03 · What You Need to Know
The Number of Papers Matters Less Than What Each Paper Contributes
Authoritative guidance does not impose a universal paper limit
ICMJE explicitly recognizes that editors may receive multiple manuscripts based on the same dataset, whether from the same research group or different groups. Such manuscripts may differ in their analytical methods, conclusions, or both.
Its central standard is qualitative rather than numerical: manuscripts based on the same dataset should add substantially to one another to warrant publication as separate papers, and previous publications from the same dataset should be appropriately cited for transparency.
That means there is no general "one dataset, one paper" rule. Nor is there an authoritative rule saying that a dataset becomes exhausted after some predetermined number of publications.
A large dataset can legitimately answer multiple substantial questions
Consider a longitudinal cohort containing demographic information, educational experiences, psychological measures, health indicators, institutional variables, and repeated outcomes over ten years. Requiring all meaningful analyses to appear in one article could produce an unreadable manuscript and prevent important questions from receiving appropriate methodological treatment.
Separate papers may therefore be justified when they address genuinely different research questions, use distinct analytical approaches, or generate substantively different conclusions.
ICMJE specifically notes that different analytical approaches to the same dataset may be complementary and equally valid.
Each paper should be able to explain why it exists separately
A useful test is to forget publication counts and ask what scientific reason justifies separation.
If two analyses answer different substantive questions, require different theoretical framing, employ materially different analytical approaches, and lead to different interpretations, separate articles may be sensible.
If the only justification is that the authors can split ten closely related outcomes into ten manuscripts, the case is much weaker.
Legitimate multiple publication
Separate papers answer substantively distinct questions or provide meaningfully different analyses while transparently identifying their shared dataset.
Excessive fragmentation
A coherent body of closely related findings is divided into minimally distinct papers whose separate publication adds little scientific value.
Ask whether readers would be better served by one integrated paper
The question is not simply whether two papers can technically be distinguished. Almost any dataset can be subdivided if researchers define questions narrowly enough.
Instead, ask whether separating the findings improves scientific communication. Would the reader understand the phenomenon better if the outcomes were considered together? Does splitting them hide important relationships? Will each paper contain enough evidence and interpretation to stand as a meaningful contribution?
If combining the analyses would produce a clearer, more coherent scientific account without making the article unmanageable, fragmentation may be difficult to justify.
Substantial added value is more important than a different title
Changing the title, dependent variable, order of authors, or theoretical vocabulary does not automatically make two papers substantively distinct.
ICMJE's guidance focuses on whether manuscripts based on the same dataset add substantially to one another. That is a much higher standard than merely avoiding identical prose.
For example, one paper examining overall treatment effectiveness and another examining a theoretically motivated moderator of that treatment may make distinct contributions. Two papers reporting nearly the same primary association with slightly different covariate sets may be harder to justify separately.
Shared participants are not themselves the problem
A single cohort can support numerous legitimate publications over many years. Large trials may also be planned from the beginning to produce separate papers addressing different research questions while using the same original participant sample.
ICMJE explicitly acknowledges this possibility. In large trials, numerous separate publications may be planned using the same participant sample when they address separate research questions. Its emphasis remains transparency.
Accordingly, using the same participants in multiple papers does not by itself establish duplicate publication.
Transparency about the shared dataset is essential
A reader should not have to conduct bibliographic detective work to discover that several apparently independent papers analyzed the same participants or dataset.
ICMJE recommends appropriate citation of previous publications from the same dataset. For secondary analyses of clinical trial data, it recommends citing the primary publication, clearly stating that the paper contains secondary analyses or results, and using the same trial registration number and persistent dataset identifier where applicable.
This transparency helps readers interpret the evidence correctly and prevents related papers from being mistaken for independent samples.
Watch Out
Changing author order, article titles, or terminology does not make overlapping samples independent. If papers use the same dataset or participant pool, make that relationship discoverable rather than leaving readers to infer it.
Why hidden overlap can distort reviews and meta-analyses
Suppose three papers analyze the same 500 participants but do not clearly disclose that relationship. A systematic reviewer might interpret them as three independent samples containing 1,500 participants.
The result could be inappropriate weighting or double-counting of evidence. ICMJE identifies this as a central reason duplicate publication of original research is problematic.
Transparent cross-citation is therefore not merely editorial housekeeping. It affects how later researchers understand and synthesize the evidence base.
Secondary analysis is legitimate when it adds something substantive
A secondary analysis asks a question that differs from the primary analysis using data already collected. There is nothing inherently improper about this. Indeed, extracting additional knowledge from responsibly collected data can be scientifically valuable.
The important questions are whether the analysis is methodologically justified, whether its relationship to the original study is clear, and whether the new paper contributes sufficiently beyond previous publications.
For clinical trials, ICMJE specifically advises secondary analyses to cite the primary publication and clearly identify themselves as secondary analyses or results.
One study should not be artificially fragmented into the smallest publishable units
The concern commonly called salami slicing arises when researchers divide a coherent study into multiple minimally differentiated papers rather than reporting related findings together.
COPE's guidance on redundant publication recognizes "salami publishing" as a form of minor overlap or redundancy that editors may need to address. The concern is contextual: repeated methods or legitimate subgroup analyses can be defensible, while artificial fragmentation that adds little may not be.
Different outcomes do not automatically require different papers
A dataset may contain ten outcome variables. That does not mean it naturally contains ten articles.
Outcomes that address a common research question may be more informative when analyzed together. Conversely, an outcome representing a distinct theoretical domain with its own hypotheses, analytical strategy, and implications may warrant independent treatment.
Scientific coherence should determine the grouping more than the number of variables available in the spreadsheet.
Different statistical models do not necessarily create different contributions
Running regression in one paper and structural equation modeling in another does not automatically justify two publications if both analyses answer essentially the same question and produce substantially the same interpretation.
Different analytical approaches can justify separate publications when they reveal genuinely complementary or distinct insights. ICMJE recognizes this possibility. But methodological difference should translate into substantive added value rather than serving as a cosmetic distinction.
Planned publication strategies can improve transparency
For large projects, developing an analysis and publication plan before results are known can help distinguish legitimate secondary questions from opportunistic fragmentation.
The plan might map major research questions, primary and secondary outcomes, proposed analyses, expected manuscript relationships, and cross-citation requirements. It can also help research teams avoid two subgroups unknowingly preparing overlapping papers from the same dataset.
Such planning does not make every proposed paper automatically legitimate, but it makes the intellectual rationale easier to evaluate.
