Publishing papers in academic journals is the standard process through which new research findings and knowledge are shared with the scientific community. There have been some cases where authors have failed to fully disclose important details about the research samples used in their published studies. Failing to disclose relevant sample information undermines the credibility and reproducibility of published research.
One type of undisclosed sample information is the omission of details about the size and characteristics of research participants. Accurately reporting sample sizes, demographics, and selection criteria is important for evaluating the generalizability and external validity of research findings. Omitting or obscuring these details makes it difficult for other researchers to judge how representative the studied sample was and whether the findings could reasonably extrapolate to broader populations.
For example, in one case, a published paper reported conducting a clinical trial with over 300 participants but the authors failed to disclose that roughly half the participants came from a single site where the lead author was employed. This lack of transparency about the sample composition and potential bias undermined trust in the research. While multi-site studies naturally have some participants clustered at collaborative research locations, disproportionate clustering can skew results and should be acknowledged.
In another example, a psychological study of emotion responses claimed to test hypotheses on a sample of 100 university students but did not mention that half of the participants were the researchers themselves or lab members. Failing to disclose this self-sampling raises serious concerns about experimental control and objectivity that are impossible for readers to assess without full transparency. Results based partially on self-report data from invested researchers lack the credibility of findings from truly independent samples.
Beyond participant characteristics, researchers also have an obligation to fully disclose unavailable samples or portions of collected data that were not ultimately included in published analyses and results. Excluding or failing to mention certain subsets of originally studied participants can misrepresent what was actually examined and distort conclusions.
In one case, a brain imaging genetics study worked with over 500 participants but only reported results for a subset of around 250 individuals for whom a complete genetic and imaging dataset was obtained. While it is reasonable for analyses to focus on fully measured cases, omitting any mention of the roughly 50% of participants who were screened out or missing data painted an inaccurate picture of the actual studied scope. Other researchers cannot replicate or build upon the findings without full transparency about screening procedures and sample attrition over the course of data collection and cleaning.
In another example, authors of a cognitive neuroscience study tested the hypotheses on a pool of 70 participants but excluded 10 participants due to measurement errors. Rather than fully reporting this 10% sample exclusion, the authors only noted working with a sample “over 60” in their introduction and methods sections without stating how many were ultimately included. Ambiguous or incomplete reporting of exactly how many people comprised the final analyzed sample undermines confidence that the findings are based on the true scope and limits of data available rather than a selectively disclosed subset.
Beyond failing to disclose exclusions, some cases have involved papers publishing results from samples that were never actually examined at all. In a highly publicized instance, a genetics research group found that the statistical analyses and results reported across several published papers could not possibly be based on the actual recruitment totals and specimen availability declared in the corresponding methods sections. It appeared fabricated samples had been presented rather than disclosing limitations that prevented testing all pre-registered hypotheses. Falsifying examined sample sizes destroys credibility in scientific work and misleads the field about what had really been studied.
While honest errors and oversights in sample reporting are inevitable to some degree given the multiplicity of details involved, intentional obfuscations or misrepresentations cross an important ethical line. Full transparency about sample characteristics, inclusions, exclusions, and actual scope of data analyzed is needed to ensure published research is built upon solid evidentiary foundations that can withstand scrutiny and replication attempts by other scientists. Misleading or ambiguous disclosures threaten to distort research directions and conclusions in the field if based on selective or incomplete portraits of what was truly examined.
There are several ways the scientific community works to promote full sample transparency and prevent undisclosed limitations. Peer reviewers have a responsibility to scrutinize methods sections for completeness and accuracy in reporting sample descriptions and processing. Editors must enforce guidelines requiring explicit detailing of participant demographics, selection criteria, screened exclusions, missing data protocols, and precise Ns for all analyzed samples. Statistical reviewers can check if reported sample sizes and groupings align with documented recruitment totals.
Post-publication sharing of deidentified data and materials also increases accountability. When other scientists can evaluate the actual sample behind published claims, intentional misrepresentation is less feasible. Some scientific communities and funding organizations now mandate data and resource sharing as a condition for publication. Pre-registration of study protocols helps establish an a priori record of planned recruitment goals and analysis plans independent of outcome pressures that could slant selective disclosure decisions.
Overall, truth and credibility in research depend upon full candor from authors about sample limitations as well as positive results. Undisclosed exclusions or distortions threaten the evidence base and mislead other scientists’ efforts. While mistakes will always occur, the scientific process aims to promote transparency and reproducibility as safeguards against such threats over time. With collective vigilance through peer and statistical review, data sharing, and pre-registration, we can work to establish a research culture maximizing integrity and trust through complete sample disclosures.
Andrew Stroud