Saturday, January 31, 2009

No Way But Up

Researchtrends, a newsletter published by Scopus, has an interesting item on the THE-QS World University Rankings. It suggests that the rankings can be used to assess the recent research performance of countries as well as institutions.

This is not a bad idea in principle but researchtrends has failed to examine the rankings closely enough. Its writer observed that universities in some countries have improved their position quite dramatically. Two Indian institutions in the top 200 have enjoyed a net change in rank of 248 between 2007 and 2008, eleven from the Netherlands a change of 230 and seven Swiss universities one of 217 while three Israeli institutions rose 194 places between them. Researchtrends describe the Indian achievement as "astonishing" and "testament to the continued development of research in India. "

The writer concludes

"This suggests that national improvements in ranking may be at least
partially the result of individual universities taking a more strategic
approach: targeting international publications, aided by bibliometric tools and building and promoting library collections"

Are they?

Looking at the scores on the citations per faculty part of the rankings, we find that The Indian Institute of Technology (IIT) Bombay had exactly 1 point in 2007 and 43 in 2008, apparently a truly remarkable increase in research or rather citations of research.

The IIT Delhi went up from 1 to 47, the Free University of Amsterdam from 1 to 38, Technion Israel Institute of Technology from 1 to 79 and the Ecole Polytechnique Federale EPF) Lausanne from 29 to 77.

A couple of American universities also apparently enjoyed spectacular rises in the citations score, Washington University in St. Louis (WUSL), already discussed in this blog, and Stony Brook University from 1 to 75.

In reality, there has of course been no spectacular increase in research or citations. What has happened is that in 2007 QS had problems with identifying certain universities or got them confused with others. They, or THE, took a while to decide whether there was one IIT or several and almost certainly got confused between WUSL and the University of Washington. One reader of the blog has suggested they got the Free University of Amsterdam mixed up with the University of Amsterdam. Something similar could have happened with the EPF Lausanne.

So, there may have been an improvement in research productivity over the last few years in India, Switzerland, Israel and the Netherlands. But that is nothing to do with the ascent of some universities from these countries in the rankings. That is testament only to errors committed by QS or -- let us give them some credit -- the correction of errors.

Thursday, January 22, 2009

Top Universities Guide

I have finally forced myself to buy and read the second edition of QS's book, retitled Top Universities Guide.

The first edition, which was called Guide to the World's Top Universities, was an unqualified disaster, full of errors the worst of which was getting every single student faculty ratio wrong as a result of somebody moving every university three rows down while copying data. Others included putting the Technical University of Munich in the profiles twice in positions 82 and 98 and listing an "Official University of California, Riverside". It also started to fall apart within a few days.

The new edition, however, tuned out to be a pleasant surprise. The previous errors seem to have been corrected and there do not appear to be any new ones. There is a new dignified purple cover and so far not a single page has fallen out.

This does not mean that the THE-QS rankings are faultless. Far from it. The peer review remains extremely biased, the international components are largely meaningless and the student faculty ratio is misleading and easily manipulable. Still, the book does show that in a technical sense, QS has improved quite a bit recently.

The change in title should be noted. There may be another reason, but this would appear to be a clever attempt to distance QS and the authors from the first edition without admitting that there was anything wrong with it.

And I wonder why Blackwell is no longer distributing the book.