Preprints: continuing the conversation

Preprint platforms have been common in physics and mathematics but the preprint landscape is changing rapidly with new platforms emerging across various disciplines. This raises opportunities for discussion across communities and for all those involved: preprint platforms, journals, authors, funders and institutions.

COPE has facilitated this discussion previously via an earlier forum discussion and with respect to peer review. To continue the conversation, COPE has developed a new ‘Preprints’ discussion document that sets out some of the benefits and challenges associated with the posting of preprints.

We have received valuable feedback since posting the discussion document and we’d very much welcome your thoughts and feedback as we discuss the key issues further:

• What are the risks if a preprint with a potential impact on public health is interpreted as established evidence?
• Could subject-specific differences influence the adoption of preprint servers, particularly in clinical fields?
• If preprints have DOIs, does that qualify them as prior publications?
• What standards exist for correcting preprints? Who is responsible? Could changes link to final published articles?
• Could researchers post a preprint that has been rejected from a journal?
• Do preprint servers address publication bias?
• Do preprint servers increase or decrease transparency?
• Who is responsible for checking if appropriate permissions have been received for the data and figures in a preprint?
• What if there is an authorship dispute on a preprint?

This will be discussed at the start of the next COPE Forum on Monday 30 April 2018. Please do leave any comments below, whether or not you are planning on joining the meeting

Comments are reviewed and, on approval, added below.

Please login as a member of register as a non-member to post comments


  • Posted by Foppe van Mil, 24/4/2018 9.50am

I can understand the critical position of the publishers well, unless you post the final paper (after peer review) as open access. Publishers current economic model includes 'selling' articles. With a paper already on a preprint platform, it is likely that other researches use the preprint version. Additionally, such a preprint version has usually not been peer-reviewed, and may be faulty or erroneous from a scientific perspective. I would not consider a preprint document a solid contribution to the kowledge-base, and therefore would rather not see such platforms publicly accessible.

  • Posted by Gunther Eysenbach, 24/4/2018 8.51pm

You may want to add the JMIR Preprints platform to your list in the appendix: It covers all manuscripts submitted to JMIR journals and has an emphasis on technology in health care/medicine (direct submissions to the Preprint server is also possible).

  • Posted by Trevor Lane, 30/4/2018 11.07am

Just some thoughts:

Ethics: Preprint platforms contain research information and are now part of the research/publishing cycle, so there need to be some governance and standards related to research ethics, rather than just providing cloud server space for researchers to self-publish drafts. Some requirements would be universal such as disclosing conflicts of interest and funding, and all authors agree to posting; some would be discipline-specific such as human ethics approval, participant consent, and trial registration.

Public accessibility: Because preprints have not been peer reviewed but are part of the open research movement and are open access, there should be clear warnings to users who are not researchers (not just saying “not yet peer reviewed” as the public may not understand that; perhaps add “For research use only”). Or users could be restricted to people with an ORCiD account. However, research theses/dissertations are also not yet formally published and sometimes publicly available too. If the funder mandates open access or public domain status, then the preprint would have to adhere to that too.

Media use: Some concerns are that the mass media may not understand that preprints are for research use only and they may contain mistakes and exaggerations, or lack limitations and caveats; unlike journal articles, preprints lack press embargoes; unlike conference reports and presentations (ICMJE guidelines say don't give any more information to reporters), preprints may be the full and nearly final version; and the altmetrics trend may encourage researchers to widely publicize and share preprints.

DOIs: Having DOIs just gives a persistent link to any object, and are useful as a link in a citation in an article. To distinguish the preprint from a research article, a reference should say Preprint (as recommended by the NLM). But the in-text citation would need to say it too, such as “According to a preprint by X et al”, rather than implying a published study. There would be opportunity to revise that if the preprint is published before publication of the paper, but if not, the preprint would have the DOI link to the published version. Linking would be a joint responsibility between publisher and preprint platform, with the latter also closing comments for the preprint version at that point. If there were any preprint comments that improve a paper between article acceptance and final publication, they could be used in a Correction.

Authors/Institutions need to know more about preprints as an additional avenue for research discussion, but also need to be aware of avoiding “research/science by press conference” and the impact of publicizing discoveries too early on patent applications.