A preprint is a full draft of a research study shared online before it goes through the peer review process. It’s typically published on a dedicated preprint site (typically hosted by journals, research institutions or open access/open science networks) where other researchers can leave comments in a sort of community peer review. These papers are usually eventually submitted for traditional peer review and publication. Just as presenting research at a conference as an abstract or poster provides researchers an opportunity to share early results and get feedback before peer review, publishing a preprint gives authors a chance to get (hopefully) constructive criticism and do revisions before submitting it for peer review. Preprints are assigned digital object identifiers (DOI numbers) as published papers are so that they can be cited.
Deeper dive
Preprints aren’t new. One of the oldest preprint servers, arXiv, hosted by Cornell University, dates back to the early 1990s. But they’ve typically remained in the physical sciences until recently. Nature Publishing Group ran the preprint server Nature Precedings from 2007-2012. Other than that, one of the earliest biology preprint servers still in use, bioRxiv, hosted by Cold Spring Harbor Laboratory, only dates to 2013. But preprints in clinical/medical research are new and increasing in popularity (PrePubMed, the source of the previous link, aggregates and indexes preprints just as PubMed aggregates and indexes peer-reviewed publications, but the two are not related.)
Preprints.org, a multi-disciplinary, open access nonprofit, is one of many sites today that exclusively hosts preprints. (See their How It Works page to see how preprints work specifically on their site.) Often, journals have their own online section for preprints, such as PeerJ Preprints. Strangely enough, probably the most comprehensive listing and description of preprint sites is on Wikipedia’s preprints entry.
At first glance, it would seem that publishing a preprint online would violate the Ingelfinger rule since the paper is appearing in full somewhere other than the journal it’s submitted to. But it turns out many journals don’t appear to worry about Ingelfinger anymore, at least not in the setting of preprint servers. The journal that first articulated the rule, NEJM, still will not publish studies that have been preprints, nor will a handful of other journals, including JAMA Network journals and Journal of Clinical Oncology. But the vast number of journal networks — including behemoths Wiley and Elsevier — and top journals — such as Science, PNAS, Cell and The Lancet — do accept preprints. So do all journals published by BMJ and Nature Publishing Group. (Again, the most comprehensive list, with citations, for journals’ preprints policies is on Wikipedia.)
Preprints offer authors visibility and credit. (Since 2017, the NIH allows preprint citations in grant applications.) But the big draw is feedback, as oxytocin and heart rate researcher Dan Quintana explained in a Twitter story (with lots of fun gifs). “Would you rather write a weak preprint, which can be publicly evaluated and corrected before journal submission, or write a weak paper (that sneaks past peer review), which will be publicly critiqued AFTER publication, leaving no room for revision?” he asked and then answered, “Easy choice.” He adds in a later tweet, “Our current publication system is poorly set up to correct papers. Until this changes, it’s much easier to correct errors *before* submission, because errors are commonly missed by reviewers and editors — even the ‘classic’ papers.” Check out more FAQs on preprints at AJE Scholar (one of my primary sources for this entry), paleorXiv and Open Science Framework.