Major methodological criticism usually concerns the credibility or interpretability of the study rather than presentation alone. The appropriate assessment is design-specific, but seven recurring problems deserve attention before submission.
1. The research question is insufficiently defined
The population, exposure or intervention, comparator, outcomes, timing, and intended inference should be identifiable. If these elements shift between the abstract, methods, and discussion, reviewers cannot determine what the study actually tested.
2. The design cannot support the stated claim
Cross-sectional and uncontrolled observational data generally cannot establish the causal or temporal claims available from an appropriately designed experiment. Authors should distinguish description, association, prediction, diagnostic performance, and causal inference.
3. Selection and measurement processes are unclear
Eligibility, recruitment, exclusions, exposure measurement, outcome ascertainment, follow-up, and missing observations influence bias and generalizability. A participant flow or explicit denominator trail is often essential.
4. The analysis does not match the data structure
Repeated observations, clustering, time-to-event outcomes, competing events, sparse outcomes, non-linearity, missing data, and multiple comparisons require appropriate handling. Merely naming statistical software does not establish analytical validity.
5. Effect magnitude and uncertainty are underreported
Interpretation should not depend only on whether a P value crosses a threshold. Effect estimates, confidence intervals, denominators, absolute risks when meaningful, and clinical context are needed to judge magnitude and precision.
6. Conclusions exceed the results
Unsupported causal wording, selective emphasis on favorable outcomes, overinterpretation of subgroup findings, and claims of equivalence based on non-significance are common reasons for major revision.
7. Reporting and reproducibility are incomplete
Key definitions, model specifications, variable handling, sensitivity analyses, software information, and analysis decisions should be sufficiently clear to understand what was done. The appropriate reporting guideline should be selected by study design rather than used as a substitute for methodological judgment.
Authoritative resources
This educational article provides general guidance and is not a methodological, statistical, clinical, ethical, or legal assessment of any particular study.