Open the limitations section of ten master’s dissertations and you will find nearly the same paragraph: “Due to limited time and funding, the sample was small and drawn from one location, so the results may not be generalisable. Future research should use larger samples.” It is true of every project, which means it says nothing about any of them.
Reviewers read the limitations section to answer one question: does the author understand the real weaknesses of this study, and how far could they distort the conclusions? Done well, this section raises the credibility of the paper. Done generically, it shows the author has not thought hard enough.
Limitations are not delimitations
Delimitations are boundaries you chose on purpose: first-year students, not final-year students. Limitations are unintended or unavoidable weaknesses that could bias results: a low response rate, self-reported measures, no control group. Listing delimitations as limitations is a common way to make the paragraph longer without making it more useful.
Which limitations to write about
You do not need every weakness you can imagine. Choose the two to four most likely to affect your main conclusion. One way to find them is to walk through the usual sources of bias:
| Source | Self-check question | Concrete example |
|---|---|---|
| Sampling | Who is likely missing from the sample, and how do they differ? | Students who had already dropped out were not surveyed |
| Measurement | Does the instrument capture the concept? Any response bias? | Self-reported study hours are prone to inflation |
| Design | Does the design support the claims being made? | Cross-sectional data cannot establish direction of effect |
| Confounding | Which unmeasured factor could explain the result? | Initial motivation was not measured |
| Analysis | Which analytic decision would change the result if made differently? | How missing data were handled |
| Context | Are results tied to the particular time or place? | The survey ran straight after exams |
A four-step structure for each limitation
- State the specific limitation, not a generic label.
- Give the likely direction of bias: could it inflate or underestimate your result? This is what separates a strong paragraph from an average one.
- Describe what you did to mitigate it, if anything: sensitivity analyses, comparison with other data, statistical adjustment.
- Judge the remaining impact and suggest a specific design that would address it.
A rewritten example
Generic: “The sample size was small, so the results may not be representative.”
Rewritten: “The response rate was 38 per cent, and respondents’ grade point averages were about 0.3 points higher than the cohort as a whole, according to registry data. Because lower-achieving students tend to struggle more with online learning, satisfaction may be overstated here. Weighting by grade point average lowered the estimates slightly without changing the main conclusion (Appendix C). Future studies could use administrative attainment data to avoid relying on volunteers.”
The second version is longer, but every sentence carries information: what the limitation is, which way it pushes the results, how the author checked, and what should come next.
Small samples: when to mention them and how
“Small sample” is a limitation worth stating only in specific situations: when a result is non-significant and the confidence interval is too wide to rule out a meaningful effect; when subgroup analyses rest on very few people; when the model is complex relative to the number of observations. State the consequence precisely: “With 42 participants per group, the study had adequate power only for differences of around 0.6 standard deviations or more; smaller effects may have been missed.” In qualitative research, a small sample is not a statistical limitation; the real limitation is more often a lack of diversity among participants or the researcher’s position.
Where it goes and how long it should be
In a journal article, limitations usually form a subsection of the Discussion, after interpretation and before the concluding paragraph. One to three paragraphs is typical. A thesis may give them a longer dedicated section. Some biomedical journals ask for a short “strengths and limitations” box after the abstract; there, brief bullet points are the norm.
Make sure the limitations do not contradict the rest of the paper. If you write that a cross-sectional design cannot support causal claims, your conclusion and title cannot use words such as “impact” or “effect”.
Mistakes to avoid
- Listing limitations without saying how they affect the conclusions.
- Citing your own time or funding constraints; readers need consequences for the data, not your circumstances.
- Undermining the whole study with lines such as “the findings are only indicative”.
- Hiding the biggest weakness, which reviewers will certainly notice.
- Ending with “more research is needed” without saying what research.
Useful phrasing
- “Because…, our estimates may overstate/understate the true value.”
- “To assess the influence of…, we…; this showed that…”
- “This limitation is unlikely to alter the main conclusion because…”
- “A longitudinal design with… would allow researchers to establish…”
Do this now: open your limitations section and strike through every sentence that does not state a direction of bias or an effect on the conclusions. For each one, either rewrite it using the four steps above or delete it. What remains will be shorter and far more convincing.
Câu hỏi thường gặp
What should be included in a limitations section?
Include the two to four weaknesses most likely to affect your main conclusions, explain which way they could bias the results, describe what you did to mitigate them and suggest designs that would address them in future.
What is the difference between limitations and delimitations?
Delimitations are boundaries you deliberately set, such as studying only one group. Limitations are unintended or unavoidable weaknesses that could bias results, such as a low response rate.
Will listing limitations make reviewers think less of my study?
Specific, well-analysed limitations usually increase credibility. What weakens a paper is a long generic list without analysis, or hiding a major weakness that readers will easily spot.
Where does the limitations section go in a research paper?
It is usually a subsection within the Discussion, after the interpretation of results. Some journals require a separate strengths and limitations section, so check the author guidelines.
Should I mention lack of time or funding as a limitation?
Not as a limitation in its own right. Readers need the consequence for the data and conclusions, for example insufficient power to detect small effects, rather than the circumstances of the research team.
