Find a study's sample size and how many people it actually analysed
A paper can report several different numbers for its sample: how many people were screened, how many were enrolled, how many were randomized and how many were finally analysed. The gap between them matters, and so does whether the authors planned the size in advance. Ask Search+ "How many participants were enrolled and analysed, and how was the sample size decided?" and the answer cites each figure, so you can check them against the paper's flow diagram or tables.
Creating an account needs no payment.
Last updated October 2026
How to check a study's sample size with Search+
- Upload the paper
Add the article. Participant numbers often sit in a flow diagram or a table, so compare the cited excerpt with the figure itself before you use a count.
- Ask for each stage of the count
Ask how many were screened, enrolled, allocated and analysed, as separate questions or one question that asks for each stage. Mismatched numbers are easier to spot that way.
- Ask how the size was justified
Ask whether the authors report a power calculation or another rationale, and open the cited methods passage to read the assumptions.
Questions to ask about sample size
How many participants were enrolled, and how many were included in the final analysis?
Do the authors report a sample size or power calculation, and what effect size did they assume?
How many participants dropped out or were excluded after enrolment, and for what reasons?
How many participants were in each group or arm?
For an interview study, how many people were interviewed and how did the authors decide when to stop recruiting?
Across the papers in this workspace, which studies have fewer than one hundred participants in the analysis?
What to look for when checking a sample size
Totals can appear in the abstract, the methods, a flow diagram and the first results table. They should agree; when they do not, ask which figure applies to the main analysis.
A power calculation states how many participants were needed to detect an expected effect. If recruitment fell short, the authors should say so and discuss it.
A small sample produces wide confidence intervals and can miss real effects. A large sample can make trivial differences statistically significant. Read the size together with the effect and its interval.
Interview and case studies are not sized by power calculations. Look for how the authors justify their sample, often by saturation or the depth of each case.
Search+ answers with inline citations to the excerpt that reports each number, with a page or section reference when the paper has one.
Ask one sample size question across every study in the workspace and the answer cites the figure in each paper.
Sample size wording you may see, and what to ask next
| Wording you may see | What it usually signals | A follow-up question |
|---|---|---|
| "were assessed for eligibility" | The number screened before enrolment | How many of those were enrolled, and why were the rest excluded? |
| "were randomized" | The number allocated to groups | How many in each group completed the study? |
| "power calculation" or "80% power" | A planned sample size | What effect size and significance level did the calculation assume? |
| "intention-to-treat" | Everyone randomized is analysed in their assigned group | Was a per-protocol analysis also reported, and did it differ? |
| "lost to follow-up" | Participants with missing outcome data | How was missing data handled in the analysis? |
| "until saturation" | A qualitative stopping rule | How did the authors judge that saturation was reached? |
What is sample size in a research paper?
Sample size is the number of participants, cases or observations a study includes. Papers usually report how many were enrolled and how many were analysed, and quantitative studies often explain how the size was planned with a power calculation.
Sample size questions
Can Search+ find the sample size of a study?
Will it tell me if the sample is big enough?
Why do I see different numbers in the same paper?
Can it read numbers from a flow diagram?
Can I compare sample sizes across many papers?
Are the reported numbers always correct?
Count who was really in the study
Start a workspace, upload the paper, and ask how many were enrolled, analysed and lost along the way.
Start a workspace