Impact will count for 20 per cent of scores in the REF, the mechanism for allocating about ?1.6 billion in annual quality-related research funding.

But conversations at the UK Research Conference in London last week were dominated by doubts about the strength and eligibility of various sources of impact.

Geoff Rodgers, Brunel University's pro vice-chancellor for research and a supporter of the impact agenda, said that although he had strong ideas about what constituted a good case study, he had "no idea" whether the REF panellists would agree with him.

His experience so far suggested that scientists and engineers tended to underestimate their impact and struggled to track the uptake of their research, while humanities academics were unsure whether mass outreach was worth more than a narrower, targeted approach.

ÖйúAƬ

He added that his institution would have some strong case studies in social science if only the government would legislate in the way its academics had recommended.

Ann Dowling, head of the department of engineering at the University of Cambridge and chair of one of the main panels in the REF, emphasised that the panels would be looking for evidence that the claimed impact was linked to specific research.

ÖйúAƬ

She said case studies involving outreach events would require evidence of how they had changed the audience's thinking, such as questionnaires completed afterwards.

But she admitted that her panel had not yet resolved every issue around impact.

Eric Thomas, president of Universities UK, said that demonstrating impact would be vital if the academy was to make a credible case for continued research spending.

"Sustained public investment in tough economic times has made us more visible and more accountable, and it could be argued that we must now strive harder to retain the respect once naturally granted," he said.

ÖйúAƬ

"The impact agenda is therefore something that should be internalised as a pivotal part of the sector's tacit pact with the public."

paul.jump@tsleducation.com.