Preprints: What Role Do These Have in Communicating Scientific Results?

The traditional definition of a preprint is a document that has been uploaded to a preprint server, is freely accessible to the public, and has not previously been published in a journal. Posting generally occurs the day of submission or the next day. There is typically no formal peer review of the article before it is posted online; however, it is checked for plagiarism and offensive/dangerous content. The only requirement is that the article be scientific in nature. There is no typesetting or editing before posting online. Because they are not journals, preprint servers have no Impact Factor and authors retain copyright of their articles. Each online version of an article allows comments to be posted by the public. Once posted and publicly available, most preprint servers will not allow the author to withdraw or delete their article.

The original purpose of preprints was to make new scientific knowledge freely available to the public before traditional validation via peer review. However, the term “preprint” is a misnomer because there may never be a future print version. In fact, some authors intend for the preprint version to be the first and final version of their article. Types of scientific works published as preprints include original research, review articles, editorials, etc. The reasons for publishing preprints are varied (Bourne et al., 2017):

• Authors feel that the information is of importance (e.g., for public health reasons) and desire to share with the public prior to formal peer reviewed publication in a journal.

• Authors may wish to have public discussion on their work before it is formally peer reviewed and published in a journal.

• The paper has been submitted to an open access journal and has already been peer reviewed but is awaiting formal publication.

• The paper was submitted to one or more journals and was rejected.

• Authors wish to make negative data available to the public and don’t feel that it would qualify for publication in a traditional journal.

• The article may be a critique of a previously published paper rather than a full article in its own right.

• Authors are frustrated by the slow speed of traditional publishing and wish to accelerate dissemination of their results.

• Authors wish to establish priority of scientific findings.

Importantly, preprints and peer-reviewed journal articles serve different purposes. Peer-reviewed journal articles are reviewed by experts in the field to ensure the article’s quality and are used by many to gauge a researcher’s success. Most would agree that the peer review process generally improves the quality of work and promotes reliability and reproducibility, so this should be the final goal of most scientific research. But there are hurdles to peer reviewed publication and the time it takes to publish is a major concern for some. The median review time for journals has increased from 85 days to more than 150 days during the past decade (Powell 2016). Posting a preprint after the journal peer review process allows the information to be publicly available without delay.

A few of the more popular online preprint servers include arXiv, bioRxiv, and ASAPbio but there are others and the field is increasing. The first preprint server, arXiv (https://arxiv.org), was started in 1991 at Los Alamos National Laboratory by Paul Ginsparg for the purpose of immediate access to new advances in the field of physics. In 2001, it was moved to Cornell University and the fields expanded to include mathematics, computer science, quantitative biology, quantitative finance and statistics. In 2013 the preprint server bioRxiv (https://www.biorxiv.org) was introduced by Cold Spring Harbor Laboratory and publishes articles in the biological and chemical sciences. In 2016 a variety of funding agencies (http://asapbio.org/summary-of-the-asapbio-funders-workshop) started ASAPbio (http://asapbio.org) which publishes articles in the life sciences. No matter the scientific discipline or the topic, there is most likely an available preprint server.

If an author is considering publishing data as a preprint or just wishes to keep track of current scientific information, then there are various advantages and disadvantages to first consider and some of these are listed in Table 1 (Berg et al. 2016). Although data published as preprints does not undergo formal peer review and thus may not be wholly accurate, authors should be cautioned that poorly conceived or written work may have a negative effect on their reputation. The proponents of preprints hope that the scientific community will police itself in this regard.

Table 1.

Advantages and Disadvantages of Preprints^*

Advantages

Work is immediately available to the public and free of cost.

Democratizes the flow of information (equal access for everyone).

Information is globally available.

Authors can obtain feedback from a wider audience compared to a few peer reviewers.

Documentation of the history of ideas. Establishes priority of discovery since the preprint creates a “date stamp” and thus attributes the scientific findings to researchers.

May help early career scientists in terms of hiring and promotion.

Can use preprint citations in grant applications, progress reports, and curriculum vitae.

Can be used by review panels as evidence of author’s most recent work.

Clinical research benefits from open and timely access to new data.

Judgement of one’s work can be based on the merit of the data alone rather than the name or impact factor of the journal.

Reviewers can read public comments and researcher responses thus helping to make the subsequent formal review process more efficient and result in an improved final manuscript.

Preprints receive DOI (digital object identifier) numbers and can are citable.

Disadvantages

No guarantee that the study was designed appropriately or that the conclusions reached are supported by the data presented because there was no formal peer review.

Selection of articles to archive is based solely on the requirement that they be scientific (i.e. research doesn’t have to be current or unique or add in a positive way to the body of published literature).

Not all journals accept preprint posting before or after peer review.

Preprints are not given as much weight as peer reviewed manuscripts for those applying for promotions or new jobs therefore may not work for younger colleagues, who need to show a track record of journal publications to help advance their careers.

Annual reports from funding agencies focus on peer-reviewed research outputs.

Grant reviewers must review the entire preprint themselves because there was no formal peer review and therefore there may be a bias against applicants who cite material that has not yet been published.

No current mechanism to link preprints with subsequently published peer reviewed articles.

There could be premature posting of research to “stake claims” and establ0sh discovery priority.

Once a peer reviewed article is published, the preprint may be cited instead since there is no mechanism to link the two.

Data may be posted that raises ethical concerns.

If used by the press or public, may be promoted as “published work” rather than non- peer-reviewed data.

No method to screen preprint submissions for conflicts of interest.

Preprints are not indexed by services like PubMed, Scopus or Web of Science.

^*Some information adapted from Berg et al., 2016

There is a growing concern that preprints should be linked to the final published versions of manuscripts. This type of service is available for some preprint servers and may be available for all servers in the future. Also, peer reviewed articles are available in PubMed but preprints are not. To address this issue, PrePubMed was developed (PrePubMed, 2018) and provides a means to search for articles that are not indexed by PubMed. Articles can also be found with common search engines, such as Google, if sufficient information is known about the author’s name and/or article topic.

Some publishers and journals offer authors the option of posting preprints but the specific guidelines might vary. Toxicologic Pathology’s publisher, SAGE, has a policy on working papers or versions of papers posted on preprint servers and this policy is available online (https://us.sagepub.com/en-us/nam/prior-publication). According to this policy, “authors should alert the Editor when submitting their paper if they have posted it on a preprint server. Authors should not post an updated version of their paper on the preprint server while it is being peer reviewed for possible publication in the journal.” Importantly, SAGE does not support posting of revisions that respond to editorial input and peer review or the final published version to preprint servers. As long as these guidelines are followed, authors submitting to Toxicologic Pathology are welcome to seek prior public input on their research by first posting to a preprint server.

Preprints appear to be the future of biology (http://asapbio.org) and have already been embraced by many authors and scientific disciplines. As long as the goals of posting data are acceptable (e.g., not because of multiple journal rejections) then preprints can serve a valuable role in disseminating information and obtaining feedback from the scientific community. However, preprint publications should be an interim step in the process and the ultimate goal should be to publish in a journal that has a clearly stated and rigorous peer review process and follows best practices promoted by professional scholarly publishing organizations.