Educational research has had a curious year. Following decades of neglect, it is now in danger of becoming newsworthy. Sometimes it is because of particular research findings - English pupils do badly at maths (their superior performance in science does not rate the same headlines). Occasionally it arises from the expectations on teachers in schools to engage in research or the pressure on teachers in universities to be “research-active”. And once in a while it is down to a good old-fashioned row, such as when leading professors of education fall out with the Chief Inspector of Schools on the proper conduct of research.

One matter that gets overlooked in all this is the way in which the conduct of research is organised. This might seem an esoteric issue of concern only to professional researchers. It is much more immediate, however. Not only does it bear critically on the national capacity to produce relevant research findings, but it also impinges directly on how higher education gears up to meet the needs of the 21st century.

The Centre for Educational Research and Innovation at the Organisation for Economic Co-operation and Development has been examining these issues across its member countries. For devotees of coherence, its conclusions are disturbing. A report in 1995 found the organisation of educational research between and within countries was so diverse as almost to defy categorisation and, moreover, bore little discernible relationship to the underlying educational systems.

This country is a particularly rich example of diversity. Apart from bodies like the National Foundation for Educational Research and the Scottish Council for Research in Education, research is conducted by university staff, LEA staff, practitioners such as teachers and psychologists, various official bodies, management consultants, market research organisations, individuals and so on. One could regard this as a creative response to the demands of information-gathering in a complex system where intelligence needs and responsibility for action are highly diversified.

An alternative view would be to regard it as a glorified cottage industry, with lots of small and idiosyncratic producers, some of them undoubtedly doing excellent work, but hardly constituting a coherent system of knowledge acquisition, distribution and utilisation. It’s all very well letting a thousand flowers bloom - but not if you want a top-class garden. Bad gardeners who cannot tell their weeds from their flowers or their annuals from their perennials and who never know when something will come into bloom or what colour it will be when it does, stand to be constantly surprised, but their gardens are likely to be a hotchpotch of colours where everything flowers in June and nothing in September, and there are lots of bare patches and other bits that are overcrowded.

If we want coherence and consistent quality from research, if we see it as articulating a cumulative understanding of the education system, it may be necessary to sacrifice some individual enterprise and accept a degree of planning in the disposition of the limited resources available. The case for organising research in a business-like way is taken for granted in agriculture, aerospace, defence, medicine, science and so on - and there is no reason to believe that these sectors lose out on creativity because the conduct of research in them is tightly structured.

A big part of the problem in educational research is that it is full of sacred cows, some of them giving all the appearance of sanity. Top of the list for culling are the notion that teaching and researching are integrally related and its corollary that high-quality teaching depends on having a direct engagement in research, that is, doing it oneself as opposed to using others’ findings. These go along with an implicit devaluation of teaching as an activity: unlike the practice of law, medicine or other knowledge-based crafts, excellence in teaching does not seem to be enough on its own without the added patina of research involvement.

This is not, by the way, to dismiss practitioner research. It is simply to insist that teaching and research are different skills, and competence in one is no guarantee of competence in the other.

Much of the difficulty here stems from the prevailing assumption in higher education about the interdependence of teaching and research. This assumption emerged in the German universities in the 19th century and gradually became established as a defining characteristic of universities, particularly in North America and Britain. In point of fact, universities were founded originally to transmit established bodies of knowledge and to provide vocational training - for lawyers, doctors, theologians and so on.

This concept of the university needs to be set aside, for a number of reasons but in particular because of the problems its creates with regard to organising research coherently. First of all, it is not affordable. Higher education is expanding throughout the developed world - by more than 50 per cent in the UK from 1989 to 1994 - but the scope for doing research will not expand in step. This disparity alone will make it impossible to have a close alignment between higher education teaching and research activity. Second, despite numerous studies, there is no convincing evidence to say that academics who engage in research make better teachers. Third, universities of the 21st century are going to have to focus much more sharply on their teaching role. A recent study for the Committee of Vice-Chancellors and Principals emphasised the need for universities to respond more effectively to the changing demands of students.

The main point, however, relates to the conditions for conducting high quality research on a sustained basis. These include:

* a critical mass of dedicated researchers, sufficient to provide both coherence and a range of disciplinary perspectives;

* career progression, underpinned by staff development;

* sustained links with the relevant policy and practitioner communities;

* dissemination capacity;

* an infrastructure to promote efficiency.

There is no reason why these conditions cannot be met in universities, but it would be unrealistic to expect them to be met for every discipline by every university. And as a matter of fact, they are frequently not met at the moment. John Goddard has drawn attention to “the poor terms and conditions of service for many contract researchers, lack of investment on research infrastructure, opaque funding regimes within universities . . . the lagging commitment in many universities . . . to staff development and training” (The THES, March 15 1996).

If we are concerned to establish and maintain a national research capacity in a cost-effective way, there is no alternative to dedicated research centres. This is the only way of meeting the requirements cited in terms of coherence, critical mass, training, infrastructure and so on. Such centres could be in universities or independent of them; in an ideal arrangement, there would be some of each to capitalise on the respective advantages of independent and higher education status.

This concentration of the research effort needs to be matched by a rethinking of the training requirement for research and how it is best met. Doctoral programmes have been seen as the traditional route into research, and public funding for them is based on the assumption that they prepare people for research careers.

Their adequacy for this task must be questioned, however. Research institutions in the social sciences - the problem is not confined to education - find it necessary to provide a great deal of training for newly appointed staff. The question has to be asked if the present arrangements are the best way of providing a professional cadre of skilled career researchers.

Doctoral programmes can be excellent in deepening individuals’ understanding of a subject area, introducing them explicitly to the structure of knowledge in it, and equipping them with heuristic frameworks for tackling new issues. But if the aim is to prepare people for a research career, perhaps the relative isolation of the PhD will not be enough. Just as in teacher education and most other areas of professional training, it may be that substantial periods of time have to be spent in a structured, professional environment. Such environments are best provided by research centres. It does not matter whether these are university based or independent. What does matter is that they be reasonably large and have a strong focus on the business of doing research.

Research is only one strand in bringing about educational reform but it is a key one which we ignore at our peril. The available research capacity will always be limited and it is important that we make best use of it. This requires greater coherence in how we organise research and a targeted uplift in training for career researchers. Then perhaps educational research can make its unique contribution to the enhancement of learning, whether in schools and other institutions or in society at large.

Dr Seamus Hegarty is director of the National Foundation for Educational Research. This is an edited version of the paper he delivered at the NFER’s recent 50th anniversary conference