Skip to content

Sample Size Calculator for Health Research

Work out how many participants your study needs. The calculator below estimates a single proportion, such as the prevalence of a condition in a cross-sectional survey. Choose another design from the list if you are comparing groups.

Best guess from a pilot or earlier studies. Use 0.5 if unknown, because it gives the largest (safest) sample.

Half-width of the interval you want, as a proportion. 0.05 means plus or minus 5 percentage points.

0.95 is the usual choice. It sets the two-sided significance level (alpha = 1 - confidence).

1 for simple random sampling. Use 1 + (m - 1) x ICC for cluster sampling, where m is the average cluster size.

Total size of the source population. Leave 0 if it is large or unknown; the finite population correction then is not applied.

Proportion expected to be lost or refuse, e.g. 0.10 for 10%. The sample is inflated by 1 / (1 - rate).

When to use this calculator

  • You are planning a cross-sectional survey or prevalence study and want a sample that gives a chosen margin of error.
  • You are comparing two groups, two proportions or two means, and need enough participants for a stated power.
  • You are planning a case-control, cohort, survival, paired or repeated-measures study and want the calculation written up for a protocol or an ethics application.

How to use it

  1. Choose the design that matches your main research question. The calculators listed below cover the common designs.
  2. Enter the expected value from a pilot study or earlier published work. If you have no estimate for a proportion, 0.5 gives the largest and therefore safest sample.
  3. Set the confidence level and, for group comparisons, the power. 95% confidence and 80% power are common conventions, not rules.
  4. Add allowances for the design effect, a finite population and non-response if they apply.
  5. Copy the sentence and formula into your protocol and keep the references the calculator shows.

How to read the result

The result is the smallest number of participants who need to provide data. Increase it for expected dropout or non-response, which the single-proportion calculator can do for you.

A sample size is only as good as its inputs. Small changes to the expected effect or proportion can change the result a lot, so try a few plausible values and report the one you chose and why.

Ask a statistician to check the plan if your design is complex, for example clustered sampling or adjustment for several confounders.

Frequently asked questions

What do I need to calculate a sample size?
Usually the expected size of the effect or the proportion, the confidence level or significance level, the margin of error or the power you want, and the variability of the outcome. Add the dropout or non-response you expect.
What power should I use?
Many studies aim for 80% or 90% power. Higher power needs more participants. State your choice and the reason in the protocol.
How do I allow for dropout or non-response?
Divide the calculated sample size by one minus the expected loss. With 15% expected loss, divide by 0.85. The single-proportion calculator does this when you enter the non-response rate.
Why is the expected proportion set to 0.5?
The product p(1 - p) is largest at 0.5, so assuming 0.5 gives the biggest and safest sample when you have no estimate. Use a better estimate if you have one.
Can this replace a statistician?
No. It follows standard published formulas and shows its assumptions, but it is a planning aid. Complex designs should be reviewed by a statistician.

Guides

Related calculators

Use this in a full research protocol

A free ToProt account lets you save calculations to a study, build your protocol and objectives, manage references and generate your data dictionary in one place.