- Why are you using double-blind reviewing?
- Do you really think blinding actually works? I suspect reviewers can often guess who the authors are anyway.
- Couldn't blind submission create an injustice where a paper is inappropriately rejected based upon supposedly-prior work which was actually by the same authors and not previously published?
- Some evidence about bias in merit reviewing been gathered below.
- What exactly do I have to do to anonymize my paper?
- I would like to provide supplementary material for consideration, e.g., the code of my implementation or proofs of theorems. How do I do this?
- Is there a way for me to submit anonymous supplemental material which could be considered by a reviewer before she submits her review (rather than potentially non-anonymous material that can only be viewed afterward) ?
- I am building on my own past work on the WizWoz system. Do I need to rename this system in my paper for purposes of anonymity, so as to remove the implied connection between my authorship of past work on this system and my present submission?
- I am submitting a paper that extends my own work that previously appeared at a workshop. Should I anonymize any reference to that prior work?
- Am I allowed to post my (non-blinded) paper on my web page? Can I advertise the unblinded version of my paper on mailing lists or send it to colleagues? May I give a talk about my work while it is under review?
- Will the fact that CSF is double-blind have an impact on handling conflicts-of interest? When I am asked by the submission system to identify conflicts of interest, what criteria should I use?
- What should I do if I if I learn the authors' identity? What should I do if a prospective CSF author contacts me and asks to visit my institution?
- The authors have provided a URL to supplemental material. I would like to see the material but I worry they will snoop my IP address and learn my identity. What should I do?
- If I am assigned a paper for which I feel I am not an expert, how do I seek an outside review?
- What criteria do I use for identifying potential conflicts of interest?
A: Our goal is to give each a reviewer an unbiased "first look" at each paper. Studies have shown that a reviewer's attitude toward a submission may be affected, even unconsciously, by the identity of the author. We want reviewers to be able to approach each submission without such involuntary reactions as "Barnaby; he writes a good paper" or "Who are these people? I have never heard of them." For this reason, we ask that authors to omit their names from their submissions, and that they avoid revealing their identity through citation. Note that many systems and security conferences use double-blind reviewing and have done so for years (e.g., SIGCOMM, OSDI, IEEE Security and Privacy, CCS).
A key principle to keep in mind is that we intend this process to be cooperative, not adversarial. If a reviewer does discover an author's identity though a subtle clue or oversight the author will not be penalized.
For those wanting more information, see the list of studies about gender bias in other fields and links to CS-related articles that cover this and other forms of bias below.
A: Studies of blinding with the flavor we are using show that author identities remain unknown 53% to 79% of the time (see Snodgrass, linked below, for details). Moreover, about 5-10% of the time (again, see Snodgrass), a reviewer is certain of the authors, but then turns out to be at least partially mistaken. For POPL'12, which used light DBR, author identities remained unknown to the PC 59% of the time, and PC members guessed wrong 23% of the time. So, while sometimes authorship can be guessed correctly, the question is, is imperfect blinding better than no blinding at all? If author names are not explicitly in front of the reviewer on the front page, does that help at all even for the remaining submissions where it would be possible to guess? Our conjecture is that on balance the answer is "yes".
A: I have heard of this happening, and this is indeed a serious issue. In the approach we are taking for CSF'15, author names are revealed to reviewers after they have submitted their review. Therefore, a reviewer can correct their review if they indeed have penalized the authors inappropriately. Unblinding prior to the PC meeting also avoids abuses in which committee members end up advancing the cause of a paper with which they have a conflict.
For authorsQ: What exactly do I have to do to anonymize my paper?
A: Your job is not to make your identity undiscoverable but simply to make it possible for our reviewers to evaluate your submission without having to know who you are. The specific guidelines stated in the call for papers are simple: omit authors' names from your title page (or list them as "omitted for submission"), and when you cite your own work, refer to it in the third person. For example, if your name is Smith and you have worked on amphibious type systems, instead of saying "We extend our earlier work on statically typed toads (Smith 2004)," you might say "We extend Smith's (2004) earlier work on statically typed toads." Also, be sure not to include any acknowledgements that would give away your identity.
A: Proofs and other supplemental textual material can be provided as clearly marked appendices, assuming the material is blinded, as with the main body of the submission.
In addition, on the submission site there will be an option to submit supplementary material along with your main paper. This supplementary material need not be anonymized; it will only be revealed to reviewers after they have submitted their review of your paper and learned your identity. Reviewers are under no obligation to look at this material. The submission itself is the object of review and so it should strive to convince the reader of at least the plausibility of reported results; supplemental material only serves to confirm, in more detail, the idea argued in the paper. Of course, reviewers are free to change their review upon viewing supplemental material (or for any other reason). For those authors who wish to supplement, we encourage them to mention the supplement in the body of the paper. E.g., "The proof of Lemma 1 is included in the non-anonymous supplemental material submitted with this paper."
Q: Is there a way for me to submit anonymous supplemental material which could be considered by a reviewer before she submits her review (rather than potentially non-anonymous material that can only be viewed afterward) ?
A: There is no official channel for doing this: the submission site only accepts potentially non-anonymous material. That said, there is nothing stopping an author from releasing a TR, code, etc. via an anonymous hosting service, and including a URL to that material in the paper. We point out this option not to encourage authors to exercise it, but to make them aware it exists, since we know of others who have used it. We emphasize that authors should strive to make their paper as convincing as possible on its own, in case reviewers choose not to access supplemental material.
Q: I am building on my own past work on the WizWoz system. Do I need to rename this system in my paper for purposes of anonymity, so as to remove the implied connection between my authorship of past work on this system and my present submission?
A: No. The relationship between systems and authors changes over time, so there will be at least some doubt about authorship. Increasing this doubt by changing the system name would help with anonymity, but it would compromise the research process. In particular, changing the name requires explaining a lot about the system again because you can't just refer to the existing papers, which use the proper name. Not citing these papers runs the risk of the reviewers who know about the existing system thinking you are replicating earlier work. It is also confusing for the reviewers to read about the paper under Name X and then have the name be changed to Name Y. Will all the reviewers go and re-read the final version with the correct name? If not, they have the wrong name in their heads, which could be harmful in the long run.
A: No. But we recommend you do not use the same title for your CSF submission, so that it is clearly distinguished from the prior paper. In general there is rarely a good reason to anonymize a citation. One possibility is for work that is tightly related to the present submission and is also under review. But such works may often be non-anonymous. When in doubt, contact the PC Chair.
Q: Am I allowed to post my (non-blinded) paper on my web page? Can I advertise the unblinded version of my paper on mailing lists or send it to colleagues? May I give a talk about my work while it is under review?
A: As far as the authors' publicity actions are concerned, a paper under double-blind review is largely the same as a paper under regular (single-blind) review. Double-blind reviewing should not hinder the usual communication of results.
That said, we do ask that you not attempt to deliberately subvert the double-blind reviewing process by announcing the names of the authors of your paper to the potential reviewers of your paper. It is difficult to define exactly what counts as "subversion" here, but some blatant examples include: sending individual e-mail to members of the PC about your work (unless they are conflicted out anyway), or posting mail to a major mailing list announcing your paper. On the other hand, it is perfectly fine, for example, to visit other institutions and give talks about your work, to present your submitted work during job interviews, to present your work at professional meetings (e.g. Dagstuhl), or to post your work on your web page. PC members will not be asked to recuse themselves from reviewing your paper unless they feel you have gone out of your way to advertise your authorship information to them. If you're not sure about what constitutes "going out of your way", please consult directly with the Program Chair.
A: Using DBR does not change the principle that reviewers should not review papers with which they have a conflict of interest, even if they do not immediately know who the authors are. To be clear, here is a definition of conflict of interest:
A conflict of interest is defined as a situation in which the reviewer can be viewed as being able to benefit personally in the process of reviewing a paper. For example, if a reviewer is considering a paper written by a member of his own group, a current student, his advisor, or a group that he is seen as being in close competition with, then the outcome of the review process can have direct benefit to the reviewer's own status. If a conflict of interest exists, the potential reviewer should decline to review the paper.
As an author, you should list PC members (and any others, since others may be asked for outside reviewers) which you believe have a conflict with you. While particular criteria for making this determination may vary, please apply the following guidelines, identifying a potential reviewer Bob as conflicted if
- Bob was your co-author or collaborator at some point within the last 2 years
- Bob is an advisor or advisee of yours
- Bob is a family member
- Bob has a non-trivial financial stake in your work (e.g., invested in your startup company)
Also please identify institutions with which you are affiliated; all employees or affiliates of these institutions will also be considered conflicted. If a possible reviewer does not meet the above criteria, please do not identify him/her as conflicted. Doing so could be viewed as an attempt to prevent a qualified, but possibly skeptical reviewer from reviewing your paper. If you nevertheless believe that a reviewer who does not meet the above criteria is conflicted, you may identify the person and send a note to the PC Chair.
A: If at any point you feel that the authors' actions are largely aimed at ensuring that potential reviewers know their identity, you should contact the Program Chair. Otherwise you should not treat double-blind reviewing differently from regular blind reviewing. In particular, you should refrain from seeking out information on the authors' identity, but if you discover it accidentally this will not automatically disqualify you as a reviewer. Use your best judgment.
A: Contact the Program Chair, who will download the material on your behalf and make it available to you.
A: PC members should do their own reviews, not delegate them to someone else. If doing so is problematic for some papers, e.g., you don't feel completely qualified, then consider the following options. First, submit a review for your paper that is as careful as possible, outlining areas where you think your knowledge is lacking. Assuming we have sufficient expert reviews, that could be the end of it: non-expert reviews are valuable too, since conference attendees are by-and-large not experts for any given paper. Second, if you feel like the gaps in your knowledge are substantial, submit a first cut review, and then solicit an external review. This is easy: after submitting your review the paper is unblinded, so you at least know not to solicit the authors! You will also know other reviewers of the paper that have already been solicited. If none of these expert reviewers is acceptable to you, just check with the PC Chair that the person you do wish to solicit is not conflicted with the authors. As a last resort, if you feel like your review would be extremely uninformed and you'd rather not even submit a first cut, contact the PC Chair, and another reviewer will be assigned.
A: The conference review system will ask that you identify
conflicts of interest when you get an account on the submission
system. Please see the related question applied
to authors to decide how to identify conflicts. Feel free to
also identify additional authors whose papers you feel you could
not review fairly for reasons other than those given (e.g., strong
More information about bias in merit reviewing
Note that this information was put together by the program chair; not all program or external review committee members are necessarily persuaded by it.
Michael Hicks introduced light double-blind reviewing for POPL 2012. Following the conference, he polled both paper reviewers and those submitting papers, and found all were significantly in favor of the process. Since then, POPL has adopted this form of review on a more permanent basis (as per its Principles of POPL document). Details about the surveys and other elements of the process can be found in the POPL'12 Chair's report (also available here, with supporting materials).
Kathryn McKinley's editorial makes the case for double-blind reviewing from a computer science perspective. Her article cites Richard Snodgrass's SIGMOD record editorial which collects many studies of the effects of potential bias in peer review.
Here are a few studies on the potential effects of bias manifesting in a merit review process, focusing on bias against women. (These were collected by David Wagner.)
- There's the famous story
of gender bias in orchestra try-outs, where moving to blind
auditions seems to have increased the hiring of female musicians by
up to 33% or so. Today some orchestras even go so far as to ask
musicians to remove their shoes (or roll out thick carpets) before
auditioning, to try to prevent gender-revealing cues from the sound
of the auditioner's shoes. Here is a
recent article from The Guardian.
- One study found bias in assessment of identical CVs but with names and genders changed. In particular, the researchers mailed out c.v.'s for a faculty position, but randomly swapped the gender of the name on some of them. They found that both men and women reviewers ranked supposedly-male job applicants higher than supposedly-female applicants -- even though the contents of the c.v. were identical. Presumably, none of the reviewers thought of themselves as biased, yet their evaluations in fact exhibited gender bias. (However: in contrast to the gender bias at hiring time, if the reviewers were instead asked to evaluate whether a candidate should be granted tenure, the big gender differences disappeared. For whatever that's worth.)
- The Implicit Association Test illustrates how factors can bias our decision-making, without us realising it. For instance, a large fraction of the population has a tendency to associate men with career (professional life) and women with family (home life), without realizing it. The claim is that we have certain gender stereotypes and schemas which unconsciously influence the way we think. The interesting thing about the IAT is that you can take it yourself. If you want to give it a try, select the Gender-Career IAT or the Gender-Science IAT from here. There's evidence that these unconscious biases affect our behavior. For instance, one study of recommendation letters written for 300 applicants (looking only at the ones who were eventually hired) found that, when writing about men, letter-writers were more likely to highlight the applicant's research and technical skills, while when writing about women, letter-writers were more likely to mention the applicant's teaching and interpersonal skills.
- There's a study of postdoctoral funding applications in Sweden, which found that women needed to be about 2.5 times as productive (in terms of papers published) as men, to be ranked equivalently. Other studies have suggested that the Swedish experience may be an anomaly. (For instance, one meta-analysis I saw estimated that, on average, it appears men win about 7% more grant applications than women, but since this is not controlled according to the objective quality of the application, it does not necessarily imply the presence of gender bias in reviewing of grant applications.)
- This study reports experience from an ecology journal that switched from non-blind to blind reviewing. After the switch, they found a significant (~8%) increase in the acceptance rate for female-first-authored submissions. To put it another way, they saw a 33% increase in the fraction of published papers whose first author is female (28% -> 37%). Keep in mind that this is not a controlled experiment, so it proves correlation but not causation, and there appears to be controversy in the literature about the work. So it as at most a plausibility result that gender bias could be present in the sciences, but far from definitive.
Snodgrass' studies includes some of these, and more.