Let there be preprint citations!

Historically, in the biological sciences, deference has been solely given to peer-reviewed publications. This meant that when writing new manuscripts, there was only ever a need to reference peer-reviewed and published manuscripts. With the explosion in popularity (for really great reasons!) of preprint servers, like bioRxiv, arXiv, and peerJ preprints (among others), I think it’s time we take a look at our current citation model.

The NIH recently ruled to allow preprints to be referenced in grant proposals. This shift comes on the heels of a larger shift valuing quicker dissemtination of information. While the debate has been settled for now in grants, I have heard less discussion about how to handle preprint references in peer-reviewed manuscripts.

To me, the solution seems simple. Preprints should be allowed to be cited in peer-reviewed manuscripts. (This is currently allowed by many but not all journals.**) This allows for the quickest dissemination of information, credit to be given where credit is due, and readers to have access to the most recent findings. I simultaneously argue that there is value in the peer-review process. Accordingly, there must be some simple way in manuscripts to differentiate between preprint and peer-reviewed references, a practice, to my knowledge, not regularly handled by journals that do currently allow preprint citations.

To differentiate clearly but unobtrusively in in-text citations journals could simply use numbers for peer-reviewed literature (as is typically done currently) and letters for preprints. For example:

If your sentence is referencing a peer-reviewed paper, the citation would be as always [1].

However, when you want to reference a preprint, maybe the citation would simply be a capital letter [A].

This model would then allow for separation in the references section. First, a section for the numbers (peer-reviewed research). Then a section for the letters (preprints). For example:


Peer-Reviewed Research
1. Granger, Hermione. Title. Journal. August 11, 2017.
2. Kringle, Kris. Title. Journal. Dec 25, 2016.

A. Quimby, Ramona. Title. Preprint Journal. August 11, 2017 [cited 2017 Sep 12].
B. Prince, Diana. Title. Preprint Journal. September 4, 2017 [cited 2016 Sep 5].}

Okay, so we’ve got a suggested update to the old model. What about problems that may arise as a result of these suggestions. Here are a few with some possible solution…

Problem #1: Some journals’ in-text citations include the authors’ name and publication year, rather than a numeric: [Ellis et al, 2017]
Solution #1: If it is a preprint being referenced, maybe a simple superscript would do the trick: [Ellis et al., 2017P], where P denotes preprint.

Problem #2: Preprints often become peer-reviewed research. How will we keep track?
Solution #2a (if this occurs during while the manuscript is being drafted): It is up to the author(s) to ensure that their manuscript is as up to date as possible upon submission for publication.*
Solution #2b (if this occurs during while the manuscript is under review): Yup. Still the authors’ responsibility*, but a reminder or tool to ensure accuracy from the publisher would be more than welcome.
Solution #2c (if this occurs post-publication): If the preprint was cited and it later goes through peer review, that does not matter. You cited the preprint. You meant to reference the preprint.
*I currently use Zotero to track citations, but I would certainly switch my source manager to whichever best handles this new model should it be implemented. The citation manager that seamlessly identifies manuscripts that have since been peer-reviewed and has the capability to update my in-text citations and references section accordingly will gain a user….unless of course that manager is Zotero, in which case they keep a user!

As we move forward to a world that increasingly values preprints, we must be able to reference preprints in manuscripts. While preprints are wonderful, the peer-review process is also valuable, and references in manuscripts must differentiate between the two. Finally, going forward, a system that detects when preprints have been peer-reviewed and updates citations and references in manuscripts accordingly will be invaluable.

** Text added to clarify that many journals do currently allow preprint citation. Thanks to Valentine Svensson for pointing out this ommission on Twitter.