Research methods guide
How to deduplicate literature search results
Put DOI, PMID, normalized title, year, first author, and database provenance in one table. Resolve high-confidence duplicate reports first and review near title matches or conflicting identifiers manually.
A worked example includes duplicate database records of one article and separate abstract and full-text reports of one trial. Deduplicate record copies and link distinct reports at study level.
How to decide duplicate candidates
Preserve original sources, retained identifier, merge reason and reviewer.
| Item | Worked example | Verification |
|---|---|---|
| DOI match | https://doi.org/10.1234/demo.7 and 10.1234/DEMO.7, synthetic identifiers. | Normalize prefix and case, then compare title and publication type and retain provenance. |
| No DOI | Same title, lead author and year with a pagination difference. | Check online-first versus issue records, PMID and journal before merging. |
| Same trial, distinct reports | An abstract reports week 12 and a full paper reports week 24, with different article IDs. | Retain both, link to one study ID and avoid repeated samples in synthesis. |
Apply the method to your materials
Keep raw and deduplicated data
Freeze raw exports. Merge confirmed identifier matches and use fuzzy titles to generate review candidates.
Reconcile record flow
Reconcile raw counts, removed copies and screened records. Keep a record-ID mapping for the PRISMA flow.
Questions that arise in practice
Can exact titles and years be removed automatically?
Review disagreements in author, publication type or identifier, especially corrections, abstracts and version updates.
Verified sources
The official sources below were checked on 2026-11-12. Access conditions and update dates are rechecked before publication.
RigorMed workspace
Start with your research materials
Bring your research question, available materials, and the steps you want to check.
