Find the sample size you need for surveys and A/B tests.
Computed with standard statistical methods in your browser.
Sample size n₀ = z²·p(1−p) ÷ e², where z comes from your confidence level, p is the expected proportion and e is the margin of error. For a known population we apply the finite population correction.
Run the rest of your analysis with our other free stats tools.
Using n = z²·p(1−p)/e², adjusted for population size when known.
95% is standard; 99% needs a larger sample, 90% a smaller one.
The ± range around your result; 5% is common for surveys.
It maximizes variance, giving the most conservative (largest) sample size.
Only for small populations; for large ones the correction barely matters.