Research directors will be tipped off about their chances of success before deciding whether to enter the next research assessment exercise, under plans being developed for the funding councils.
Universities would be ranked according to measures such as number of publications, external research income and number of PhD students in each unit of assessment.
It would be up to each institution to decide whether to accept a basic research grant based on these metrics or to apply for a higher level of funding through peer review but risk losing the lower grant.
The proposals emerged last week from the review of research assessment being conducted by Sir Gareth Roberts, president of Wolfson College, Oxford, on behalf of the funding councils. He outlined the plans at a briefing at London's Senate House and said that the latest model offered something for everybody.
Sir Gareth said: "The method of assessment should be chosen to match the character of the research. There should be greater emphasis on international benchmarking and involvement of panel members. We need greater transparency in the selection of panels and an obligation for them to conform to their published guidelines for assessment.
"Assessment should cover and reward a wide range of criteria, leading possibly to a research excellence profile. Institutions should submit all staff whose contracts oblige them to do research. Young people should not be disadvantaged by the assessment process."
A third of units are anticipated to be funded through the "research quality assessment" route, with the rest getting cash through the "research capacity assessment" route. However, the RQA route would attract five times as much money as the RCA route.
Subject areas would be grouped so that the number of units of assessment fell from 69 to no more than 20. This could be achieved by establishing sub-panels to assess, for example, all health-related subjects.
If a department or faculty were to be peer reviewed, all staff would be assessed. Assessment panels would decide how many departments were internationally excellent and should be funded.
To access research cash, institutions would also have to show that they met standards for how research informs teaching, how contract researchers and postgraduates are developed and managed, and how knowledge is transferred to those outside the institution.
About 10 per cent of research cash would go to centres of excellence, which would be peer reviewed outside the RQA route. The centres involve collaboration between departments or institutions. An institution the size of Imperial College London could expect to have a dozen centres of excellence.
Formal proposals will be published for consultation from May to September.
The funding councils will decide in October whether to adopt them.
It is hoped that the RQA will continue to be identical in England, Scotland, Wales and Northern Ireland, even though the different funding bodies have chosen to fund the results differently. The first assessment would be in 2008, and it would be repeated every six years.