Skip to content

Sample size calculator (survey)

Find out how many people to survey based on margin of error and confidence level.

%
%
Required sample
385
Without population adjustment385

Use 50% when you don't know the expected answer (worst case, needs the largest sample). Enter the population only if it is small (e.g., students in a school).

How the calculation works

The base sample is n₀ = Z²·p·(1−p) ÷ E², where Z depends on the confidence (1.96 for 95%), p is the expected proportion and E the margin of error. For small populations, apply the correction n = n₀ ÷ (1 + (n₀−1)/N).

Examples

  • 95% confidence, 5% margin, 50% proportion: ~384 responses (large population).
  • For a 3% margin instead of 5%, the sample nearly triples (~1,068).

Frequently asked questions

What is the margin of error?

It is how much the result may vary up or down. E.g., 60% ± 5% means between 55% and 65%.

What proportion should I assume if unknown?

Use 50% — the most conservative case, requiring the largest sample.

Is a bigger sample always better?

It reduces error, but with diminishing returns and higher cost. Doubling precision needs ~4× the sample.

How many people do I need to survey?

For a large population, 385 responses give ±5% at 95% confidence; 1,068 give ±3% and 2,401 give ±2%.

Does population size matter?

Very little. A city of 100,000 and a country of 200 million need nearly the same sample size.

Why is a smaller margin of error so expensive?

Because sample size grows with the square of precision: going from ±2% to ±1% quadruples the interviews.

Does a large sample guarantee a good survey?

No. The formula assumes random sampling and cannot fix selection or non-response bias. A biased sample of 10,000 is worse than a random one of 400.

Related calculators

More Statistics calculators

See all →

Updated on June 30, 2026 · by Rafael Rossi · Methodology & sources