Ranking Web (Webometrics) of Universities

[Loet Leydesdorff]

> Using a mix of heterogeneous indicators will always permit one to say that the mix bag represent the “overall” activity of an institution. It is like saying that the sum of temperature and humidity gives the “overall weather” as opposed to temperature.

Dear Yves: I agree. We need a multi-variate model, including co-variates, etc. For example, the “document type” can make a difference. Furthermore, most distributions are skewed and thus non-parametric statistics are needed. “Average impact” is not good enough. 

A bientot (in Montreal),

Best,
Loet

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20120901/2c3e807f/attachment.html>