From: <[log in to unmask]>
Date: Fri, 14 Sep 2012 09:58:10 +0000

I responded to Chen's first article (for those of you who subscribe to
Serials Review:

http://dx.doi.org/10.1016/j.serrev.2011.01.003).

My point is: Google Scholar is an addition to subscription-based
databases, certainly not a replacement.

Google Scholar's interface has several disadvantages:

1. Google cannot handle complexity very well. Usually, a literature
search question is complex. Such a question will consist of a
combination of semantic components, and these components can be
weighed as well. A complex question can yield hundreds or sometimes
thousands of relevant references in subscription-based databases.
Because Google Scholar cannot handle complexity, vocabulary control or
term weighing in a proper way, results will lack recall.

2. Google's excellent coverage diminishes precision in search results.
A simple search in Google Scholar, say "rheumatoid arthritis," will
result in some very relevant references, but also in tens of thousands
of items which are not relevant at all. So results will lack
precision.

3. Google Scholar has no instruments to control semantics. For
example, Google Scholar cannot distinguish between depression (mental
disease) or depression (physiologic phenomenon). It also cannot
distinguish migration (geographic mobility) and migration (cellular
biology).

4. Google Scholar's gigantic coverage can be misleading. Now and
again, for very specific simple questions, recall of Google Scholar
can be incomplete. Recently, I searched for a very specific text
string: Health-F2-2008-223404. Google Scholar will yield 22
references, whereas in all combined subscription based databases, a
total of 64 references were found.

Dedicated databases are still very valuable, not only in science but
also in patient care and education.

Cordially,

Jan W. Schoones
Walaeus Library LUMC, Leiden, The Netherlands
http://www.lumc.nl/JanWSchoones