A group of academics who oppose the drive to measure the impact of research have accused Research Councils UK of taking a "dogmatic stance" over its refusal to reassess the issue.
Earlier this month, representatives of the group of 52 prominent academics, which includes 10 Nobel laureates, met a number of RCUK senior staff, including chief executive Alan Thorpe, to urge them to abandon the "impact" section of research council grant applications.
The group's coordinator, Don Braben, honorary professor at University College London, said that predicting impact was impossible and claimed that the requirement to do so would discourage scientific creativity and risk-taking.
"We do not know of any major scientific discovery that followed a predictable course. Therefore, RCUK's impact initiative encourages speculation and introduces unscientific criteria into the research selection processes," he wrote in a letter to Professor Thorpe ahead of the meeting.
Professor Braben said RCUK had declined to express the "scientific basis" of its position on impact. "In my long experience of dealing with officialdom, I have not come across such an overtly dogmatic stance. Alan Thorpe and his colleagues were completely unmoved by scientific arguments, and indeed, refused in effect to discuss them," he said.
He added that the group has been granted a meeting with David Willetts, the universities and science minister, on 26 July.
Last week, Mr Willetts announced a one-year delay to the implementation of the research excellence framework to allow the Higher Education Funding Council for England to "figure out whether there is a method of assessing impact that is sound and that is acceptable to the academic community".
At a conference earlier this month Mr Willetts said that predicting impact was much harder than assessing it in retrospect, as institutions will have to do under current plans for the REF. But he declined to call on RCUK to review its policy on impact and said he had delayed the REF because the issue had demanded an immediate decision from him.
In a statement, RCUK denies that the standardised "pathways to impact" section of grant applications requires applicants to predict the outcome or impact of research, but rather "to describe how they intend to open up the pathways to achieve impact".
It says the primary criterion for funding remains scientific excellence, with impact one of several "secondary criteria" that can be employed in borderline decisions.
RCUK has taken the concerns of Professor Braben and his colleagues "very seriously", the statement adds. "We are disappointed and concerned that there are still some erroneous perceptions surrounding RCUK's position on impact," it says.