Research Assessment in the UK

Science  03 May 2002:
Vol. 296, Issue 5569, pp. 805
DOI: 10.1126/science.296.5569.805

In the 1980s, the United Kingdom's University Grants Committee responded to demands for greater accountability by introducing a Research Selectivity (later called Assessment) Exercise. The RAE ran over five cycles between 1986 and 2001. Panels in 69 discipline-based units of assessment (UoAs) awarded grades against performance data and strategy statements from every university and college funded for research, using criteria based on comparative national and international excellence. RAE grades not only assigned prestige but also weighted funding allocations. These powerful incentives led to measurable changes in efficiency. Resources, staff, and research students became concentrated in departments with world-class research profiles. Despite growing international competition, UK research improved relative to world benchmarks, reversing declines in the early 1980s. Average research performance, on the basis of publication and citation data, gained progressively from 1986 through 2000.

This policy for research assessment has driven the quality of university research so successfully that UK research funding has run into some unanticipated problems. The RAE may now have reached its limits. It was designed to drive core funding by rewarding high achievers in a system with a spread of performance, so academics worked even harder to achieve excellence because higher grades led to better resourcing. Now, with some 55% of UK academics in the top-rated category, the funding differentials are being flattened, and the resource pot hasn't grown as fast as the improvements. Shortfalls in expected funding are likely to affect morale, so the United Kingdom now needs a replacement assessment and allocation system to maintain the momentum of research excellence. It may also need to rethink its broader justification for the core funding of research in universities, which are set for further major expansion.

Freezing the present state of excellence is untenable, and it may not be feasible to differentiate among top performers in the present system or to assess them differently from the rest. The static grade criteria of the RAEcould be avoided by progressively raising the bar and applying grade normalization to each UoA in successive RAEs, but there is then the challenge of comparing excellence among subjects.* Any new system must continue to provide accountability for the use of public funds, must be at least as selective in distributing those funds to the places that will use them most effectively, and must be flexible enough to treat sciences and arts in a comparable way. It also needs to remain dynamic, so that funds can be reallocated between units and institutions as the peaks of excellence develop and shift. The dynamic nature of university research in the United States, the United Kingdom, and elsewhere is one reason why the scientific enterprise in these countries compares well with the national systems of research institutes in France, Germany, and Japan.

Many alternatives have been floated, but it is unclear that any would serve UK science well. Institutional self-assessment is unlikely to build confidence in the process, and if institutions were to assess one another, then disciplinary and regional politics might bias the outcome. The system could shift from cycles to a rolling assessment, across either institutions or disciplines, but this could be an expensive logistical nightmare. (The RAE has been relatively cheap to operate: The attributable costs are only 1% of the allocation.) The United Kingdom could simplify its dual support of core and project research funding by gearing the core against project grants, but the close correlation between funding streams disguises a crucial residual variance: There can be twofold grant income differences for institutions with similar core funding because of differences in disciplinary portfolios.

The RAE created incentives that channeled natural research competitiveness into a pervasive driver of excellence. It overcame initial opposition in the community by showing that those who deliver good research have little to fear and much to gain from well-structured scrutiny. So why is such an assessment system a rarity internationally? Other countries could benefit from the challenge that a dynamic allocation system presents.

  • *J. Adams, Nature 396, 615 (1998).

  • †R. M. May, Science 281, 49 (1998).

  • ‡J. Adams et al., The Role of Selectivity and the Characteristics of Excellence (report to the Higher Education Funding Council for England, HEPU, University of Leeds, UK, 1999).

Navigate This Article