Sir

The aim of blinding in peer-review is to improve the quality of reviews by removing any bias that might arise from knowing the identity of the authors. As your Editorial 'Working double-blind' (Nature 451, 605–606; doi:10.1038/451605b 2008) acknowledges, reviewing is not genuinely blind. You cite a study in which blinded reviewers identified at least one author on some 40% of papers (M. K. Cho et al. J. Am. Med. Assoc. 280, 243–245; 1998). The recognizable authors were presumably well-known, and anecdotally, these are the authors favoured by single-blinding. So the effect of double-blinding may be to increase the bias towards well-known researchers.

We need definitive data on this score. Meanwhile, journals are plentiful and varied and (although most use single blinding) they offer a choice from totally open to double-blind reviewing. Switching to a different practice might help some journals to attract good manuscripts from scientists influenced by the type of review process.

Diversity is rated highly in ecology and economists lecture us about the benefits of free markets. Why not apply the same norms to the way we decide which papers to publish?