It may have provided juicy soundbites for newspapers and politicians, but the spin the ESRC put on our findings about grammar in schools does no one any favours, warn Christopher Brumfit and Rosamond Mitchell. It is widely held that if you take money direct from a government agency you cannot complain at simplistic attempts to "use" your work. But the Economic and Social Research Council is not a government agency, though it is committed to a policy of "ensuring that research continues to be relevant to the needs of users". Does that entail allowing research to assist government publicity?

Monday June 17 saw a string of headlines: "Grammar is foreign to English children" (Daily Telegraph), "Teachers 'failing on grammar"' (Guardian), "Teaching of English fails grammar test" (Independent). These were followed by the headline "Back-to-basics school exam" in the London Evening Standard. The Telegraph included articles on its favourite theme of English teachers' "failure" to teach grammar for most of the week.

The origin of this interest was a press release by the ESRC based on research which we at Southampton performed as part of the "Quality of Teaching and Learning" programme. We reported to the ESRC in January 1994, and since then have produced occasional papers in books and journals. Two years later, in June 1996, newspapers presented accounts based on that same research but which misleadingly described it as a "recent report for ESRC", a "survey" and a "report for Gillian Shephard". The very next day the Secretary of State was reported to be introducing "tests" of spelling, grammar, etc. and in the Telegraph of June 20 both she and David Blunkett cited the "Southampton report" as evidence to justify their back-to-basics position.

The way in which this work has been published raises serious issues for the relationship between the ESRC and government, and for the integrity of genuine, knowledge-creating research in the social sciences. The problem for the ESRC is that it is expected to produce performance indicators which show that it is having "impact" and influencing policy.

³ÉÈËVRÊÓƵ

While all researchers hope their ideas will affect practice, the notion that the prime function of research is to have a direct effect on policy, even in areas like education, may well lead to the cart preceding the horse. Our research was designed to address issues relating to the topics which were taken up by the newspapers. But our prime concern in our bid for research funding was that there was very little empirical data available, and a great deal of assertion. The publicity has just added to the assertion. We made a bid for work in five areas, four of which we produced documentation on: 1. To document the understandings of secondary English and foreign language teachers regarding the nature of language, their beliefs about the role of explicit knowledge about language in language education, and their reactions to national curriculum policy proposals in this area; 2. To document and compare the handling of explicit knowledge about language in English and foreign language classrooms; 3. To explore the models of language held by 13 to 14-year-old pupils and their origins, in the classroom and outside it; 4. To document how such pupils make use of explicit knowledge about language in the course of English/foreign language classroom activities and assignments.

In writing the report, we were conscious of the possibility of distortion of our evidence. We phrased every sentence relating to grammar, discussion of language, and the practice of English teachers, extremely carefully.

³ÉÈËVRÊÓƵ

We did not conduct a survey (as asserted in the press), but case studies in three schools. We limited these to detailed analysis of eight-week periods in each, in which we recorded sequences of lessons in modern languages and English for 13-year-olds.

At no point did we set out formally to explore the competence of learners in these fields in comparison with the past, or with external measures. We did document the ways in which they made use of linguistic concepts.

To the public, it appears that the major funding council for social science, with a need no doubt forced upon it by the climate of the times, has to attract substantial publicity for the work it is doing in education to show its usefulness. Thus it becomes tempting to enter a debate for which this specific research did not equip it. This in turn raises serious questions for the whole researching profession about the ways in which the ESRC sees its role.

A cynical view might be that any publicity is good publicity, but social science researchers have to work with practitioners, and it does no service to imply that research supports positions which are politically derived, in order that the Government may enforce its own agenda. There are differing views on how much grammar should be taught in classrooms, but there is no argument for extrapolating out of a series of limited case studies over a short period of time, global judgements about the practices of teachers throughout the country.

³ÉÈËVRÊÓƵ

Our research showed clearly that there was a great deal of discussion of language undertaken by English teachers with 13-year-olds, though it tends to be at a stylistic rather than sentence grammar level. Our findings were scarcely surprising. Whether there is a case for continuing teaching of basic English grammar is an issue about which we may well have views, but this particular data does not offer evidence on that question; nor did we at any stage make any claims about that.

We did make some statements about the need for teacher education, particularly in ensuring that teachers' own knowledge about language should be fully developed, but this was in the context of substantial evidence of the concern of both foreign language and English teachers for understanding of language. There is no way in which our research report could be identified with an attack on the willingness of teachers to address linguistic issues.

Because the ESRC press release had to find a "story" for journalists, it managed to imply that a fairly short-term project which was only a small part of a five-year programme of ten projects was important enough to be direct evidence for the Secretary of State for Education. This enabled Gillian Shephard to report in the Telegraph that Southampton University researchers have found that levels of grammatical knowledge are inadequate. This judgement has been imposed on data which could not possibly support such a generalisation. But if ESRC puts out press releases in which 11 out of 14 paragraphs are on this one topic, generalising more substantially than we possibly could from our own material, the Secretary of State's misunderstanding makes some sense.

Either way, there are major concerns for ESRC committees. Above all, we should be worried by the tendency for any research council to produce material which, even accidentally, feeds directly into government agendas, if only because examples of this will force researchers away from significant issues that have direct bearing on a practice, because the negative side effects on the profession are bound to be dangerous.

³ÉÈËVRÊÓƵ

Practitioners will become more suspicious of working with researchers, researchers will shy away from potentially controversial issues (precisely those that need more light and less heat), and researchers will be afraid to discuss concerns about the research councils and politics because they want to be funded by them (as so much in the Research Assessment Exercise depends on such funding) and they will not want to rock the boat.

We have seen enough examples from civil servants and politicians of the misuse of research findings, and enough examples of misunderstanding, to ask that the major body responsible for funding social and educational research, in spite of the pressures, should act with care and responsibility.

³ÉÈËVRÊÓƵ

Christopher Brumfit is professor, and Rosamond Mitchell, reader in education, University of Southampton.