You might think the official rankings of institutions for research carried out by the Higher Education Funding Council for England (Hefce) would be a fairly clear-cut business. But Professor Williams shows they aren't. It all depends what data you feed in and what kind of information you want to get out.
After the 1996 research assessment exercise (RAE) he looked at the different ways that league tables could be constructed and found that the table showing Reading in the worst light was the one ranking institutions according to the average grade of all staff members entered for the exercise. This was the table published by several newspapers, including the Independent the day after the results of the RAE were announced. It showed Reading in 31st position, after removing specialist institutions such as Cranfield and the School of Oriental and African Studies.
He then constructed another table, allowing for the fact that universities entered differing proportions of their staff, and averaging grades for all staff, regardless of whether they were entered for the RAE. This calculation showed Reading in 28th position.
Third, he looked at a table showing research power as calculated by the newsletter Research Fortnight. It added up all the departmental scores for an institution and ranked them after standardising Oxford at 100. By this criterion, Reading came in with a very respectable 19th position.
After the RAE results were turned into research grants, another table was compiled showing the money universities received for research. Again Reading came 19th in a league table.
Finally Professor Williams ranked universities by the percentage of their total Hefce grant received that went into teaching and research. This, he says, could be called "research intensity". By this measure, Reading came 12th.
So, the university is 31st, 28th, 19th or 12th, depending on what information you feed in. Professor Williams believes no single league table can show definitively how universities perform. It all depends on the indicator you use. "I am wary of simplistic indicators," he says. "And I think some of these tables around today are too simplistic."
Lucy HodgesReuse content