Which is more thorough: completely manual review of a topic within the
periodical literature, or completely automated review, using MEDLINE search
terms?

The reasonable answer is "both," particularly when conducted by experienced
librarians. It's my hunch, though, that, at least in the real world, each
method introduces a degree of error. Manual searches rely on distractible,
inconsistent, and mistake-prone mortals (though ones' experience in this
matter may minimize the degree of error). Completely automated searches will
miss references enjoying less-than-perfect indexing and so on.

Anybody aware of research that's attempted to quantify these error rates? I
have some hunches, but I'd rather start with hard numbers and go from there.


Best, L


Larry Rudiger, Ph.D.
Medical Content Developer
PKC Corporation
www.pkc.com
[log in to unmask]