Two research groups on different continents are chasing the same question. Group A posts a preprint the day the manuscript is finished, then submits to a journal. Group B waits for peer review, which takes nine months. By the time Group B’s paper appears, Group A’s preprint has been cited a dozen times and the field regards Group A as having got there first, even though the two journal articles came out only weeks apart.
A preprint is a manuscript made public on a dedicated server before, or in parallel with, journal peer review. In physics, mathematics and computer science it has been routine for decades; in biology, medicine and the social sciences it has spread rapidly over roughly the past ten years. That does not make it the right default for every paper.
What you gain
- A timestamp. A preprint carries a public posting date and usually a DOI, evidence that you had the result at that point.
- Earlier readers. Your findings reach the community months ahead of the journal process.
- Free feedback. Colleagues read it and email corrections or missing references before formal reviewers see it.
- Something to show. Many funders and hiring committees accept preprints as research outputs, usually listed separately.
- Open access at no cost. Anyone, anywhere, can read the full text without a subscription.
What you risk
- No peer review yet. Any mistake is public. You can post a new version, but earlier versions remain available.
- Anonymity is broken. For double-blind journals, reviewers may find the preprint and know who you are. Most journals accept this, but be aware of it.
- Medical research. Unreviewed results can be reported by media as settled fact. Medical preprint servers screen more carefully and add warning labels, but the responsibility for cautious wording remains yours.
- Journal policy. Most journals accept previously preprinted work; a minority still restrict it or attach licence conditions. Check first.
- Patents. Public disclosure before filing can affect patentability. If the work has commercial potential, talk to your technology transfer office first.
Choosing a server by field
| Server | Main fields | Notes |
|---|---|---|
| arXiv | Physics, mathematics, computer science, statistics, quantitative economics | Moderated; first-time submitters may need an endorsement |
| bioRxiv | Life sciences | Basic screening before posting, DOIs assigned |
| medRxiv | Medicine and health sciences | Closer screening; requires ethics and trial registration statements where relevant |
| SSRN | Economics, law, management, social sciences | Widely used for working papers |
| OSF Preprints and community servers built on OSF | Psychology, sociology, education and many others | Integrates well with shared data, code and preregistration |
The rule is simple: post where your community reads. An education paper on a physics server is effectively invisible to the people who need it.
The licence: a small decision with a long tail
When posting, you choose a licence, typically the server’s own non-exclusive licence, which only permits the server to distribute the work, or a Creative Commons licence such as CC BY. Once chosen for a version, a licence generally cannot be withdrawn. If your target is an open access journal publishing under CC BY, a CC BY preprint rarely causes problems. If you are unsure where the paper will end up, a narrower licence keeps more options open. Check the journal’s policy in its author guidelines or in a policy database such as Sherpa Romeo.
After the journal accepts the paper
- Update the preprint to the accepted manuscript if the journal allows; do not post the publisher’s typeset version unless the article is open access under a licence that permits it.
- Add the DOI of the published article to the preprint page. Many servers link automatically, but not all.
- On your CV, merge the preprint and the published article into a single entry so it is not counted twice.
When not to post a preprint
- Early release could seriously mislead, especially in health.
- The target journal does not accept preprinted work.
- The data contain sensitive information not yet fully anonymised.
- The results are part of a pending intellectual property application.
- Co-authors have not agreed. Posting a preprint is a form of publication; every author must consent, just as for journal submission.
In practice: a pre-posting checklist
- All co-authors have agreed in writing.
- The target journal accepts preprints and your chosen licence.
- The manuscript has been proofread like a journal submission: numbers, figures, references.
- Competing interests, funding and ethics statements are included.
- Data and code are anonymised, or their sharing status is noted.
- The abstract is cautious and free of hype, particularly on topics likely to attract media interest.
- The server matches where your community reads.
When citing other people’s preprints, label them as preprints and check whether a published version exists before you submit your own paper; reviewers often ask.
Next step: open the author guidelines of the journal at the top of your shortlist, find the section on preprints and licensing, and decide from there. If everything checks out, posting the preprint on the day you submit to the journal is the simplest and most common approach.
Câu hỏi thường gặp
Do preprints count as publications?
A preprint has not been peer reviewed, so it is usually listed separately on a CV. Many funders accept preprints as research outputs, but assessment rules vary between institutions.
Can I submit to a journal after posting a preprint?
Most journals allow it, but some restrict preprints or set licence conditions. Check the author guidelines or Sherpa Romeo before posting.
Which preprint server should I use?
Use the one your field reads: arXiv for physics, maths and computer science; bioRxiv and medRxiv for life sciences and medicine; SSRN for economics, law and management; OSF-based servers for many social sciences.
Can someone steal my idea from a preprint?
A preprint has a public date and usually a DOI, so it is itself evidence of priority. The risk of being scooped is often greater when results circulate privately among a few people.
Should I update my preprint after publication?
Yes, update it to the accepted manuscript if the journal allows and add the DOI of the published article so readers can find the final version.
