work provides a two dimensional map of the material contained in the studies. But, from a user perspective more work is needed to illustrate as it were, the topology of the material; to show its contours and reveal the texture of the landscape. The Review Group asked itself, in effect, would simple lists of this and that aspect of the research constitute a synthesis that would be meaningful to practitioners? We knew they would not. Our aim was not just to explore whether collaborative CPD had on impact, or indeed the nature of that impact. We wanted also to address practical questions about how the CPD worked and the links between processes and outcomes. In other words the data had not only to be described but to be interrogated and made meaningful.

We used two lenses for this interrogation. First we tried to explore the significant patterns and categories emerging from the data and to expose connections between them. One example of a finding emerging from the data rather than our hypotheses or the protocol is the extent of the use of specialist expertise. Our focus had been on professional collaboration not expertise. Yet this has emerged as a positive element within all of the CPD processes reported in our review.

Second we looked at the data against questions identified by practitioner and research colleagues in our advisory group. For example the review group was interested to know what the studies had to offer in relation to teacher readiness. In fact there was very little exploration of this concept but this exploration did uncover an important strand of related evidence about teacher ownership, motivation and differentiation. So far, so enlightening.

At the other end of the spectrum, however, the synthesis involved a dialogue with EPPI which focused on issues relating to the nature of our question and the reliability of findings. As we have said, our question raised both "whether" and "how" issues. Clearly there is no point in looking at how CPD had an impact if it were not first firmly established that it did in fact have an impact. However the data extraction had resulted in judgements about reliability and, we thought, answered the 'whether' question. Fifteen studies were judged to provide sound evidence of impact. Now, the review group wanted to concentrate on synthesising across the process data to discover as much as possible about what worked for the teachers and pupils in the CPD interventions - not to spend more time and effort in distilling methodological information. The dialogue about this process and EPPI's continued emphasis on method has, we believed resulted in a rigorous review of the evidence as to whether collaborative CPD has an impact. But, we believe, the whole process would be greatly strengthened and made more interesting and accessible to users if at least as much attention is paid in future to the analysis of how effective interventions take place, the detailed processes involved and the detailed outcomes.

A further issue causing dialogue around synthesis concerned whether or not the analysis of the studies in relation to our hypotheses should appear as discussion or as part of the descriptive layer of the synthesis i.e. whether it should appear in section 4 or 5. At first glance this may seem to be a trivial issue. For practitioner-related research, however, we believe it to be an important one. There needs to be a clear line drawn between the data themselves (in whatever lists, patterns or tables they are presented) and the construction which the reviewers have put upon them. There needs to be a clear boundary between description and interpretation. Having a model report structure is helpful. The rigid standardisation of reports however can lead to a diminution of 'fitness for purpose' and we believe there is room for greater flexibility depending on the nature of the investigation and the aims of the review. CPD, for example, is a 'third order' activity - the end beneficiaries, the pupils, are two steps removed from the intervention and its impact on teachers. This has implications for structuring the descriptions and interpretation of patterns in the data.

Our conclusions about the dialogue around the synthesis are that first of all, more structured guidance about the nature of the process would ensure that this important stage is handled consistently between reviews and would enable review groups to plan their efforts effectively and efficiently. The specification of a minimum quality standard would be helpful in making the nature of the EPPI contribution to the process clear and in enabling groups to make informed choices about where to put their efforts in these complex closing stages of a review.

4.6 The research report: writing for a teacher audience

The professional development literature and the literature about teacher engagement with research and evidence consistently highlights the importance of presentation (Desforges (2000), Guskey (1986), Huberman (1993), Mitchell (1999), Wikeley (1998), Cordingley & Bell.(2002),Of course, as Joyce and Showers (2002) point out, strategies for awareness raising and strategies aimed at behaviour change are two very different things, but studies of potentially useful practice need to be skillfully reported if they are to engage the attention and interest of a teacher audience. In particular, we know from the research and from our own experience of writing for a teacher audience that research is more likely to interest teachers if it:

In the face of this evidence, we found the EPPI Centre's insistence on adhering to a rigid reporting format difficult. We understood that the "Guidelines for Reporting" were just that, Guidelines. But EPPI view the report format itself as a key part of their quality assurance process. In constructing the report format as a QA rather than as a communication tool we were obliged to ignore our knowledge and experience of practitioner needs in relation to reporting. Practitioners want to know what the review was looking for and what it found out before exploring the evidence base. At the very least we wanted to highlight the findings in the summary of the review at the front of the report. However EPPI's emphasis on enabling 'read across' between health, education and other fields and between reviews led them to reject even these proposed variations at least in the report which was destined for the REEL database. Since this database is intended to be a source of reliable research information for practitioners it seems a pity not to use what we know about making research attractive and accessible to teachers.

Although (very short) summaries of the reports are written by 'users', including teachers, no summary can do more than act as a taster for the main report; a means of raising awareness about the research in such a way as to draw people to look for more detail. We would argue that these summaries, however well presented cannot act as a substitute for paying attention to user friendliness in the reports themselves. The summaries EPPI look for are only 3-4 pages long so cannot engage users with the material in any depth. The style and formats of the reports is a critical factor in sustaining practitioner interest in the review process and outcomes.

Our experience of preparing the final report included an enormous amount of detailed communications with EPPI. The team were anxious to help and made extensive efforts to do so although, perhaps inevitably, the "many hands" approach to resolving what we often saw as minor technical queries sometimes created more work than it saved. But what was a surprise was the extent of both support and editorial intervention from the EPPI team. Whilst we were able, through extensive discussion to resolve all disagreements this was a time consuming process. We believe that given EPPI's very hands on approach to supporting Review Groups, particularly in first reviews, their own efforts and contribution should be subject to the same tests of transparency as the work of the review groups themselves. A detailed account of EPPIs contribution would have the effect of increasing transparency, clarifying issues related to intellectual property and also informing future review groups about the detailed operational matters in order to improve planning and resource management.

Finally, we believe that research communication is a strategic enterprise. It requires a multi-faceted approach, one facet of which is making the best possible use of appropriate media. The Times Educational Supplement is (according to a 2001 teacher survey) read by the overwhelming majority of teachers. Sadly, our attempts to interest the TES in our EPPI-style report, met with little success. We were told that other EPPI Review reports had met the same fate. It appears that the process, format and style "drains the research of all its flavour" from a media perspective. More encouragingly however, the TES have expressed an interest in the issues raised in the review, following attendance at the launch seminar for practitioners, and have commissioned a 4 page feature on the related CPD issues for the Friday magazine.

5. Recommendations

Our experience and the evidence from the literature shows that there are differences between health and education research which accords with Hammersley's comments on the differences in the types of information different users want and the kind of language used. Throughout the review we have described here, we have learned a great deal. We believe the EPPI tools and methods to have the potential to act as a strong developmental tool for systematic reviewing in education, but only if the practitioner perspective is more fully acknowledged. Our experience (supported by the literature) indicates that reviewers in education need to pay particular attention to context and process data if their work is to be of interest to practitioners.

Our core recommendation in the light of this is that the tools and guidelines should be reviewed and that a more even balance should be struck between method and process, particularly at the data extraction and synthesis stage . Methodological data are important for establishing the reliability of the research but we question whether it requires the extensive recording of such data at the expense of more content and process-based data.

Other recommendations which we believe would help to resolve some of the challenges peculiar to educational reviewing are:

Core key words need to be revisited to allow finer grained education specific searching.

Specific guidelines for synthesis to match those provided for other stages should be developed to enable more efficient and effective planning.

More specific identification of EPPI's own contributions to the review process should be provided in generic terms. Since EPPI take a very active role in relation to review development and quality assurance there is also a case for a specific account of their contribution to specific reviews since, at present this is the only aspect of the process that is not completely transparent.

EPPI needs to set a minimum explicit quality standard for reviews. The current emphasis on the continuing pursuit of ever higher standards whist laudable makes somewhat open ended demands on resources.

Reporting format and style and media need to be given much more strategic attention in terms of their capacity to attract practitioner interest.

Bibliographic References

Britt MS, Irwin KC, Ritchie G (2001) Professional conversations and professional growth. Journal of Mathematics Teacher Education Netherlands 4: 29-53

Brown DF (1992) The Development of Strategic Classrooms in Two Secondary Schools. Unpublished research report; Wellington, New Zealand: Ministry of Education

Bryant DP, Linan-Thompson S, Ugel N, Hamff A (2001) The effects of professional development for middle schools general and special education teachers on implementation of reading strategies in inclusive content area classes. Learning Disability Quarterly 24: 251-264

Cordingley, Philippa; NTRP. Teachers perspectives on the accessibility and usability of research outputs. BERA Conference, 2000 [Online] Url: http://www.tta.gov.uk/assets/itt/providers/research/school/bera.doc 

Cordingley, P. & Bell, M. (2002) Literature and evidence search: teachers' use of research and evidence as they learn to teach and improve their teaching. London, TTA.

Da Costa JL (1993) A study of teacher collaboration in terms of teaching-learning performance. Paper given at the American Educational Research Association Annual Meeting April 1993, Atlanta

Day, C (1999) Developing Teachers: The challenges of lifelong learning. London: Falmer Press.

Desforges, Charles. How does experience affect theoretical knowledge for teaching? Learning and Instruction,1995, Vol 5 pp.385-400.

Desforges, Charles. Putting educational research to use through knowledge transformation. Keynote lecture to the Further Education Research Network Conference; Coventry. Learning and Skills Development Agency, December 2000.

Ertmer PA, Hruskocy C (1999) Impacts of a university-elementary school partnership designed to support technology integration. Educational Technology Research & Development 47: 81-96

Flecknoe M (2000) Can continuing development for teachers be shown to raise pupils' achievement? Journal of In Service Education 26: 437-458

Figgis, Jane; Zubrick, Ann; Butorac, Anne; Alderson, Anna. Backtracking practice and policies to research. In: The impact of educational research Australia: Commonwealth Department of Education Science and Training, 2001.

Gersten R, Morvant M, Brengelman S (1995) Close to the classroom is close to the bone: coaching as a means to translate research into classroom practice. Exceptional Children 62: 52-66)

Guskey, Thomas R. Staff development and the process of teacher change. Educational Researcher, May 1986, pp.5-12.

Hammersley, Martyn ., and Foster, Peter. (2003) A Review of Reviews: Structure and Function in Reviews of Educational Research. In Educational Research Policymaking and Practice.126-147 Sage publications

Hargreaves, D. H. (1993) A common-sense model of the professional development of teachers. In: Elliot, J. (ed) Reconstructing teacher education: teacher development. London: Falmer.

Harland J & Kinder K (1997) Teachers' continuing professional development: framing a model of outcomes. British Journal of In-Service Education, 23 pp.71-84.

Harvey S (1999) The impact of coaching in South African primary science INSET. International Journal of Educational Development 19: 191-205

Harwell SH, Gunter S, Montgomery S, Shelton C, West D (2001) Technology integration and the classroom learning environment: research for action. Learning Environments Research 4: 259-286

Huberman, Michael. Changing minds: the dissemination of research and its effects on practice and theory. In: C. Day; J. Calderhead & P. Demicolo (eds) Research on teaching thinking: understanding professional development. London; Falmer Press, 1993.

Joyce B. & Showers B. (1988) Student achievement through staff development. London: Longman.

Joyce, B. & Showers, B. (2002) Student achievement through staff development. In: Joyce, B. & Showers, B. Designing training and peer coaching: our needs for learning. VA, USA: ASCD

Kimmel H, Deek FP, Farrell ML, O'Shea M (1999) Meeting the needs of diverse student populations: comprehensive professional development in science, math, and technology. School Science and Mathematics 99: 241-249

Kirkwood M (2001) The contribution of curriculum development to teachers' professional development: a Scottish case study. Journal of Curriculum & Supervision 17: 5-28

Kohler FW, Ezell HK, Paluselli M (1999) Promoting changes in teachers' conduct of student pair activities: an examination of reciprocal peer coaching. The Journal of Special Needs 33: 154-165

Mitchell, Ian. Bridging the gulf between research and practice. In: Loughran, John (ed) Researching teaching: methodologies and practices for understanding pedagogy. Falmer Press, 1999.

Nutley, Sandra; Walter, Isobel and Davies, Huw. From knowing to doing: a framework for understanding the evidence into practice agenda.Discussion paper 1, RURU publications, University of St Andrews, March 2002 [Online] Url: www.st-and.ac.uk/~cppm/home.htm [18/06/03].

O'Sullivan MC (2001) Communicative approaches to teaching English in Namibia: the issue of transfer of western approaches to developing countries. International Journal of Early Years Education 9: 51-61

Parke HM, Coble CR (1997) Teachers designing curriculum as professional development: a model for transformational science teaching. Journal of Research in Science Teaching 34: 773-789

Ross JA, Rolheiser C, Hogaboam-Gray A (1999) effects of collaborative action research on the knowledge of five Canadian teacher-researchers. The Elementary School Journal 99: 255-275

Saxe GB, Gearhart M, Nasir NS (2001) Enhancing students' understanding of mathematics: a study of three contrasting approaches to professional support. Journal of Mathematics Teacher Education 4: 55-79

Wikeley, Felicity. Dissemination of research as a tool for school improvement. School Leadership and Management, 1998, 18(1) pp.59-73.

Wilkins CW (1997) Effects of a resident mentor teacher on student achievement in mathematics. Unpublished report. Mississippi, USA

Williams, D., & Coles, L.The Use of Research by Teachers: information literacy, access and attitudes Final Report 2003 ESRC Robert Gordon University.

APPENDIX 1

Continuing Professional Development Review Group

Review Protocol for the Initial Review, December 2001 to November 2002
1. Initial Review Research Question:
The question for the first review has been the subject of much discussion and refinement, in meetings of the Review Group; through consultation with teachers; through consultation with members of the Advisory Group and through informal contact with knowledgeable 'specialists' in the field of Continuing Professional Development. Practical considerations such as manageability within the time frame were considered alongside the likely availability of good quality studies and the likely interest and use of the search to a teacher audience. The Review Group acknowledges that this first review is in many ways a pilot - at least part of the goal is building skills (particularly teacher skills) in undertaking such reviews. Teacher involvement has been a core concern of the Review Group at every stage of the planning and design of the first review. Teachers will also be involved at all the stages of the review itself, from the initial search process to data extraction and synthesis.

The question we are proposing for the initial review is:

How does collaborative Continuing Professional Development (CPD) for teachers of the 5-16 age range affect teaching and learning?

i.e. In what ways does CPD involving working on a sustained basis with one or more professionally relevant colleagues for the purposes of meeting identified learning objectives affect teaching and learning?

Boundaries
The review will focus on studies involving more than one teacher, across the
5-16 age range, reported after 1988 and before 31st October 2001 in order to:

Definitions

Collaborative CPD includes teachers working together; teachers working with LEA or HEI or other professional colleagues. It does not include individual teachers working on their own.

By specifying CPD on a 'sustained basis' we are deliberately excluding one-off, one-day or short residential courses with no planned classroom activities as a follow up and/or no plans for building systematically upon existing practice. We are looking for studies where there is evidence about planned opportunities for teachers' learning prior to, during and/or after specific interventions to enable teachers to relate inputs to existing and future practice. We believe the continuing nature of professional development will be an important factor in creating evidence about impact.

2. The Review Process

The Review will proceed through a series of graduated filters. We have developed a series of explicit criteria for including or excluding studies at each stage. We will apply these criteria successively to (i) titles and abstracts and (ii) full reports. Titles and abstracts will be imported into a Biblioscape database (BD1). We will obtain full reports for those studies which appear to meet the criteria or where we have insufficient information to be sure.

These reports will be entered in to a second Biblioscape database (BD2). We will reapply the inclusion and exclusion criteria to the full reports and exclude those which do not meet these initial criteria. The remaining texts will be keyworded and our stage 2 criteria applied, by pairs of teachers and researchers, in order to identify potentially sound studies for data extraction. All the keyworded studies will be entered on a database as BD3. Keyworded studies will be added to the larger EPPI database, REEL, for others to access via the website.

The review will use the EPPI core-keywording framework, plus additional keywords which are specific to the context of the review. Thus, for example, where EPPI uses only one word (teacher) to identify the population focus of the study, our review group has consulted with potential users of the research and compiled a detailed and differentiated list which will help to identify the relevance of the study for particular user groups. This list includes Heads of Department, SENCOs, Heads of Year, Headteachers etc. The list is currently being piloted with a group of teacher-reviewers and will be attached to the published version of this protocol.

Studies identified as meeting both substantive and methodological concerns following the application of stage 2 criteria, will be analysed in depth, separately, by pairs of researchers and/or teachers using EPPI's detailed data extraction software, EPPI Reviewer. The data will then be synthesised and a final report prepared. The synthesis will bring together the studies which answer the review questions and which meet the quality criteria relating to appropriateness and methodology. The Review Group has decided not to map the literature field from the keyworded studies but to proceed directly to the data extraction, quality assurance and synthesis. We believe this to be the more useful activity, given the tight timescale for the review; and the deadlines for the 'deliverables' specified in the EPPI contract. We propose to produce summaries of the report for end users, tailored to the needs and concerns of particular interest groups. The final report and the summaries will be available on the EPPI website. However the Group is not confident about this as the sole means of dissemination and will be taking a number of other steps to bring the messages from the review to a wider audience.

There are a number of studies which will not meet the criteria but which are important commentaries or theoretical contributions to the field of knowledge about continuing professional development. An introductory or contextual section of the review will be prepared to report on key issues raised in this literature, together with the policy and professional developments which form the background for the review.

3. Methods for Finding Studies

Initial search procedures and resources will include:

i Electronic databases (including theses and grey material); these will include Ingenta, ERIC, ESRC (Regard), CERUK, BEI, Education-online and OCLS;

ii Key Journals in the field will also be hand searched. These will include:

Journal of Teacher Education
Teachers and Teaching: Theory and Practice
Teaching and Teacher Education
Teacher Development
Harvard Educational Review
Teacher College Record
British Journal of In-Service Education
Journal of Education for Teaching
Journal of In-Service Education
Professional Development Today
Teacher Development: An International Journal of Teachers' Professional Development
European Journal of In-Service Education

iii Websites

The group will also scrutinise websites which are likely to contain relevant material, such as ACER, SCRE, NFER, OFSTED, DfES, BERA, AERA, AAER selected LEA and university websites

iv Recommendations from Review and Advisory Group members, known specialists and overseas correspondents, practitioners and others;

v Following up citations contained in published and unpublished
research, and especially in research reviews and published literature searches.

4. Search Terms

The search terms for the initial search will include combinations and permutations of the following key terms, grouped to indicate (1) forms of CPD (2) age range 5-16 (3) outputs

practice

school

teachers

action research

primary school

teacher knowledge

enquiry

secondary school

teacher understanding

professional-

early years

learning

development

key stage 1

teacher attitudes

reflective-

key stage 2

pupil motivation

practice

key stage 3

teacher skills

peer coaching

high school

subject knowledge

evaluation

middle school

thinking

intervention

first school

cognition

teacher research

elementary school

teacher behaviours

In-service

pupil referral unit

pupil learning

education

special school

pupil self-esteem

collaborative

professional-learning

team teaching

5 Inclusion and Exclusion Criteria

We will apply the initial search criteria to titles and abstracts and then to full reports. First, we will identify studies which:

We will collect copies of full reports/studies/articles selected through the
initial search process, and re-apply the initial criteria. We will then apply the following, stage 2 criteria to identify potentially sound studies for in-depth review. Studies will have to meet all the criteria below to be selected for in-depth review:

The Review Group's current knowledge of the literature leads us to expect that there will be relatively few studies of CPD which go beyond the teacher to examine the impact on pupil learning. Deciding whether to exclude studies which do not attempt to measure impact on pupils in some way has been problematic. We have decided to include only those studies for in-depth review which set out to ascertain the impact on learning at the first stage. If this excludes too many studies, we will consider including studies which assess impact on teaching but not upon learning, where there is other reliable evidence about the effectiveness of the interventions involved.

Decisions to exclude or include studies will be sampled for consistency regularly, internally. The process as a whole will be subject to formative and summative EPPI quality assurance processes.

Types of Study

The EPPI categories of studies will be used to identify study type. At the first stage of reviewing no study type will be excluded. The EPPI categories are:

Because of the way we have framed the research question and the stage 2 criteria for this review the study most likely to meet substantive and methodological criteria are:

Studies of methods and reviews (systematic or otherwise) may also be particularly relevant at this stage.

Data

Pupil assessment data in the studies reviewed is likely to have been collected via, for example:

  • directly administered assessment tools;
  • analyses of pupil work;
  • observations of pupil activities;
  • cumulative , recorded assessment of pupils' work over time;
  • peer or pupil perception data;

Teacher assessment data will be collected via, for example:

  • systematic observation or recording of teaching practice;
  • out-of-class assessments of teachers skills and knowledge and
  • triangulated teacher self assessment data;
  • repertory grids to explore beliefs and understandings;
  • systematic interviews;
  • peer or pupil perception data;

Data about interventions that we are targeting include:

  • Inputs (data about teachers' knowledge, beliefs, attitudes and understanding; teacher learning needs; skills and knowledge of facilitators/providers)
  • Processes (observation of intervention processes)
  • Design (plans, implementation, strategies)
  • Outcomes ( pupil evaluation or assessment of knowledge, skills, performance or attitudes; pre and post assessments of changes in teacher skills/attributes/behaviours/knowledge)

We shall target, at least in the first stage, studies that collect data about teachers' learning prior to and on completion of interventions. If this excludes too many studies we shall consider at the second stage including studies that make only post hoc assessments of impact of teachers' learning.

Members of the Review Group

Janet Sturgis NUT Chair and Co-ordinator
Hazel Hagger University of Oxford
Philippa Cordingley Research and CPD Consultant to NUT
Janet Friedlander NUT Information Officer
Miranda Bell CUREE Assistant Director
John Bangs NUT Education and Equal Opportunities Secretary
Colin Biott European Education Research Assoc (Unconfirmed)
Lesley Saunders General Teaching Council Research Officer
Member National Teacher Research Panel
Secretariat Richard Stainton and Pamela Collins, NUT Headquarters

Members of the Advisory Group:

Anne Edwards University of Birmingham
Michael Eraut University of Sussex

David Jackson Director for Research, National College of School Leadership

Richard Harrison DfES CPD Lead Officer
Campbell Russell Teacher
Ray Waterhouse Teacher
Jeremy Kraus Cheshire LEA
Chris Day University of Nottingham

APPENDIX 2

Evidence for Policy and Practice

Information and Co-ordinating Centre

The EPPI-Centre is part of the Social Science Research Unit, Institute of Education, University of London

Structure for a review report

PROTOCOL

The methodology for undertaking the review which forms the basis of several parts of the review report as colour coded in the report structure below

Cover sheet: Title, name of review authors/team/organization/advisory group

Background: Authors and potential conflicts of interests

Aims of the review and review question

Review methods

User involvement
Identifying and describing studies
Identifying and describing studies quality assurance
In-depth review including assessing quality of studies
In-depth review quality assurance

References

-----------------------------------------------------------------------------------------------------------------

REPORT

TITLE PAGE

PREFACE PAGES

AUTHORS
REVIEW TEAM MEMBERSHIP
ADVISORY GROUP MEMBERSHIP

ACKNOWLEDGEMENTS

The title/authorship pages need to include institutional bases for the review, date submitted, acknowledgements, including funding sources (which should include acknowledgement of any HEFCE funded time that was contributed though not directly funded). Declaration of any potential conflicts of interests of authors, Review Team members and Advisory Group members.

LIST OF ABBREVIATIONS

SUMMARY

Background
Aims
Review questions
Methods
Results: Identification of studies/ Systematic map/ In-depth review and synthesis
Conclusions: Strengths, Limitations, Implications for policy, practice and research

1. BACKGROUND

OUTLINE OF CHAPTER

Section 1.1 introduces the basic principles that are discussed in more detail in the rest of the chapter

1.1 Aims and rationale for current review

Brief introduction to the theoretical, policy, practice and research background to the review, and the rationale for the methods used both in the identification and broad characterisation of the studies overall (mapping) and any in-depth review stages. This should be taken from the publicly available peer-refereed review protocol and indicate any key departures.

Sub-headings if any will be review specific

1.2 Definitional and conceptual issues
Theoretical background. Sub-headings if any will be review specific

1.3 Policy and practice background
Sub-headings if any will be review specific

1.4 Research background
This may include some discussion of previous (systematic and non-systematic) literature reviews Sub-headings if any will be review specific

1.5 Authors, funders, and other users of the review

The reasons why the review is being done by these people, funded in this way, done at this time, for different audiences (users of the review). This should include backgrounds of authors, Review Group Members and Advisory Group members

1.6 Review questions
Sub-headings if any will be review specific

2. METHODS USED IN THE REVIEW

All the details necessary to allow replication
Outline of Chapter

2.1 User involvement
2.1.1 Approach and rationale
2.1.2 Methods used

Which users involved in what way for what part of the review process.

2.2 Identifying and describing studies
2.2.1 Defining relevant studies: Inclusion and exclusion criteria
2.2.2 Identification of potential studies: Search strategy
2.2.3 Screening studies: applying inclusion and exclusion criteria
2.2.4 Characterising included studies (EPPI-Centre and review-specific keywording)
2.2.5 Identifying and describing studies: quality assurance process

2.3 In depth review

2.3.1 Moving from broad characterisation (mapping) to in-depth review
2.3.2 Detailed description of studies in the in-depth review: (EPPI-Centre and review-specific data extraction)
2.3.3 Assessing quality of studies and weight of evidence for the review question
2.3.4 Synthesis of evidence
2.3.5 In-depth review: quality assurance process

3. IDENTIFYING AND DESCRIBING STUDIES: RESULTS

Outline of Chapter

3.1 Studies included from searching and screening
Figure 3.1 Filtering of papers from searching to map to synthesis
Figure similar to the attached

3.2 Characteristics of the included studies (Systematic Map)

Sub-headings if any will be review specific, dependent on the purposes of creating a map of the research. The coding classification in the keywording provides a means for developing the map. Textual details can be kept relatively brief by referring to Appendix 3.1. (Figures and tables can be produced directly from keywords in databases such as EPPI-Reviewer.)

3.3 Identifying and describing studies: quality assurance results

4. IN-DEPTH REVIEW: RESULTS

Outline of Chapter

4.1 Selecting studies for the in-depth review
(IF APPLICABLE)

4.2 Comparing the studies selected for in-depth review with the total studies in Systematic Map
(IF APPLICABLE)
Sub-headings if any will be review specific

4.1 or 4.3 Further details of studies included in the in-depth review
Sub-headings if any will be review specific, dependent upon the in-depth review question. Textual details can be kept relatively brief by referring to Appendix 4.1. (the coding classification in the keywording and data extraction provides a means for describing the studies Figures and tables can be produced directly from EPPI-Reviewer).

4.2 or 4.4 Synthesis of evidence
Methods for and extent of synthesis and thus chapter sub-headings will be review specific. This is likely to include differing degrees of: I) a priori conceptual distinctions specified in the conceptual framework and protocol; (ii) new conceptual distinctions arising from the primary research studies considered. When synthesis includes statistical meta-analysis the balance is usually toward a priori conceptual distinctions.

4.3 or 4. 5 In-depth review: quality assurance results
4.4 or 4.6 Nature of actual involvement of users in the review and its impact

This may involve judgement by the authors

5. FINDINGS AND IMPLICATIONS

Outline of Chapter

5.1 Summary of principal findings
5.1.1 Identification of studies
5.1.2 Mapping of all included studies
5.1.3 Nature of studies selected for in-depth review
5.1.4 Synthesis of findings from studies in in-depth review
Should refer to Background, especially to previous research/reviews of research including any systematic reviews

5.2 Strengths and limitations of this systematic review
Sub-headings if any will be review specific but can include comparison to other reviews on this topic

5.3 Implications
5.3.1 Policy
Including involvement of non policy maker users in these processes
5.3.2 Practice
Including involvement of non practitioner users in these processes
5.3.3 Research
Including involvement of non researcher users in these processes

6. REFERENCES
6.1 Studies included in map and synthesis
Studies selected for in-depth review could be marked with asterisks
6.2 Other references used in the text of the report

APPENDICES

Appendices should be numbered by chapter. The numbering is only indicative but the appendices should normally include:

APPENDIX 1.1 Advisory Group membership

APPENDIX 2.1: Inclusion and exclusion criteria

APPENDIX 2.2: Search strategy for electronic databases

APPENDIX 2.3: Journals hand searched

APPENDIX 2.4: EPPI Keyword sheet including review specific keywords

APPENDIX 3.1: Details of studies included in the systematic map.

Nature and extent of detail will be review specific but the detail should be in the appendix rather than in the text of the report. Using databases such as EPPI-Reviewer allows simple production of tables.

APPENDIX 4.1: Details of studies included in in-depth review

Sub-headings if any will be review specific. Nature and extent of detail will be review specific but the detail should be in the appendix rather than in the text of the report. Ensuring standardized styles for text entries into EPPI-Reviewer allows automatic production of structured abstracts on each study

APPENDIX 4.2 etc: Syntheses tables

Figure 3.1: Filtering of papers from searching to map to synthesis

Key for Figure 3.1 (see above):

*Criteria for exclusion are not mutually exclusive. But if applied sequentially the data will be mutually exclusive which can be informative. For example, if Criterion 1 is 'Being on topic', then papers excluded on basis of Criterion 2 must be on topic (as not excluded on those grounds). This process may not be fully accurate when screening on title and abstract only.

** It is preferable for criteria for exclusion to be recorded but requires resources. If quickly excluding papers as soon as any criterion is breached may mean that not all criteria are considered. It is also labour intensive to record the full references of all papers so quickly rejected. Compromises include: (i) to only record the full reference for papers nearly meeting the inclusion criteria that may be of interest to further reviews in the topic area. This should be done systematically according to specific inclusion/exclusion criteria; (ii) to only record the number of papers excluded for different reasons from each source such as an issue of a journal or a section of a web site

*** In some cases, one paper will describe more than one study. In this case, the numbers of studies and papers will seem inconsistent, so a note should be added to this effect.

This document was added to the Education-line database on 03 October 2003