I HAVE GREAT sympathy with Dini Power (TESS, October 2) in her frustration with educational research, which she sees as necessary and potentially useful to teachers, but not living up to its promise in practice.

Research into any aspect of education that is really important is complex and difficult. In consequence, on any major issue - such as class size, ability grouping, class organisation or methods of teaching - one can find contradictory conclusions from individual studies. Often it is almost impossible to separate variables and to compare in a valid way different approaches or practices.

For example, in studies of class size it is not sensible to use exactly the same methods of teaching and have the same degree of teacher-pupil interaction in very small and very large classes (if they were the same, then the whole purpose of having small classes would be lost).

When comparisons are made it is difficult to know if differences in performance are due to class size, teaching style or teacher-pupil interactions.

Rather than despair of ever having the “perfect” research that settles things once and for all, it is necessary to accept that any one piece of work in these areas should never be considered on its own. Dini Power was quite right to challenge what she read in the TES Scotland about evidence on the effect of setting, if it was the case that the research was carried out with “a tiny sample . . . subjective and designed round a series of leading questions”.

Having carried out a review of research into setting and streaming, I am well aware of the enormous complexity of this matter and of the vast range of different findings, from which the only conclusion that could be drawn was that the research at primary level provides “no evidence that achievement of pupils is raised either by streaming or setting within the school” (Setting and Streaming: a Research Review by Wynne Harlen and Heather Malcolm, SCRE 1996).

However, my point here is not to argue this specific matter but to consider what can be done to reduce the feeling that teachers must have of being swayed one way and then another by inconsistent research findings.

I believe that what is needed is for each piece of research to be seen not as standing on its own but as adding to what is already known on the matter under study. It is the responsibility of researchers to assist in this by contextualising their findings within existing knowledge and research findings. It means that new research should be accompanied by a sound, serious and comprehensive review of relevant existing work.

Of course, a review of work in the field is often the starting point of a research proposal or report, but it is usually of a ritual nature, rarely going beyond summaries of the most easily accessible research and providing no basis at all for drawing conclusions from a wide range of relevant and high quality evidence.

By contrast a thorough review of research will be based on a search of reports and other literature, focused by decisions about what is included or excluded that are made explicit and are justified. When the relevant research studies have been located they will almost certainly be found to be varied in quality and should not all be given the same weight in drawing conclusions to guide practice and policy.

The meaning of “quality” has to be decided in relation to the research questions of each new study and, again, made explicit. Where there is a range of good quality research it is appropriate to draw on this and ignore the rest. Where there is little of quality, then shortcomings should be pointed out and conclusions stated cautiously.

Research is a means through which a profession learns. So, just as the learning of individuals has to begin from existing ideas, it is appropriate that we should begin further learning as a profession from what is already known. There is nothing new about that.

What would be new, however, would be for researchers, in reporting each new study, to discuss their findings in the context of existing knowledge and explicitly to update that existing knowledge with the outcomes of their work. They should show where it lends support to, or conflicts with, the existing picture and attempt to account for any differences.

If researchers were to report their research as helping to build understanding of what are necessarily complex matters hopefully the media would also communicate it in this way. Then practitioners would be able to see that research is adding to what is known, and does not ignore what has been found before, even if it does not agree with it.

Wynne Harlen is director of the Scottish Council for Research in Education.