In defense of signed reviews

Scientific publication goes through a peer review process. After submission of an article, a journal editor invites experts (reviewers) to comment on it and provide informed advice on the opportunity of publication. Most reviewers do not sign. In many journals, authors are anonymous as well but, in my experience, this practice is less pervasive. Anonymity on the part of the reviewer is supposed to preserve objectivity. The reviewer may indeed be biased by fear of punishment for unfavorable reviews or hope of rewards from favorable ones. The author's anonymity is meant to prevent evaluations on the basis of factors that are irrelevant to his or her work.

I am regularly invited to review papers. In spite of these norms, I decided a few months ago to sign my reviews. Here are the reasons why.

1. Anonymous reviewing is often unfair
2. It contributes to publication bias
3. It increases the likelihood of low quality or disrespectful reviews
4. It is ethically questionable
5. It contributes to an excessively high rejection rate
6. It fuels suspicion towards colleagues
7. It conveys a stereotype of scholars as not trustworthy and insincere
8. It limits the benefits of the considerable work involved in reviewing papers

Below, I detail each of these reasons.

1. First, whereas the reviewer can easily remain unidentifiable, the author's identity is much more difficult to conceal.  It can often be betrayed by citations of work in press or in preparation, by the location of the studies or simply by allusions to their previous work. Thus, the practice of anonymous peer review reinforces the imbalance of power between authors and reviewer that is inherent to their role.

2. This is especially disturbing when considering publication bias (the tendency to publish "significant", rather than null, results): The fragility of replications in (social) psychology is well known. When a "new" effect is found, it is much easier to publish than later unsuccessful attempts at replication. Often, papers reporting these attempts are sent to the investigators who published the primary effect. Those may have a vested interest in rejecting a paper questioning their findings. As they are not accountable to the authors, it is much easier for them to act in line with their interests (and criticize the paper) if they are anonymous (and may pass as "unbiased" reviewers) than if they aren't. Thus, it is very plausible that anonymous reviewing contributes to publication bias.

3. From an ethical point of view,  being identifiable to the person(s) whose work one is judging (in ways that can be highly consequential for this person) is certainly preferable.

4. Reviewing papers is a relatively unrewarding task that academics do voluntarily in addition to their many other duties. Obviously, there is a risk that the quality of their reports may suffer from these many other responsibilities. Whether we like it or not, being accountable to the person whom we are judging increases the likelihood that our criticism will be cogent and well articulated but also that the tone of the review will be respectful. Traffic norms may be an appropriate analogy: When drivers are isolated in their cars, they dare to engage in behaviors they would never display if they were face to face with other drivers. I am not immune to this influence: Since I started signing my reviews, I am much more prudent before clicking on the "submit" button. Do I want my name to be associated with the prose I just produced?"

(Postscript: While this argument makes sense to my imperfect mind, an inspection of the evidence from few studies on this is relatively mixed. Some studies find a significant, but very small increase in quality for signed reviews, others find no effect at all) 

5. Of course, reviewers are accountable to editors. So, preserving their reputation vis-à-vis editors may demand to write relatively thoughtful and well articulated reviews. However, given the high rejection rates (typically 85 to 90%) and limited journal space in (social) psychology, editors may often need from reviewers more good justification for rejecting than for accepting papers. Thus, accountability to editors places reviewers in a very different situation than if they were accountable to authors. Their accountability to editors may actually encourage them to provide arguments for rejecting papers. Besides, editors are sometimes so desperate for timely reviews that they may not "sanction" disrespectful behaviors (e.g., by mentioning in the decision letter that they disapprove of the tone or content of the review). And indeed, studies suggest that reviewers are more likely to reject papers when anonymous than identifiable.  

6. Something I noticed: When authors (including myself) receive anonymous reviews and dislike the reviews or suspect malicious intent on the part of the reviewer, they will often attribute it to a specific colleague or range of colleagues. This will be based on such cues as a citation made by the reviewer, the proximity between the arguments proposed and those echoed by a colleague at a conference, or even the style of the review. This very human tendency may fuel a negative feelings towards these colleagues. This may be especially unfair when these reviewers are misidentifed (this has happened to me several times). The atmosphere of suspicion and resentment generated by such reviews may be very difficult to dispel as it would involve breaking the taboo of revealing authors' identities. Obviously, resentment may also exist when reviewers are identifiable. But it will not be misdirected and the conflict can potentially be resolved. 

7. Most importantly, anonymous reviewing perpetuates a belief that reviewers can't be trusted when the masks are down. Just like riding in Brussels may convey the norm that drivers are dangerous egoists, anonymous reviewing may play a role in promoting an image of science as a "rat race" full of people who are fundamentally dishonest. According to the well known "false consensus effect", perceiving that a norm is shared may encourage us to endorse it (even if it is not actually shared), leading to a form of self-fulfilling prophecy. 

8. I believe that signing reviews may not only prevent some of these unfortunate consequences of signed reviews, but that it may also have advantages of its own. Anonymity does not prevent some reviewers from crafting extremely thoughtful and constructive reviews, full of ideas to improve the authors' work. Usually, these reviews go unacknowledged (except in the editor's letter) and the work involved in writing them may appear quite useless (as its only visible consequence may be to justify a rejection). Through signed reviews, authors and reviewers can connect. Thus, even after a rejection, an author could acknowledge the reviewer's contribution in a revision, contact the reviewer, ask him or her to elaborate on the comments made, make suggestions as to how to address them, dispel misunderstandings that may otherwise be attributed to the reviewers' bad intents, etc. This could potentially generate fruitful exchanges and even collaborations. Whereas the fruits of anonymous reviews often rot on the ground, those of signed reviews can generate new trees...

Thus signing reviews may not only contribute to the reliability and validity of the reviewing process: It can improve the wellbeing of the scientific community. This may, indirectly, have positive effects on the production of valid and reliable knowledge. 

In view of these advantages of signed reviews, their drawbacks should be considered as well. As all journal editors will tell you, it is quite difficult to find reviewers. If they are forced to disclose their identity, they may be even more reluctant to accept reviews (this assumption is not unwarranted, as this study finds). Note that this argument can be viewed as predicated on the assumption that anonymity is a veil that masks the bad quality of reviews. "Given that I don't have time to make thoughtful reviews, I don't want to project an image as an incompetent or unprofessional scholar". But, even if we accept this argument, it would still be possible to consider anonymity as an option, but not necessarily the default one (as it is now).

Another counterargument is that signing reviews may have the opposite effect of encouraging reviewers to downplay their critiques and be more positive than they actually feel. I accept this argument but my intuition is that signing reviews will mostly affect the tone of the review, and possibly the recommendation, rather than its substance, which should matter most to the editor. The study I just mentioned confirms this as well.

Undoubtedly, there are cases in which dissimulating one's identity as a reviewer may be perfectly legitimate because the reviewer can truly expect his or her judgment to be biased by self-interest. So, in the future, I may not sign all of my reviews. Yet, in my experience, these cases represent only a tiny minority. I haven't encountered any since I started signing (but you can't trust introspection, can you?). And, as I have suggested, anonymous reviews are also biased by the reviewer's self-interest in ways that are much more perverse.