Last month, the 2018 Asian-Pacific Conference of the International Society of Managing and Technical Editors was held in Singapore, a city with both heritage districts and futuristic vistas. One of the sessions on day 1 was about online preprint platforms and their relationship to the international research landscape and to traditional research publishing. On day 2, I was one of three Committee on Publication Ethics (COPE) representatives to host an interactive workshop, where preprints were mentioned again. In fact, COPE has a new discussion document on preprints and is welcoming feedback from all stakeholders. Today is World Intellectual Property Day, so it is an appropriate day to look at preprints.
What are preprints?
A preprint is nothing to do with printing (although it can be printed out of course). It is actually a draft research article that is posted to a website by an author, and the website must be a special preprint server or platform (eg, not an author's own site, for-profit general repository, blog, or social media). It can be revised as many times as needed, but the earlier versions usually must remain on the preprint server. Most preprint servers allow the uploading of only research reports, with full methods and results.
Uploaded preprints are drafts that have not yet been published in a peer-reviewed academic journal. If a draft has been submitted to a journal, the preprint revision process must stop; the author should not keep revising it independently or include any intellectual property from the journal peer review process, such as revisions that were requested by the peer reviewers and/or journal editor.
If the article gets published, many journals require that the preprint version must then include a statement about publication status and must link to the final publisher version ("version of record"), via the publisher version's Digital Object Identifier (DOI). Usually, the final version or the penultimate version (which was peer reviewed and revised but not yet edited and/or formatted in the journal style, sometimes called a "postprint") cannot be posted on a preprint server, with or without the associated preprint. Even if a preprint server or repository allows this, the author must check the publisher allows it, depending on the author's rights policy, usage rights, as well as copyright ownership.
When are preprints made?
Preprints are early drafts, revisions, and even final drafts of completed research that are uploaded to preprint servers and not yet subjected to formal journal peer review (or accepted provisionally by a journal or publishing platform that uses a postpublication peer review model, such as F1000Research). By uploading a version before journal submission, researchers can reach their peers internationally and receive early feedback to improve the work, as well as gain offers of collaboration. Potentially, the feedback could take place faster and at a greater scale than traditional peer review, and the (preliminary) results are disseminated faster to other researchers.
Preprint commenting on a preprint platform can be compared to presenting a poster, paper, or slide presentation at a conference and receiving attendee feedback or networking with peers. As with previously published conference abstracts, and sometimes conference papers, peer-reviewed academic journals allow authors to submit the full paper for formal publication later; authors are asked to disclose prior presentation at submission and seek copyright permission if necessary. Some journals require and some request that any related or overlapping material be submitted for reference. However, conference materials have theoretically already been peer reviewed and are preliminary in nature.
Preprints are not viewed as being formally peer reviewed (only informally commented on) or published, but, in a final revised state or even as the only preprint version, they may be exactly the same as the version submitted to a journal. As such, some journals do not allow submissions to have appeared on a preprint server, especially if the journal eventually owns the copyright of published papers, does not allow any prior publicity (known as the Ingelfinger rule), and wants to avoid science primarily or only by media channels ("science by press conference").
The SHERPA/RoMEO website and a Wikipedia page lists the policies of major journals and publishers, respectively. If in doubt, authors should ask their future target journal before posting a preprint; if they have posted a preprint they should declare this when making a journal submission. Similarly, authors should check with a conference organizer if prior preprint posting is allowed.
Where are preprints?
Preprint servers or platforms now exist in most disciplines; a list can be found in the COPE Preprint Discussion Document. In general, they are public and are free to upload to and to use, and are non-profit. They also assign persistent identifiers such as DOIs so preprints are archived, can be version-controlled, and are citable. Preprint servers are not journals, are not indexed by Clarivate Analytics, and do not receive impact factors. Preprints are not indexed in Medline or SCOPUS but are searchable in Google Scholar and preprint aggregators such as OSFPreprints.
Most preprint servers state they allow only preprints, but one of the earliest, arXiv, set up in 1991 and pronounced archive, calls itself an e-print archive--that is, "a highly-automated electronic archive and distribution server for research articles" (not just preprints). In general, preprint platforms perform a screening process that lasts 1-2 days, to check for plagiarism and minimal criteria such as libel before posting. Currently, unlike journal submission, authors can post the same preprint to more than one preprint server.
In general, preprint platforms have a commenting function (notably not arXiv), but authors are not obligated to use the feedback. Some commenting sites allow peer commenting on preprints from different platforms in a convenient way and in one place, such as Academic Karma and Peer Community in, and the online journal club PREreview. In biology, preLights is a preprints highlight service run by a select team.
Why make preprints?
The main perceived advantage of preprints is the fast and wide dissemination of results so as to speed up research. Researchers may also upload their preprints to establish primacy/priority of an invention or discovery, and to avoid getting scooped by competitors, although journals may or may not recognize this purpose of preprints. Still, being assigned DOIs (or similar code for arXiv) and preprint server date/time-stamps allows authors to officially record the timing of their research outputs. Indeed, some funding bodies such as the National Institutes of Health in the United States and Wellcome and Medical Research Council in the United Kingdom have recently started to allow researchers to cite preprints as legitimate research outputs in project reports and grant applications.
Some preprint platforms make journal submission easier, by having agreements with journals and publishers to allow ingest of files and metadata into their manuscript submission systems. Some journals even assign editors as scouts who are on the lookout for potential submissions to the journal. However, authors should remember that the international rule is submitting a paper to one journal at a time.
Not all preprints get eventually published, and some authors may treat their preprint as the final version. It has been argued that allowing researchers to upload to preprint servers as the final destination means they can archive and share work that would otherwise be lost and thereby contribute to publication bias (the bottom drawer, filing cabinet, file drawer, or drawer effect/problem/phenomenon). Examples are being too busy or having generated weak, negative, or replication data. Another argument is that results obtained with public funds simply deserve to be given back to society for accountability and for the greater good.
Who owns preprints?
Preprint servers let authors retain copyright and usually allow them to choose the type of creative commons (CC) license dictating reuse rights for readers, ranging from quite liberal (CC BY) to non-commercial use and no distribution of any derivatives (CC BY-NC-ND). For the purpose of informing, non-commercial peer use, and commenting, the reader can access the preprint for free.
Technically, reuse of the preprint in a journal requires compatibility of CC licenses, and copyright should remain with the authors. However, after peer review, revision, and possibly reformatting, editing, and artwork, it can be argued the final version is a different, value-added product, or derivative work, and authors may assign copyright of that derivative to the journal or only grant a license to publish while keeping the copyright of the derivative. Some journals allow authors to retain copyright of the final version, whereas the journal owns copyright of only the paricular compilation of articles in the overall publication as a collective work.
Preprint servers are intended for research use only, and some have criteria for who is allowed to post a preprint. Usually though, the preprints can be accessed by the general public and also journalists, so most if not all preprint servers and platforms add a watermark or include a front-page notice warning users of the lack of formal peer review. Because work is made public in this way, even though not yet published in a recognized academic journal, intellectual property has technically been published and become "prior art" and hence affects future patent applications. Different jurisdictions have different "grace periods" to still allow patenting, but authors should check with their institution's technology transfer office or legal representatives.
How do preprints 'work'?
Non-profit preprint platforms may receive funding from an institution, donations, and funders. They work on the principles of open access (that is, free public access) and open science/research (that is, early and worldwide sharing for maximum research impact), in ways that both compete with, complement, and are compatible with academic journals and publishing platforms. Beyond initial screening of submissions, assigning of persistent identifiers and licenses, and hosting the server with or without a commenting system, there are few internationally standardized or agreed on practices. Although preprints are not widely regarded as formal scientific publications, they do share characteristics with them and resemble them.
In the near future, a new medical preprint platform (MedRxiv) will come online, containing medical information that may affect public health and patient behavior. Medical journals have strict rules and ethics, such as those in the guidelines of the International Committee of Medical Journal Editors and COPE, so it seems sensible that preprint platform providers more closely (re)consider the overall research landscape and implications of essentially a form of self-publishing. Some preprint platform providers already quote COPE guidelines, but it is unclear how those guidelines are or can be put into practice or if they need to be adapted for preprint platforms.
Join the discussion
Some emerging issues and concerns are outlined in the new COPE Preprint Discussion Document, and there is an opportunity in the COPE website to add your opinions and influence how the preprint movement and preprint ethics will develop.
Dr Trevor Lane
Education and Engagement Consultant
Edanz Group
Note: Edanz Group is a corporate associate member of COPE, and Dr Trevor Lane is an individual associate member and Council Member of COPE; the above represents personal views only.
