2026.09.22
2026.09.28

How to Write Limitations of the Study: Examples and Tips

A limitation explains how far your findings extend. It affects interpretation of valid results; it does not make those results wrong.
Every strong limitation does 4 things: states the constraint, explains its consequence, gives context, and points to a next step.
Most journal articles need only 3-5 limitations in 150-350 words, placed near the end of the Discussion.
Generic wording such as “small sample size and limited generalizability” costs credibility. Specific, quantified wording earns it.

Table of Contents

What Are the Limitations of a Study?

A limitation is any feature of your sample, design, method, or data that limits how far your conclusions extend. Limitations shape how results are interpreted; they do not make the results invalid.

Almost every researcher hesitates at this section, because it feels like handing reviewers a list of reasons to doubt the work. That instinct is misplaced. Reviewers already suspect the weak points. The only question is whether you name them first, on your own terms.

A well-written limitations section signals 3 things about the author:

You understand your own evidence well enough to know where it stops.
You can distinguish what the study establishes from what it merely suggests.
You can specify the study that should come next.

How Do Limitations, Delimitations, and Errors Differ?

Limitations are constraints you could not avoid, delimitations are boundaries you deliberately chose, and errors are mistakes that must be fixed before submission rather than disclosed.

Feature	Limitation	Delimitation	Error
Origin	Imposed by circumstances, resources, or access	Chosen by the researcher to define scope	Introduced by mistake during design or execution
Where it belongs	Discussion section	Methods or introduction	Nowhere. Correct it before you submit
Example	Only 62 eligible patients existed at the single recruiting site	The study covered public hospitals only	A survey item was reverse-coded incorrectly
Effect on claims	Narrows the conditions under which claims hold	Defines the population the claims target	Invalidates the affected result until fixed

A quick test: if unlimited time and money would have changed the decision, it is a limitation. If you would make the same choice again with unlimited resources, it is a delimitation.

A further distinction matters. “There is limited prior research on this topic” is a gap in the field and a justification for your study. It is not a limitation of your study, and it belongs in the background section or introduction.

Why Journals Require a Limitations Section

Journals require limitations because readers need to know the conditions under which a finding holds. Reporting guidelines such as CONSORT, STROBE, PRISMA, and COREQ all contain a limitations item.

Editors weigh 4 functions, roughly in this order of importance:

Interpretive accuracy: readers learn which conditions the result depends on, such as setting, population, or time period.
Replication value: the section tells the next research team which design features to keep and which to change.
Credibility: an author who spots a subtle confound before the reviewer does earns trust for the rest of the paper.
Guideline compliance: many journals now screen submissions against the relevant reporting checklist before peer review.

Placement of the Limitations Section

In most fields, limitations are placed in a subsection near the end of the Discussion, just before the conclusions. Check the target journal’s author guidelines first, because conventions vary by discipline.

Placement	Common Fields	Notes
Subsection at the end of the Discussion	Medicine, psychology, social sciences	The default convention. Usually placed before implications or conclusions.
Woven into the Discussion	Humanities, some qualitative traditions	Each caveat sits beside the claim it qualifies, which suits interpretive work.
Standalone section after the Discussion	Engineering, computer science	Often titled “Limitations and Future Work” and paired with next steps.
Dedicated chapter section	Theses and dissertations	Expanded, and normally followed by a separate recommendations section.

There are also 2 secondary placements worth knowing. Some journals require 1 clause in the abstract, such as “findings are based on a single-center cohort and require external validation.” A brief note in the methods can also justify a specific choice at the point where you describe it.

Types of Limitations to Check For

Work through these 7 categories systematically rather than relying on memory. Most studies have 3-4 real limitations, and researchers routinely find 1 or 2 they had not consciously registered.

Sample and Participant Limitations

Sample size too small to detect effects of the size you expect.
You used non-probability sampling, such as convenience, snowball, and self-selected samples.
Narrow demographic, geographic, or institutional range.
Non-random attrition, low response rate, or survivorship bias: These mean that your final sample isn’t representative, because the missing cases went missing for a reason connected to what you’re studying. Attrition: dropouts differ from stayers (e.g., the participants finding an exercise program hardest quit). Low response rate: only people with strong opinions bother replying. Survivorship bias: failed cases never entered the data at all, like studying only startups that are still running after 5 years and concluding that stock options improve job satisfaction.

Design Limitations

Cross-sectional design that prevents causal inference or ordering of variables.
No control group, no comparison condition, or non-random assignment.
Follow-up period shorter than the natural timeline of the outcome.
Blinding of participants or assessors that was not feasible.

Measurement Limitations

Self-report measures exposed to recall bias, social desirability, or common-method bias.
Instruments validated in a different population, language, or era.
Proxy variables standing in for the construct you actually care about.
Single-item measures for multidimensional constructs, or inter-rater reliability below the accepted threshold.

Data Limitations

Secondary data collected for another purpose, with variables that map imperfectly onto your constructs.
Missing data, and the assumptions behind whatever imputation strategy you used.
Records filtered by an administrative process you cannot observe.
A collection window disrupted by an external shock such as a pandemic or a policy change.

Analysis Limitations

Statistical assumptions that were violated or only approximately met.
Multiple comparisons handled without correction, or with a correction that costs substantial power.
Model specification choices that other reasonable analysts might make differently.
Effect sizes estimated with wide confidence intervals.

Scope and Theoretical Limitations

A flawed or extra narrow theoretical framework that foregrounds some mechanisms and obscures others.
Findings are constrained by a specific cultural, regulatory, or temporal context.
Constructs defined in a way that does not match every use in the wider literature.

Researcher and Practical Limitations

Researcher positionality and its influence on qualitative interpretation, which is a genuine limitation rather than a confession.
Language barriers during the literature search or data collection.
Funding, time, or access constraints that shaped the design.
Restricted access to proprietary, classified, or clinical records.

How to Write the Limitations Section

Every effective limitation 1) states the constraint, 2) explains its consequence, 3) provides context, and 4) points toward a solution. Authors often skip 2) and 4), and their absence is what makes a section read like you’re just checking off boxes or shoving in boilerplate text.

Step	What It Does	Example Wording
1. State	Names the constraint plainly, without hedging or apology	“The design was cross-sectional, with all measures collected at a single time point.”
2. Explain	Specifies what conclusion is now off limits, ideally with a number	“Power analysis indicated 80% power to detect effects of d = 0.72 or larger, so smaller differences may have gone undetected.”
3. Contextualize	Justifies the choice or reports a check that partly reassures	“Our findings are reflective of the population available at a single specialist clinic, where prevalence is over 10%.”
4. Point forward	Names the specific design that would resolve the issue	“Multi-site recruitment with a 6-month follow-up would detect the more modest effects plausible in routine care.”

Weak Versus Strong Wording

A weakly described limitation restates something the reader can already see in your methods. You instead should tell the reader what to do with the result.

Issue	Weak Version	Strong Version
Sample size	“A limitation of this study is that the sample size was small.”	“The final sample of 62 gave 80% power to detect effects of d = 0.72 or larger, so smaller but clinically meaningful differences may have gone undetected.”
Self-report	“Self-report data may be biased.”	“Because burnout and workload were reported in the same instrument, common-method variance may inflate their association. Harman’s test showed no dominant factor, at 34% of variance.”
Design	“The study was cross-sectional.”	“With 1 measurement point, the direction of the autonomy and satisfaction relationship cannot be established; satisfied employees may reinterpret existing autonomy more favorably.”
Scope	“Results may not generalize.”	“Recruitment was limited to 3 firms in 1 metropolitan region with well-established remote-work norms, so the relationship may differ in sectors requiring on-site presence.”
Future work	“Further research is needed.”	“A cross-lagged panel design with supervisor-rated outcomes across at least 2 sectors would test the causal ordering proposed here.”

Common Limitations by Study Type

Each subsection lists the constraints reviewers look for first.

Quantitative Survey Research

Cross-sectional timing that blocks causal claims.
Common-method variance from a single self-report instrument.
Sampling frame restricted by sector, region, or employer.

Experimental Research

Blinding that was not feasible, and the demand characteristics that follow.
Follow-up shorter than the outcome’s natural timeline.
Laboratory conditions that trade ecological validity for control.

Qualitative Interview Research

Transferability rather than statistical generalizability.
Volunteer bias among participants with strong views.
Researcher positionality and its effect on what went unsaid.

Mixed-Methods Research

Sequencing that narrows the questions asked in the later phase.
Overlapping participant pools that weaken corroboration.
Unequal weighting of the 2 strands during integration.

Systematic Reviews and Meta-Analyses

Search restrictions by language, database, or publication date.
Publication bias inflating the pooled estimate.
Heterogeneity that makes the pooled effect an average across unlike interventions.

Secondary Data Analysis

Constructs available only as coarse proxies.
Missing data and the assumptions behind your imputation.
Collection purposes that differ from your research question.

Phrases That Work and Phrases to Avoid

Function	Use	Avoid
Opening the section	“Several limitations should be considered when interpreting these findings.”	“Unfortunately, this study suffers from a number of problems.”
Naming a constraint	“The cross-sectional design precludes conclusions about causal ordering.”	“The study failed to establish causality.”
Stating consequence	“This may attenuate the observed association.”	“This makes the results questionable.”
Giving context	“A sensitivity analysis excluding these cases produced similar results.”	“We did the best we could with the time available.”
Pointing forward	“Replication in primary care settings would establish whether the effect persists.”	“More research is needed.”

Delete the words “unfortunately,” “sadly,” “we regret that,” and “due to the researcher’s inexperience” wherever they appear. All keep you from sounding confident and capable.

What Mistakes Should You Avoid When Writing Limitations?

The most common mistakes are listing constraints without explaining their consequences, over-disclosing trivial issues, disguising errors as limitations, and recycling generic template wording that could appear in any paper.

Mistake	Why It Hurts	Fix
Listing without interpreting	Reads as box-ticking and leaves the reader to guess the implications	Attach at least 1 clause of consequence to every constraint
Over-disclosing	Buries the constraints that actually matter under generic ones	Include a limitation only if a knowledgeable reader would conclude differently knowing it
Undermining your contribution	After 4 caveats the reader concludes the study shows nothing	Follow each concession with what the finding still supports
Disguising errors as limitations	Invites rejection, because an error is a repairable defect	Correct the error and re-run the analysis before submitting
Introducing new results	Breaks the structure reviewers expect and looks like post hoc rescue	Report sensitivity checks in results, then reference them here
Contradicting the abstract	Reviewers notice when the abstract claims something that the limitations make impossible	Align claim strength across abstract, results, and conclusions
Recycling a template	Signals that little thought went into the section	Make the wording specific enough that it could not be pasted into another paper

A Step-by-Step Writing Workflow

Keep a running log from week 1 of data collection, noting every compromise, workaround, and access problem. You will not remember them at write-up.
Read your own methods adversarially. Take each design decision and ask what a hostile reviewer would say about it.
Screen your study against all 7 categories listed above, rather than relying on the 2 or 3 that come to mind.
Rank the candidates by consequence. Start with the limitation that would most change a reader’s interpretation, and drop anything trivial.
Draft each surviving limitation following 4 steps: state, explain, contextualize, point forward.
Check consistency across the manuscript. If the abstract or conclusion overclaims, adjust the claim rather than softening the limitation.
Ask a colleague outside the project to read only the discussion, then tell you what they think the study shows.

Frequently Asked Questions

What Is an Example of Limitations in a Research Paper?

A typical example: “Because the design was retrospective, exposure was drawn from routine clinical notes rather than measured prospectively, so some patients were likely misclassified. Records kept for treatment rather than research also left smoking status unrecorded for 40% of the cohort, preventing adjustment for a known confounder.”

This wording names the constraint, identifies the mechanism, and states how the results and their implications are affected. Compare it with “the study was retrospective,” which adds nothing a reader hasn’t already learned from the Methods section.

How Do You Start a Limitations Paragraph?

Open with a neutral framing sentence such as “Certain limitations should be considered when interpreting these findings,” then move straight to the most important limitation. Avoid apologetic openers like “Unfortunately this study has several limitations”.

What Is the Difference Between Limitations and Delimitations in Research?

Delimitations are boundaries you chose, such as studying 1 age group or 1 country. Limitations are constraints you could not control, such as low response rates. Delimitations belong in the methods; limitations belong in the discussion.

How Many Limitations Should a Research Paper Have?

Aim for 3-5 in a journal article. Fewer than 3 usually means you have not looked hard enough, and more than 5 dilutes the ones that matter. 5-8 limitations are acceptable in a dissertation because examiners are assessing methodological judgment in more depth.

Should Limitations Be Mentioned in the Abstract?

Only when the journal requires it or when the abstract would mislead readers without it. A single clause is enough, for example: “findings are based on a single-center cohort and require external validation.” Save the detail for the discussion.

Is a Small Sample Size Always a Limitation?

Not automatically. A small sample is a limitation when it leaves you underpowered for the effects you care about, or when it restricts the range of participants. In qualitative work, a sample of 15-20 may be entirely appropriate, and the relevant issue is saturation rather than size.

How Do You Write Limitations in a Qualitative Study?

Frame them around transferability rather than generalizability. Address volunteer bias, the range of settings sampled, saturation, and your own positionality. Describe the steps you took to manage each, such as reflexive memos, second coding, or member checking.

Do Limitations Reduce Your Chances of Publication?

Only concealing your limitations makes it more likely to get rejected. Peer reviewers and editors expect constraints in every study, and a frank, specific section signals methodological competence. What harms a submission is a serious limitation that a reviewer discovers and frames less charitably than you would have.

AI Model and Tool Updates: What Researchers Need to Know

Quota Sampling in Research: Steps, Types, Examples