[Sigmetrics] The THE World University Ranking for Arts and Humanities is thoroughly unprofessional and should be retracted in toto

[Andreas Strotmann]

It was to prevent precisely the type of "research" represented by the Times
Higher Education World University Ranking for Arts and Humanities from
being published - or at least from being taken seriously - that I gave my
presentation on "an 80/20 data quality" requirement in professional
bibliometrics at ISSI 2015 in Istanbul.
According to its methodology description, roughly half the score for a
university in the Arts&Humanities ranking of THE is based on the
university's research performance as represented in a single database,
namely, Scopus. Publications (2/3 of this) count only if they appeared in
journals indexed in Scopus, and Citations (another 1/3 of this) only if
they occur in publications published in journals indexed in Scopus.
As I pointed out two years ago, bibliometric indicators are likely unable
to produce any meaningful results at all(!!!!) unless the data source used
is roughly 80% complete and accurate - for all entities studied, that is.
But for both publications and citations in most of the social sciences and
the arts and humanities, bibliometric databases are generally considered
excellent if they cover even (a highly skewed) 50% of publications in the
field; significant social science research communities have even been
measured at a mere 10% or worse coverage in these databases, and it is
unlikely that the arts and humanities fare much better than the social
sciences in this regard.
This makes the ranking worse even than random: half of it is skewed
completely by effects of institutional and national culture and the degree
to which they enforce acceptance of the primacy of the journals indexed in
citation databases as publication venues.

https://www.researchgate.net/publication/280924322_An_8020_Data_Quality_Law_for_Professional_Scientometrics
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20170916/dd8a78a2/attachment.html>