Reviewing a paper again for another journal?

In our two most recent "how can we help you?" threads, two comments were posted (by the same reader?) on whether to accept a reviewer assignment for a paper that they already recommended rejecting at another journal:

I've been asked to review a paper in a journal that I've already reviewed in a different journal and recommended rejection. I haven't seen the new paper but the title is the same. I was quite sure the paper was not ready for publication then and its been about 3 months since then that I've received the request and so I think its unlikely the paper has substantially improved. Should I still take up the review request?

----

Should you review a paper that you've reviewed before for another journal that you recommended rejection there? I'm not sure if it seems unfair to them that they have to be reviewed by the same reviewer that already didn't like their work the first time and maybe deserve to have someone else read the work?

Good questions. Another reader submitted the following reply:

I have refused to review for a second time. Reason is that I believe that I may be mistaken. And being on the other end, my currently most cited paper (and among the top cited papers in the relevant debate) was rejected by a particular reviewer several times at different journals, always giving the same set of comments. (Almost prompting me to leave academics.) I think it would be bad to exacerbate one's mistake.

This is my policy as well, and for the same reasons. Peer-review is such a crapshoot, and referees so different in their judgments, that I think an author deserves a shot with different reviewers in different places. 

What do you all think?

Comments

[David Thorstad]

You've had your say. It's time to give someone else a chance.

[JDF]

I have the same policy, for the same reasons.

[one rejection is enough]

Agreed: rejecting twice for two separate journals is excessive. If you've already rejected the paper once, the right thing to do is to decline a second invitation to review.

Here's a story: I once rejected a paper for Journal A, and then was asked to review it for Journal B. I declined the invitation from B, the paper was eventually published in B (and, in my defense, in a far more cleaned up and clear version), and now has quite a few citations. I stand by my initial decision to reject the first version I saw, but for the sake of the profession and the author, I'm glad I didn't go after this paper a second time.

[Chris]

It sounds like the main reason given - that I may be mistaken - applies even if I recommended “yes” for the first journal. Should I still decline to referee for the second journal even when I recommended publish the first time around (on grounds that I may be mistaken?)
In one case where this happened I told the editor I’d do it for the second journal and I gave the same recommendation but they published it.

[Daniel]

Small comment: I don't see why OP is confident that the paper is unlikely to have been substantially improved in 3 months. It's not uncommon for me to substantially revise a paper in that time frame. (Of course I admit to not knowing the specific details of the case at issue.)

[one rejection is enough, but one acceptance may not be enough]

@Chris, FWIW, I would most definitely consider writing a second referee report for a second journal if I recommended publication the first time only to have the paper get rejected. It would depend at least somewhat on how much faith I have in the paper, and I've never been in this position, but if I felt the paper was strong and it was rejected by Journal A, I would certainly accept an invitation to referee again for Journal B.

It seems to me that the "I may be mistaken" reason has force when the outcome is negative, like a "shadow of a doubt" in legal deliberation, and with the acknowledgement that referee work (almost) always entails a shadow of a doubt. I've certainly rejected papers according to my best-formed judgment while acknowledging that my judgment is fallible. When I believe a paper warrants publication, I take my job to be doing what I can, within reason, to help it get published.

[op]

Thanks all, apologies for posting it twice. The first time I mistakenly posted it on the February thread and was worried that it might have been overlooked.

@daniel Perhaps my judgment is unfair (perhaps another reason not to review the paper again) but the first time I reviewed it, it was pretty badly organised and I could not even see any new or interesting point the paper was making over the existing literature.

[Daniel Weltman]

If the paper is unchanged then I'm not sure but I lean towards what others have said.

If the paper has been updated or if you are unsure if the paper has been updated (because you only see the abstract or title) then what I've done is email the managing editor or whoever saying that I've reviewed the paper before and recommended rejection and I'd be willing to review it again but if it hasn't changed I'd be recommending rejection. The one time I've done this the editor has said "thanks, we'll find another reviewer."

I quite thoroughly echo Daniel (who is not me) with respect to the possibility of the paper changing in 3 months. OP, it seems inconceivable to me that you could know a priori that someone can't change a bad paper into a good one in three months. A brand new paper can be written in 3 months. Surely a bad one can be rewritten to be good in 3 months.

[D.]

The important question is whether the paper has changed significantly from the previously rejected version. If it hasn't then i'd say decline to review. if it has, then you're not really refereeing the same paper. the problem is that it is often impossible to check this before accepting/declining an invitation to referee.

When I get asked to referee a paper that I recognize as a paper that i already rejected for a different journal, my policy in general is to flag to the editor that i already reviewed the paper (or at least a previous version of this paper), and let them decide what to do. Generally, they let me off the hook and ask somebody else.

[UK Postdoc]

I think another reason to not review for a second time (apart from the fact that one might be mistaken) is that the initial negative impression might linger even if the paper has improved. It feels to me like it's simply easier to notice flaws in a paper if one has seen an earlier version that was worse. So, it might be fairer to give over another reviewer who isn't biased in the same way and can read the paper at face value.