Posting a preprint — sharing a manuscript publicly before or during journal peer review — has real, well-documented benefits: faster dissemination, an earlier public priority date, and open access to work that might otherwise sit behind a review process for months. But “post it and it’s all upside” is not an accurate picture. Before an author or a research office decides to preprint a given manuscript, it’s worth weighing five real, well-documented downsides against those benefits — not to argue against preprinting, but so the decision is made with the actual trade-offs in view.
1. Scooping and priority-claim risk
The most commonly cited author fear is that posting a manuscript publicly, before it has gone through journal review, exposes unpublished ideas or methods to be used by another group who then publishes first. In practice, most major preprint servers assign a timestamped, citable DOI at the moment of posting, which is itself evidence of priority — and this is one of the reasons researcher communities in physics, mathematics, and increasingly biology have moved toward preprinting specifically as protection against being scooped, not as a source of scooping risk. Empirical surveys of preprinting behavior generally find that authors’ fear of scooping is more common than documented instances of it actually happening. Still, the concern is genuine enough that it shapes real decisions: for a result an author considers highly time-sensitive or competitive, some choose to hold off on a preprint until submission, or coordinate posting timing with the target journal.
2. No independent peer review at the time of posting
This is the structural disadvantage underneath most of the others. A preprint server’s screening process checks for plagiarism, basic ethical compliance, and topical scope — it does not evaluate methodological soundness, statistical validity, or whether the conclusions are actually supported by the data. That means a preprint can contain errors, incomplete analyses, or overstated conclusions that would normally be caught (or required to be corrected) during formal peer review. For readers, this shifts the burden of quality assessment onto the individual reader rather than a journal’s editorial process. For authors, it means anything posted is visible in its rawest form — including any flaws a reviewer would otherwise have flagged privately before the work reached a wider audience. COPE’s discussion document on best practices for preprints addresses this directly, recommending clear “not peer reviewed” labelling on every preprint server page specifically because readers (and, at times, the media) don’t reliably make that distinction on their own.
3. Citation and version-tracking confusion
A preprint and its eventual peer-reviewed, published version are not always identical — text, figures, and even conclusions can change during review. Most preprint servers version postings explicitly (v1, v2, v3…) and keep prior versions accessible, which is good practice but also means a citing reader needs to know which version they’re pointing to. Confusion arises in a few concrete ways: a citation to the preprint DOI that never gets updated to point to the published version once one exists; a reader who cites conclusions from an early preprint version that were revised or removed by the time of formal publication; or duplicate/near-duplicate records accumulating in citation databases and reference managers because the preprint and the published article carry separate DOIs and separate metadata records. ICMJE’s guidance on overlapping publications places the responsibility for linking a preprint to its eventual published version on the authors, not the journal or the server — in practice, this is a step that’s easy to forget once a paper is finally accepted. See our guide on how to cite a preprint for the citation-format mechanics this creates.
4. Journal and publisher preprint-policy variance
Most major publishers today explicitly state that a preprint does not count as prior publication — but “most” is not “all,” and policies are set at the individual journal or publisher level, not universally. ICMJE’s own position is that preprint posting “does not itself constitute prior or duplicate publication,” but explicitly notes it “may influence a journal’s interest in or priority for peer review and publication of that work” — meaning even a permissive baseline policy leaves room for a specific journal or editor to treat a widely-circulated preprint differently. Springer Nature, for example, publishes a unified cross-journal policy stating preprints will not jeopardize consideration; other publishers document their preprint stance at the individual journal’s “Guide for Authors” level, which is more work for an author to check and easier to miss. This variance is a direct descendant of the historical strict “no prior disclosure” norm sometimes called the Ingelfinger rule (after a 1969 NEJM editorial policy), which most journals have moved away from but which a minority of specific outlets or article types can still echo in narrower form. We cover this specific risk, and how to check a target journal’s actual policy before you post, in a dedicated guide: Preprint Server Policies: Avoiding a “Prior Publication” Conflict Before Journal Submission.
5. Media and public misinterpretation of unreviewed findings
A preprint is, by definition, unreviewed — but once it’s publicly posted and indexed, nothing stops it from being picked up, summarized, or reported on as though it were a finalized, vetted result. This became a widely-discussed issue during the COVID-19 pandemic, when preliminary preprinted findings were repeatedly covered in general media and on social platforms without the “not yet peer reviewed” caveat being preserved through to the final coverage. For an individual author, the risk is that a flawed or preliminary early version becomes the version the public, policymakers, or even other researchers remember and cite informally — even after a more careful peer-reviewed version supersedes it. COPE’s preprint guidance specifically recommends authors discussing a preprint publicly or with journalists make the “not yet peer reviewed” status explicit every time, precisely because it does not travel with the finding by default.
Other considerations worth weighing
- Retraction and correction visibility. When a peer-reviewed article is corrected or retracted, that action is typically indexed and flagged prominently (e.g. via Crossmark, database retraction notices). Preprint servers vary more in how visibly a withdrawal or correction to a preprint is surfaced to someone who only ever saw the original posting or a downstream citation of it.
- Institutional and funder policy variance. Most major funders (NIH, Wellcome, and others) now explicitly recognize preprints as citable, fundable research outputs — but tenure and promotion committees, and some national research assessment exercises, don’t uniformly weight a preprint the same as a peer-reviewed article of record. An author or their research office should check local institutional policy rather than assume preprints are treated identically everywhere.
- Loss of reviewer anonymity conditions. Posting a preprint before submission can compromise a journal’s double-blind review process, since a reviewer can often find the preprint (and therefore the authors’ identities) with a simple search, even where the journal itself tries to keep authorship concealed during review.
Weighing the trade-off, not avoiding preprints outright
None of this is an argument against preprinting — it’s an argument for making the decision deliberately rather than by default. For most fields and most manuscripts, the speed, priority-claim, and open-access benefits of preprinting outweigh these risks, which is why preprint use has grown substantially across disciplines over the past decade. The practical response to each disadvantage above is usually procedural, not avoidance: check the target journal’s actual preprint policy before posting (see our journal-policy guide), label and link versions clearly, update the preprint record once a published version exists, and be explicit about “not yet peer reviewed” status in any public or media discussion. For a fuller side-by-side of what changes between a preprint and its eventual published version, see Preprints vs. Peer-Reviewed Articles: What’s the Real Difference?. For the mechanics of choosing where to post in the first place, see how to choose a preprint server and arXiv preprints: what they are and how to use them.
Frequently asked questions
Does posting a preprint count as prior publication?
Not under ICMJE’s guidance or the policies of most major publishers (Springer Nature, PLOS, and others explicitly say so) — but individual journals can and do vary, and a small minority of outlets still treat certain kinds of prior public disclosure more restrictively. Check the specific target journal’s author guidelines before submitting, don’t assume a blanket rule applies. See our guide to this specific conflict.
Can I take a preprint down after posting it?
Most servers allow a withdrawal, but the original posting typically remains visible with a “withdrawn” label rather than being fully deleted, since the DOI and any citations already pointing to it need to resolve to something. Policies on this differ by server — check the specific platform’s withdrawal policy before posting if this matters to your decision.
Does a preprint hurt my chances of getting into a peer-reviewed journal?
For the large majority of journals and publishers today, no — most explicitly state that preprints do not affect consideration. The genuine risk is narrower and procedural: forgetting to disclose the preprint at submission, or a specific journal with a stricter policy than the field norm. Both are avoidable by checking the target journal’s policy in advance.
How do I make sure people cite the right version of my work?
Update the preprint’s record to link to the published version once it exists (most servers support this directly), and use the published article’s own DOI in your own subsequent citations and CV/ORCID record rather than continuing to point to the preprint. See how to cite a preprint for format specifics across the interim.
