Can a preprint turn out to be wrong even after it gets published?

Yes. Publication and peer review lower the odds of serious error but do not remove them entirely. Papers get corrected or retracted after publication fairly often, usually because other researchers later find a flaw that reviewers missed.

Why do some good studies stay as preprints and never get published?

Some authors never submit them to a journal, often because of time, cost, or a decision that the result is not significant enough to chase through review. Others are submitted and rejected everywhere they try, sometimes for valid reasons and sometimes due to factors unrelated to quality, such as a journal's narrow focus.

Does peer review catch deliberate fraud?

Rarely, and that is not really its job. Reviewers assume the data they are shown is real and focus on whether the methods and logic are sound; fabricated results are usually only uncovered later, often by other researchers trying and failing to reproduce the work.

How long does peer review usually take?

It varies enormously by field and journal, from a few weeks to well over a year, especially if the paper needs multiple rounds of revision or the reviewers are slow to respond. Preprints exist partly to get findings out during that wait.

Should news articles ever report on preprints?

It can be reasonable if the story makes clear the work has not yet been independently checked, but caution is warranted for strong claims, especially in medicine, since the findings could still change before publication.

Preprints vs Peer Review: What Actually Changes

1 October 2026

A preprint is a scientific paper that its authors have posted publicly before anyone outside their lab has formally checked it. A peer-reviewed paper has been read by independent experts who can demand changes, request more evidence, or reject it outright. The gap between the two versions is usually fixes and caveats rather than a different story, but it is a real gap and it is worth knowing where it sits.

What a preprint actually is

A preprint is a manuscript that researchers upload to a public server, such as bioRxiv for biology or medRxiv for medicine, before it has gone through a journal's review process. It gets a permanent web address and a timestamp, so anyone can read it, cite it, and see exactly when it appeared.

The servers do basic screening. Someone checks that the file is actually a scientific paper, that it is not plagiarised, and that it is not obviously dangerous or false on its face. Nobody checks whether the methods were sound, whether the statistics hold up, or whether the conclusions follow from the data. That part has not happened yet.

Researchers post preprints for speed. Journal review can take months or years, and in fields moving fast, such as a new outbreak or a competitive area of physics, authors want their findings visible straight away. A preprint also stakes a claim: if two labs are chasing the same result, the preprint shows who got there first.

What peer review adds

When a paper is submitted to a journal, an editor reads it and decides whether it is worth sending out. If it is, the editor picks two to four independent experts in the same field, known as reviewers, who read the manuscript closely and write a report.

Reviewers look for specific things: whether the sample size was large enough to support the claims, whether the statistical tests were the right ones for the data, whether a control group was missing, whether the conclusions overreach what the results actually show. They send this back to the editor, usually with a recommendation to accept, reject, or revise.

Most papers that eventually get published go through at least one round of revision. Authors rewrite sections, run extra analyses, add missing details, or soften claims that the reviewers thought were too strong. This back and forth can take anywhere from a few weeks to well over a year, and a meaningful share of submitted papers are rejected outright and never reach this stage at a given journal.

What changes in the text

Compare a preprint with its later published version and the core finding is usually still there. What changes is mostly around the edges: clearer descriptions of methods so another lab could repeat the experiment, corrected statistics, extra control experiments that reviewers asked for, and references to other relevant work that the authors had missed.

Language often gets more careful. A preprint might say a treatment "reduces" a symptom; the published version might say it "was associated with a reduction" once a reviewer points out that the study design cannot prove cause and effect. Limitations sections tend to grow, listing things the study could not rule out.

Occasionally the changes are bigger. A key result might be removed because it did not hold up under closer statistical scrutiny, or the headline conclusion might be narrowed from a broad claim to a more specific one. This is less common than small fixes, but it does happen, which is exactly why the review step exists.

How much trust each deserves

Treat a preprint as a claim that has not yet been stress tested by people whose job is to find its weak points. That does not make it wrong. Many preprints turn out to be solid and are published with only minor edits. But nobody outside the authors' own circle has yet tried hard to break the argument.

Peer review does not guarantee correctness either. Reviewers can miss errors, and fraud has slipped through peer review at serious journals before, sometimes not caught until years after publication. Review raises the average reliability of what gets through; it does not make every paper airtight.

The practical difference is who has checked the authors' homework. With a preprint, nobody independent has checked it yet. With a peer-reviewed paper, at least a few qualified strangers have tried to find holes and the authors have had to answer for them. That is a meaningful difference even though it is not a guarantee.

Reading either one safely

Check whether a preprint has since been published in a journal. Most servers show this directly on the page, often with a link to the final version, so you can read the more scrutinised text instead.

Look at how long the preprint has been up. A preprint that has sat for a year or more with no journal version can be a sign that it struggled in review, was never submitted, or is still being revised. It is not proof of a problem, but it is worth noticing.

Be cautious when a single preprint is the basis for a big news story. Strong claims resting on one unreviewed paper deserve more scepticism than the same claim appearing in a published paper that other labs have since built on or replicated.

Common questions

Can a preprint turn out to be wrong even after it gets published?: Yes. Publication and peer review lower the odds of serious error but do not remove them entirely. Papers get corrected or retracted after publication fairly often, usually because other researchers later find a flaw that reviewers missed.
Why do some good studies stay as preprints and never get published?: Some authors never submit them to a journal, often because of time, cost, or a decision that the result is not significant enough to chase through review. Others are submitted and rejected everywhere they try, sometimes for valid reasons and sometimes due to factors unrelated to quality, such as a journal's narrow focus.
Does peer review catch deliberate fraud?: Rarely, and that is not really its job. Reviewers assume the data they are shown is real and focus on whether the methods and logic are sound; fabricated results are usually only uncovered later, often by other researchers trying and failing to reproduce the work.
How long does peer review usually take?: It varies enormously by field and journal, from a few weeks to well over a year, especially if the paper needs multiple rounds of revision or the reviewers are slow to respond. Preprints exist partly to get findings out during that wait.
Should news articles ever report on preprints?: It can be reasonable if the story makes clear the work has not yet been independently checked, but caution is warranted for strong claims, especially in medicine, since the findings could still change before publication.

Lucent, once a week

We read the new research and write up what actually holds, in the same plain language as this page. No jargon, no hype, three papers a week.

Read the latest issue