I would be grateful if any of you could comment on current thinking regarding articles sent out to reviewers anonymously, that is, the reviewer is not aware of the author’s identity. Have any of you done this, or had articles reviewed in this way? What are your thoughts, both positive and negative?

John Stirling  
______________________________
I have never had an article sent out for review where my identity was NOT anonymous (blind review). Reviewers are also anonymous. We do this in our journal as well. In fact, when anonymity is inadvertently broken, we sometimes ask the reviewer to withdraw and always query the reviewer about their connection to the author.

I think it is a good idea, even though it is sometimes hard not to know (or guess) who wrote the article or review. However, it is also sometimes awkward, because, as an author, you should try to protect the blindness of the review by removing information that would identify you.

For example, I recently submitted a piece that was part of a series of papers. Although both the analyses and data were unique to that paper, it overlapped with other datasets and questions in such a way that I was afraid that the reviewers might worry about duplicate or salami publications. However, clarifying how this paper fit in with the others would make it obvious who I, the author, was.

It turned out my concerns were justified and the reviewers and editor did query. In my revision, I did add the material that placed this paper in context, and wrote in my letter to the editor about why I had left it out earlier. This satisfied everyone.

Nancy Darling
_______________________________
In contrast to Nancy, I'm skeptical about the merits and effectiveness of double blind/masked peer review.

In many cases, the reviewers can successfully guess the identity of the authors. The pattern of citation and the way that other work is referred to generally makes this very clear. I think that it becomes a bit farcical to think that the reviewers won't know the identity of the authors.

I think that there is a necessity for authors to declare competing interests, contributions, acknowledgments, and compare their work to their previously published and unpublished work. We want reviewers to decline to review if they have a connection to the authors that the editor missed—under blind review, they may feel less obligation to declare this to the editor if they guess the identity of the author. Conflicts of interest, salami slicing and duplicate publication are more likely to be missed under blind review. I had a case in which a reviewer's investigation into the institution of the authors due to some concerns about their patient recruitment revealed a major conflict of interest, and more analysis in light of this uncovered several flaws in the work. We rejected the manuscript, but we might have published under blind review as the reviewer would not have necessarily done that investigation.

Although we don't have this (yet) on our biology journals, in the medical journals in the BMC series we have entirely open peer review. The reviewers agree for their names to be known to the authors, the exact opposite of double blind review. It is the case with anonymous peer review that reviewers often give away who they are by their comments or by insisting upon citations to their own work, so it is pseudo-anonymous, and reviewers can use anonymous review as a shield from behind which they can hurl slings and arrows.

Open peer review does tend to make comments more polite and well considered, which I think counts greatly in its favour. I know that PLoS Medicine recently pulled back from encouraging open peer review on the pragmatic grounds that they found that many reviewers were uncomfortable with it. I'd note that being driven too much by what authors or reviewers want is not necessarily a good idea—authors would probably rather not declare competing interests, but all medical journals should now insist on this. The PLoS Medicine editors also noted that "we have found a bias in that reviewers are more likely to sign positive reviews", but this was under an uncontrolled system wherein the reviewers themselves chose whether to sign or not after completing their review, not a system where open peer review is compulsory, as with in the medical journals in the BMC series. We see many critical reviews returned under open peer review.

Even under anonymous peer review, anonymity should not be sacrosanct. Should reviewers wish to name themselves in their review, as I do see on occasion, I believe that the editor should allow them to do so.

Matt Hodgkinson
_____________________________
The following publication may be relevent to the ongoing thread:
Cho MK, Justice AC, Winker MA, Berlin JA, Waeckerle JF, Callaham ML, Rennie D.

Masking author identity in peer review: what factors influence masking success? PEER Investigators.
JAMA. 1998 Jul 15;280(3):243-5. Erratum in: JAMA 1998 Sep 16;280(11):968.
http://jama.ama-assn.org/cgi/content/full/280/3/243

David Schriger
______________________________
Although blinding the reviewers to the authors sends a good message (the journal wants the review to be fair and unbiased), such blinding has been well-studied and found not to affect the quality of the reviews or editorial decisions. And it is tough to do because authors often refer to their own work. To really blind a reviewer, one has to remove all citations to prior work.

Bill Tierney
_____________________________
There were a number of good quality studies of this issue presented at the Peer Review Congress and published in peer reviewed journals in the 1990s.

It is true that the studies found that blinding did not appear to affect quality of the review, which was simply a quality score assigned by the editor. What none of them investigated, however, was whether it affected the fairness or accuracy of the review, which is the crucial issue. Authors are not concerned that reviewers might do a sloppy review if they have a conflict of interest, they are concerned that their study will be distorted, misrepresented, or judged with undue harshness.

This aspect of the issue has not been objectively assessed to my knowledge. Even if it had been and found not to be present, the perception of authors might be unchanged. Reviewer blinding is like conflict of interest. A potential or assumed conflict is really almost as much of a problem as a real one.

Michael Callaham
_____________________________
This is not an uncommon practice and indeed many journals currently use this method of peer review. However, research conducted on this issue has shown that there is no significant difference between the results of review when we use blind or open review. You can find several of such presentations on the Web site of JAMA (The fourth and fifth Congress on Peer Review). Nonetheless, since most of the research was conducted in developed countries with a large scientific community, I'm not sure about its applicability to some developing countries with small medical communities, hence small numbers of researchers and referees. And that's why, for instance, I believe with all its shortcomings, perhaps, it is better to still use the blind peer review system (reviewers and authors do not know each other) in settings where we have access to a small number of referees within a small scientific community. If you can have a large database of referees, that's another thing.

Farrokh Habibzadeh
_____________________________
Of course this is a case of blind peer review and has been exploited in large scopes (not a current thinking). This kind of reviewing could prevent possible bias towards the writer. Unfortunately in daily experiences, we see that the name of the writer and his/her relation with the reviewer affects the reviewer's decision. I think it depends largely with the culture of the region in which the related journal is released. It is possible that an article written by an outstanding person, even of insufficient quality and data, would be accepted merely because of the writer's popularity and, the paper of a competing colleague be rejected unjustly. A query of the relationship between the writer and reviewer by the Editor-in-Chief can prevent such problems in open peer review systems.

MB Rokni
Associate Editor, Iranian J Publ Health
Associate Editor, Iranian Journal of Parasitology
_____________________________
We do this whenever an author requests it. It makes little difference to our review process. Most good reviewers can guess at the authors identity if they choose to do so.

AnnBarry Pless
_____________________________
I totally agree with AnnBarry Pless. We have the same procedure. Moreover, in a specialty journal (like ours), the number of researchers and the reviewers from the same country remains limited so a reviewer has  a good idea about the author even if we at the editorial board conceal the identity.

Muhammad Irfan
Associate Editor, Journal of Pakistan Psychiatric Society