There is something of a fashion for knocking educational research at the moment. This often follows the time-honoured technique of finding a couple of daft - or daft-sounding, it doesn’t matter - examples of research projects and concluding on the strength of them that the whole business is a waste of time.

This is a bit like pointing to a few badly-done school inspections and proclaiming that the entire inspection system is rotten. An easy trick but a cheap one. Education debates - and the students they concern - deserve better.

Some of the criticism comes from impatience with bad or irrelevant research. As such, it is entirely legitimate and must not be gainsaid. After all, who wants bad research, or bad inspections, or bad anything? One might just observe that the most trenchant criticism of such research comes from within the research community.

There is another strand of criticism, however, that is rather less welcome. Fed by anti-intellectual prejudice and built on a profound misunderstanding of the nature of educational research, it takes the research community to task on spurious grounds. A common example of this is when research findings are judged purely in terms of their immediate classroom application. Any effort to take a broader perspective - or, Heaven forfend, focus on theory - is dismissed as airy-fairy nonsense.

Classroom relevance is important but it is not the sole criterion of worth. The purpose of educational research is to improve teaching and learning, but it cannot achieve this by focusing exclusively on classroom practice. Classrooms and the students in them do not exist in isolation from the rest of the world. Educational research must therefore look at the classroom in context. It must take account of the multiple environments - family, cultural, socio-economic, media - within which young people learn. It must investigate the many policies and structures which have an impact on schooling. And it must have regard for theory, since otherwise research risks producing a jumble of unrelated facts.

In any case, it is not the function of research to tell teachers what to do. Teaching is not a mechanical business, which is to be enhanced by tips from researchers. This is an impoverished view of both teaching and research, and quite misunderstands the relationship between them.

Teaching is a form of intelligent, expert action. As such, it entails applying general principles to practical situations, which are both specific and changing, and it requires the use of expert-driven tools such as curriculum frameworks, resource materials and tests. The job of research is to underwrite these principles, by articulating them clearly and testing them out in diverse situations, and to ensure that appropriate tools are available.

Let me give a couple of examples. Research has shown that explicit instruction in thinking skills can improve students’ learning - the Cognitive Acceleration through Science Education project is a good example. Given this general finding, it is for teachers to apply it appropriately in the diverse learning challenges they face.

Researchers cannot do that task for them - nor should they attempt to. It is for teachers to integrate relevant research findings with their own expertise, their personal repertoire of teaching skills, and the constraints and opportunities of the specific situation. Important though the research findings may be, they are only one element within the pattern of intelligent teaching behaviour.

The use of tests provides another example. When teachers use tests to diagnose a pupil’s learning difficulties or to monitor the progress made by a class, they are drawing on the research which has to underpin any sound assessment instrument.

If the relationship between teaching and research is complicated, that between research and policy is tortuous. Research can impact on policy when policy is being formulated, when it is being implemented and when it is being evaluated. There is a distinctive relationship at each stage in terms both of what is possible and of what tends to happen in practice.

Educational and social policies are determined by a mix of factors - tradition, resource and system constraints, conviction and so on. Research evidence does make a contribution but generally a modest one. It is rare for research findings to determine educational policy, but they can inform the decisions taken by policy-makers. Research has a particularly important role in the evolution of policy, both by supporting implementation and by evaluating outcomes.

Each of these strands can be exemplified by reference to inclusive education, where there have been major policy changes over the past 30 years. These changes have stemmed from many sources but research has been present throughout in policy formulation, implementation and review.

The prevailing segregation policies were challenged initially on ideological grounds. Research boosted the process by undermining the efficacy claims of segregated schooling and by offering pedagogical legitimacy to integration. Neither research nor ideology was as persuasive on their own as they were together, and this conjunction both focused and speeded up policy development.

The implementation of inclusive education policies has benefited enormously from research. By careful documentation and analysis of practice, many studies have informed the decisions that schools and authorities must take. Such studies identify the factors critical to successful programmes, display the consequences of different courses of action and make examples of good practice accessible.

As for policy evaluation, there is no alternative to rigorous and independent research. Inspection and internal review have a part to play, especially in quality assurance. Policies must ultimately be validated, however, by systematic evaluation which analyses their impact on practice. Inclusion policies have benefited greatly from formative and summative evaluations, which while modifying them have helped to anchor them in mainstream policies.

Research is an essential part of an effective education system. This is why we must not tolerate obscurantist tirades against it any more than we should put up with bad research. Fashioning a high-quality education system for the 21st century requires more research, not less. The task facing us is to build on what we have and ensure that the system has an adequate research capacity for the challenging years ahead Seamus Hegarty is director of the National Foundation for Educational Research.