Find the limitations a research paper admits, and the ones it does not
Every study has weaknesses, and good papers name them: a small or unrepresentative sample, a design that cannot show cause, measures that rely on self-report, a short follow-up. The limitations are usually a paragraph late in the discussion, written briefly. Ask Search+ "What limitations do the authors acknowledge, and how might they affect the main result?" and the answer cites the passages, so you can weigh each weakness against the finding it qualifies.
Creating an account needs no payment.
Last updated October 2026
How to review a paper's limitations with Search+
- Upload the paper
Add the article to a workspace. Limitations sometimes appear in the methods too, for example when the authors explain a compromise in recruitment.
- Ask for stated limitations first
Ask what the authors themselves list as limitations. Open the cited paragraph to read how they frame each one and whether they dismiss it.
- Then ask about likely unstated ones
Ask targeted questions, such as how participants were selected or how the outcome was measured. The cited methods text lets you judge weaknesses the authors did not discuss.
Questions to ask about limitations
What limitations do the authors acknowledge in the discussion?
Do the authors say how any limitation might change the size or direction of the main finding?
Who or what was studied, and do the authors say whether the result applies beyond that group?
Does the study rely on self-reported or proxy measures, and do the authors discuss it?
How much data was missing or lost to follow-up, and how was it handled?
What to look for in a limitations section
Most limitations fall into a few groups: who was studied, how the study was designed, how things were measured, how data was analysed and how long participants were followed. Ask about each group by name.
Authors sometimes name a limitation and then argue it does not matter. Read the cited passage to see whether the argument rests on evidence or on assertion.
A limitations paragraph lists what the authors chose to discuss. Comparing it with the methods can reveal weaknesses they left out, such as an unblinded assessment.
A finding from a convenience sample of volunteers does not automatically hold for the general population. Read the conclusion with the stated limitations beside it.
Search+ cites the supporting excerpt for each limitation it reports, and includes a page or section reference when the paper has one. Search+ does not make up locations.
Limitation wording you may see, and what to ask next
| Wording you may see | What it usually signals | A follow-up question |
|---|---|---|
| "cross-sectional design" | Data from one point in time | Do the authors still describe the result as causal anywhere? |
| "convenience sample" or "volunteers" | A sample that may not represent the population | How were participants recruited, and who was missing? |
| "self-reported" | Measures that rely on participants' own accounts | Was any part of the outcome measured objectively? |
| "residual confounding" | Other factors may explain the association | Which confounders were adjusted for, and which were not? |
| "loss to follow-up" | Participants dropped out | How many dropped out, and did they differ from those who stayed? |
| "underpowered" | The sample may be too small to detect an effect | Do the authors report a power calculation? |
What are the limitations of a research paper?
The limitations of a research paper are the weaknesses in its design, sample, measurement or analysis that restrict how far its results can be trusted or applied, usually acknowledged by the authors in a paragraph of the discussion.
Limitations questions
Can Search+ find the limitations in a paper?
Can it point out limitations the authors did not mention?
Will it say whether a limitation invalidates the result?
Can I compare limitations across studies?
Is this the same as a formal critical appraisal?
Are answers about weaknesses reliable?
Weigh the weaknesses before you rely on the result
Start a workspace, upload the paper, and ask what limits its conclusions.
Start a workspace