All work to be counted in RAE evaluation

Every research paper published by every academic in each university department will be graded for quality under the system planned to replace the research assessment exercise, The Times Higher has learnt, writes Zoe Corbyn.

Funding chiefs have developed the proposal to help judge research quality for science, engineering and technology subjects. It is a radical departure from the current situation in which academics are individually selected for entry to the RAE and submit only their four best pieces of work for peer review. The change forms the basis of the bibliometric component of the system, which will determine how billions of pounds for universities' research infrastructure will be distributed after the 2008 RAE.

The idea forms the backbone of advice delivered by the Higher Education Funding Council for England to John Denham, the Secretary of State for Innovation, Universities and Skills. If Mr Denham approves, Hefce will consult the sector on the proposal from late November.

Every publication produced by a department would be placed in a band according to the impact factor of the journal in which it was published. Scores in each band would be totalled to make up a departmental work portfolio.

"There would be a full profile of publications for the department within the university," said Paul Wellings, vice-chancellor of Lancaster University and chairman of Hefce's research committee.

"It would be a distribution of publications against their impact. There would be a spectrum in each department. Once you have the bibliometric component in place, you could run the model as often as you wanted to."

The plans were drawn up by a group at Leiden University in the Netherlands after Hefce asked it to find a quality indicator based on research publications or citations that could help determine funding allocations. Professor Wellings said the system would take into account disciplinary differences in publishing, but exactly which publications would be included in the bandings was a level of detail "too fine-grained" for the consultation to consider.

Hefce's next task is to combine the three measures - research income, postgraduate numbers and the bibliometric indicator - that the Government wants to see as part of the research quality indicators for SET subjects.

Hefce's current thinking is that it will set a common algorithm for all SET subjects as the default, The Times Higher understands. Seven peer review panels would then advise on whether it is appropriate in their subjects.

"Once we have agreement on the bibliometrics the trick will be how to sew together the other indicators to get the right framework," Professor Wellings said.

The new system for SET subjects will be introduced gradually between 2010- 11 and 2013-14. For non-SET subjects, light-touch peer review remains; the start date will be 2013.