First, don’t panic—discrepancies between a preprint and its published version are common, and the published article is almost always the authoritative record. That said, your next steps depend on what differs and why, so here’s a practical playbook.
1. Confirm the published version is final
Check the journal’s website (not just a PDF you downloaded) for the version of record. Some journals post “accepted manuscripts” that look final but haven’t been copyedited or peer-reviewed post-acceptance. Look for a DOI, publication date, and volume/issue numbers. If the preprint is older, the published paper may include revisions from peer review—this is a good sign, not a red flag.
2. Categorize the disagreement
Not all differences carry equal weight:
- Methodology or results changed (e.g., different sample size, statistical test, or conclusion): This usually reflects peer-review feedback. Trust the published version, but read the preprint’s methods to understand what was revised.
- Numbers or figures are slightly off: Likely a typo or updated data. Check the published article’s supplementary materials for corrections.
- Author list or affiliations changed: Could mean an author was added/removed after review—verify the published version for credit and citation purposes.
- Entirely different findings: Rare but possible if the preprint was a preliminary analysis. Treat the published paper as the definitive science, and note the preprint as a historical artifact.
3. Check for an erratum or corrigendum
Journals publish corrections for errors in the version of record. Search the article’s DOI or PubMed/Crossref for “erratum” or “corrigendum” linked to it. If the published paper itself has a flaw, this is where it’ll be flagged.
4. Use the preprint to understand why
Preprints often include raw data, extended methods, or exploratory analyses cut from the final paper. If the published version is more concise, the preprint can fill gaps—just don’t cite it for claims the published version contradicts. For deep reading, WisPaper’s Scholar QA can compare both documents side-by-side and trace each claim back to the exact page, so you can see whether the discrepancy is substantive or cosmetic.
5. Decide what to cite
Always cite the published, peer-reviewed version in your own work, unless you’re explicitly discussing preprint culture or replication efforts. If you must reference the preprint (e.g., for a very recent finding), state the date and note it’s not yet peer-reviewed.
6. When to contact the authors
If the discrepancy is unexplained and affects your own research (e.g., you’re replicating their experiment), email the corresponding author. Be specific: “I noticed the preprint reports X, but the published version says Y—could you clarify which is current?” Most researchers appreciate the catch.
Bottom line: Treat the published version as ground truth, use the preprint as context, and verify corrections before moving forward. Disagreements are part of science’s self-correcting process—your job is to navigate them without propagating errors.
