How Secret Should Peer Review Be?

Between Labor Day on Monday and a problem with our server yesterday, we missed a big British study on peer review released Tuesday. Nature took note, emphasizing in particular an interesting if not wholly surprising aspect of the report: the 4000 scientists who were surveyed like their peer review nice and secret:

The surveyors were also asked to weigh in about what they thought could make peer review better. The idea of "open peer review," where reviewers names are made public, scored just 20% on the survey, while a whopping 76% of researchers thought that "double blind" peer review, where the names of authors and reviewers are hidden from each other, was a good idea. That contrasts with the last time the survey was done in 2007. Back then, 27% of survey participants supported open peer review, while just 71% wanted the reviews to be done double-blind. Incidentally, most Nature-brand journals don't use double-blind peer review.

Neither does Science, for what it's worth--peer-reviewers are anonymous to a paper's authors, but the peer-reviewers know who the authors are. Theoretically, open peer review would add new accountability to the process, which can appear unfair or biased against unestablished authors. Nature actually tried a four-month experiment with open peer review in 2006. The result?