A UK university hoping to be a world leader in research needs about 7,500 staff, according to an analysis for Times Higher Education.
Paul Whiteley, professor of politics at the University of Essex, has examined the relationship between the average score achieved by 121 institutions in the 2008 research assessment exercise and the total number of staff in each of the institutions according to data from the Higher Education Statistics Agency.
His analysis suggests that while larger institutions have higher average RAE scores, the relationship holds only up to a point. Once they have more than 7,500 staff, further increases may even be counterproductive.
|Critical mass: what’s the optimum number%3F|
|Institution||Total number of staff||Average RAE score|
|Thames Valley University||1,930||1.67|
|University of Teesside||2,760||1.99|
|University of Exeter||5,835||2.62|
|University of Liverpool||7,715||2.54|
|University of Nottingham||11,710||2.67|
|The Open University||15,515||2.52|
"It is certainly true that smaller institutions don't do well, but as they get bigger it doesn't necessarily mean they do better. There is no relationship between institutional size and RAE scores over the range 7,500 to 10,000 staff. Actually the chart suggests that RAE performance declines when an institution gets too big," Professor Whiteley said.
"The implication is that once you get bigger than 7,500, it doesn't help (RAE performance) very much."
He said the results suggest that the government policy of concentrating research funding in a small elite of large, research-intensive universities - which currently have about 10,000 staff - may be misguided.
"The claim that you have got to concentrate research in larger universities doesn't really hold up," he said. "If a place gets too big, actually research suffers."
He suggested that reasons for the effect may include the impact of greater amounts of bureaucracy.
Professor Whiteley acknowledged that his analysis was a departure from his personal research field, and that there were a number of complicating factors that his analysis had not considered, including the resources an institution devoted to research, the number of staff it had submitted to the RAE as research active, and the research culture of the university.
"It is necessary to take account of a variety of other factors when evaluating the relationship between institutional size and performance ... but nonetheless even if one does take this into account the size effect remains," he said.