Educational researchers who really want to influence policy should become politicians.

Pamela Munn, dean of education at Edinburgh University, said in a setpiece address that researchers were in greater favour with ministers than at any time in the past but were still subject to political whim. “If research does not chime with the current ideology, it will not influence policy,” Professor Munn said.

There was ample evidence in the 1950s and 1960s about the unreliability of IQ tests and the inappropriateness of selection but it took the political commitment of the Labour government in 1964 to introduce a comprehensive system. Since then the political culture had changed and selection had returned.

“Policy-making is inherently political,” Professor Munn said.

Labour in Britain had invested more in educational research than ever before in an effort to link thinking and policy. Some pound;2 million had been pumped into projects in Scotland through the cross-institution applied educational research scheme (AERS).

Professor Munn, a key researcher on school ethos and discipline, advised fellow researchers to accept that what policy-makers wanted to know was what works. “I am not convinced we do that,” she said.

She appealed to them to collaborate in studies with related disciplines and not to make exaggerated claims for their findings. They should avoid over-complication and make sure their results were easily communicated to a wider audience. They had to be far clearer in their claims.

“We also need to talk to HMIE. No inspector would claim they do research but they do collect interesting information across the country. They are able to say what is going on in schools across Scotland,” Professor Munn said.