Does Pre-Print compromise anonymity for a later peer-review?

Question

I'm from a social scientific field, so pre-prints are just becoming more and more "normal." Therefore, I'm a little in the dark regarding certain things. I'm planning to publish a pre-print of one of my manuscripts before sending it out to a journal. I tried finding answers to my questions, but they were pretty much vague or never to be found:

1. If I publish a pre-print and then I send it out to a journal, wouldn't the reviewers know my (and my co-authors) identity? Assuming they have read the pre-print by some other means or accidentally come across it. [please note that in our field, double blind-review is still the norm].
2. If my journal article is published, can I update the pre-print version with the updates that were made during the peer review?
3. Would the journal consider my pre-print as something that has "already been published"?
4. It is my understanding that I would be able to link my pre-print with the journal article once it is published. In that case, what would people likely cite: the article or the pre-print?

Answer 1

I think that most journals use only "single-blind" reviewing in which the author doesn't know the reviewer(s), but the reviewers have the author's name and affiliation. In such a case, pre-prints are no problem.

For those journals that use double-blind reviews it can be an issue. Reviewers can't be expected to avoid reading new work in their field, even if it is only available on pre-print servers. But some journals, not necessarily the same ones, won't accept a paper submission if it has appeared previously in any form, including a pre-print.

So, know the journal you want to submit to and act accordingly. If you expect double blind, then it isn't a good idea to "publish" prior to submission.

For the citation issue, it is hard to predict. People probably prefer to cite the official version, but may not want to pay for access, in which case they might cite the pre-print. Academics usually have ways to avoid paying (library access...), but sometimes just do the easiest thing.

Answer 2

If your submission is easily discoverable, through pre-print or because it’s somehow publicly available, then indeed referees could infer the names and affiliations of the authors. That’s the “risk” you take, against the benefits of staking your claim and possibly getting some pre-submission comments.

On arXiv, it is possible (and even recommended) to update submissions. In particular, it is possible to include details of the journal version (journal name, doi etc) as stand-alone or as part of an update. I don’t see a good reason not to update, but different sites may have different update policies.

Preprints are not usually considered publications as they are not peer-reviewed, unless preprints are posted on a site that will have some sort of copyright restrictions. I don’t know of such sites with restrictions and posting there would defeat the purpose of producing “preprints”. Some journals (Nature for one) frown upon pre-publications as they feel it decreases novelty, and might not accept submissions already posted on preprint servers; such journals are the exception rather than the rule.

People can choose to cite both preprint and journal versions, sometimes one and not the other. Good journals often ask authors to update citations to journal articles rather than preprints.

Answer 3

Regarding 1 and 3: the best way to be sure is to directly ask the editors of the journal if a submission would be compatible with the existence of a preprint. Different journals will have different policies. Note that double-blind peer review is not necessarily incompatible with a preprint: some venues (eg NeurIPS, a leading conference in machine learning) uses double-blind peer review, encourages authors to submit a preprint, and discourages reviewers to check them.

Regarding 2 and 4: most preprint servers indicate if a preprint is now available as a published paper (see for instance on bioRxiv) and will provide a link to it.