Inconsistent standards and wasteful efforts dogged last year's research assessment exercise, according to Universities UK. Moreover, it told the funding councils that "the funding outcome from RAE 2001 is not acceptable".
The RAE - which informs how much money each university gets for research - came in for "strident criticism" after the funding councils invited organisations to contribute to a review of research assessment.
UUK used its preliminary response to attack the RAE.
UUK expressed concern that there was a lack of consistent application of standards between panels. This meant that in some subject areas everyone did well, while in others there was a much more mixed outcome.
For example, pre-clinical studies research across the country was deemed to be brilliant all round: there were entries from five institutions; two were awarded the top-rated 5* and three gained a 5.
At the other end of the scale, 34 institutions entered staff in the environmental sciences category. Two received the 5*, two got 5, ten got 4 - at which funding stops in England - two got 3a, seven got 3b and 11 got 2.
UUK also expressed concern that there was a diminishing rate of return from the amount of effort.
Its preliminary response states: "At the boundaries between grades, more and more effort was being put into differentiating between outcomes that were felt by many to be to a large extent already determinate.
"Arguments have been made that the same conclusions could be reached with much less effort and the extent of the subsequent redirection of funding is of the same order as the cost of the assessment process, which is hardly cost-effective."
UUK states that research funding should be allocated using a selective method that is transparent, accountable, proportional, consistent and targeted.
The closing date for the initial consultation by the national funding councils is November 29.
The research assessment steering group, which is chaired by Sir Gareth Roberts, president of Wolfson College, Oxford, will then devise a number of different models that will be presented to the funding councils in April.
A formal consultation on the shape of future research assessment is due to begin in May 2003.