Teachers need to be part of university studies, former minister tells the British Educational Research Association conference.
Academics will have to work more closely with teachers if they are to influence education practice in the future, according to Estelle Morris, a former education secretary.
She wants researchers and politicians to develop a better relationship, but warned that academics would also need to get more in touch with the classroom.
The former PE teacher said the Department for Education and Skills had been most successful during her tenure when civil servants, academics and teachers worked together on policy.
And she told delegates to the British Educational Research Association conference at Warwick university last week: "If what Government says is right about handing over more power to schools and making schools the leaders of change, then researchers will increasingly have to look to teachers to implement research and not Government.
"I am a great believer in teachers becoming researchers and unless they do, you (academics) might affect policy development but there is no way you will have any impact on implementation."
Baroness Morris told The TES that the Government should have put teacher research at the centre of the next stage of school reform instead of spending the six months trying to change structures.
She said that as a minister she had got cross with researchers and said there was "friction and a lack of understanding" between the political and academic worlds.
On issues such as private schools, mixed ability teaching and faith schools most politicians' positions were fixed.
She said: "The most successful ministers are those who know what they want to do the minute they get through the door of the DfES.That sits uneasily with an openness of mind that says ,'Can I see the evidence? Can I see the research so far'.
"As politicians our values sometimes makes it difficult for us to look at research. We come from a different culture. We think in a different way.
Media pressure, the elections politicians have to win, which dictated timings for targets, and manifestos, could also conflict with research."
She said the best example in her era was the commitment to limit class sizes for five to seven-year-olds to 30. She had been aware of a local authority where schools had classes of 35 that were staffed by two adults.
"Evidence and common sense meant we should have said, 'We have drawn this policy in the best interests of children and there is no point in pushing it in this local authority'," she said. "But it is very difficult to do that."
Policies that were evidence- based, such as the numeracy and literacy strategies, had worked and it was important to develop a common vision and understanding between Government and researchers.
"The evidence of what can happen when you get that right is immensely powerful."