Best of the bright sparks

Martin Ince dissects the data to identify centres of science-based excellence

The Times Higher World University Rankings (November 5) provoked a global debate among academics and the media about overall institutional excellence.

In these tables, we begin to extend the analysis to faculty level. This week, we examine data on science and on engineering, including information technology.

The principal measure we have used is to ask our peer-review college of 1,300 active academics across the world for their opinions about the top universities in subject areas and regions of which they have expert knowledge. These data went to QS, a UK-based firm with expertise in this activity and a background in the assessment of MBA courses. Its sample was weighted by area and subject, and the resulting scores were normalised to a maximum of 200 for the top institution - Cambridge University for science, and the University of California, Berkeley, for engineering and IT.

We also display a measure of the influence of the research done at these top-rated institutions. This information comes from the Essential Science Indicators database compiled by Thomson Scientific (formerly the Institute of Scientific Information) and was put together for The Times Higher by its European affiliate, Evidence Ltd, based in Leeds. It covers the years from 1994 to 2003.

Because there is less data available for specific faculty areas, we have published information for 100 institutions rather than the 200 listed in the full World University Rankings. In addition, we have not aggregated the two measures we collected, partly because citations data are not available for all institutions. Instead, we have listed them in order of their peer-review success with their citations per paper listed alongside.

If the peer review and citations data for science had been aggregated at an equal weighting, Harvard University would push Cambridge into second place, with Berkeley, the Massachusetts Institute of Technology and Oxford University following.

The top three in our engineering and IT table would be unchanged, but the Indian Institutes of Technology would lose their fourth position because of their low number of citations, with their place being taken by the California Institute of Technology.

In coming months, we will publish data for biomedicine - omitted from this table for fear that its ferocious publishing and citation culture would overwhelm other areas of science - and on the arts, humanities and social sciences.

Links to tables in the statistic section:

World's top 100 universities in science  
World's top 100 universities in engineering and IT  
World's top non-universities in engineering and IT

These tables were compiled by Martin Ince (martin@martinince.com), contributing editor of The Times Higher . Thanks are also due to Nunzio Quacquarelli of QS (www.qsnetwork. com), Jonathan Adams of Evidence Ltd (www. evidence. co.uk) and their colleagues.