A preprint is a complete manuscript posted publicly before peer review. It is real research made available early, and it carries no claim of having been checked by anyone other than its authors.
A preprint is a complete manuscript posted publicly before peer review. It is real research made available early, and it carries no claim of having been checked by anyone other than its authors. The format is explicit about that, even where coverage of it is not.
Review takes months and sometimes years. Posting early establishes priority, allows comment from a wider group than two reviewers, and gets findings into circulation when timing matters. In several fields it is now the normal first step rather than an alternative route.
Reviewers mostly catch unclear methods, unsupported inference and missing analyses. They catch fabrication rarely and statistical error inconsistently. So the gap between a preprint and a reviewed paper is real but narrower than the word unreviewed suggests.
Preprints are revised, and the version a reader found may not be the current one. Servers keep version histories, and comparing the first version with the published paper sometimes shows which conclusions survived review and which were softened.
A preprint with no subsequent journal version may have been rejected, abandoned, or simply left as it is. The absence is not proof of a problem, and it is a reason to check before citing one as though it had completed the process.
Secondary reporting frequently drops the distinction entirely, describing a preprint as a study without qualification. The finding may be perfectly sound; the reader has simply been given no way to apply the appropriate discount.
Preprint servers link to the published version when one exists, and the digital identifier of the journal article will differ from the preprint's. A search on the title usually resolves the question in under a minute.
Read preprints, cite them as preprints, and weight a finding by its methods rather than by where it sits in the publication process. A well-designed preprint is better evidence than a weak reviewed paper.
The claims on this page about research published before review, and what that changes about reading it can be checked against sources outside this site: