The limitations of a study are the weaknesses in its design, sample, data or analysis that restrict what its results can show. Every study has them, from a survey of your classmates to a multisite clinical trial. The limitations section is where you name yours, explain how they affect your findings and say what a stronger study would do differently. Written well, it makes your paper more convincing, not less.
Below you'll learn how to find your study's real limitations, how to write limitations that read as judgment rather than apology, and how they differ from delimitations. You'll also get sample paragraphs for seven types of student research and a template. The advice draws on the USC Libraries research guide, APA's reporting standards and peer-reviewed papers, all linked so you can check them.
A limitation is a weakness in how a study was designed or carried out that may influence its outcomes and conclusions. That definition comes from Ross and Bibler Zaidi's 2019 paper "Limited by our limitations" in Perspectives on Medical Education. The USC Libraries guide to the limitations of the study frames it in terms of what you can claim: limitations constrain your ability to generalize from your results, apply them to practice or interpret what they mean.
Some limitations are built into the design you chose. A one-time survey, for example, can't show cause and effect. Others appear while you work: a low response rate, a site that withdraws access, a question you wish you had asked. Both kinds belong in the section if they change how a reader should read your results. This applies to any empirical research project, whether you collected numbers, interviews or observations.
Graders look for it because it shows you understand your own evidence. The USC guide is blunt about this: it is far better to identify your limitations yourself than to have your professor point them out and lower your grade because you seemed not to notice them.
Published research is held to the same standard. APA's Journal Article Reporting Standards for quantitative research ask authors to interpret results while taking into account sources of potential bias, threats to internal and statistical validity, imprecision of measurement and the adequacy of sample sizes, and to discuss how far the findings generalize. Several widely used reporting guidelines give limitations their own checklist item: item 19 in STROBE for observational studies, item 30 in the CONSORT 2025 checklist for randomized trials, and items 23b and 23c in PRISMA 2020 for systematic reviews.
Most research limitations fall into a few categories tied to the stage of the study where they arise. Go through this table with your methods section open and note every row that applies to you.
| Category | Typical examples | What it limits |
|---|---|---|
| Sample | Small sample, convenience or volunteer sample, low response rate, dropouts | Who the results describe and whether smaller effects could be detected |
| Design | Cross-sectional or correlational design, no control group, short follow-up | Whether you can claim cause and effect or long-term effects |
| Measurement | Self-reported answers, an instrument never tested on your population, one question for a complex idea | How accurate the data are in the first place |
| Data | Missing responses, secondary data collected for a different purpose | Which variables you could analyze and how well they match your concepts |
| Analysis | Unplanned post-hoc tests, many comparisons, statistical assumptions only roughly met | Confidence that a significant result isn't a coincidence |
| Setting and time | One school, one clinic, one semester, data gathered during an unusual period | Whether findings hold in other places or at other times |
| Researcher | Limited access, language barriers, your own expectations or background | What you could collect and how you interpreted it |
Sample problems are the easiest limitations to spot and the easiest to write badly. In quantitative work, a small sample makes significant relationships hard to find, as the USC guide notes, but size matters less in qualitative research, where the question is whether you gathered enough depth. Composition often matters more than size. Ross and Bibler Zaidi point out that even with appropriate sampling, a study can end up limited to the people who chose to take part, which is self-selection bias. If the students who answered your survey care more about the topic than those who ignored it, your results lean their way.
A correlational or cross-sectional design can show that two things move together but not which one drives the other, and Ross and Bibler Zaidi note that observational studies often fail to remind readers of this. Self-reported data can rarely be checked independently. USC lists four biases to watch for in it: selective memory, telescoping (recalling an event as happening at a different time), attribution (crediting yourself for good outcomes and outside forces for bad ones) and exaggeration. Two more are social desirability bias, where people give the answer they think the researcher wants, and the Hawthorne effect, where people change their behavior because they know they're being observed. Our guide to internal and external validity covers the threats behind many of these problems.
Limitations are weaknesses you didn't choose, while delimitations are boundaries you set on purpose. Theofanidis and Fountouki's review "Limitations and Delimitations in the Research Process" describes a limitation as an imposed restriction that is essentially out of the researcher's control, and delimitations as limits the authors consciously set so that the study's aims stay achievable. Ross and Bibler Zaidi draw the same line: delimitations are the deliberate inclusion and exclusion decisions you make while planning a study.
| Limitations | Delimitations | |
|---|---|---|
| Who controls it | Usually outside your control, or the unavoidable cost of a design choice | You choose it |
| When it arises | During or after the study | While planning, before data collection |
| Examples | Low response rate, dropouts, self-report bias, no access to a key group | Only first-year students, one state, sources from 2015 to 2025, English-language articles |
| Where it usually goes | Discussion | Introduction or methods, often under "Scope" |
| What you write | The weakness, its effect on your results and what would fix it | The boundary and why you chose it |
A quick test is to ask whether you decided it in advance. Theofanidis and Fountouki describe delimitations as less about "why I did this" than "why I did not do it like this," so a good delimitation statement names the options you rejected and gives a reason.
The two terms often describe the same decision from opposite sides. Say you study only first-year students at your own university. That boundary is a delimitation, and you justify it in your methods. Its consequence, that your results may not describe juniors or students at other schools, is a limitation, and you discuss it in your discussion.
Delimitation (methods): We surveyed only first-year students because the research question concerns the transition into college-level coursework.
Limitation (discussion): Because all participants were first-year students at one public university, the findings may not describe students later in their degree or at colleges with different advising systems.
Placement varies by program. In theses and dissertations, scope and delimitations often appear in the first chapter. Colorado State University's School of Education, for example, lists "Delimitations" and "Assumptions and limitations" among the Chapter 1 headings in its suggested dissertation proposal format. Use your own program's template if it has one.
To write limitations well, give each one four parts: name it, explain its effect, say what you did about it and point to what would fix it. This follows Ross and Bibler Zaidi, who argue that a meaningful limitation should describe the problem, explain its implication, offer alternative approaches and describe steps taken to reduce it. The USC guide adds that you should explain why each limitation couldn't be overcome with the methods you chose.
Imagine your survey of undergraduates found that students who reported more sleep also reported higher GPAs. Here is a self-report limitation with all four moves, numbered so you can see each one:
(1) Both sleep and GPA were self-reported, so the data may reflect what students remembered or hoped was true rather than what actually happened. (2) If students with lower grades rounded their GPA up, or if students overestimated their sleep, the association we found could be stronger or weaker than the true relationship. (3) To ease the pressure to give flattering answers, the survey was anonymous and collected no names or student ID numbers. (4) A follow-up study could pair sleep-tracker data with registrar-verified GPA, with participants' consent, to test whether the association holds with objective measures.
Limitations usually go in the discussion section, either at the start, so readers know them before your interpretation, or at the end, where they lead into suggestions for future research. That's the USC guide's advice, and it warns against burying them in the middle of the discussion unless a limitation applies only to the point you're making there. In that case, restate it at the end.
| Paper type | Where limitations usually go |
|---|---|
| APA-style research paper or journal article | A paragraph or "Limitations" subsection in the discussion |
| Thesis or dissertation | Often scope, delimitations and expected limitations in Chapter 1; the limitations that affected your results in the final discussion chapter, often next to recommendations |
| Lab report | Discussion, usually framed as sources of error and their effect on your measurements |
| Literature review or systematic review | Discussion: limitations of the studies you reviewed and of your own search process |
| Research proposal | Anticipated limitations, usually in the methods or a short section of their own |
Watch your tenses. USC recommends the past tense because you identified the limitations after completing the research: "The sample included only..." Switch to the present tense for what they mean now: "This means the results may not apply to..." If you're still drafting the earlier sections, our guide on how to write a research paper covers the full structure.
Each example below is a sample limitations paragraph for a common type of student research. The studies are invented, so swap in your own details and numbers. What you should copy is the structure: each paragraph names the weakness, explains its effect and says what would fix it.
This study has two main limitations. First, participants were recruited through a link shared in three introductory psychology courses, so the sample over-represents first-year psychology students and may not reflect the wider student body. Second, the cross-sectional design measured social media use and anxiety at a single point, so it cannot show whether heavy use raises anxiety or whether anxious students turn to social media more. A longitudinal study that follows a sample drawn from the full enrollment list across a semester would address both issues.
Students were assigned to the flashcard or practice-test condition by class section rather than at random, so the groups may have differed before the study began. The practice-test section met in the morning, and time of day could explain part of the difference in quiz scores. Pretest scores were similar across sections, which reduces but does not remove this concern. The quiz was also given one week after the study session, so the results say nothing about retention at the end of the term. Random assignment within sections and a later follow-up quiz would strengthen future work.
The 12 participants volunteered after a call posted in a first-generation student group, so the findings likely reflect students who already feel connected enough to campus to respond; the most isolated students may be missing from the data. As a first-generation student myself, I may have heard some accounts as more familiar than they were. To check this, I kept reflexive memos and had a second coder review a third of the transcripts. Readers should judge how well these findings transfer by how closely their setting resembles a mid-sized public university with an active first-generation program.
Notice the last sentence. Qualitative studies speak of transferability rather than statistical generalization, and APA's reporting standards for qualitative research ask authors to identify the study's strengths and limitations and to describe the limits of the scope of transferability.
This analysis used a public county-level dataset collected for another purpose, so median household income served only as a rough proxy for access to healthy food. Counties with missing values were dropped; if missing data were more common in rural counties, the sample may lean toward urban areas. Most important, the associations reported here are correlational and cannot show that income causes the differences in obesity rates. Future work could combine grocery store location data with individual-level surveys.
As a single-case study of one nonprofit's move to remote volunteering, this research describes how the change unfolded in an organization with its own leadership, funding and volunteer base. It does not estimate how common these outcomes are across nonprofits. Access was also limited: the executive director approved interviews with staff but not with board members, so the account of why the decision was made comes mainly from the staff who carried it out. Comparative case studies that include board perspectives would test whether the same pattern appears elsewhere.
The search covered two databases and English-language articles published from 2015 to 2025. These boundaries kept the review manageable, but relevant studies in other languages or indexed elsewhere may have been missed. Within the included studies, most relied on self-reported screen time and few followed participants for more than a year, so the evidence itself limits what can be concluded about long-term effects. If studies with null results were less likely to be published, this review may also overstate the strength of the association.
This example separates limitations of the evidence from limitations of the review process, the same split PRISMA 2020 makes in items 23b and 23c. The date range and database choices are delimitations; the studies they may have missed are the limitation.
The calculated specific heat of the metal was lower than the reference value. The most likely source of error is heat lost to the room, both while the hot sample was moved into the calorimeter and through the open top of the foam cup, which would lower the final water temperature and therefore the calculated specific heat. The thermometer read only to the nearest 0.5 °C, limiting precision for small temperature changes, and each measurement was run only twice because of lab time. A lidded calorimeter, a digital probe and at least three trials would reduce these errors.
Strong limitations tell the reader what to do with your result, while weak ones only repeat what your methods section already said. Compare these pairs:
| Weak | Strong |
|---|---|
| "The sample size was small." | "With 19 students in each group, the study could reliably detect only large differences, so smaller effects may have gone unnoticed." |
| "Self-reported data may be biased." | "Because GPA was self-reported, students with lower grades may have rounded up, which could weaken the observed link between sleep and grades." |
| "The results may not generalize." | "All participants attended one small private college, so the findings may not describe students at large public universities or community colleges." |
| "The study was cross-sectional." | "Because data were collected once, the study cannot show whether stress reduced sleep or poor sleep increased stress." |
| "More research is needed." | "A study that follows the same students through a full academic year would show whether the pattern holds as workloads change." |
Only claim what your analysis supports. If you didn't run a power analysis, describe the sample's size and makeup instead of stating what it could detect.
The University of Manchester's Academic Phrasebank lists dozens of phrases under "Recognising the limitations of the current study." A few that suit student papers:
Leave out "unfortunately," "we regret," "due to my inexperience" and "the study failed to." They make you sound apologetic without telling the reader anything about the results.
The biggest mistake is listing problems without saying how they affected your results. The USC guide warns that a bare list leaves the reader wondering whether, or how, those limitations changed your findings. These other mistakes also cost points:
Use this template as a starting point, then cut any sentence that doesn't fit your study:
These findings should be read with several limitations in mind. First, [name the limitation, with specifics]. Because of this, [explain the effect: the direction of any bias, or who the results do and don't apply to]. [State a step you took to reduce it, or why it couldn't be avoided.] Second, [limitation]. This means [effect]. [Step taken.] Notwithstanding these limitations, [what the findings still support]. Future research could [a specific design that addresses the main limitations].
Before you submit, check that:
Want a second opinion before you hand it in? A tutor can review your draft limitations section and give feedback on whether each point is specific and supported, or walk you through the logic of your research design. For a full paper, our research paper editing service can check that your limitations line up with your methods and conclusions.
Common examples include a small or non-random sample, a low survey response rate, self-reported data, a cross-sectional design that can't show cause and effect, no control group, missing data, a short study period and limited access to participants or records. In qualitative work, volunteer bias and the researcher's own position are typical. The label matters less than explaining how that specific weakness affected your results.
No style guide sets a number. Include each limitation that would change how a careful reader interprets your results, and leave out minor ones. In many class papers that's a short paragraph covering a few points, while a thesis or dissertation usually discusses more, in more depth. Ross and Bibler Zaidi note that redundant or irrelevant limitations waste word count, so if your rubric sets a length, write to it.
No. In quantitative research, a small sample is a limitation when it leaves the study unable to detect the effects you're testing or narrows who the results represent. In qualitative research, the USC Libraries guide notes that sample size is generally less relevant if you explain it in the context of your research problem, because the goal is depth rather than statistical generalization. Say why your sample size is or isn't a problem for your question.
Scope describes what your study covers: the population, setting, time period and variables you chose. Those deliberate boundaries are your delimitations. Limitations are the weaknesses that restrict what your results can show within that scope, such as a low response rate or biased self-reports. A "Scope and Limitations" section, common in theses, covers both: what you studied and why, then what may weaken your conclusions.
In most research papers they go in the discussion section, either at the start or at the end, where they lead into recommendations for future research. The USC guide advises against burying them in the middle. Some papers and theses use a separate "Limitations" subsection or a "Limitations and Future Research" heading near the end. Follow your assignment instructions, since some instructors ask for limitations in the conclusion instead.
Use the past tense for what happened in your study, since you're reporting weaknesses you identified after the research was done: "The sample included only transfer students." Use the present tense for what those weaknesses mean now: "The results may not apply to first-year students." In a research proposal, where no data exist yet, describe anticipated limitations in the future tense: "The study will rely on self-reported data."
Yes. Qualitative studies aren't judged on statistical generalizability, but they still have limitations, and APA's qualitative reporting standards ask authors to identify a study's strengths and limitations and the limits of its transferability. Typical examples are volunteer bias, a narrow range of settings, the researcher's own position shaping interviews or interpretation, and limited time in the field. Describe what you did to manage each one, such as keeping reflexive memos or using a second coder.
Post your assignment, compare tutor bids and work with an expert who explains each step.
Professional Research Paper Editing
Get affordable and top-notch help for your essays and homework services from our expert tutors. Ace your homework, boost your grades, and shine in online classes—all with just a click away!
Fast, secure, and handled by vetted experts.
Learn why an ALEKS Knowledge Check removes topics from your pie, what the drop means, and a step-by-step plan to win lost topics back faster next time.
Read More
Learn how to cram for a test with a smart plan: triage the material, use active recall and practice questions, protect your sleep and review on test day.
Read More
Each country's growth revolves around the ⭐10 PRINCIPLES OF ECONOMICS⭐ Get to know each of the recommended ten principles by economist Gregory Mankiw
Read More