Monday, April 4, 2011

The injustice of rankings

I've just seen the Top 100 Chemists list, as collected by ISI. In it, they report the impact these chemists have made by dividing the total number of citations by the number of papers in the years 2000-2010. So far, so good.

However, by doing the ranking myself, I noticed that the scientist with the most impact (Charles M. Lieber, 240 citations per paper on average) in fact does not belong to the top. He is not even in the top 100!!!
(according to Essential Science Indicators, he's actually number 123)

What is wrong? Am I looking at different things? No.
ISI has looked only at chemists with at least 50 papers.
This is reminiscent of the h-index, where not the total number of citations count, but also the number of well-cited papers; i.e. an author with one well-cited paper (J.S. Cottrell, or D.M. Creasy, or D.J.C. Pappin, all with 2715 citations for 1 paper) has a h-index of 1, while a scientist like Charles Lieber (75 papers, 18200 citations) has probably at least a h-index of 75 (according to WOS, in total his is 107).
In itself this filtering makes some sense, but can have some unwanted side-effects. For instance, what to think of K.L. Kelly with an average of 366 citations per paper? This might still be debatable, because 10 is still a small number of papers.

However, injustice is being done to (at least) two chemists: Mohamed Eddaoudi and George M. Sheldrick. The former obtained 13196 citations for 39 papers (338 on average!), while the latter obtained 10084 citations for 47 papers, with an average of 215 per paper. This would have made him fourth (counting Eddaoudi) or third (not counting Eddaoudi), were it not for the fact that 47 is just below the magic threshold of 50...

George and Mohamed, if you want to complain to ISI, you've got my support!

No comments:

Post a Comment

Robin, Talent of the year?

Summer 2015 . My wife's brother came to visit us for holidays, together with their six kids and the grandparents, for the first time s...