🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
salami_slicing
Salami slicing is the practice of dividing the results of a single study into multiple publications, each presenting a thin slice of the overall findings. This inflates the apparent volume of evidence, padds publication records, and can mislead systematic reviewers who may count each slice as an independent study. It also fragments information, making it difficult for readers to see the full picture, and can enable selective emphasis on favorable subsets of the data.
A research team conducts one large survey of 5,000 workers on job satisfaction, stress, burnout, and compensation. Instead of publishing a comprehensive analysis, they publish four separate papers — one on each variable — in different journals. A meta-analyst later treats these as four independent studies, quadrupling their weight in the pooled estimate.
A clinical trial tests a new drug measuring blood pressure, cholesterol, weight, and sleep quality in 800 patients. Rather than publishing one comprehensive report, the lead researcher submits four separate journal articles over two years — each targeting a different journal, padding their publication list and making the single trial appear to be four independent studies.
A PhD student collects data on teenagers' social media use, mental health, academic performance, and sleep habits in one survey. Under pressure to publish, her advisor splits the findings into three papers: one on screen time and anxiety, one on grades, and one on sleep — tripling the apparent output of a single data collection effort and consuming three peer-review slots with what could have been one paper.
Binary (yes/no) questions an LLM must answer to identify this aspect:
Does this publication appear to report a subset of results from a larger study?
Type: binaryAre there other publications by the same authors using apparently the same dataset or study population?
Type: binaryCould the findings have been more informatively reported as part of a single comprehensive publication?
Type: binaryDoes the fragmentation across multiple papers obscure important context or create an inflated impression of the evidence base?
Type: binarySalami slicing is the practice of dividing the results of a single study into multiple publications, each presenting a thin slice of the overall findings. This inflates the apparent volume of evidence, padds publication records, and can mislead systematic reviewers who may count each slice as an independent study. It also fragments information, making it difficult for readers to see the full picture, and can enable selective emphasis on favorable subsets of the data.
Academic incentive systems reward publication quantity, so researchers benefit from maximizing the number of papers from each dataset. Reviewers and editors may not realize a submission is a slice of a larger study, and readers rarely check for duplicate or overlapping publications.
Check for other publications by the same authors using similar methods or populations. Look for trial registrations that reveal the full scope of the original study. Meta-analysts should screen for overlapping datasets. Journals should require disclosure of all publications from the same dataset.
Common across academic disciplines, particularly in biomedicine where a single clinical trial can spawn dozens of publications, each reporting a different outcome, subgroup, or time point.
Studies with statistically significant or positive results are more likely to be published, while null results remain unpublished. This distorts the published literature and inflates apparent effect sizes in meta-analyses.
Selective sharing of research findings based on the direction or significance of results.
Presenting post-hoc hypotheses as if they were formulated before seeing the data.
Searching through large datasets for any statistically significant pattern without a prior hypothesis. Found patterns are presented as confirmatory when they are actually exploratory and likely to be spurious.
The statistical error of performing many tests without adjusting for the increased probability of false positives. With a significance level of 0.05 and 20 independent tests, there is a 64% chance of at least one false positive. Failure to correct for this inflates the apparent number of 'significant' findings.
Use these tools to detect, analyze, or train this aspect.