How to Write the Limitations Section of a Study (Without Undermining It)

Naming a limit is not a weakness

How to write limitations of a study is often taught as a defensive exercise: name what your study did not do, and hope the panel does not press further. That framing produces limitations sections that read as apologies. Examiners read them as apologies and follow up harder, because apology signals uncertainty about what the study can support.

Ross and Bibler Zaidi (2019), publishing in Perspectives on Medical Education, recommend a specific four-part template for limitations that flips the tone from apology to judgement. The template shows the panel that the writer knows what the study can and cannot support, has considered the alternative approaches, and has taken specific steps to mitigate the effects of the limit. The rest of this piece walks the template and applies it to the three limit types every thesis has to handle.

The four-part template

Every load-bearing limit in a study is handled in the same shape.

Describe the limit specifically. “The sample of 214 participants was drawn from three sites in one country, which restricts the generalisability of the finding to comparable settings.”

Explain the implication. “The pattern observed may reflect features of the three sites rather than the phenomenon under study more broadly. Findings should be treated as suggestive rather than definitive for non-comparable settings.”

Offer alternative approaches that would address the limit. “A multi-country replication with matched sampling at each site would test whether the pattern holds beyond the studied context.”

Describe the mitigation steps taken. “Site selection was designed to include one urban, one suburban, and one rural setting, which strengthens external validity within the sampled range. The analysis reports site-specific effects alongside the pooled estimate.”

All four parts matter. Describe alone reads as a confession. Describe with implication reads as fatalistic. Describe with implication and alternative shows that the writer knows what the field would do next. Adding the mitigation shows that the writer has already done what was possible inside the study’s constraints. Ross and Bibler Zaidi (2019) argue that all four together demonstrate the calibrated judgement examiners are actually testing for.

Limit type 1: Sample limits

Every study has a sample, and every sample has boundaries. The limit is not the sample itself. The limit is what the sample cannot support.

• Describe: the sample of X participants from Y setting.
• Implication: findings apply to comparable settings; caution warranted for settings that differ on Z.
• Alternative: sampling across more sites, or across a maximum-variation range that includes Z.
• Mitigation: what site selection or recruitment strategy was used inside the study to broaden the range.

A thesis that names the sample limit at this level of specificity closes the “how generalisable is this?” question in the room. A thesis that says “the sample size is a limitation” and stops leaves the question open.

Limit type 2: Design limits

Every design rules some questions out. A cross-sectional design cannot answer change-over-time questions. An experimental design cannot answer questions about ecologically valid settings. A qualitative design cannot answer prevalence questions.

• Describe: the design (cross-sectional, experimental, qualitative, etc.).
• Implication: what the design rules out. State this in the language of the question the design cannot answer.
• Alternative: the design that would answer the ruled-out question, named specifically.
• Mitigation: any features of the study that reduce (but do not eliminate) the limit, such as retrospective questions in a cross-sectional design, or careful reporting of context in an experimental one.

Trafford and Leshem (2008) note that examiners return to design limits repeatedly in the viva. A design-limit paragraph that has already run the four-part template makes the return question a check rather than a challenge.

Limit type 3: Measurement and analytic limits

Every measure has boundaries: what it captures, what it misses, and how sensitive it is to change. Every analytic approach makes assumptions that may or may not hold in the data.

• Describe: the measure or analytic assumption, specifically.
• Implication: what conclusions this rules out, or what confidence in a conclusion has to be adjusted.
• Alternative: the measure or analytic approach that would address the constraint.
• Mitigation: any sensitivity analyses, robustness checks, or triangulations that were run inside the study to test the constraint.

Cohen, Manion and Morrison (2018) emphasise that a strong methodology chapter includes a limitations paragraph that reads as evidence of judgement, not as a list of things the study did not do. The four-part template produces exactly that reading.

What to place in the limitations section, and what to keep out

Limitations sections should include the constraints that affect the interpretation of the findings. They should exclude two things: minor issues that any reader could recover from, and issues that are not actually limits.

Include: sample constraints, design constraints, measurement constraints, analytic assumptions that may not hold, ethical constraints on design, and any decisions that reduced the study’s scope.

Exclude: typos, minor editing inconsistencies, and problems that were mitigated so completely that they are no longer limits. Also exclude “the study is limited by resources and time,” which is a constraint on doctoral work generally, not a limit of this study specifically.

Including too much reads as defensive. Excluding too much reads as unaware. The balance is: name the limits an experienced reader would identify, at the specificity a critical review would demand.

Where the limitations section sits in the thesis

Two placements are common: at the end of the methodology chapter and at the end of the discussion chapter. The best-argued theses often include both, with different emphases.

In the methodology chapter: the paragraph focuses on what the design and sample cannot support. This lets the reader move through the results chapter with an accurate sense of the study’s reach.

In the discussion chapter: the paragraph revisits the limits in light of what the findings actually showed. Some limits will have been less consequential than expected; others will have shaped the results more than the design anticipated. This is where the “next study” section connects.

The Thesisroom method treats the limitations paragraph as a whole-chapter audit target. Thesisroom’s structural critic reviews the limitations section for the four-part template and flags passages where the template is incomplete.

What Thesisroom does with this

Thesisroom’s structural critic reads the limitations section and returns comments on the specific passages where the four-part template is incomplete: description-without-implication, implication-without-alternative, alternative-without-mitigation. The Mock Viva Coach generates the examiner questions those passages invite, so the writer can rehearse the response before the panel arrives.

The tool does not decide which limits to include. That is the writer’s judgement. It reviews how the limits that are included are written.

Frequently asked questions

How long should the limitations section be?

Long enough to run the four-part template on each load-bearing limit. Most doctoral limitations sections run 500 to 1,500 words in a discussion chapter, plus a shorter paragraph in the methodology chapter. Longer than this usually signals defensive writing; shorter usually signals limits that were not fully worked through.

Should I name limits my supervisor did not raise?

Yes, if you know they are there. Examiners find limits that supervisors missed, and volunteering them in the thesis is stronger than being caught by them in the viva. Wisker (2012) treats limit-naming as one of the specific signals of doctoral maturity that panels reward.

Can naming limits weaken my contribution claim?

Only if the limits contradict the claim you are making. If your contribution claim is scoped correctly, the limits describe what lies outside the scope, not what undermines the contribution inside it. A well-scoped small claim with named limits reads stronger than an overclaimed large one with no limits.

What are the most common limitations examiners look for?

Sample generalisability, design constraints on the causal question, measurement validity, and analytic assumptions. Every panel checks these four. A thesis that has already named them at the four-part level of detail has answered most of the limitations discussion before it starts.

Do I have to name a mitigation for every limit?

If one exists in the study, yes. Not every limit has an in-study mitigation, and it is honest to say so: “no mitigation was feasible within the design.” What is not honest is naming a limit and implying a mitigation that was not actually done.

Can Thesisroom review my limitations section?

Yes. The Structural Critic reads the section and flags passages where the four-part template is incomplete. It does not decide which limits should be there, because that is the writer’s judgement. It reviews how the limits are handled once named.

One takeaway before you close this tab

The strongest limitations sections are not confessions. They are demonstrations of judgement, run through the four-part template of Ross and Bibler Zaidi (2019): describe, implicate, offer an alternative, name the mitigation. Every load-bearing limit gets the full pattern. When you want the section audited chapter by chapter, Thesisroom’s structural critic is where the review lives.

References

Cohen, L., Manion, L., & Morrison, K. (2018). Research Methods in Education (8th ed.). Routledge.

Ross, P. T., & Bibler Zaidi, N. L. (2019). Limited by our limitations. Perspectives on Medical Education, 8, 261-264.

Trafford, V., & Leshem, S. (2008). Stepping Stones to Achieving your Doctorate. Open University Press.

Wisker, G. (2012). The Good Supervisor (2nd ed.). Palgrave Macmillan.