Every nugget of quality deserves our attention

Universities must ensure that Government recognises and rewards the best research wherever it is found, says Paul Wellings.

It is a disturbing thought that, for the non-science units of assessment, the audit period for the 2014 research assessment exercise started just over a fortnight ago. This, combined with the close of play on the 2008 RAE, the forthcoming Sainsbury review and the new Department for Innovation, Universities and Skills, is likely to trigger a prolonged period of institutional lobbying as universities jockey for position.

It is a fair bet that the university groups will set out their stalls in response to the Sainsbury review, encouraging institutions to decide whether they are business-facing or research-led. The Government will continue to fret about the cost of evaluating research and the potential to further concentrate research. Most universities will want to defend their positions, fuelling the debate on how to measure excellence.

UK research culture has changed dramatically in the past two decades, with much greater emphasis being placed on quality. More recently, we have been encouraged to think about conducting fine research with an eye to its application. It seems likely that this policy will accelerate. The creation of DIUS will lead to more support for programmes designed to increase innovation, and, for universities, a greater emphasis on applied research and engagement with the business community. In the process we appear to have moved towards an agreement that institutional scale is of second-order importance compared with the quality and utility of the research outputs.

At present, it is possible to evaluate institutions over several dimensions of research activity. As input measures, we know much about the total research income per academic and the relative Higher Education Funding Council for England institutional investment in research and teaching. As outputs, we can evaluate quality through the RAE, the vitality of the postgraduate research environment and the volume and relative impact of publications per academic.

The "top 20" rankings of research universities against estimates of each of these six measures can be derived from a number of national databases. When you conduct this exercise, about 40 universities appear in the lists. Half a dozen - Bristol, Cambridge, Imperial College London, Oxford, St Andrews and University College London - appear in the top 20 for all six measures, and 17 others appear in at least three. This latter group includes two specialist institutions - the London School of Hygiene and Tropical Medicine and the Institute of Cancer Research - while the rest are members of the 1994 Group and Russell Group. Scotland and each of the English regions has at least one university performing well, using these measures.

One can draw a number of conclusions from this. First, it is clear that research excellence is not a wholly owned subsidiary of any one lobby group. It is widespread, by geographical region and by institutional mission. Second, research excellence embraces a broad set of institutions. For example, by some measures only nine of the top 20 in the 2001 RAE appear in the list of the top 20 for research impact. This indicates that we need a wide set of metrics relating to research excellence if we wish to capture data to inform a complex policy environment.

While labels such as "research-led" or "world class" are handy as part of international marketing and in lobbying national government, they mask the way in which different stakeholders engage with higher education institutions. For example, most university-business interactions are driven by the quality of research teams within a department and the relevance of the work to a specific company. In contrast, measures of disciplinary esteem are normally located at the departmental level. This differs from the quality of the graduate school environment, which is driven by institutional-level policies and interventions.

All this suggests that, in the future, our measures of research excellence need to be pluralistic and scaled to the appropriate level. We cannot avoid the complexity of these problems. However, in the face of growing international competition for research funds, our worst response would be to abandon processes that measure and differentially reward outstanding research and its application.

So this autumn, as the lobby groups and individual universities respond to the Sainsbury recommendations and new DIUS and Treasury policies, there should be a dominant sectoral message. We should argue for the development of a system with multiple measures designed to reward research excellence wherever we find it.

Paul Wellings is vice-chancellor of Lancaster University and chair of the Higher Education Funding Council for England's research committee.