8

I'm writing a contribution for a double-blind submission system.

Since I would like to supply more data than I could show on a small number of pages, I'm looking for a way to reference supplemental data in my manuscript, such as software scripts, figures or databases.

I could upload on GitHub, RPubs, Figshare and Zenodo, but all of these would expose my real user name (and I don't feel like making a throwaway e-mail account with a new repository for every anonymous publication that I intend to write).

Is there any way to get the above services to do what I want? Or is there a similar service that would allow anonymous hosting of research data?

It would be even better if I could later (e.g. after publication) declassify the information, with the link from the manuscript still being intact.

  • This sounds like a question whose answer is "arxiv". But I assume you are in a science which does not see that as an option? I think you should specify the science then, as it does matter. – darij grinberg Jan 23 '15 at 16:12
  • It's computer science, more specifically multimedia and user experience. Can you elaborate on why arxiv would not be an option? – slhck Jan 23 '15 at 16:37
  • I am not saying that! I am saying that if arxiv is an option (and arxiv is an option at least in the mathematical parts of computer science, so this is worth a try), then arxiv is an answer. You can upload your ancillary along with the actual paper on arxiv. – darij grinberg Jan 23 '15 at 20:14
4

What you could do is to host your data on Dropbox. When you try to share a file, you will get a link that you can put into the paper. It seems as if the recipient of the link has no way to see the dropbox user account name/e-mail address of the person who shared it. After acceptance of the publication, you could move the file to your institutional website.

Equally important, when using Dropbox it will be clear for the reviewer that you cannot see any server logs -- if you referred to a link on a server that you manage on your own, you could use the information to de-anonymize the reviewer (to some extent).

Other cloud service providers may also be suitable. I have no experience with them, though.

2

Most journals allow (and even expect) you to upload supplementary data with your submission, which can be accessed by the reviewers. If you are submitting to a journal, this is the best way to make it accessible as part of the reviewing package, as well as to ensure long-term storage of an archival copy of the relevant data.

On the other hand, if you are submitting to a conference (which typically have strict page limits and discourage supplementary information), it is often considered sort of cheating to have your paper depend critically on supplementary information. The reviewers would be acting reasonably if they did not choose to read your supplementary information in any case. I would thus recommend for a blinded submission to not include it, but instead say "[link omitted for blinding]" in place of the archival URL you will use for the final submission.

  • In this case, it's a conference. As mentioned in my question, I consider the data supplemental. It's not strictly needed to understand the paper or any of the conclusions therein, but just a nice-to-have addon for both interested readers and reproducible research. – slhck Jan 23 '15 at 15:27
  • @slhck Since it's not needed for the reviewers, I would recommend leaving the links out for the initial submission, then putting them in for the camera-ready. – jakebeal Jan 23 '15 at 15:29
2

If an author is very concerned about keeping their data private and anonymous until the last moment before publication, the workflow would be like this:

  1. The authors deposit their data in a private repository while they are finalizing the manuscript, before submission. They ensure that the documents and files they upload to the repository do not have their names on them.

  2. The authors obtain a 'view-only' link from the repository. This link will give anyone access to their private repository in a way that hides the repository metadata so that the authors of the repo are not visible to people using this link.

  3. The authors insert this view-only link into the text of their manuscript, for example in the data availability statement. Then they submit their manuscript for peer review. Their data remains private on the repository.

  4. Anyone with the manuscript, such as the editor and peer reviewers, can click on the 'view-only' link in the paper to download the data. The peer reviewers can anonymously inspect the data on the repository, and they do not see the names of the authors on the repository.

  5. Once the paper is accepted, the authors switch their data repository from private to public, and the 'view-only' link is removed from the manuscript, and replaced with the DOI to the public repository of the data. The editor or copyeditor checks that the paper now has a DOI, because the 'view-only' links typically expire after a few months.

The repositories that I know support this workflow are:

There may be others I don't know about!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.