Written and maintained by CASRAI Editorial Board
Last updated
Every study has limitations — constraints on sample size, design, measurement, generalizability, or execution that bound how much weight a reader should put on its findings. Naming them clearly is not an admission that the work failed; it is what lets a reader, reviewer, or future researcher calibrate confidence in the results correctly. A Limitations section that is vague, defensive, or missing entirely is one of the more commonly cited specific reasons reviewers request revisions — not because reviewers expect a flawless study, but because an author who cannot or will not name their own study’s constraints raises real doubt about whether they understood them in the first place.
This guide covers how to identify, categorize, and write about a study’s limitations, whether the target journal wants them in a dedicated subsection or folded into the Discussion. For the section they most often live inside, see CASRAI’s How to Write a Discussion Section; for the section-by-section logic of the whole paper, see the IMRaD Structure entry.
Last verified 25 August 2026 against ICMJE’s Recommendations for manuscript preparation and standard editorial guidance from major biomedical and scientific journals.
Where Limitations Belong
Journals genuinely differ on placement, and both are structurally acceptable: some require an explicit “Limitations” heading, usually near the end of the Discussion or as its own subsection before the Conclusion; others expect limitations addressed within the flow of the Discussion without a separate heading at all. What is not acceptable under either convention is omitting them, or reducing them to a single generic sentence appended as an afterthought. Check the target journal’s author guidelines for its specific formatting expectation before drafting — some reporting guidelines (e.g., CONSORT for trials, STROBE for observational studies) explicitly require a dedicated limitations discussion as part of their checklist.
Categories of Limitation
Most study limitations fall into a small number of recurring categories. Working through each category deliberately, rather than trying to freehand a list, is the most reliable way to avoid missing a real constraint on the study’s conclusions.
Design Limitations
Constraints inherent to the study design itself — an observational design that cannot establish causation the way a randomized trial could, a cross-sectional design that cannot capture change over time, a lack of a control or comparison group, or a single-arm design without blinding.
Sample Limitations
Sample size too small to detect a real effect (underpowered), a sample drawn from a narrow or non-representative population, self-selection in recruitment, or a specific population (e.g., a single institution, a single country, a particular demographic group) that limits how far findings generalize.
Measurement Limitations
Instruments or measures with imperfect validity or reliability, reliance on self-report where objective measurement would be stronger, missing data, or a measure that captures a proxy for the construct of real interest rather than the construct itself.
Confounding and Bias
Variables that could not be controlled or measured and might explain an observed association, response bias, recall bias, or selection effects introduced by how participants entered or remained in the study.
Practical and Resource Constraints
Constraints imposed by funding, time, access, or feasibility — a shorter follow-up period than ideal, a narrower set of outcomes measured than the research question would ideally call for, or a study stopped early for a documented, disclosed reason.
How to Write Limitations That Read as Credible
Name the Specific Claim Each Limitation Weakens
A limitations discussion that lists limitations generically, without connecting them back to the paper’s own claims, reads as boilerplate. Instead of “the sample size was small,” state what that small sample specifically limits: “the sample size limits statistical power to detect effects smaller than [X], and the confidence intervals around the primary estimate should be read accordingly.” Specificity about which claim is weakened, and roughly how much, is what separates a credible limitations discussion from a defensive one.
Don’t Explain Limitations Away
Following every limitation with a rationalization for why it doesn’t really matter undermines the section’s purpose and often reads as evasive to an experienced reviewer. It is appropriate to note a mitigating factor if a real one exists (e.g., a sensitivity analysis that produced consistent results despite the limitation) — but the default posture should be to state the limitation, not to argue it away.
Distinguish Limitations from Simple Errors
A limitation is a real constraint inherent to the study’s design, resources, or context — not a mistake that could have been corrected within the study as conducted. If something went wrong that was correctable (a protocol violation, a data-entry error), that is a methods or reporting issue to address directly, not something to file under “limitations” as if it were an unavoidable constraint.
Keep Proportion in Mind
A limitations discussion should be proportionate to the study’s actual constraints — neither a single dismissive sentence for a study with real, substantive constraints, nor an exhaustive list of every conceivable hypothetical weakness padded out to seem thorough. The goal is calibration, not a checklist exercise.
Common Pitfalls
Burying Limitations in One Generic Closing Sentence
The most frequent complaint reviewers raise is a single boilerplate sentence (“as with any study, there are limitations”) that names nothing specific. This is functionally equivalent to omitting the section.
Treating Limitations as an Apology
A limitations discussion is a scientific accounting of what the study can and cannot support, not an apology for imperfection — every published study has limitations, including the strongest ones in the field. Framing them neutrally and specifically, rather than defensively, is what a reviewer expects.
Omitting a Limitation the Authors Clearly Knew About
A limitation visible from the Methods section (a small sample, a single-site design, a short follow-up) that goes unmentioned in the Discussion reads as either an oversight or an evasion — reviewers routinely check the Methods against the Limitations discussion for exactly this gap.
A Short Checklist Before Submission
- Have design, sample, measurement, confounding/bias, and practical-constraint categories each been considered, even if not every category applies?
- Does each named limitation connect to a specific claim it weakens, rather than sitting as a generic statement?
- Are limitations stated plainly rather than argued away?
- Does the section avoid both a single dismissive sentence and an exhaustive, padded list?
- Do any limitations visible from the Methods section (small sample, single site, short follow-up) appear here?
- Does the target journal or reporting guideline (CONSORT, STROBE, etc.) require a specific placement or format for this discussion?
Frequently Asked Questions
Do I need a separate Limitations section?
Some journals require an explicit “Limitations” heading; others expect limitations addressed within the Discussion’s flow without a separate heading. Either is structurally acceptable as long as limitations are specific and not omitted — check the target journal’s author guidelines.
Where does the Limitations discussion go in a paper?
Almost always within or at the end of the Discussion section, typically before any Conclusion. It should not appear in the Methods (which describes what was done, not its constraints) or in the Results (which reports findings neutrally).
Is naming a study’s limitations a weakness in a manuscript?
No — the opposite. A limitations discussion that specifically names what the study cannot support is a mark of scientific rigor and is what allows readers to correctly weight the findings. Reviewers are more skeptical of a study that reports no limitations than one that names real, specific ones.
How long should a Limitations discussion be?
Proportionate to the study’s actual constraints — there’s no fixed length. A study with several genuine, substantive limitations should discuss them specifically rather than compress them into one sentence; a study with few real constraints should not pad the section to appear thorough.
What is the difference between a limitation and a methodological error?
A limitation is a constraint inherent to the study’s design, resources, or context that could not reasonably have been avoided. A methodological error (a protocol violation, an avoidable data problem) is a different kind of issue that should be addressed directly in Methods or as a correction, not folded into the limitations discussion as though it were unavoidable.
References
- International Committee of Medical Journal Editors, “Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals” — Preparing a Manuscript for Submission.
