HOW SINGAPORE EDUCATION WORKS · Writing university research limitations
Find your next learning step
Write a research limitation by naming a specific feature of the study, explaining how it restricts interpretation and stating what remains uncertain. Then propose a change that addresses that uncertainty. “The sample was small” is only a starting point; the reader needs to know which claim becomes less secure and why.
Full chapter index · Practice and checking · Sources and next reading · How Singapore Education Works
A limitations section should help readers use the findings responsibly. It is not a list of apologies or a place to invent problems that did not occur. Follow your assignment's structure and your discipline's methods. The original exercise below demonstrates reasoning, not a compulsory template for every university.
The choices at a glance
| What you notice | Possible issue | Useful next move |
|---|---|---|
| Lists only a small sample | The consequence for the claim is missing | Specify whose experience the evidence describes |
| Calls a pre–post difference a cause | Alternative explanations were not separated | State the observed change and design limit |
| Proposes only more research | The unresolved question is unclear | Match a follow-up to one uncertainty |
Full chapter index
Define the claim
Work through the evidence
Recall and transfer
Before listing limitations, write the proposed finding in one sentence. Identify who was studied, what was measured, when it was measured and what comparison supports the sentence.
A claim that participants scored higher immediately after a workshop differs from a claim that the workshop caused lasting improvement. The second requires evidence about both cause and persistence that the first does not supply.
Read the methods alongside the findings. A limitation often becomes visible when the wording of the claim reaches beyond the study's design.
If your question is still too broad, our earlier guide to designing university questionnaire questions addresses a different stage of the work. A polished limitations paragraph cannot repair an unclear question after data collection.
CHAPTER 2 OF 8 · Define the claim
2. Separate a limitation from an error and a boundary
A limitation is a constraint on what the evidence can support. An error is something that went wrong, such as a miscalculated total. Correct a known calculation error rather than leaving it in place and calling it a limitation.
A boundary describes what the project deliberately includes. Studying one institution may be an appropriate boundary for a local question. It becomes relevant to limitations when someone wants to apply the result to other institutions.
Do not assume that every project needs the same sample, comparator or statistical method. Ask what the chosen design can reasonably answer. A close analysis of one document and a population survey have different purposes.
Use a short reasoning chain: study feature, consequence for the claim, unresolved question. A proposed improvement should respond to that chain rather than merely sounding more ambitious.
CHAPTER 3 OF 8 · Work through the evidence
3. Worked case: eight volunteers and an immediate quiz
Here is a fictional pilot study. Eight volunteers attend a short learning workshop. Each completes the same ten-question quiz immediately before and immediately after it. There is no comparison group, no delayed test and no random selection from the wider adult population.
For simplicity, every item is scored correct or incorrect, with one point per item. Each participant's post-workshop score is two points higher.
| Participant | Before, out of 10 | Immediately after, out of 10 |
|---|---|---|
| A | 4 | 6 |
| B | 5 | 7 |
| C | 5 | 7 |
| D | 6 | 8 |
| E | 6 | 8 |
| F | 7 | 9 |
| G | 7 | 9 |
| H | 8 | 10 |
Checking answer: the before scores total 48, giving a mean of six. The after scores total 64, giving a mean of eight. The mean increase is two quiz points. The mean percentage score changes from 60% to 80%, a rise of 20 percentage points.
That observed increase is real within the fictional data. It does not, by itself, establish that the workshop caused the change, that the change will last or that all adults would show it.
The relative increase in the mean score is one-third, approximately 33.3%. Presenting that number as “a 33% improvement in adult memory” changes a narrow quiz result into a much broader claim. The quiz has not established a general measure of memory.
CHAPTER 4 OF 8 · Work through the evidence
4. Turn three constraints into useful paragraphs
First consider selection. The participants volunteered. Their interests, availability or prior experience may differ from those of adults who did not volunteer. A defensible statement is that the results describe these eight volunteers; application to a wider population remains uncertain.
Next consider the repeated quiz and missing comparator. Participants may have benefited from seeing the same questions twice, alongside anything learned in the workshop. Without an appropriate comparison, the design cannot separate those explanations.
Finally consider timing. The second quiz occurred immediately afterwards. The study therefore does not establish whether the higher scores persist a week later or appear in a different task.
These constraints affect different claims. Recruiting more volunteers would not automatically remove the repeated-quiz problem. Adding a delayed test would address persistence, but would not by itself establish what caused the initial change.
If a paragraph says only “small sample,” ask what inference is affected. Is the concern precision, coverage of different experiences or transfer to other settings? Specify the concern supported by the actual design.
If the paragraph says “the results are invalid,” check whether the conclusion is too sweeping. The study may accurately describe what happened in the observed group while supporting only limited explanation or generalisation.
If every proposed repair is “more research,” identify the next uncertainty to resolve. A suitable comparison task, a broader recruitment approach or a delayed measurement may answer different questions.
If you discover a factual mistake in your methods account, correct it and reconsider the analysis. Do not conceal an error inside cautious wording. A limitation paragraph should make the evidence easier to assess.
For guided practice, complete this original reasoning statement: “Because the post-test was immediate, the study does not establish ____. A follow-up could ____.”
Checking answer: it does not establish whether the higher quiz scores persist over time. A follow-up could test later, using a justified measurement approach. A delayed test addresses the timing question; it does not automatically resolve selection or causal attribution.
For independent practice, evaluate this proposed conclusion: “The workshop caused every adult to remember one-third more information.” Identify three separate unsupported extensions.
Checking answer: “caused” exceeds the design's ability to distinguish explanations; “every adult” extends eight self-selected volunteers to a population; “remember one-third more information” treats relative change on this quiz as a general memory measurement.
Rewrite it more carefully: “The eight volunteers' mean score on the same ten-item quiz rose from six to eight immediately after the workshop. The design does not isolate the workshop's effect or establish persistence and wider applicability.”
A concise rewrite can preserve the finding while removing claims the study did not test.
CHAPTER 7 OF 8 · Recall and transfer
7. Recall later and transfer to a different design
Tomorrow, reconstruct the three main uncertainties without looking at the table: who the result applies to, what explains the change and whether it persists. Then reopen the case and check your account.
Transfer to a student interview project. The relevant limitations may concern whose experiences were heard, the interview context, interpretation and the researcher's role. Do not paste the quiz study's limitations into a different methodology.
The IPSA-NUS Methods School's introduction to qualitative data analysis connects research design, analysis and the assessment of conclusions within a specific course. That is a local example of methods education, not a requirement for every Singapore degree.
The University of Manchester's Academic Phrasebank discussion guide identifies limitations and future work as functions of research discussion. Its international writing guidance is useful when treated as a way to express sound reasoning, rather than as ready-made sentences that replace it.
Does acknowledging limitations weaken my project?
It narrows the conclusion to what the evidence supports. A modest, justified finding can be more useful than a confident claim with no defensible boundary.
Must I list every possible problem?
Prioritise constraints that materially affect the question and claims. Explain them specifically. Follow the assignment's length and structure rather than filling a paragraph with remote possibilities.
Is a single setting always a weakness?
No. A single setting may suit a local or in-depth question. Explain when and why moving beyond that setting would require additional evidence.
Can I use this structure in a workplace report?
Yes, as a reasoning aid: explain the data boundary and its practical consequences. Use the report's actual requirements. The exercise does not certify an evaluation design or replace professional methodological review.
Sources and next reading
Reviewed 23 October 2026. The pilot, participants, scores, calculations and rewritten findings are entirely fictional and original.
IPSA-NUS Methods School: Introduction to Qualitative Data Analysis supplies the institution-specific Singapore connection. University of Manchester Academic Phrasebank: Discussing findings supplies the international writing connection.
Explore the How Singapore Education Works hub for research and lifelong-learning guides.
Learning at other stages
Continue with a different learning task in this series. Each guide has its own worked cases, practice and limits.
Comparing container capacity in kindergarten