EXECUTIVE SUMMARY

1 Introduction.
2. A Brief Outline of the Case Studies
3 General Points About the Growth and Development of Teachers' Centres.
4 General Issues Relating to the Teachers' Centre
5 Observations made at the Teachers' Centres
6 Observations made in schools
7 Conclusions and Recommendations

1 Introduction.

1.1 Between 1997 and 1998, a team from the University of Leeds School of Education made a study of teachers resource centres for the Department for International Development (contract No. - CNTR 95 5578A). This paper provides a brief summary of the detailed report of this study. The major aim of the study was to attempt to assess the effectiveness of teacher resource centres (TRCs) as a strategy in attempting to improve the quality of education in schools in developing countries.

The study was in part a summative evaluation. The basic question that drove this aspect of the study was: 'To what extent do TRCs help to improve the environment for learning in schools and the quality of teaching and learning in classrooms?'

The study was also intended to be a formative evaluation in that it sought to shed light on what might be done with existing TRCs to make them more effective. The basic question that drove this part of the study was: 'What are the issues surrounding TRCs; how are TRCs affected by these issues; and how do they react to them?¹

1.2 The research methods employed in the study were a literature review followed by case studies in four countries: Kenya and Zambia in Africa, Andhra Pradesh, India and Nepal in Asia. Each of the projects in the case studies has on-going, British assisted educational development programmes, which include teacher resource centres of different forms and functions. Brief details of each of the projects and the roles of the TRCs within the projects are given in Section 2 below.

The team from the University of Leeds spent two, 2 week periods in their respective countries with which they were very familiar through previous and, in some cases, extended periods of work. The visits were separated by 3-5 months. The team were greatly supported by host country colleagues who accepted roles as in-country counterparts. The counterparts added greatly to the work by putting the research into context because they were part of the education system and could gain access to relevant people within their respective systems. They also engaged in an on-going process of observation beyond the time in-country of their UK colleagues.

In each country, on the basis of a number of agreed focuses, a wide range of interviews were conducted and observations were made at TRCs and in schools. In particular, by applying a tracer technique, the study attempted to establish if the ideas and materials available to teachers through the activities and services of the TRCs were actually being implemented in schools and classrooms. However, because of the nature of the research, the methodologies and procedures adopted in each country had to be determined to some extent by the opportunities afforded to team members and were adapted accordingly along agreed lines to allow the research to maximise data collection while allowing comparisons wheer appropriate.².

1.3 In the summary that follows the development of TRCs is put into context before mention is made of some of the general issues that were identified in relation to the use of TRCs as a strategy for teacher development. In Sections 5 and 6 we report and comment on some of the findings and observations made in the TRCs and in the schools. Finally we consider ways in which the TRCs studied were effective in that aim and we suggest a number of conditions which we feel would allow TRCs to function more effectively.

2. A Brief Outline of the Case Studies

Country	Project	Previous Project/s	Funding	Role of the Teachers' Centre in the Project.
India - Andrah Pradesh	District Primary Education Project (DPEP)	The Andhra Pradesh Primary Education Project (APPEP)	DflD	In-service education courses are held elsewhere. The teachers' centre is a meeting place, usually a classroom in a school, for a cluster of 7-13 schools. Teachers who have attended the central training courses attend 6 mandatory meetings each year.
Nepal	Basic and Primary Education Project (BPEP)		ADB	There are around 700 Resource Centres (RCs) serving over 9, 000 primary schools. Each centre serves a cluster of 10-25 schools. These are multipurpose centres for: training, materials development, teachers' library, exam centre, parents' meetings and a community hall. A Resource Person (RP) looks after 2 RCs. Their main work is school support.
Nepal	Secondary Education Development Project (SEDP)	Science Education Development Project (SEDP)	DflD in a consortium of funders including ADB and World Bank	There are 25 Secondary Education Development Units (SEDUs) each serving 1-4 districts. These play a major part in in-service teacher training programme for secondary teachers. SEDU Master Teachers (SMTs) are principally training course managers, they are also trainers and they make follow-up visits to schools.
Zambia	Action to Improve English Maths and Science (AEIMS)	Self Help Action Plan for Education (SHAPE)	DfID	Each of the 16 provincial Resource Centres is run by 3 subject coordinators. The district centres have a co-ordinator who is an experienced primary teacher. Each district has three subject trainers. The centres have a role in a cascade system of in-service training and as resource centres.
Kenya	Teachers' Resource Centres TRCs	English Language Programme (ODA)	Gov. of the Republic of Kenya & a levy on parents	TRCs are centres for English teachers in secondary schools. They have not had donor funding since 1992. The main role of the 25 TRCs is to loan out class sets of English readers and reference books for teachers. There is also a small in-service programme based on local needs analysis.
Kenya	Strengthening Primary Education Programme SPRED II	Strengthening Primary Education Programme SPRED I	Gov. of the Republic of Kenya & DflD	System of Teacher Advisory Centres (TACs) developed in 1978 for primary teachers. TACs are staffed by TAC tutors with offices usually in a primary school. District TACs carry reference and other materials but few Zonal TACS have any resources. TAC tutors are full-time advisers who visit teachers in then-group of 10-15 schools, work with the zonal inspector and facilitate workshops. Workshops are either zonal or school based. Zonal workshops are held in convenient school in the Zone. After each workshop the TAC tutor provided follow-up visits to the teachers.
Kenya	School Improvement Programme (SIP) Mombasa	School Improvement Programme (SIP) Kisumu	Aga Kahn Found-ation Education Service (AKES). Mombasa Municipality &DfID	SIP TACS have a TAC tutor and a Programme Officer (PO). Each TAC services about 12 primary schools within 3-5 kms of the TAC. The TAC facilities - safe storage for materials and a meeting room for workshops - must be provided by the community. Resources are provided by the project. POs spend three days each week working intensively in 3-4 schools. A different group of schools is targeted each year. On these days the TAC tutor visits the other schools in the cluster. The POs meet together each week to plan and prepare in-service workshops, they also work with the tutor at their TAC to identify needs and prepare for workshops.

3 General Points About the Growth and Development of Teachers' Centres.

Teachers' centres began in Britain in the 1960s. Over time, they came to be regarded as a very successful way of supporting professional development for teachers and of providing access to a range of educational resources. However, the effectiveness of the teachers' centre as a strategy, in terms of its impact on schools, was never properly assessed in Britain. Towards the end of the 1980s, with the advent of a National Curriculum and the need for rapid and efficient retraining of all teachers, effectiveness became much more relevant and teachers' centres began to decline. More efficient and effective strategies were developed which did not see individual professional development as a priority, but instead focused training more the school development plans and the groups of teachers implementing those plans.

The teachers' centre as a model has influenced teacher development in other countries around the world. In the early 1970s, teachers' centres were enthusiastically sold abroad as an effective way of responding to teachers' needs and of ensuring professional growth (Kahn 1982, Gough 1989). However, some doubt is expressed in the literature about whether a model, developed in Britain in a very ad hoc way in response to particular demands, could be used as a strategy for improving education elsewhere (Gough 1989, Hawes 1977, Hoppers 1996).

There are a number of underlying ideas commonly identified with teachers' centres, the most enduring one being that they should address teachers' needs. However, what teachers' centres actually become and what they can achieve seems to depend very much on the context in which they are used. The variety of use we encountered in the literature and in the projects in our study, illustrates that the term 'teachers' centre' certainly doesn't represent one model. More importantly perhaps, for a wide range of contextual and ideological reasons, the model developed and hailed as successful in Britain is generally not the one that has survived the influence of donor aided projects in the developing world.

The literature concerning teachers' centres and the project histories in our study shows that the teachers' centres in developing countries have changed over time. In many countries they began in a small way to address local needs. More recently, with the moves to dramatically increase education provision and improve the quality of education, successive donor aided projects have taken over the running of teachers' centres. These projects have rejuvenated the centres and often increased their number, and then they have used them for their own purposes. Under each project the role of the teachers' centre and the contribution it has made to teacher development and to improved practice in schools has changed. All of the teachers' centres in our case studies had been taken over by a recent project or were part of a new phase of an established programme.

4 General Issues Relating to the Teachers' Centre

4.1 Function.

A disparity appears to exist between what teachers' centres are felt to do best and what teachers actually get out of them. This was true in Britain where, on one hand teachers' centre wardens and bodies such as the Schools Council emphasised the value of teachers engaging in local level curriculum development at the teachers' centre, while on the other hand teachers were more interested in attending practical courses and having access to a wide range of resources. Reviews of centres in Britain show that there was actually very little systematic curriculum development and that only small groups of teachers and head teachers engaged in such activities. It was concluded that material production and curriculum development were very difficult tasks for teachers to achieve in limited sessions at a teachers' centre (Weindling et al. 1983).

The underlying idea that the teachers' centre would encourage professional interaction and a certain level of curriculum development and materials production, which in turn would result in improvement in the quality of teaching was present in all our case studies. However the extent to which this ideal is realised varies considerably. In Andhra Pradesh and Zambia, teachers were supposed to reflect on practice, exchange experiences and develop materials at the teachers' centre. However, in both projects the centrally determined programme failed to support these ideals. In Zambia, the only strategy that appeared to be giving teachers the opportunity to think about teaching was the newly established, school based Teacher Workshop programme. The tightly structured meetings at the centres in Andhra Pradesh allowed few opportunities for teachers to consider problems and exchange ideas. On top of this lack of formative opportunity, there was a scarcity of the skilled personnel needed to facilitate a more critical level of reflection.

The emphasis at the teachers' centres in Nepal was on training and dissemination with little provision for individual or school needs. Only in Kenya did we find that teachers' centres were providing on-going support for what the teachers were doing. Workshops focused on exams and syllabuses and in the secondary TRCs provided essential resources although only to those teachers and schools within easy reach of the centre.

With the exception of the SIP project in Kenya, there was considerable contrast between the British centres and the centres in our case studies in terms of resources and personnel to run the service. If teachers in Britain with the existing levels of resource availability found it difficult to produce materials and develop the curriculum for the local situation then it is entirely predictable that teachers in the developing world must find such activities many times more difficult to accomplish successfully.

4.2 Issues Relating to Training.

One of the main role for the teachers' centres in the projects of our study related to the provision of in-service courses or workshops for teachers. In fact it was impossible to look at what was happening at the teachers' centres without considering some of the wider issues relating to in-service training in each of our projects. Within the constraints of our study it was not possible, and never intended, to provide an in-depth study of in-service provision. However, we identified a number of issues relating to in-service training which influenced the effectiveness of the teachers' centres in the four countries and their projects we looked at.

4.2.1 Approach to training.

The most effective approach to teacher training is much debated in the literature on this subject. Influential factors for choice of approach are identified as the teachers and their experience and knowledge, the context in which they teach, the demands of syllabuses, exams and the pressures of innovation.

The teachers in Britain were relatively well trained, very well resourced and, possibly most importantly, had a high level of autonomy in terms of curriculum and classroom practice. Also the courses they attended at the teachers' centre were not proposing that they should make any radical changes in their practice. Given the very different conditions in the developing world, there is discussion in the literature about whether the teachers' centre is necessarily the most appropriate place for in-service training.

The courses provided by the centres in Britain were not attempting to disseminate or interpret centrally directed policies. They were mainly intended to support and extend what teachers were already doing. In contrast to this, the in-service courses in Andhra Pradesh, Nepal and Zambia, were tied to pedagogical philosophies, principles and methods predetermined by the projects. Courses were planned and prepared centrally by high ranking national personnel and overseas consultants.

The aim of the courses of the projects in these three countries was to change what teachers did in the classroom and thereby to improve the quality of teaching and learning in schools. One of the main points that reoccurs through the discussions of observations and findings in our study is that the impact of these centrally developed programmes of INSET was, in fact, very limited. It was felt that the lack of impact was due partly to a perception that the ideas being transmitted were too abstract or too general so that it was not easy for teachers to apply them in the classroom.

4.2.2 The cascade as a system of training.

When we considered the approach each project had taken towards training, the cascade model adopted by two of the projects, in Andhra Pradesh and in Zambia, gave particular cause for concern. One of the problems identified with this approach was that, to be effective, the cascade needed to be regularly fed with well considered, high quality input. This was very demanding of the personnel responsible for preparing the courses because they rarely had the necessary combination of imagination, experience in the classroom and experience of the limitations of the cascade system.

A further problem resulted from the way that the training systems in these two projects attempted to transmit the same content to all teachers with minimal transmission loss. To achieve this they used the same course for each level in the cascade. However, it was found that at each level the training was inadequate for the needs of the participants at that particular level. As a result, in Andhra Pradesh the tutors at the teachers' meetings, who needed to make appropriate interventions in response to model lessons, didn't receive any training to help them to support the teachers. And in Zambia, while the course prepared the higher levels for training, it did not have obvious links with the training for the teachers and it did not have real practical application for the classroom.

Another problem was that the conditions at each level of the cascade limited what was possible at every other level. This meant for example that although video facilities were available centrally, which would probably have helped the trainers at the top of the cascade to consider practical applications in the classroom, videos could not be used in the cascade because facilities were not available at the level at which the teachers were trained. Similarly, though it would have been useful to demonstrate or try out methods with groups of children at the bottom of the cascade, the lack of experience and expertise at the top made this impossible. The emphasis in the courses was therefore on underlying theory rather than practical and relevant application. One positive outcome of this system of training, which was also noted in a study of a similar system in England (Kinder and Harland 1991), was that the trainers benefited in terms of increased skills and knowledge, though impact on teachers was limited.

4.2.3 The relevance of in-service courses.

Research into methods of achieving change in teacher behaviour suggests that teachers are more likely to adopt practices which have similarities with what they are already doing and which are easily applied in their classrooms. Beeby (1986), for instance, recommends that it is much better to help teachers to do what they do already more effectively than to try to get them to switch to a radically different concept of education. The World Bank Review in 1995 found that input from a teachers' centre was not used unless the inputs fitted local conditions and unless teachers were confident about how to use them.

The problem of lack of relevance of in-service courses and teachers' centre workshops was noted in three of our projects. In the teachers' centre meetings in Andhra Pradesh the teachers were asked to plan and demonstrate one-off, model lessons which made use of the six educational principles being targeted by the project. However these lessons did not form part of the teachers on-going planning and teaching at the school. The lack of sequential planning and the lack of relevance to what teachers normally did seemed to ensure that this type of planning did not go back to the school with the teachers. In Zambia the abstract approach to ideas and skills during training sessions succeeded in raising awareness, but it did not appear adequately to develop skills for the classroom, nor did it help teachers to engage in critical reflection about their teaching. There seemed to be a lack of valid inputs leading to clear and effective changes in the classroom.

In contrast to this, in Kenya the limited in-service courses at the TRCs were at least mainly trying to help teachers to work with the English texts in the exam syllabus.

4.2.4 Stages of teacher development.

One often repeated claim in the literature is that a staged approach to INSET is most effective, especially where change in practice is required. In particular the work of Joyce and Showers (1980: 379) is much quoted. Joyce and Showers found that after training teachers quickly reverted to their usual tried and tested routines rather than persisting with new behaviours that they felt awkward with. They suggest that a 'coaching' stage in the training process helps to get teachers past the feeling of being 'temporary novices'. The implication of their findings is that following training sessions trainers need to support teachers in the classroom.

Some form of in-school support was intended in the projects studied in each of our case studies. However, in practice this strategy not only worked in quite different ways in each of the projects, but it also had considerably different levels of impact. In Nepal the trainers from the teachers' centres were supposed to visit the trainees who had attended their courses, but they found this difficult to achieve because of the number of schools, the distances involved and the lack of transport.

The newest phases of the DPEP project in Andhra Pradesh intend to have a level of personnel who go into schools to support the training, but the vast numbers of schools involved in the project will make real impact from this strategy very difficult to achieve. The training in Zambia was in reality becoming increasingly school rather than teachers' centre based. However, this involved teachers meeting in groups in the schools rather than receiving support in the classroom. In the primary SPRED II project in Kenya, teachers' centre tutors did visit schools after training workshops, but because of the newness of this phase of the project the effectiveness of these occasional visits was difficult to assess.

Only in the Kenya SIP project was the support in schools a major ingredient of the training. This represented a response by the donors to the feeling that teachers' centre workshops and the sporadic support in school from TAC tutors was not achieving the level and pace of improvement in the quality of education that they had hoped for. By putting in another level of trained, experienced personnel to target schools in an on-going way, the donors hoped to have a greater and more lasting effect on what happened in the classroom.

4.3 Provision of Resources

Resources were not a major concern of the teacher's centres in our projects. The TRCs in Kenya were the main exception to this. Their main function was to provide reference books for teachers and to lend out class sets of the books required for the English syllabus.

The level and type of resources available at the centres in the other projects varied enormously. However, the main idea for most of the teachers' centres was that they should attempt to increase the scant resources available for teaching and learning in schools by helping teachers to make their own materials. It was noted that in Zambia, where the facilities for producing materials were very much under used, there was no culture of teachers making learning materials. This was understandable considering the size of the teaching groups and the cost in terms of time and money.

5 Observations made at the Teachers' Centres

The teachers' centres in each of our case studies had their own identity, range and type of activities and their own level of effectiveness. A comparison between the British model and the various teachers' centres we observed clearly demonstrates how difficult it is to define this intervention. The funding, the roles of the staff, the facilities available and the activities and resources associated with the teachers' centres show a bewildering level of variation.

5.1 Funding

The teachers' centres in Britain received regular and reliable funding from the local education authority. In each of our case studies, the funding and consequently the direction of the teachers' centres had changed in the last ten years. In Nepal, it was felt that the changes of priority and direction of successive projects, and the state of flux in the education system generally, needed to be taken into account when trying to critically analyse the work of the teachers' centres.

Where donor funding had been discontinued as a project ended, even where this had only been for a short time, the functioning of the teachers' centres had been affected. In Kenya, donor funding for the TRCs had stopped in 1992, and the teachers' centres were found to be functioning at a fairly low level of activity. The TRC tutor was still paid by the government, but the levy from headteachers was not always forthcoming and only about 60% of eligible schools participated in the programme. Where financial support was lacking or only half hearted the tutor was found to be struggling to maintain a useful level of service. In Andhra Pradesh, in the last year of the previous project, there was substantial evidence of meetings lapsing as a result of breaks in continuity of funding from the government. While the amounts of money involved were very small they did seem to be vital to the functioning of the teachers' meetings.

In contrast to this, when the SIP programme in Kenya funded and trained extra expert staff at the Teacher Advisory Centres (TACs) in order to achieve intensive school input, the TAC workshops had a noticeable positive impact on effectiveness in the schools being targeted by the project.

The projects in our case studies were employing a number of different strategies to make teachers' centres more self sufficient. However, adequate and reliable funding seemed to be necessary for the teachers' centre to really have an impact on teaching. When financial support from donors ended, we felt that the teachers' centres would inevitably change their focus and level of activities. The difficulties being faced by the centres remaining from the Kisumu phase of the SIP project in Kenya, described in detail in the main study, suggest that it would certainly be useful for any project to anticipate and possibly help teachers' centres to prepare for eventual changes of function.

5.2 Staffing

The role of the teachers' centre warden in Britain was identified as vital by most commentators. Wardens had status and a fair level of autonomy. Their success was partly attributed to their flexibility and to the way they were quick to respond to new demands and challenges. They had responsibility for the co-ordination of INSET but the LEA advisory staff mainly provided the actual courses. The centres they managed provided stimulating environments for teachers and were generally found to expand and adapt impressively under their guidance.

In the literature about teachers' centres abroad, the role of manager or co-ordinator of the teachers' centre was also seen as important to effectiveness. However, this role rarely afforded status, often being allocated to a teacher seconded from a local school. The person running the teachers' centre usually had very limited autonomy. Besides being managers and co-ordinators, they were often also expected to be the key providers of in-service courses usually without specialised training on how to do so or where to obtain experts to contribute to programmes.

The role and effectiveness of the teachers' centre managers in our case studies varied considerably. In the two projects with cascade training, in Andhra Pradesh and Zambia, the teachers' centre co-ordinators were primarily facilitators and trainers in the cascade. However, in Andhra Pradesh, the task of facilitating the meetings at the teachers' centre demanded much more of the Resource Person than the cascade training programme had provided. As well as supporting the initial training in the six APPEP principles, they needed to interpret these principles in the light of local circumstances and the skill of participating teachers. The lack of skilled intervention at the teachers' centre meetings seemed to result in little progress for the teachers in terms of the development of professional skills. As training became increasingly school based in Zambia, the teachers' centre was losing its role and purpose. The co-ordinators at the teachers' centres needed to initiate and adapt in response to this change, but so far there was not much sign of this happening.

The TRC tutors in Kenya were senior teachers with a number of other obligations. The basic resource management at the TRCs was very well executed and the tutors were found to open the TRC more than was required of them by the committees. However, their achievements were limited because of their other duties and the lack of time. In the SPRED II TAC system the TAC tutors worked at school level more than at centre level. They provided a limited INSET programme and had few resources to manage. In the SIP TACs the TAC tutors also worked mainly as advisers in schools. They had the added support of the professional officers (POs), who planned workshops in a co-ordinated programme and helped them to deliver INSET.

The Nepal Resource Person seemed to have a position more like the wardens in Britain, but they also had responsibility for training and a school advisory role. The centres in Nepal were found to be well run and to provide stimulating environments for the teachers, but the position of resource person did seem to be very demanding.

Providing adequate and appropriate training for teachers' centre wardens was identified as a problem in Britain (Weindling et al 1983). The wardens attended a range of courses, but generally found that meeting together to discuss problems and ideas most effectively supported their developments at the teachers' centre. Most of the co-ordinators and tutors in our study required training to develop managerial, training and mentoring skills. This training seemed to be particularly successful in Kenya where the SIP project not only encourages professional development for its officers, in the form of in-country post graduate award-bearing courses from institutions of Higher Education, but also requires the Programme Officers to meet together regularly with the project Director to share ideas and plan in an on-going way. These meetings appeared to contribute to the enthusiasm and commitment of the POs and they helped keep the workshop programme tuned to the needs of the teachers and children.

The essential role of teachers' centre co-ordinator or manager seems to have been an aspect of the model in Britain which has been difficult to replicate in other countries, particularly those with more rigid hierarchies. In many countries the personnel trying to run teachers' centres appear to lack the necessary level of status, training and the freedom and confidence to take initiative.

5.3 Facilities.

The facilities provided for teachers' centres also varied greatly. The centres in Nepal were purpose-built and had a number of rooms and functions. The stark contrast between these centres and the poor school facilities in Nepal was possibly contributing to lack of impact. In contrast to these centres, in Andhra Pradesh the meetings were mainly held in very basic facilities, usually in the best and biggest room in a school, but the teachers appeared to make the best of what was available. In Zambia the facilities were good, but under used because they were not always appropriate for the training programme. In Kenya the meetings were held in schools, in the case of the SIP TACs, a classroom improved for this purpose by the local community.

In the SIP programme, the level of community ownership in the TAC facilities appeared to be contributing to the success of the project. We found evidence of increased local interest in the work of the centres and the schools. The project hoped that the financial commitment to the TACs would help to sustain the work of the centres once the project funding ended.

The number of teachers' centre facilities in the primary projects leads us to wonder whether the respective governments would be able or willing to take over the funding required for meetings and centre maintenance once an externally funded project finished.

In each country, the apparent reliance on successive donor funded projects to keep the centres functioning lent support these doubts.

5.4 In-service Education.

It was impossible to fully analyse and evaluate the training provided by the centres during the short visits of our study, and we cannot claim to have done this. We can only comment on the activities that we observed. However, in most of the projects there seemed to be aspects of the training which affected, either aided or interfered with, impact on schools.

5.4.1 Focus of workshops and training courses.

The focus of the training was thought to influence the level of impact. The fact that little of the training observed related directly to the text books used in the schools was felt to be particularly relevant. The text book was the main resource that teachers had in the classroom. Nevertheless, in the cascade training in both Andhra Pradesh and in Zambia it was noted that the focus on theory and principles in the training largely ignored the need to consider application to text books and their effective use in promoting learning. The teacher was left to interpret and reconstruct for themselves, and more often than not they did not seem to do this.

An example of teachers' centres helping teachers to work with a new text book and teachers' guide was found in the RCs in Nepal. The centres distributed the books and provided a 3-4 day 'orientation' course which was very popular with the teachers. However, other courses in Nepal paid little attention to the text books. For example, the SEDUs provided a course to help secondary teachers become familiar with the new science equipment. After the course a number of teachers complained that it had not helped them to see how the equipment could be incorporated into the syllabus, how it related to activities in the textbooks or how it could be most effectively used in the classroom.

The teachers in all our case studies worked to an exam syllabus or a national curriculum which focused on content. It is therefore not really surprising that many teachers did not make the effort required to adopt new pedagogical approaches, particularly as the in-service training was generally not helping them to restructure existing practice nor did it appear to convince them that the effort of restructuring would be worthwhile.

The courses in Kenya were based on a needs analysis, but in the TRCs and in the TACs this analysis was directed at exam results rather than on what the teachers said they needed, wanted or lacked. The SIP TAC staff in Kenya were working with teachers to make their teaching more effective and there were obvious signs that this intervention was having an effect on the contributions of both learners and teachers. However, it was felt that tutors were trying to make improvements on too many fronts and that more could be gained by concentrating on a smaller number of locally agreed initiatives, with concomitant materials production.

5.4.2 Sphere of influence.

Impact on schools also seemed to depend on who the teachers' centre was targeting and whether the school had systems in place to receive and make use of the messages from the centres. The criteria for selecting participants for courses varied considerably. In Andhra Pradesh untrained teachers were present in large numbers in the schools but these teachers were not targeted by the project and so lacked any kind of training. In Nepal on the other hand, training for the untrained teachers was a priority. However, the limited impact on schools as a result of this training could partly be explained by established teachers, trained many years ago, not being open to innovations brought back by trainees. This seemed particularly likely if the innovations were in conflict with the established teachers' own tried and tested routines.

The impressive training at the TRCs in Nepal was thought to lack effect partly because it was targeting teachers as individuals, it failed to help the teachers to think about ways of sharing ideas with colleagues back at school. Lack of impact was also partly due to the schools having no systems in place in terms of subject panels or curriculum development meetings to receive and adapt the ideas brought back from courses at the centre.

The TAC workshops in Kenya target teachers from subject panels in schools and these teachers were supposed to go back to the school and cascade the training to the other members of the subject panel. It was not possible to assess the effectiveness of this approach. However, as concern for the effectiveness of INSET gathered pace in Britain in the 1980s, the problem of teachers not sharing ideas from courses with colleagues in school became an issue and staff focused INSET increased. Similarly the project in Zambia was increasingly side stepping the teachers' centre and taking the training directly to groups of teachers in schools. However, these groups did not consist of teachers with similar needs so the training remained rather general.

The most effective way of ensuring that what happened at the centre had an impact on teaching and learning appeared to be for the trainers to follow the teachers into the schools. It was not possible to tell whether the occasional visits of the TAC tutors in SPRED II would eventually be sufficient to have noticeable impact, however, our observations of the SIP project persuaded us of the effectiveness of this strategy.

5.5 Resources.

The wide range of resources and reprographic faculties made available at the teachers centres in Britain, combined with the freedom teachers had to select materials, gave teachers lots of reasons to visit a centre. In the projects we observed, resources did not have the same level of significance at the centres.

Teachers in Kenya appreciated the availability of resource books and library boxes at the centres because there were few libraries in the schools. However, the main resources at the TRCs were sets of books and readers related to the KSCE exam syllabus in English. We have questioned whether a teachers' centre was necessary for the storage and supply of this type of books.

The facilities and resources at the TRCs in Zambia had not been integrated into the cascaded training programme and teachers made little use of them. The centres in Nepal distributed resources to the schools, but these were depressingly under used.

The only resources in most of the TACs in Kenya and in the Andhra Pradesh centres were those that the teachers produced. In Andhra Pradesh these were mainly replicas of teaching or learning materials that teachers had met during training courses. Except for a few notable exceptions, the teachers were not attempting to apply ideas encountered during courses to the making of materials for different contexts and purposes. The materials that teachers had made in Zambia were often displayed at the centre rather than being used in schools. The only materials developed at a centre which appeared to be really having an impact on schools were the materials made during the workshops at the SIP TACs. These materials had direct application for the work in progress at the schools.

It seemed that schools were generally so badly resourced that for some time to come the resource that really influenced teaching was the text book. Ideally it was felt that the donor aided projects should strive to improve the quality and relevance of the textbooks, in particular by making sure the books supported any principles or skills being targeted by the training. This had been a successful strategy in the much quoted New School project in Colombia (Wolf et al 1994). If text book renewal was not possible, then at least the teachers' centres could try to ensure that adequate copies of the current text book and the teachers' books were available and that the workshops at the teachers' centre helped teachers to develop more effective ways of using the text books that were available.

5.6 Accessibility

In all the projects the accessibility of the centres seemed an important factor in their use. For many teachers in Zambia and Kenya the TRCs were inaccessible. It was felt that a centre needed to be within 5km of a teacher's home or school for teachers to consider using it and even then most teachers needed to be encouraged to 'drop-in'. There seemed to be more hope of teachers' centres being effective in densely populated areas.

5.7 Attitude

One strikingly similar finding in both Britain and in our case studies was that teachers were generally very positive about teachers' centres. Although in Zambia teachers rarely attended the centre and hardly ever used it to drop-in for help or access to resources, there was still the feeling that teachers' centres were a good thing. Teachers in Kenya were convinced that they had changed as a result of the TAC tutor's work with them. However, there was no observable difference in their practice or in their classrooms. In Andhra Pradesh teachers liked the meetings, but could not say how these affected then-work in schools and some even said that they did not often try to apply the principles because this involved too much work. In Nepal teachers found the courses interesting, awareness was raised, but practise largely seemed to remain unchanged.

6 Observations made in schools

It was not easy to interpret what we saw in schools in terms of what was seen at the teachers' centre. There seemed to be a separation, operationally and conceptually, between what teachers did at the teachers' centre and what they did in the classroom. We partly attributed this to the many other constraints and influences the teachers had to work with in schools. While we also recognise that our presence in classrooms made it likely that teachers would not teach in their normal way, and that it is possible that the teachers we were observing may not have had time to try to integrate new ideas into their teaching, in general, we found very little evidence of teaching and learning having changed as a result of work at the teachers' centres.

6.1 Conditions in Schools.

One major obstacle to teachers implementing ideas which they had met at the teachers centre could well have been the conditions in which they worked. The literature on improving the quality of schools in developing countries suggests that the conditions in many schools and classrooms make it unlikely that teachers would be able to apply new approaches encountered during training outside of the schools (Lockheed et al 1991, Levin &Lockhead 1993). In many ways our study supports this view.

The school facilities in Nepal and Andhra Pradesh were very poor. In Andhra Pradesh, besides there being no furniture, water, toilets or electricity, there were insufficient classrooms. The children had little space on the floor and when it rained the classes being taught outside were sent home. In Nepal, in contrast to the teachers' centre which was often on the same site as a school, the classrooms were cramped, uncared for, and not at all stimulating.

In Kenya and Zambia, the fabric of the schools was better, but there were clearly other limitations on what happened in the classroom. Zambian urban schools could not cope with the number of children so there were two or three shifts each day. Private schools often used the same facilities in the evening and, because of low levels of pay, teachers were keen to teach in both systems. In rural schools there was a large number of untrained teachers. Facilities varied and in some schools classes were very large, in one instance a school had a class of 80 children because a classroom was being used as a resource centre. In Kenya it was difficult to try to trace evidence of training in schools because both INSET courses and schools had been disrupted by political unrest surrounding a general election.

We found impressive schools in other countries, but only in the SIP schools in Kenya was the obvious pride in the school environment so noticeable. The community were helping to extend and maintain the facilities and, according to one headteacher, the staff competed to have the most attractive classroom.

Considering the pedagogical aims of the various projects in our study countries, which stressed the use of group work, activity based learning and the use of teaching aids, it is not really surprising that teachers found these ideas difficult to implement in the cramped and under resourced environments in which they worked. In addition, during the training, the teachers were often not helped to consider how new methods and principles could be applied in their particular teaching situation.

6.2 Evidence of impact

We looked for evidence of impact from the work at the teachers' centre in terms of what happened in the classroom and which resources were being used in the school. We also looked for the specific messages and resources being targeted by the recent training at the teachers' centre.

6.2.1 Impact on schools

Evidence to indicate that the teachers' centre had been effective in improving the quality of education in schools was very difficult to find. In Zambia the project was obviously influencing the way teachers worked together in schools. Whilst the Teacher Groups for the cascade training were not leading to much impact on practice, the subject based Working Groups which had recently begun to function did seem to be having noticeable effects, mainly because the meetings had immediate relevance to what was happening in the school. However, there was little direct evidence of the teachers' centre being instrumental in these changes at school level.

In contrast, in both Nepal and Andhra Pradesh there was little evidence of teachers working together in the schools. Although teachers planned together for model lessons in the meetings in Andhra Pradesh, there was no evidence of teachers planning together or making materials outside of these meetings. In Nepal, there was no management structure for planning and implementing curriculum across the school. There were no subject coordinators and the text book was the sole curriculum guide, and as the teachers' centres treated teachers as individuals there was no guidance on how their training could be used to help the teaching in the school.

6.2.2 Impact on what happened in classrooms.

Teaching and learning in the classrooms mainly followed the traditional approach of teacher talk, questioning and chorus response. The focus was on transmitting content. There was little observable transfer of the pedagogic messages relating to active learning, discovery methods, group interaction, or differentiated tasks being advocated at the teachers' centres. Exceptions in individual classrooms in Nepal, Andhra Pradesh and Zambia did suggest that teachers should have been able to transfer ideas from the training. However, as neither the text books nor the exams rewarded teachers for using such approaches it is not really surprising that the training had achieved little change in most teachers' practice.

In Zambia and Andhra Pradesh, where particular skills or principles were the main focus of INSET, there was some evidence of transfer from the training sessions. However, the transfer seemed to be mainly at a superficial level of technique. A number of instances were observed in which ideas from the training were being used inappropriately, which suggested the teachers had not really understood the underlying principles. Teachers seemed to find group work particularly difficult to use effectively so that even when children were arranged in groups they were generally taught as a whole class. Examples of teacher-led innovation were rare.

In contrast to the problems with new approaches, some of the traditional teaching was effective and seemed like a sensible reaction to the situation in which the teachers found themselves. Harvey (1998) mentions similar findings. Helping teachers to make the best of their traditional practice, and the resources easily available in their present circumstances, seemed to be more useful than expecting teachers to take on completely new pedagogy in circumstances which did not yet offer them the necessary support from resources like text books.

While it was necessary to search for evidence of impact in the schools in other projects, in the SIP project in Kenya there seemed to be evidence of impact at every level; from the SIP TAC management committee to the head and deputy and into the classroom. The enthusiasm and involvement of staff, pupils and members of the community was impressive. We saw many examples of transfer of ideas from the workshops at the teachers' centre to the classrooms.

6.2.3 Impact on the availability and use of resources.

Text books and the blackboard were the major resources in the classroom. The text books were mainly supplied by the government, but the necessary books were not always available. A range of textbooks was being introduced in Zambia, which allowed schools some choice. The Teachers' centres could have jumped at the opportunity to display these books, as the centres did in Britain. Workshops could have helped teachers to select the most appropriate books for their context and to work out the best way of using them in the classroom, but so far they had not adopted this supportive role. The TRCs in Kenya were helping teachers by providing class sets of readers and box-libraries.

In several of the schools, potentially useful resources were available, but teachers were not making use of them. Science equipment in Nepal was locked away in cupboards or not even unpacked. In Zambia we saw atlases and other materials in the staffroom, but no evidence of teachers using these when they would clearly have been useful. In Andhra Pradesh, teacher-made materials were displayed on the walls, but did not seem to have any purpose, while materials made in schools in Zambia were displayed at the teachers' centre rather than contributing to teaching and learning.

Although increasing the resources for teaching and learning was one of the main purposes of many of the teachers' centres in our study, there was little evidence in the schools to suggest that this aim was being achieved.

6.3 Support For Schools.

In three of our case studies, the staff at the teachers' centres also had a role in schools. The Resource Person from the primary resource centres in Nepal should have been spending lot of time in schools, but the effectiveness of their work was limited. Besides the very real practical problems, they did not have a clear role in schools. Also, because there was no focused plan to develop skills or content, the advice they gave was not necessarily very useful. They tended to just observe individual teachers who had attended course and comment on their lesson. There appeared to be no system of helping teachers to plan for future teaching and no involvement with other teachers who were teaching the same subject. The resource person did not appear to concern themselves with the quality of education in the school.

The newest phases of DPEP in Andhra Pradesh and SPRED II in Kenya have school follow-up. It is too early to judge the effectiveness of this strategy, but the huge numbers of schools and teachers involved and the distances between schools and the centres make it doubtful that this support could be very effective. Establishing a rolling programme of targeted schools, as the SIP programme has in Kenya, may make better use of the expertise available.

The project in Zambia was increasingly targeting teachers in school. Although this did not involve the teachers' centre in any way as yet, the formation of Teachers Groups in the school seemed to be having an effect on school policy and on what happened in the classroom. The staff from the teachers' centre could usefully lend their expertise to the groups in the future.

When teachers went to the centre to attend a course during school hours, there was rarely anyone to cover their classes. Therefore, a disadvantage of teachers going away from the school for training was that this added to the widespread problem of teacher absenteeism. Working with teachers in the schools might also be a way of ensuring that the children have a teacher in the classroom.

In the light of previous studies (Harvey 1998, Kinder & Harland 1991) and our observations in schools, it did not seem sufficient that staff from teachers' centres only visited teachers as a follow-up to courses. The apparent success of the SIP project suggests that for real impact, long term, intensive input is necessary, whereby sympathetic and dedicated experts have clear roles and work intensively with groups of teachers. This is a high cost, slow, labour intensive strategy, but if it brings about the desired increase in quality of education it may be money well spent.

7 Conclusions and Recommendations

According to any of the criteria used in our studies regarding impact on schools, teachers' centres alone did not appear to be an effective strategy for improving the quality of teaching and learning in the classroom. Neither did they make a noticeable difference to the availability and use of teaching and learning materials in the classroom.

The SIP project in Kenya, which seemed to be achieving the most success, supported the teachers' centre with a number of other initiatives. The trained and supported staff, the community involvement and the long term, intensive work in individual schools all appeared to contribute to the level of impact on teachers and schools. Maintaining the momentum of change once the donor funding is withdrawn will be a challenge for both the staff at the teachers' centre and in the schools. The relative collapse of the SIP programme in Kisumu points to the pitfalls for the Mombasa SIP project once the level of financial support is reduced.

7.1 Other indicators of effectiveness.

Although we found little evidence of direct impact on classroom we did find that effectiveness could be interpreted in a number of other ways.

· The presence of teachers' centres as much as their use seemed to improve morale to some extent and give teachers a feeling of status. They allowed teachers, and to an extent the community, to feel part of and contribute to the energy of the project.

· Established schedules for meetings did provide regular opportunities for professional interchange and as such could be seen as an important prerequisite for the development of a professional culture. At the moment this culture is mainly being shaped by the donor driven projects but for the future, self steering strategies would need to be encouraged.

· Courses at the teachers' centre appeared to begin the process of change in teachers by raising awareness and providing ideas and materials to replicate. Studies which have looked at achieving change in teacher's practice suggest that this level of change is necessary before other strategies could help the teacher to develop deeper professional understanding and innovative practice (de Fetter et al 1995 cited in Harvey 1997: 109, Harvey 1997, Joyce & Showers 1980, Kinder & Harland 1991). The teachers' centre would thus need to be recognised as a useful intervention, but one with a limited purpose.

· The in-service trainers and the staff working in the centres benefited from the opportunities available in terms of increased knowledge about teaching and increased skill in working with teachers. Hence there were significant changes in the formal education system for teachers, particularly in terms of the teacher training curriculum. The focus and relevance of this curriculum was clearly becoming more finely tuned to teachers' needs.

· Where teachers' centres had involved the community there was an awareness in that community that things were changing in education and consequently expectations may have been raised.

As a result of our research, we feel that there are a number of conditions which need to exist in the schools and at the teachers' centre if centres are to have more impact on what happens in schools.

7.2.1 In schools, the environment and the systems operating seemed to be fairly critical factors in determining whether messages from the TRCs had any effect on teaching or learning.

i. While it was not possible to say that teachers' centres are more effective if conditions in schools are conducive to learning, conditions in many of the schools were very poor and the teachers' centre needs to face up to this reality. The physical condition of schools and classrooms and the number of children in a class often did not help the teacher to implement ideas from the teachers' centre. The centres in turn did not seem to be helping the teacher to confront the conditions by helping them to find solutions to their problems.

ii. Teachers needed to be helped to implement innovations. This needed to be done in a number of ways. Management of any innovation was necessary at school level before it seemed to be effective. It seemed to help if mechanisms, such as subject panels, were in place to receive ideas from the teachers' centre. A headteacher with a positive attitude to professional development, enthusiasm, commitment and active involvement in the activities of the school, and who showed appreciation for the efforts of staff and children, was found to be influential in achieving changes in classroom practice. Working as a staff seemed to be motivating for teachers, while on-going support for a group of teachers from a sympathetic adviser seemed to be particularly effective.

iii. Teachers needed the support of basic resources, in particular the support of appropriate text books. Help with coherent forward-looking planning, rather than for particular one-off lessons or problem areas, seemed to be particularly helpful for teachers as they adapted to new approaches. However, restructuring new ideas to match established text books and exam syllabuses was very difficult. The benefits for the teacher and learner of text books, teachers' books and exams which matched and supported the messages from the teachers' centre were noted, but situations where this was found to happen were rare.

iv. Teachers needed to be encouraged to develop or create solutions to local pedagogical problems in their own schools. The teachers' workshops in Zambia appeared to be a successful way of empowering teachers.

v. It was felt that by working intensively to change conditions in a limited number of schools, and by encouraging teachers to take responsibility for running meetings and workshops, it might be possible to help these schools to become self-sustaining centres of innovative practice. Targeted schools could in turn become model schools for others to learn from.

i. Enthusiasm and commitment were essential at every level. The components of this commitment include the centres receiving adequate and reliable financial and material support and appropriate levels of expertise. They also include recognition of the efforts of tutors and all those involved in the work of the teachers' centre and recognition by tutors of what teachers and headteachers are trying to achieve in schools.

ii. A major condition for impact at school level seemed to be that courses needed to help teachers with their day to day teaching. In particular they needed to help teachers to develop more effective ways of using the limited resources they have in the conditions in which they work. It was felt that if input was subject specific, or text book and teachers' guide related, there could be an on-going continuum into which the gradual development of new skills, methods and content could be integrated.

iii. Where new approaches were considered appropriate and relevant it was felt that courses needed to begin with where the teachers were rather than with an end view of what 'good practice' involves. It was felt that donor-driven projects with predetermined agendas were often trying to achieve too much through cascaded training courses in too short a time. Workshops which presented new ideas in ways that related to the teachers normal practice and gave suggestions for ways of moving on in small incremental steps were felt to have more chance of impacting on schools than sessions which required the teacher to make radical changes in their practice.

iv. However, it was felt that rather than importing ideas of good practice from elsewhere it might be more effective to identify the good practice which does already exist in the schools and devise training to build on this.

v. Teachers' centre staff needed to have a clear role in the schools they served. The projects were increasingly recognising the effectiveness of directing inputs towards schools rather than at individual teachers. However, the logistics of teachers' centre staff getting to the schools and working effectively with the large numbers of teachers involved considerably limited what could be achieved. As there was rarely an adequate number of appropriate staff for this very demanding job, the slower process of advisors spending more time in fewer schools as part of a rolling programme could achieve more tangible results.

vi. The teachers' centre needed to have more autonomy and freedom to respond to needs and changing demands. There was a feeling that neither the teachers' centre staff nor the teachers were being empowered to initiate change and seize opportunities. Project aims and time constraints dominated the work of the centres. Cascade approaches to training encouraged dependency on centralised initiatives and marginalised initiatives at the teachers' centre level. With the exception of the SIP TACs in Kenya, teachers' centres were not able to respond to the changing needs of their teachers. Encouraging local level involvement in the running and ownership of the centre was thought to be beneficial and to increase the likelihood of the centre continuing to function when donor support ended.

vii. The staff at the centres needed to have adequate training specifically for their work with teachers, not just for the purposes of a cascade. They also needed to have status as experienced and successful practitioners and a level of independence so that they were not perceived as part of the inspection service. The practice of personnel from a number of centres meeting regularly to discuss their work was a strategy which appeared to encourage initiative and the development of activities at the centre.

viii. Expectations for materials production at the centre needed to be more realistic. Facilities for making materials were underused and impact from teacher made materials was difficult to find. While exceptions were identified in a number of schools, few teachers are sufficiently able or motivated to spend time producing effective teaching materials. It was felt that teachers could be helped to produce low cost, relevant teaching resources which matched an obvious need. It was also felt that more could be done to identify and exploit creative and imaginative teachers who had the ideas, talent and skills required to develop materials for local use.

ix. The centres needed to be easily accessible for teachers. Centres were rarely used as drop-in facilities. The distances and expense involved in getting to the centres and the restricted opening times were mainly blamed for this. It was felt that centres were particularly useful in urban areas where schools and teachers' homes were no more than 3-5 kms from the centre. Trying to serve schools any further away than this was found to be very difficult so it was felt that other strategies might be more effective where schools were more dispersed.

It was felt that a more flexible approach needed to be taken when selecting strategies for school improvement, with careful consideration of the context, the needs of the teachers and their schools and the sustainability of any intervention. A number of other strategies are discussed in the main report and a number of alternative roles for teachers' centres are suggested.

FOOTNOTE

1. them = A list of subsidiary questions for each part of the research are given in Section 2 of the main report.

2. appropriate = Full details of the research instruments and procedures are given in Section 3 of the main report.