> From: "Johnson, Richard" <RICH92@psy.soton.ac.uk>
> Date: Wed, 31 May 1995 15:14:04 GMT
>
> Reading Peters & Ceci's article on peer review of journal articles, it
> seems that the peer review is "non-blind", ie. the reviewers know who
> the authors are and which institution they are from. It seems that the
> peer review process SHOULD ensure that publication is on merit only.
> Surely the fact that it is a non-blind process can only serve to open
> the door to such undesirable factors as "institutional affiliation,
> paradigm confirmation or theory support, editor-author friendship, old
> boy networks" (P&C, p188).
Some journals have blind review, some not. The reason those who don't
don't is that (1) identity can often be guessed anyway (from
self-citations, etc.), but, more important, (2) it helps to be able to see
the prior work an author has done. The idea of completely isolated
"merit" is a fiction: If you see a borderline result (the really good
stuff is not at issue, only borderline stuff is, where you're not sure
whether it should be accepted or not), it makes you more confident if
the author has successfully done something of the sort before. Peer
evaluation is an underinformed task. More information can improve the
decision; less is no help. The assumption is that the identity is a
handicap and will bias decisions against an author. It is not clear
whether the empirical evidence bears this out, though more studies are
needed.
Stevan Harnad
mailto:harnad@soton.ac.uk
This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:17 GMT