Sample Size Calculator
Estimate the completed responses needed for a survey proportion, with a finite-population option and an invitation estimate. Switch modes to calculate the approximate margin of error from an existing sample size.
Enter your values or load an example. Calculations stay in this browser.
Plan for a proportion, not every type of study
This calculator concerns a yes/no or categorical proportion under simple random sampling. It does not estimate statistical power, a mean, a regression effect or the number of participants for a clinical study. Those questions need a different design.
Keep the planning assumptions visible
Choose a two-sided confidence level, a margin in percentage points and an expected proportion. An unknown proportion can be planned at 50%. Enter a finite population only when it represents the population you are sampling without replacement.
Read the rounding and response-rate steps
The calculator rounds the required response count upward. It separately divides that count by your anticipated response rate to estimate invitations. Sending more invitations does not remove nonresponse bias or ensure that enough people answer.
Check margin of error in the same page
The second mode starts with completed responses and returns the approximate sampling margin. A census has zero sampling margin under the finite-population formula, but measurement problems and missing or incorrect responses still matter.
Related free tools
Sources and methods
Frequently Asked Questions
What does a margin of 5 mean?
It means approximately ±5 percentage points, not five percent of your reported percentage.
What if I do not know the expected proportion?
Use 50% as a conservative planning input because it maximizes p(1−p). The result still depends on the sampling assumptions.
Does a large sample make an online poll representative?
No. Sampling bias and nonresponse bias are not fixed by the number of responses. This model assumes simple random sampling.
Can I calculate margin of error here?
Yes. Switch to the margin-of-error mode and supply the response count, confidence, expected proportion and optional population.
Continue in the app
Use Statistics AI: Statikia for the full guided result after this quick check.
