Comparison of Europe's and US research performance

An interesting article appeared recently (Dec 20, 07) on the web, entitled Europe’s poor research performance, that uses the Highly Cited Scientists data of the Institute of Scientific Information (ISI), to compare research performance in Europe and the US. (The full article can be found here).

I have some reservations on the approach they use that I have posted in their site. My reservations are appended below.

I find your article, and the idea on which is based, interesting. However, I feel that your method fails to capture important aspects of the data used, thus arriving at comparisons/conclusions that are not necessarily representative of the true situation. The reason is that the ISI HCR’s index, assigns a scientist -and the totality of his output/achievements- to the latest Institution he is affiliated with, and to the country where this Institution is located. This may, sometimes, be accurate but quite often not entirely true. While such an index is certainly an indication of the quality of an institution (a good scientist would choose to go to a good place), it does not take into account the fact that the scientist may have been “nurtured” elsewhere.To give a real example I came across, Lopez-De-Silanes Florencio, an economist HCR “from” the Netherlands, has a Mexican origin (and first degree), got his Ph.D. from Harvard in 1993, was a Professor at Yale for 4 years, and has advised more than 10 governments around the world, before joining the University of Amsterdam in 2006. Seventy-eight of his 104 papers (contributing to his total number of citations) were published before he joined the University of Amsterdam. How can such a scientist be counted (solely) as a “European” in the US-Europe comparison? A similar problem of course arises with other indices used for scientific ranking, for example indices that give a substantial weight to the number of Nobel Prize winners affiliated with an institution. However a Nobel Prize winner, who is now at University X but did his award winning research at University Y, cannot be entirely “credited” to University X. Obviously, most of these indices favour the wealthiest institutions (beyond the obvious confounding effect).