How many manuscripts to review?

Question from a colleague about how many manuscripts to review:


I saw your note about being awash with review requests on Twitter. 

I'm just curious what your opinion is on the number of papers we should agree to review per month; i.e. What is professionally responsible?

I still learning the rules here on when is OK to say no to things.



My fairly off-the-cuff response (although it is something I've thought about over the years): 

Well, I don’t think there is any rule at all against saying ’no’; especially to review a paper. 

Before agreeing to review, I must:

1. Be very interested to read the manuscript

2. Confident I am qualified to critique at least 1 major aspect of the paper

3. Not be reviewing more than 2 other manuscripts already at the time (unless REALLY interested in it)

4. Feel I have a reasonable chance to be able to complete it in the timeframe they request (ie not too swamped with other stuff at the time). 

I suppose my rule of thumb-calculus is that every paper requires 2-3 reviewers (although more if submitted more than once), so to break even, we’d need to review 2-3 manuscripts for every publication — but don’t forget to divide by the number of co-authors of all your publications. So, for me, my papers have at least 2-3 co-authors almost always. Therefore, reviewing 1 paper for every publication feels fair to me. I’ve never discussed this with anyone before, so there could be some flaws in my logic.

But honestly, I don’t think about the quota, I think about #1-4. I get enough requests that agreeing to the interesting ones leads to enough (based on my rule-of-thumb calculus; which others may disagree with, of course).

If declining, I try to do so quickly, and recommend someone else.  

Also, I sometimes ask grad students to review papers that I”m asked to do. If they are new at it, I read their review, and let the editor know about that. It is good training for them, and can save a little time for me doing the full review. A few journals now have a formal process for that, I think it might be common in molecular biology.