2026.08.24

What Is a Preprint? The Complete 2026 Guide for Researchers

Key Takeaways:

A preprint is a complete research paper posted publicly, with a DOI, before peer review; it is citable, permanent, and free to read, but it has not been formally validated.
Preprints are now mainstream infrastructure: bioRxiv and medRxiv alone host more than 389,000 manuscripts, and roughly 80% of them go on to appear in peer-reviewed journals, on average 250 days after posting.
The decisions that matter most happen at the moment of posting: which server, which license, and whether a patent filing needs to come first. All 3 are difficult or impossible to reverse.
Preprints are no longer a stage before peer review; they are increasingly a place where peer review happens, through platforms such as PREreview, Review Commons, Peer Community In, and the eLife reviewed preprint model.

Table of Contents

Glossary of Key Terms

Term	What it means
Preprint	A full draft of a research paper made publicly available before peer review and journal publication.
Preprint server	An online repository that screens, hosts, and assigns identifiers to preprints, such as arXiv or bioRxiv.
Screening	The basic check a server performs before posting: plagiarism, scope, ethics statements, and obvious non-research content. It is not peer review.
DOI	Digital Object Identifier. A permanent, resolvable identifier that makes a preprint citable and trackable.
Version of Record (VOR)	The final, copyedited, peer-reviewed article published by a journal.
Author Accepted Manuscript (AAM)	The peer-reviewed manuscript accepted by a journal, before the publisher’s copyediting and typesetting.
Postprint	A loose term for the AAM. Distinct from a preprint, which precedes review.
Preprint review	Structured, often public and signed, evaluation of a preprint carried out independently of a journal.
Publish-Review-Curate (PRC)	A model in which the work is posted first, reviewed openly second, and curated or assessed third, with no accept or reject gate.
Overlay journal	A journal that peer reviews and curates papers already hosted on a preprint server rather than hosting them itself.
Endorsement	arXiv’s system requiring a first-time submitter in most categories to be vouched for by an established arXiv author.
Ingelfinger rule	A now-rare journal policy treating prior public posting as prior publication and grounds for rejection.
Interim research product	The NIH term for a preprint or other output that may be cited in grant applications and progress reports.
Scooping	Having your finding published first by another group. A preprint timestamp is protection against this, not a cause of it.
Altmetrics	Measures of online attention such as news coverage, policy citations, and social posts, distinct from citation counts.
APC	Article Processing Charge. The fee many open-access journals charge authors. Preprints carry no APC.
CC license	A Creative Commons license setting out what others may legally do with your posted manuscript.
Text and data mining (TDM)	Automated extraction and analysis of text, including the use of manuscripts to train machine learning models.
Grace period	A window in patent law during which an inventor’s own public disclosure does not destroy novelty. It varies sharply by country.
Withdrawal	Marking a posted preprint as no longer valid. The record and a notice remain; the content is not erased.

What Is a Preprint?

A preprint is a full draft of a research paper that is made freely available online before it has been peer reviewed or published in a journal. It carries a DOI, so it is citable and permanent from day 1.

Think of it as the manuscript in near-final form, complete with methods, results, figures, and references, shared openly while the formal review process catches up. Preprints are posted on dedicated repositories called preprint servers.

A preprint is not a lesser form of research output. It is an earlier one. The content is usually identical or near-identical to what will be submitted to a journal; what differs is the amount of external validation attached to it.

What a preprint is not:

It is not a summary, abstract, or conference poster. It is the whole paper.
It is not a peer-reviewed article, and no server claims otherwise.
It is not a duplicate publication under the policies of most journals.
It is not editable after posting. You add versions; you do not overwrite history.

How Do Preprints Differ from Published Journal Articles?

The core difference is validation, not content. A preprint is screened but not peer reviewed; a journal article is peer reviewed, copyedited, and typeset. Both carry DOIs and both are citable.

Feature	Preprint	Published journal article
Peer reviewed	No. Basic screening only	Yes
Freely available	Yes, always	Depends on the journal’s access model
DOI assigned	Yes	Yes
Citable	Yes	Yes
Typical time to public access	1-3 days	6 months to 2 years
Copyedited and typeset	No	Yes
Cost to the author	None	None, or an APC of roughly 1,000 to 12,000 USD (covered by the funder or author’s institution in most cases)
Considered final	No	Yes
Can be revised after posting	Yes, through versioning	Only through correction or retraction
Counts as prior publication	Almost never	Yes

How Big Is the Preprint Ecosystem in 2026?

Large, and still growing. bioRxiv and medRxiv together host more than 389,000 preprints from authors in over 170 countries, and 2025 was their busiest year on record. arXiv peaked at roughly 24,000 submissions per month.

Figures worth knowing before you decide:

bioRxiv received 49,256 new manuscripts in 2025; medRxiv received 14,973. That is more than 6,000 per month across the 2 servers.
Those servers record more than 8 million page views and 6 million downloads per month.
Manuscripts typically appear 48-72 hours after submission.
About 80% of bioRxiv and medRxiv preprints are later published in a peer-reviewed journal, on average 250 days after posting.
That 250-day figure is the practical case for preprinting: it is the length of the visibility gap you close.

The 80% conversion rate matters for a second reason. It undercuts the claim that preprinting replaces peer review. In practice it front-loads dissemination while the review process runs in parallel.

How Do You Post a Preprint?

You upload the manuscript to a server, it passes a screening check, and it goes live with a DOI, usually within 1-3 days. There is no editorial decision and no revision round.

Prepare the manuscript. Use the same version you are submitting, or about to submit, to a journal.
Confirm all coauthors agree. Posting without consensus creates disputes that are hard to unwind.
Check the target journal’s preprint policy and your funder’s requirements.
Choose a server appropriate to your field and, if relevant, your funder’s mandate.
Choose a license. This is a substantive decision, covered in detail below.
Upload, and complete the ethics, competing interests, funding, and data availability declarations.
Wait for screening. Staff check for plagiarism, scope, obvious non-research content, and, on health-related servers, potential harm.
The preprint goes live with a DOI and becomes publicly citable.
Respond to comments, requests, and review invitations.
Post new versions as the manuscript evolves, and link the journal article when it appears.

The arXiv Endorsement System

arXiv adds a requirement no other major server has: endorsement. Before your first submission to most subject categories, an established arXiv author must vouch that your work is legitimate research suitable for that category.

Endorsers are not evaluating quality or correctness. They are confirming basic suitability, which is why the request is a small ask rather than an imposition on a senior colleague.

Who Needs an Endorsement?

Most first-time submitters to a given category. You are automatically exempt in 3 situations, and arXiv will tell you at submission if endorsement applies.

You have previously submitted to that subject category.
Your email address is registered to an academic institution arXiv recognizes.
The specific category does not require endorsement.

How to Request an Endorsement

Identify an endorser active in your target category: a supervisor, collaborator, or senior colleague.
Start a submission in arXiv. If endorsement is required, arXiv generates a unique request link.
Send the link with a short explanation of the work. Include your ORCID or institutional page as evidence that you are a working researcher.
Do not send the full manuscript. At this point the paper is not yet posted and carries no timestamp, so circulating it broadly is the one genuine scooping risk in the whole process.
Once the endorser confirms, that category is unlocked permanently.

Major Preprint Servers by Discipline

The landscape now includes discipline-specific servers, multidisciplinary platforms, and regional repositories serving non-English research communities.

Discipline-Specific Servers

Server	Disciplines covered	Operated by
arXiv	Physics, mathematics, computer science, quantitative biology, statistics, economics, electrical engineering	Cornell University
bioRxiv	Biology and the life sciences	openRxiv
medRxiv	Clinical and health sciences	openRxiv, with BMJ and Yale
ChemRxiv	Chemistry and chemical engineering	American Chemical Society with partner societies
PsyArXiv	Psychological sciences	Society for the Improvement of Psychological Science
SocArXiv	Social sciences	University of Maryland with the Center for Open Science
EarthArXiv	Earth, planetary, and environmental sciences	Community-run, volunteer moderated
SSRN	Social sciences, law, economics, finance	Elsevier
RePEc and NBER	Economics working papers	Distributed volunteer network; NBER

Multidisciplinary and Regional Servers

Server	Coverage	Note
Research Square	All disciplines	Integrated with the In Review service of many Springer Nature journals
Preprints.org	All disciplines	Operated by MDPI
SciELO Preprints	All disciplines, Latin America and Iberia	Multilingual; strong in Portuguese and Spanish
ChinaXiv	All disciplines, China	Operated by the Chinese Academy of Sciences
AfricArXiv	All disciplines, Africa	Community-run; supports multiple African languages
OSF-hosted community servers	Varies by community	The generalist OSF preprint service stopped accepting new submissions in August 2025; community servers such as PsyArXiv continue

The ASAPbio preprint server directory maintains the most complete comparison of server policies and is worth checking before you commit, because policies often change.

How Do You Choose the Right Preprint Server?

Match the server to your field first, then to your funder’s mandate, then to the operational details: screening depth, license options, withdrawal policy, and indexing. Field convention usually settles it.

Ask these questions in order:

Where do the researchers I want to reach actually read preprints? Convention beats features.
Does my funder or institution require a specific server or license?
Does my target journal have a direct transfer arrangement with a particular server?
Does the server’s license menu include the license I need?
Is the content clinical? If it could influence patient care, the higher-scrutiny route is the correct one.

Screening and Speed

Server	Typical time to post	Screening approach
arXiv	1-2 business days	Moderator review by category; endorsement required for most first-time submitters
bioRxiv	48-72 hours	Checks for plagiarism, scope, and non-research content
medRxiv	48-72 hours, sometimes longer	Additional scrutiny for clinical risk, patient identifiability, and trial registration
ChemRxiv	1-3 business days	Screening for scope, safety, and research integrity
SSRN	1-3 business days	Light screening for scope and legitimacy
Research Square	1-3 business days	Screening plus optional editorial checks

Licensing, Withdrawal, and Indexing

Consideration	What varies between servers	What to check before posting
License menu	Some servers offer the full Creative Commons range plus CC0; others offer a narrow set or a restrictive default	Whether the license your funder mandates is on the menu
Default license	Defaults differ, and a default is easy to accept without noticing	The license actually selected on your submission form
Withdrawal	All major servers leave a public withdrawal notice; none delete the record	The exact wording the server will display in your place
Versioning and DOIs	Some servers mint a version-specific DOI; others keep 1 DOI with version suffixes	Which DOI you should give collaborators and cite
Indexing	Coverage in Google Scholar, Europe PMC, the Preprint Citation Index, and PubMed varies	Whether your work will surface where your readers search
Journal transfer	Some servers push directly into journal submission systems	Whether transfer saves you a resubmission step
Cost	The major servers are free to authors	Whether any optional service attached to the server carries a fee

Server policies are revised more frequently than any guide can track. Treat the tables above as a checklist of what to verify on the server’s own policy page, not as a substitute for reading it.

What License Should You Choose for Your Preprint?

Choose the most open license your funder, coauthors, and institution permit, which for most researchers means CC BY. The license is selected at upload, it is difficult to walk back, and it determines how your work can legally be reused.

This is the single most consequential decision most authors make without thinking about it. Servers present it as a dropdown; it functions as a contract.

License Options Compared

License	What others may do	Best suited to
CC BY	Reuse, adapt, translate, and mine, commercially or not, with attribution	Maximum reach; required by several major funders
CC BY-SA	As CC BY, but derivatives must carry the same license	Authors who want openness to propagate
CC BY-NC	Reuse and adapt for non-commercial purposes only	Authors wary of commercial republication, at the cost of some legitimate reuse
CC BY-ND	Redistribute verbatim only; no adaptations	Authors worried about distortion; note that this blocks translation
CC BY-NC-ND	Non-commercial, verbatim redistribution only	The most restrictive Creative Commons option; limits reach considerably
CC0	Anything, without attribution requirements	Datasets and authors prioritizing frictionless reuse
Server default, all rights reserved	Read and cite only; no reuse without permission	Authors with unresolved IP or commercial constraints

What the License Choice Actually Affects

Funder compliance. Some major funders require CC BY on the preprint itself, not just on the final article. A restrictive license can put you out of compliance.
CC BY-ND and CC BY-NC-ND block translations, which disproportionately limits readership outside English-speaking countries.
Text and data mining. Restrictive licenses limit legitimate systematic reviews and meta-analyses more reliably than they limit anything else.
Journal acceptance. Most journals accept preprints under any license, but a small number object to CC BY on the preprint because it complicates their own rights position. Check first.
Creative Commons licenses cannot be revoked for copies already distributed. You can relicense future versions; you cannot claw back the past.

The Benefits of Posting a Preprint

The advantages cluster into 5 areas: priority, feedback, visibility, open access, and career evidence. They compound, which is why researchers who preprint once tend to keep doing it.

Establishing Priority and Getting Credit

Posting a preprint creates a timestamped, publicly verifiable record of your work. Servers log the submission date, usually to within 24 hours of posting, and that date is permanently open for anyone to check.

In a competitive field where formal publication can take 250 days or more, that timestamp is the cleanest evidence of priority available to you. It is more robust than a conference abstract and far more robust than a lab notebook.

Early and Broader Feedback

Traditional peer review draws on 2 or 3 reviewers chosen by an editor. A preprint exposes the work to the entire field immediately. Readers can:

Flag methodological problems before they enter the permanent record.
Suggest datasets, controls, or analyses that strengthen the argument.
Point to related work you missed, which is often the most valuable form of feedback.
Propose collaborations that change the shape of the final paper.

Increased Visibility and Citations

Preprinting is associated with greater attention to the eventual journal article. Crossref links preprints to their published counterparts, so attention accumulated during the review period flows to the journal article rather than dissipating.

Open Access Without Article Processing Charges

Preprints are open by construction. Anyone can read them regardless of institutional subscriptions, and posting costs nothing. As APCs at open-access journals have climbed, this has become a substantive equity argument, not just a convenience.

Natalia Reinic Babic, Head of Journal Publishing and Open Science at IntechOpen, made the point directly in an Editage Insights interview, noting that with the growth of preprints, “the time taken for research to be shared is reduced.”

Benefits for Early-Career Researchers

Demonstrating output during the long gap between finishing the work and publishing it.
Building a visible record before a publication list exists.
Attracting collaborators and, in practice, attracting journal editors who scan servers for submissions.
Strengthening fellowship and job applications, since the NIH and other funders accept preprints as interim research products.
Protecting yourself. A timestamped preprint is leverage a junior researcher would otherwise not have in a priority dispute.

Preprint Review: The Layer Most Guides Miss

Most guides describe preprints as the stage before peer review. That description is roughly 7 years out of date. A substantial review ecosystem now operates directly on preprints, independently of journals.

What Is Preprint Review?

It is structured evaluation of a preprint carried out outside the journal system, usually published openly and often signed. The reviews attach to the preprint, are citable, and travel with the manuscript.

This inverts the traditional sequence. Instead of review deciding whether the work becomes public, the work is public and review becomes a layer of assessment on top of it. The shorthand for this is publish-review-curate.

Dr. Leonid Schneider, a cell biologist turned science journalist, put the case for the open version bluntly in an Editage Insights interview, saying there is “no better peer review than the open, signed and discussible review of a preprint,” and adding that he hopes more scientists will take part in it.

Where Preprint Review Happens

Initiative	What it does	Run by
PREreview	Open platform for requesting and writing preprint reviews; trains early-career and under-represented researchers as reviewers	Nonprofit community organization
Review Commons	Journal-independent peer review; the resulting reports can be transferred with the manuscript to affiliate journals	ASAPbio and EMBO Press
Peer Community In (PCI)	Communities of researchers that review and publicly recommend preprints; recommendations are free and citable	Researcher-led, nonprofit
eLife reviewed preprints	eLife reviews preprints and publishes the reviews plus an eLife assessment, with no accept or reject decision	eLife
Sciety	Aggregates evaluations from multiple groups so readers can follow reviewed preprints in one place	eLife
Biophysics Colab and similar colabs	Discipline-specific collaborative review of preprints	Researcher communities
MetaROR	Open review and curation for metaresearch and research-on-research	Research on Research Institute and partners

Why This Matters for Your Decision

A reviewed preprint carries evidence of quality that a bare preprint does not, which changes how it reads on a CV or in a grant application.
Review Commons and PCI reports can shorten the journal process, because affiliate journals accept the existing reports rather than starting over.
Requesting a PREreview is free, and early-career reviewers often engage more thoroughly than time-pressed senior reviewers.
Open review is still a minority practice. Schneider’s own diagnosis is that fear of retaliation, not indifference, is what keeps most researchers from reviewing openly.
AI-assisted feedback tools are entering this space; openRxiv began piloting structured automated feedback for authors in 2025. Treat these as a supplement, not a substitute.

Do Preprints Affect Patents and Intellectual Property?

Yes, significantly. A preprint is a public disclosure, and in most of the world a public disclosure before filing destroys patent novelty. If your work may be patentable, file first and post second.

Most preprint guides reduce this to a single line advising you to contact your technology transfer office. That advice is correct but insufficient, because the consequences differ sharply by jurisdiction and are irreversible.

Grace Periods by Jurisdiction

Jurisdiction	Grace period for the inventor’s own disclosure	Practical effect
United States	12 months under the America Invents Act	You can still file for 1 year after posting, but you lose the ability to file elsewhere
European Patent Office	None in general	A preprint posted before filing can bar patentability outright
Japan	12 months, with procedural requirements	Available, but the paperwork must be filed correctly and on time
Canada	12 months	Similar in effect to the US position
Australia	12 months	Similar in effect to the US position
China	6 months, narrow circumstances only	Should not be relied on for an ordinary preprint posting

The asymmetry is the trap. Researchers in the US often assume the 12-month grace period is universal. It is not. Posting a preprint before filing typically forecloses European patent protection permanently, even though your US position remains intact.

Practical Steps Before You Post

Ask whether anything in the manuscript is patentable. Composition, method, device, and use claims are the usual candidates.
If yes, contact your technology transfer office before posting, not after. This is the only step in the whole process that cannot be corrected later.
Consider filing a provisional application first. In the US this is comparatively inexpensive and establishes a filing date.
Check industry sponsorship agreements, material transfer agreements, and nondisclosure agreements. Many impose review periods before publication of any kind.
Check whether your institution requires invention disclosure before public release. Many do, and preprints count.

This section is general information, not legal advice. Patent law is jurisdiction-specific and fact-specific, and a qualified patent attorney is the right source for a decision about your own work.

How Do You Cite a Preprint?

Cite it like a journal article, but name the server in place of the journal, include the DOI, and label it as a preprint. If a peer-reviewed version now exists, cite that instead.

Formats by Style

Style	Pattern for a preprint citation
APA 7	Author, A. A., & Author, B. B. (Year). Title of the preprint. Repository Name. https://doi.org/xxxxx
AMA 11	Author AA, Author BB. Title of the preprint. Preprint. Posted online Month Day, Year. doi:10.xxxx/xxxxx
Vancouver / NLM	Author AA, Author BB. Title of the preprint [preprint]. Repository Name; Year [cited Year Mon Day]. Available from: URL. doi:10.xxxx/xxxxx
MLA 9	Author, First. “Title of the Preprint.” Repository Name, Day Month Year, doi:10.xxxx/xxxxx. Preprint.
Chicago (author-date)	Author, First. Year. “Title of the Preprint.” Preprint, Repository Name. https://doi.org/xxxxx
IEEE	A. Author and B. Author, “Title of the preprint,” arXiv:XXXX.XXXXX, Year.

Rules That Apply Across All Styles

Always include the DOI. Just the server URL is less stable and less useful.
Always label the item as a preprint. Readers and reviewers need to know the work is unreviewed.
Include the version number when the server uses one, because the content of v1 and v3 may differ materially.
Give the posting date of the version you actually read, not the date you accessed it.
Check the target journal’s instructions to authors. Some journals restrict preprint citations, require them in a separate list, or ask for them in the text rather than the reference list.

Citing a Preprint That Has Since Been Published

Cite the published version. It is the version of record, it has been reviewed, and its content will not change. Preprints can be revised after you cite them, which makes the citation unstable.

Cite both only when the versions differ in a way that matters to your argument, for example when the preprint contains an analysis that did not survive review. In that case, say explicitly why you are citing the earlier version.

Search for a published version before finalizing any reference list. Crossref links most preprints to their published counterparts, and roughly 80% of life-science preprints eventually have one.

How Do You Withdraw or Correct a Preprint?

You post a corrected version rather than editing the original, and full withdrawal leaves a permanent public notice in place of the content. Preprints are designed not to be erasable.

This surprises authors, and it is the reason the decision to post deserves more thought than the 5 minutes the upload form takes.

Your Options Compared

Action	What happens	When to use it
New version	A new version is posted; earlier versions stay publicly visible and citable	Revisions, added data, corrections after review, the accepted manuscript
Withdrawal	The content is replaced by a withdrawal notice; metadata and the DOI remain	The findings are invalid, or the work should no longer be relied on
Removal	The record is taken down entirely; rare and at the server’s discretion	Legal orders, privacy breaches, or serious ethical violations
Linking a retraction	The preprint is linked to the retracted journal article	The published version has been retracted

How Versioning Works

Versions are numbered sequentially and all remain accessible. Readers can compare them.
DOI behavior differs by server. Some mint a version-specific DOI; others keep 1 base DOI with version suffixes. Confirm which applies before you circulate a link.
Cite the version you read, and say so.
Post the accepted manuscript as a new version once the paper is accepted, if the journal permits it. This is a straightforward route to green open access.

If Your Preprint Turns Out to Be Wrong

Post a corrected version quickly, and state plainly in the version notes what changed and why.
If the core finding does not survive, withdraw rather than quietly revise.
Notify the journal if the manuscript is under review.
Contact authors who have cited the preprint, particularly anyone conducting a systematic review.
Do not delete social posts promoting it. Correct them, so the correction reaches the same audience.

Field-Specific Norms and Expectations

Preprint culture is not uniform. In some fields not posting is the unusual choice; in others posting still raises eyebrows. Know your field’s convention before treating general advice as applicable.

Physics, Mathematics, and Computer Science

arXiv posting at or before journal submission is standard, and in some subfields universal.
In machine learning, the arXiv version is frequently the version people read and cite, with the conference paper following.
Since October 2025, arXiv computer science requires documentation of successful peer review for review articles and position papers.
Not posting can itself be read as a signal, particularly for job candidates.

Life Sciences and Clinical Research

bioRxiv is the default for basic and translational biology.
Clinical work belongs on medRxiv, which applies heavier screening and expects trial registration numbers for clinical trials.
Content that could identify individual patients faces additional restrictions.
Several major biomedical funders now expect or require preprint deposition.
Clinical preprints carry real-world risk. Findings that could change patient care demand more caution before posting than any other category of research.

Economics, Finance, and Law

Economics has a working-paper tradition older than preprint servers, through NBER, IZA, and RePEc.
SSRN is the standard venue in finance and law, and in US legal scholarship posting is effectively expected.
Working papers in economics are routinely cited and often carry more weight than the eventual journal version.

Psychology, Social Sciences, and Humanities

PsyArXiv and SocArXiv are established, and preprinting is closely tied to the open-science reform movement in psychology.
PsyArXiv moved to moderating submissions before posting rather than after, in response to low-quality submissions.
Humanities adoption remains low, and monograph-oriented fields have no equivalent convention.
In fields where the argument rather than the data is the contribution, concerns about premature circulation carry more weight and deserve a hearing.

Regional and Multilingual Considerations

SciELO Preprints, ChinaXiv, and AfricArXiv serve communities that the anglophone servers underserve.
Most major servers are hosted in wealthy countries and dominated by English-language content, which is a real equity limitation of the current system.
If you post in a language other than English, check whether the server supports multilingual metadata, since discoverability depends on it.

Preprints in the Age of Generative AI

The most significant recent change to preprint servers has nothing to do with peer review policy. It is the volume of low-quality, machine-generated submissions, which began rising sharply in early 2025.

What Changed

arXiv moderators historically rejected roughly 4% of submissions. That figure has risen to an estimated 10-12%.
On October 31, 2025, arXiv announced that its computer science category would require documentation of successful peer review for review articles and position papers, describing the influx as unmanageable.
The generalist preprint service hosted by OSF stopped accepting new submissions in August 2025, citing the quality of incoming work.
PsyArXiv switched from moderating after posting to moderating before posting.
Moderators report that the overwhelming majority of machine-generated submissions come from first-time submitters.

What This Means for You

Expect more friction as a first-time submitter, and more scrutiny of review-style manuscripts.
Screening times may lengthen. Build that into your submission timeline.
Disclose AI assistance according to the server’s policy. arXiv and other servers have explicit generative-AI disclosure rules.
Genuine literature reviews now need a venue strategy. In arXiv computer science, that means securing peer review first.

Preprints as AI Training Data

Openly licensed preprints are among the most accessible sources of high-quality scientific text, and they are used to train and evaluate language models. Authors should understand this rather than be surprised by it.

CC BY and CC0 licenses explicitly permit text and data mining, including model training.
Restrictive licenses do not reliably prevent scraping. They mainly deter compliant researchers conducting legitimate systematic reviews.
Choosing a restrictive license to block model training generally costs you more in legitimate reach than it gains in protection.
If this concerns you, talk to your department chair or senior leaders at your institution about the question, rather than relying on a license dropdown to settle it.

How Do You Promote a Preprint?

Share the DOI rather than the PDF, write a plain-language summary, and actively solicit review. A preprint that nobody reads delivers none of the benefits described above; distribution is your responsibility, not the server’s.

Always circulate the DOI link, not a file. The DOI carries version information and download statistics; a PDF carries neither.
Write 3 or 4 sentences of plain-language summary and lead with those, not the abstract.
Post a thread with 1 clear figure. Figures travel; abstracts do not.
Email the preprint to authors you cite heavily. This is the highest-yield and most underused promotional step available.
Add it to your ORCID record, Google Scholar profile, institutional page, and, if applicable, your NIH bibliography.
Use a QR code on conference posters and slides. bioRxiv and medRxiv generate these for you.
Request a review through PREreview or an appropriate PCI community.
Watch the server’s download statistics and the Altmetric score to see which channels actually worked.
Post a new version when the paper is accepted, and announce that too. The second announcement usually outperforms the first.

Limitations and Risks You Should Understand

Preprints carry genuine costs alongside their benefits. An honest account of them is more useful than advocacy, and it is what reviewers and coauthors will expect from you.

Preprints Are Not Peer Reviewed

Screening confirms that a submission is plausibly legitimate research. It does not assess whether the statistics are sound, the controls are adequate, or the conclusions follow. That distinction is absolute and servers state it plainly. A preprint is not necessarily lower in quality than a peer-reviewed paper but it might be.

Media Misinterpretation Is a Real Risk

The COVID-19 pandemic made this concrete. Preliminary preprint findings were reported as settled science, and some were later shown to be wrong. Servers responded by adding prominent disclaimers to every preprint page.

If your findings are clinically actionable, weigh the risk of misreporting seriously.
Consider a plain-language summary that states the limitations explicitly, so journalists have the caveats in front of them.
Respond quickly to public misreadings. Silence is read as endorsement.

Journal Policies Still Vary

The large majority of journals accept manuscripts previously posted as preprints. A small number still apply the Ingelfinger rule and treat prior posting as prior publication. Verify before you post, not after.

The Open Policy Finder tool lets you look up preprint and self-archiving policies by journal and by funder. Disclose the preprint in your cover letter when you submit; this prevents plagiarism-detection software from raising false flags and helps indexing systems link the 2 versions.

Confidentiality, Sensitive Data, and Consent

Patentable material should be handled before posting, as covered above.
Data governed by sponsorship agreements, MTAs, or NDAs may not be shareable at all.
Human-subjects work requires the same consent and privacy standards as journal publication, and medRxiv screens for identifiability.
Dual-use research of concern requires institutional review before any public release.

The One Genuine Scooping Risk

Posting a preprint does not expose you to scooping. It protects you, because the timestamp is public. The real risk sits earlier: circulating a full manuscript by email while seeking an arXiv endorsement, before any timestamp exists.

Send the endorsement link and a short description. Do not send the paper.

Preprints in the Broader Scholarly Landscape

Preprints are increasingly treated as part of the scholarly record rather than as a workaround. Indexing services, funders, and evaluation systems have moved considerably in the past 5 years.

Development	What it does	Why it matters to authors
Preprint Citation Index, Web of Science	Indexes preprints alongside peer-reviewed literature; launched 2023	Preprints can appear on author profiles and accumulate visible citations
NIH Preprint Pilot in PubMed Central	Makes preprints from NIH-funded research discoverable in PMC	Substantially wider discovery for biomedical work
Europe PMC preprint indexing	Indexes life-science preprints and links them to published versions	Improves discoverability in the searches clinicians and researchers actually run
NIH interim research products policy	Permits preprints to be cited in grant applications and progress reports	Recent work can support a funding application before it clears review
Funder preprint mandates	Several major philanthropic and public funders now expect or require preprint deposition	Compliance may depend on posting, and on the license you select
Crossref preprint linking	Links preprints to their version of record	Attention accrued before publication follows through to the article

With regard to hiring and promotion decisions, preprints are widely accepted as evidence of productivity but are rarely accepted as equivalent to peer-reviewed publications.

Should You Post a Preprint?

For most researchers the answer is yes, with 3 exceptions that need resolving first: unfiled patents, restrictive sponsor agreements, and coauthors who have not agreed. Work through the list below before you upload.

Does my target journal permit preprints? Verify with Open Policy Finder rather than assuming.
Does my funder require a preprint, and does it require a specific license?
Is anything here patentable? If so, has my technology transfer office signed off?
Do sponsorship agreements, MTAs, or NDAs impose a review period before public disclosure?
Have all coauthors agreed, in writing, to posting and to the license?
Is the manuscript in a form I am willing to defend publicly, today?
Have I chosen the license deliberately rather than accepting the default?
Have I selected the right server for my field, my funder, and my content’s risk level?
Are my ethics approvals, trial registration, competing interests, and data availability statements complete?
Do I have a plan for distributing it and for soliciting review once it is live?
If the findings turn out to be wrong, am I prepared to post a correction or withdraw publicly?

Conclusion

Preprints have moved from a physics curiosity to core scholarly infrastructure in roughly 3 decades. They close a 250-day gap between finishing research and sharing it, at no cost, with a citable and permanent record attached.

The format’s remaining weaknesses are real: no validation gate, uneven field norms, and a growing volume of machine-generated noise that servers are actively fighting. None of these is a reason for a careful researcher to avoid preprinting.

What separates a preprint that helps a career from one that creates problems is the handful of decisions made at the point of posting: server, license, patent position, coauthor consent, and a plan for distribution and review. Get those right and the rest follows.

Frequently Asked Questions

Is a preprint considered a publication?

No, not in the formal sense. A preprint is a citable research output with a DOI, but it is not a peer-reviewed publication and should not be listed as one on a CV.

List preprints in a separate section labeled preprints or interim research products. Funders including the NIH accept them in grant applications, and the Preprint Citation Index makes them visible on author profiles, but conflating them with peer-reviewed papers on a CV is a credibility risk you do not need to take.

Can I submit a preprint to a journal after posting it?

Yes, for the large majority of journals. Major publishers including Elsevier, Springer Nature, Wiley, and PLOS permit submission of manuscripts previously posted as preprints, and some actively encourage it.

A small number of journals still apply the Ingelfinger rule and treat prior posting as prior publication. Check the specific journal using Open Policy Finder before posting, and disclose the preprint in your cover letter at submission.

How long does it take for a preprint to be published in a journal?

About 250 days on average for life-science preprints, and roughly 80% are eventually published in a peer-reviewed journal. That 250-day window is exactly what a preprint makes visible in the meantime.

The figure varies by field and by how many revision rounds the manuscript goes through. Papers that go through multiple journals after rejection take considerably longer, which strengthens rather than weakens the case for posting.

Do preprints count for tenure, promotion, or grant applications?

For grants, increasingly yes. The NIH permits preprints as interim research products in applications and progress reports, and other funders have similar provisions. For tenure and promotion, they usually count as evidence of productivity rather than as publications.

Practice varies sharply by institution and discipline. Ask a senior colleague on your committee how preprints are actually treated in your department, rather than relying on the written policy, which is often silent.

Can you get scooped by posting a preprint?

The risk is far lower than it feels, and posting is generally protective. A public timestamp is the strongest evidence of priority available to you, and scooping a preprint without citing it would be a clear ethics violation.

The genuine risk sits before posting: emailing a complete manuscript while seeking an arXiv endorsement, when no timestamp yet exists. Send the endorsement request link and a brief description of the work, never the full paper.

Are preprints indexed in PubMed and Google Scholar?

Google Scholar indexes preprints from all major servers. PubMed Central includes preprints from NIH-funded research through its preprint pilot, and Europe PMC indexes life-science preprints broadly.

Coverage is not universal. Preprints outside these programs may be discoverable only through Google Scholar, the server itself, or aggregators such as OSF search and Sciety. Check the indexing position of your chosen server before assuming reach.

Can I post a preprint after my paper has been rejected by a journal?

Yes. Rejection has no bearing on your right to post, and there is a strong argument for posting at that point, since the manuscript now faces a further delay before it appears anywhere.

Consider incorporating the reviewer comments first if they were substantive. Posting the improved version costs you nothing and means the public record reflects your best current work rather than the version that was rejected.

Is it acceptable to cite a preprint in a systematic review or meta-analysis?

Yes, and excluding preprints can introduce publication bias. Major evidence-synthesis guidance now treats preprints as includable, with the caveat that their unreviewed status must be reported transparently.

Search for a published version of every included preprint before finalizing, since roughly 80% eventually have one and the versions may differ. Record the version number and posting date of what you actually screened, and flag the risk-of-bias implications in your methods.

Is ResearchGate a preprint server?

No. ResearchGate is a scholarly social network with a file upload feature, not a moderated preprint repository. It can assign a DOI to work you upload, but it performs no screening and offers no preservation guarantee.

The practical differences matter. Recognized servers such as arXiv, bioRxiv, and medRxiv screen submissions, commit to permanent archiving, and are indexed by services including Europe PMC and the Preprint Citation Index. ResearchGate content can be removed, and accounts can be deleted, taking the record with them. Several funder policies also specify deposition in a recognized repository, and commercial social networks generally do not qualify. Check your funder’s wording rather than assuming.

Can I upload my preprint to ResearchGate as well as a preprint server?

Yes, and this is the recommended approach: post to a proper preprint server first, then share the link on ResearchGate for reach. Post the DOI rather than the file, so downloads and version history stay attached to the canonical record.

The risk on ResearchGate is not the preprint; it is what people upload afterward. Uploading the published version of record, or in many cases the accepted manuscript, can breach the copyright agreement you signed with the journal. Publishers have pursued large-scale takedowns over exactly this. Check the journal’s self-archiving terms in Open Policy Finder before uploading anything post-acceptance, and note that embargo periods often apply.

Do Elsevier journals allow preprints?

Yes. Elsevier’s sharing policy lets authors share preprints anytime and anywhere, and the publisher actively encourages it. Posting to a preprint server does not count as prior publication.

There are exceptions worth checking. Some society-owned titles and journals running double-anonymized peer review operate different preprint policies, and Elsevier specifically advises authors targeting Cell Press, The Lancet, and certain society titles to check the journal’s own author pages first. The publisher-level yes does not override a journal-level restriction, so read the Guide for Authors rather than assuming.

Elsevier also runs its own route: close to 1,000 journals let you opt in during submission to post the manuscript free on SSRN, and the decision to post has no effect on the editorial process or outcome. You are not obliged to use it. Posting to arXiv, bioRxiv, or another server is equally acceptable.

Two constraints apply after acceptance. Preprints must not be enhanced to resemble or substitute for the final published article, and updating an arXiv or RePEc preprint with the accepted manuscript requires switching the license to CC BY-NC-ND. Note the licensing implication: if you posted under CC BY, that update changes your terms.

For non-open-access Elsevier journals, the preprint is the only version you may share publicly on ResearchGate.

Sources and Further Reading

How to Write the Background of a Research Paper: Examples and Tips

Research Paper vs Review Paper: Key Differences, with Examples