This table shows that whatever stresses the US university system may be experiencing, its personnel cannot be faulted on their research output.

North American universities come close to a clean sweep on the measure of citations per staff member for the past decade - only ETH Zurich, at number three, breaks into the top ten. On this criterion the US also takes nine of the following ten places, with the Australian National University at 14.

This finding can be seen in a number of ways. To achieve these high impacts, academics have to be productive. It is possible for a researcher to write a single paper, such as Einstein's 1905 paper on relativity, that is an almost mandatory citation for all later authors in the field. But more often, highly cited authors are those who have produced many papers to cite. This favours institutions such as our overall citations winner, the California Institute of Technology, which has fewer than 700 undergraduates and is strongly focused on research.

It is also worth being in the right subject. Biomedicine and other areas of science clean up here because of their high publishing rate and their tendency to have many citations per article. People in this area write dozens of articles in a career, not a single big book.

ÖйúAƬ

These figures may reignite discussion of English-language bias in bibliometric measurement. Most journals indexed are in English, critics point out, and the members of their editorial boards tend to be in Boston rather than Bombay. Much has been done to counter this problem in recent years, and the superior performance of high-profile US institutions shown here is probably genuine.

Statistics section  

ÖйúAƬ

World university rankings 2004
Back to index page

Download full-colour version
(331K)