- What is the Retraction Watch database?
- How to search it for free
- Using it for hiring, grant, and co-author due diligence
- Common questions, answered
- Limits, credibility, and the leaderboard
- Implications for research offices
When a hiring committee, grant panel, or co-author vetting workflow needs to check a candidate’s publication record, the Retraction Watch Database is usually the first stop. Accessing the retraction watch database free of charge is now straightforward, but only if research offices know which of its three access routes to use — and where its coverage runs thin. This guide sets out a practical, step-by-step approach for institutional due diligence, plus the credibility caveats a screening checklist should not skip.
What is the Retraction Watch database?
The Retraction Watch Database (RWDB) is maintained by the Center for Scientific Integrity, the nonprofit behind the Retraction Watch blog founded in 2010 by science journalists Ivan Oransky and Adam Marcus. It now logs more than 65,000 retractions, corrections, and expressions of concern, up from roughly 40,000 entries at the time of its September 2023 data-sharing agreement with Crossref, which made the dataset far more widely and openly accessible.
Retraction Watch also publishes derivative lists that get searched separately: the Retraction Watch Leaderboard (most-retracted individual authors) and country-level breakdowns used in bibliometric research. These are useful signals, but they are not substitutes for a record-level check, as the sections below explain.
How to search the Retraction Watch database for free
There are three free access routes, and each suits a different due-diligence task.
| Access method | Best for | Coverage / updates | Cost and limits |
|---|---|---|---|
| Web search form (retractiondatabase.org) | Quick, single-name or single-DOI checks | Live; each search returns up to 50 entries | Free; since October 2024 you must fill at least one of article type, a date range, or a PMID/DOI |
| Full CSV download (Crossref GitLab repository) | Bulk screening of long candidate or co-investigator lists | Complete dataset, refreshed periodically | Free; no per-query limits |
| Crossref REST API | Automated checks embedded in onboarding or grant systems | Updated on working days from publisher data | Free; standard Crossref API rate limits apply |
For a single name or one paper’s DOI, the web form is fastest. For anything resembling a batch check — a full hiring shortlist, an entire grant consortium, or an author list on a multi-author manuscript — the CSV download or the API is the correct tool, because the web interface’s per-search filter requirement makes open-ended browsing impractical by design.
Using the database for hiring, grant, and co-author due diligence
Research offices increasingly build RWDB checks into standard screening, alongside conflict-of-interest disclosures and authorship verification. A workable process looks like this:
- Before a hiring or tenure decision, batch-check the candidate’s DOI or PMID list against the CSV download rather than the web form, which is not built for unfiltered browsing.
- For grant panels, check every named co-investigator’s publication list, not only the principal investigator’s — retraction risk is frequently concentrated in co-authored papers rather than sole-authored ones.
- For co-author vetting ahead of a manuscript submission, search prospective collaborators by name and affiliation, and treat a “Correction” or “Expression of Concern” entry as a prompt for further reading, not an automatic disqualifier.
- Record the query date and parameters used: RWDB entries are added and revised continuously, so a clean result today is not a permanent clearance.
- Cross-reference any hit against PubMed and the publisher’s own notice before acting on it, given documented metadata discrepancies between databases.
Common questions, answered
Is there a Retraction Watch database?
Yes. The Retraction Watch Database is maintained by the Center for Scientific Integrity and hosted at retractiondatabase.org. It logs more than 65,000 retractions, corrections, and expressions of concern, and since September 2023 has been distributed publicly through a data agreement with Crossref.
How do I search the Retraction Watch database for free?
Use the free web form at retractiondatabase.org, filtering by article type, date range, or a PMID/DOI — required since October 2024 to manage server load. For bulk or unrestricted searching, download the complete dataset as a free CSV from Crossref’s GitLab repository instead.
How do I check if a specific article has been retracted?
Search the article’s DOI or PMID directly in the RWDB, or check the publisher’s own page for a retraction notice. Tools such as Zotero and the Crossref REST API also flag retracted status automatically when a DOI is queried or a reference is added to a library.
Is Retraction Watch data comprehensive for corrections and expressions of concern?
No. Retraction Watch states its database is the most complete source specifically for retractions; coverage of corrections and expressions of concern is skewed toward items linked to existing retractions or its own reporting, so due-diligence teams should treat those two categories as indicative rather than exhaustive.
Limits, credibility, and what the leaderboard does (and doesn’t) tell you
The RWDB’s own user guide is explicit about its boundaries: expressions of concern and corrections are entered “as they relate to existing retractions, blog posts, or high-profile studies,” not comprehensively, so any counts of those two record types drawn from the database will be skewed and should not be read as population-representative.
Is Retraction Watch credible for institutional due diligence? The evidence points to yes, with caveats. The Center for Scientific Integrity’s dataset has been independently examined in the academic literature — a 2025 study in Accountability in Research compared RWD metadata accuracy against PubMed and Web of Science and found it a strong, though not flawless, source. Its integration into Crossref’s scholarly infrastructure since 2023 adds a layer of institutional stewardship beyond a single newsroom, and reference tools such as Zotero rely on it to flag retracted citations automatically.
The Retraction Watch Leaderboard — a running list of the most-retracted individual researchers, topped for years by anaesthesiologist Yoshitaka Fujii with more than 180 retractions — and country-level breakdowns used in bibliometric studies are genuinely useful for spotting patterns. But leaderboard rank reflects investigative attention and reporting history as much as underlying prevalence, and cross-country comparisons need population and output normalisation that the raw retraction watch by country counts do not themselves provide. Neither should replace a DOI-level check on the specific individual or paper under review.
Implications for research offices
The database’s growth — from roughly 40,000 entries at the time of the Crossref agreement to more than 65,000 now — reflects both rising retraction volume and improved detection, not necessarily declining research quality alone. For research administration teams, the practical implication is to treat RWDB screening as a routine, documented step in research administration workflows, sitting alongside authorship and contributorship checks rather than as an ad hoc search performed only when a concern is already raised.
As programmatic access matures through the Crossref API, expect RWDB checks to become embedded directly into hiring, grants, and manuscript systems, much as authorship verification already is. Institutions that build this into a documented, repeatable checklist — rather than a one-off Google search — will be better placed to defend their due-diligence decisions if a retraction surfaces after the fact.
