Sir

Although I agree with T. Clausen and O. B. Nielsen in Correspondence (”Reviewing should be shown in publication list“, Nature 421, 689; 2003) that peer review is a very significant factor in the quality of the scientific literature, their suggestion that peer-review activities should be shown in scientific CVs has practical problems, as they themselves note.

An alternative solution might be for journals to send a letter to their reviewers each year stating how many manuscripts they have reviewed, with some associated measure of quality. This verifiable information should become one criterion for assessment exercises, and would also improve and maintain the general standard of the peer-review system.

The efforts of reviewers should not be underestimated. Even a short and exceptionally well-written manuscript takes at least three or four hours to review properly: more commonly this task takes a day or more. Most manuscripts are revised by their authors and reviewed again by the original reviewers. Two or three reviewers are involved in every manuscript, and about two-thirds of submitted manuscripts are rejected. Hence, on average, each published article has received about 10–15 days of reviewing activity.

Good reviewers may receive one or two manuscripts a month from each journal that knows of them. It follows that these scientists are spending a large percentage of their time on reviewing manuscripts that could otherwise be spent on research, writing and so on (although of course they themselves benefit from peer-review when they submit their own papers).

Editors of journals complain that it is becoming more difficult to attract good reviewers because university researchers increasingly need to earn 'scientific credits'. What is needed, therefore, is a change in attitude from university managers, boards, agencies and others who decide about grants, tenure, promotion and so on.

It has been (and will be) mentioned many times that the current system in which quantity is taken as a measure for academic achievement should be replaced by one that gives credit to quality (Nature 422, 259–261; 2003). Reviewers are chosen because of their quality: their standing in their own discipline and their ability to think critically. An endorsement by a journal could be one way to acknowledge this ability in real terms, by incorporating peer-review activities into the professional career structure. The more prestigious the journal to researchers in the field, the more weight could be given to peer-review activities for that journal by assessment committees.

If reviewers do not receive public credit, the better scientists will eventually no longer be prepared to do this work, which would then devolve to less good reviewers, and the standard of scientific publications would fall.