Duplicate publication is the republication of substantially the same data, analysis or conclusions in more than one paper without full disclosure and cross-referencing to the original. Splitting one dataset into several papers is legitimate only when each paper answers a genuinely distinct question, adds a new contribution, and is fully disclosed to every editor involved — anything less risks being classed as salami slicing or redundant publication under COPE guidance.
Duplicate publication is one of the most common editorial integrity problems editors encounter. The Committee on Publication Ethics (COPE) treats it, alongside “salami slicing” (dividing one study into the smallest publishable units), as a form of redundant publication that misleads readers and wastes peer-review capacity. This guide uses COPE’s own case guidance to show exactly where the line sits between a legitimate multi-paper research programme and a compliance breach.
Contents
- What is duplicate publication and how does it differ from salami slicing?
- When is splitting one dataset into multiple papers legitimate?
- What do COPE case studies show about editorial decisions?
- What should editors and authors check before publication?
- Answer-first Q&A on duplicate publication
- Implications for multi-paper research programmes
What is duplicate publication and how does it differ from salami slicing?
Duplicate (or redundant) publication is the submission or publication of a paper that overlaps substantially — in hypothesis, data, methods, results or conclusions — with a paper already published, without acknowledging or citing the earlier work. The COPE case archive documents this pattern repeatedly: two journals independently publish near-identical articles from the same author group, unaware of each other, because the overlap was not disclosed at submission.
Salami slicing is related but distinct: instead of publishing the same paper twice, authors divide one coherent study into the smallest set of “publishable units”. Neither practice is automatically dishonest at the point of writing; both become misconduct at the point of non-disclosure. The ICMJE Recommendations (updated January 2024) state that authors must not submit “papers describing essentially the same content to more than one journal” and must disclose all related submissions from the same dataset at the point of submission.
When is splitting one dataset into multiple papers legitimate?
Partial publication in research is not, by itself, unethical. Large studies — multi-site trials, longitudinal cohorts, complex mixed-methods programmes — routinely generate more than one legitimate paper. The test COPE and ICMJE apply is whether each paper stands alone as a distinct, substantial contribution, not whether the data originate from a single collection exercise.
A split is defensible when the following conditions hold:
- Each paper addresses a separate, pre-specified research question with its own hypothesis, analysis plan and conclusion — not a restatement with a slightly different subgroup.
- Each paper makes a substantial, non-overlapping contribution, so a reader needs both papers to understand the full programme, but neither is redundant alone.
- Authors disclose all related and prior submissions to every editor involved, including manuscripts under review elsewhere from the same dataset.
- Papers are explicitly cross-referenced, so readers and systematic reviewers can identify the shared dataset and avoid double-counting participants.
- Any secondary or translated publication follows the ICMJE’s conditions for acceptable secondary publication — both journals’ editors informed, primary publication respected, and readers told via a footnote.
A cohort study reporting baseline characteristics in one paper and a distinct clinical outcome analysis eighteen months later is typically legitimate partial publication. A single trial reported twice with the discussion reworded is not — this is what duplicate publication bias in evidence synthesis refers to: the same result counted twice in a meta-analysis because the overlap went undetected.
What do COPE case studies show about editorial decisions?
COPE’s published case discussions are instructive because they show how editors reason through ambiguous, real submissions rather than textbook definitions. In one widely cited case, a newly appointed editor noticed that an article just published in their journal closely resembled one published months earlier elsewhere by the same author group (with two additional names added). The later paper reported a subset of the earlier paper’s results, and both reached an identical conclusion.
The authors explained that the first paper reported “preliminary findings” and the second “final results based on a larger dataset” — an explanation COPE Council members treated with scepticism. Their advice highlighted several markers editors should weigh:
| Marker | What it suggests |
|---|---|
| Sentences rearranged but conclusions identical | Deliberate reworking rather than genuine reanalysis |
| All authors signed copyright forms without flagging overlap | Plausible but unlikely genuine oversight |
| Same conclusion drawn from an overlapping subset | No new substantial contribution — redundant, not partial |
| Neither editor informed of the other submission | Breach of the disclosure obligation, regardless of intent |
The outcome was a jointly published notice of inadvertent duplicate publication in both journals — a remedy COPE’s redundant (duplicate) publication flowchart sets out for a confirmed post-publication overlap. Where overlap is caught at submission instead, COPE’s parallel flowchart directs editors to seek explanation from the authors before deciding to accept, reject, or request the manuscripts be combined.
What should editors and authors check before publication?
Editors handling a manuscript that may derive from a previously used dataset should not rely on author self-declaration alone. A structured check reduces both false accusations and missed cases:
- Ask authors directly, at submission, whether any part of the data, cohort or findings has appeared in — or been submitted to — another publication.
- Compare the introduction, methods and conclusion sections for substantive overlap, not just the abstract.
- Check whether the “new” paper’s conclusion could reasonably have been drawn from the earlier paper alone.
- Confirm that any secondary or translated publication meets the ICMJE’s disclosure and footnote conditions.
- Ask authors to state, in the cover letter, how this submission differs from and complements other linked outputs.
Clear, role-specific contributorship statements also help editors assess legitimacy quickly: where different authors led different analyses within a shared dataset, a structured contributor role taxonomy such as CRediT makes it easier to see that a paper’s intellectual contribution is genuinely distinct rather than a relabelled version of prior work. CASRAI originated the CRediT contributor role taxonomy in 2014; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022. Journals that also require full authorship criteria disclosure alongside dataset provenance statements catch most cases before peer review.
Answer-first Q&A on duplicate publication
What is a duplicate publication?
A duplicate publication is a paper that overlaps substantially — in hypothesis, data, methods, results or conclusions — with an article already published, submitted without disclosing that overlap. It differs from legitimate secondary publication because the earlier work is not acknowledged or cross-referenced.
What is the difference between redundant publication and duplicate publication?
Redundant publication is the broader category, covering any republication of substantially the same material, including salami slicing. Duplicate publication is the specific case where two papers are near-identical in content — the narrowest and most serious form of redundancy.
What is a dual publication?
“Dual publication” is a synonym for duplicate publication: the same article, or one overlapping substantially with it, published in two separate journals without acknowledgement of the first. It is treated identically under COPE and ICMJE guidance regardless of which term is used.
What should an editor do if duplicate publication is suspected?
Editors should follow COPE’s published flowcharts: for a submitted manuscript, seek an explanation from the authors before deciding to accept, reject or request combination into a single paper; for an already-published article, contact both journals’ editors, review the extent of overlap, and issue a joint notice or correction as warranted.
Implications for multi-paper research programmes
For research administrators overseeing large, multi-output programmes, the practical implication is procedural: build dataset-provenance disclosure into submission workflows before manuscripts reach a journal, not after a reviewer flags overlap. Funders and institutions increasingly expect research offices to demonstrate that outputs from a shared dataset are distinct contributions, not salami-sliced units inflating a publication count.
Analyses of retraction reasons catalogued in the Retraction Watch Database consistently place duplication among the recurring categories behind retraction notices, alongside plagiarism and data-integrity concerns — a reminder that the cost of getting this wrong is a permanent mark on the published record, not merely a desk rejection. As multi-site and multi-arm studies become more common, editors and authors who apply COPE’s disclosure-first standard consistently, rather than relying on ad hoc judgement calls, are best placed to keep legitimate partial publication distinct from redundant publication.
