Data revealing the proportion of eligible academics each university has excluded from the 2008 research assessment exercise (RAE) will not now be published after complaints that guidance on eligibility was unclear.
The news was greeted with anger by the 1994 Group of universities, whose members submit a higher percentage of their research-active academics to the RAE than some larger Russell Group institutions.
The decision has also cast doubt on whether an equal opportunities review of the RAE, due to be carried out by the UK funding bodies, will go ahead.
The Higher Education Statistics Agency (Hesa) was due to publish data showing how many staff at each university were eligible for inclusion and whether or not they were submitted, in December, at the same time as the RAE results are released.
The RAE results show only numbers of staff submitted. Large institutions with many staff could achieve high grades by submitting only their best performers, disguising the fact their overall quality may be lower than at smaller institutions submitting a higher percentage of staff.
Hesa said it was not releasing the data because of an "inconsistency" in a Higher Education Funding Council for England (Hefce) circular on the submission procedure.
"The inconsistency arises over the definition of staff eligibility for inclusion in such submissions," a Hesa spokesman said. "This inconsistency means that institutions could have reasonably submitted data that are not truly comparable across the sector."
Some Russell Group universities have obtained legal advice that says there is sufficient scope for confusion in Hefce's definition of eligible staff to render the data already collected unreliable. Hesa has stopped gathering data and will not use the statistics already collected.
The lack of Hesa data means that comparisons of RAE performance might have to rely on a volume measure based on each institution's grade-point average and the number of staff assessed. Such a comparison would put the largest universities at the top. Last year's staff data returns could also be used.
Robin Sibson, Hesa's chief executive, said: "Hesa is disappointed to find that it is no longer possible to publish information relevant to the RAE that would have been of considerable interest both within the higher education sector and more widely. However, any material that we publish has to be based on quality-assured data, and such data will not now be available."
Steve Smith, 1994 Group chairman and vice-chancellor of the University of Exeter, said Hesa's inability to publish the data was "a blow to the credibility of the sector". "There is a clear need to publish an intensity measure in the interests of transparency, and also to ensure that a clear and accurate picture of research quality and research environment is available on publication of RAE results. Publication of the quality profiles by themselves will present only a partial - and in many cases misleading - picture," he said.
A Hefce spokesman said: "We believe that the RAE should be as inclusive as possible, and we have attempted to achieve this through the guidelines we issued at the start of the process. We still intend to review the RAE for equal opportunities in due course."
The Russell Group declined to comment.