Unpublished research commissioned by the Office for Standards in Education suggests there may have been a decline over 20 years in the standard of maths in schools in England and Wales. More controversially, it suggests poor performance in maths may be linked to the way the subject is taught in primary schools.

Work undertaken for OFSTED by David Reynolds, professor of education at Newcastle, is due to be revealed on the BBC’s Panorama next month before its official publication. OFSTED is not providing any details before the programme is transmitted.

However, the conclusions reached on the basis of the review of the international literature on achievement in maths and science are likely to provoke a row over teaching methods on similar lines to the controversy over the teaching of reading.

The report is expected to suggest that teachers in English schools, particularly primaries, attempt over-complicated lessons designed to cope with children of different ability. Observation of classes in Taiwan suggests that teachers might do better by dropping group and individual work and teaching the class as a whole.

The chief inspector of schools, Chris Woodhead, has since his appointment expressed the view that teachers might usefully spend more time in front of the class and less organising work in groups.

While the research suggests such lessons might mean a few children would not be able to keep up, they could be given additional help through homework or extra sessions.

It does not suggest adopting the Taiwanese practice of classes of 40 or more on the grounds that large classes are only possible there because of the culture of obedience and the social pressure to succeed.

In an early draft, the research also suggests maths teaching might be improved with the use of good quality textbooks; that there may be a case for children doing more homework and for there to be more school days in the year.

The survey and its conclusions are likely to come under fire from other academics. John MacBeath of Strathclyde University believes there is no convincing evidence that schools should simplify their teaching methods.

He says:“To focus on whole-class teaching is to miss the point. We need to focus on how children learn. There is no logical argument to support whole-class teaching as against a more differentiated approach which takes account of the stage an individual has reached.”

However, Peter Mortimore, director of the Institute of Education at London University, who has expressed concern at the effectiveness of maths teaching, suggests it is vital that teachers examine what methods are successful.

One problem already identified, he says, is that there is a shortage of mathematicians in primary schools and staff need to be prepared to teach the subject with competence and confidence.

Other academics question whether research has proved that particular teaching methods affect maths results. Margaret Brown, professor of maths at King’s College, London, cites research from the Second International Maths Study which looked at 13-year-olds. That study, she says, did not link attainment with any particular teaching method.

There are problems, she says, with interpreting international comparisons. The surveys do appear to show that pupils in English schools are less good at number work, but that might be because schools have a broader maths curriculum.

“I am not suggesting we are doing brilliantly in maths, but we must consider carefully before any policy switch,” she says.

The Reynolds review suggests English schools are doing better in science. Schools do more practical and experimental work than found in other countries, though he suggests that finding might be the result of flaws in the way data was collected.