Sample size calculator (survey, margin of error, confidence level)
Find how many survey responses you need for a chosen margin of error and confidence level.
The sample size calculator works out how many respondents you need to estimate a proportion (such as the percentage who prefer an option) within a chosen margin of error and confidence level. Choose 80%, 90%, 95%, 98% or 99% confidence, set the margin of error and the expected proportion (50% is the safe default), and optionally enter a known population size to apply a finite-population correction.
How to use
- Enter your desired confidence level and margin of error, and an estimated proportion (use 50% if unsure — it gives the most conservative, largest sample size).
- Optionally enter a known population size to apply a finite-population correction for smaller populations.
Common use scenarios
- Planning how many survey responses are needed before fielding a study.
- Checking how much sample size requirements change with a tighter margin of error.
- Estimating required responses for a survey drawn from a small, known population.
- Comparing required sample sizes across different confidence levels.
✓Content & verification reviewed: 2026-10-03
How we verify Sample size calculator (survey, margin of error, confidence level)
How to confirm the required sample size matches what you'd expect.
A tighter margin of error requires a much larger sample
Because margin of error appears squared in the denominator of the formula, halving it roughly quadruples the required sample size — small improvements in precision get expensive fast.
50% estimated proportion gives the largest, most conservative sample size
The required sample size is largest when the estimated proportion is 50% and shrinks as the estimate moves toward 0% or 100% — using 50% when you're unsure is the safe, conservative default.
The finite-population correction only matters for a small, known population
Leaving population size blank assumes an effectively infinite population; entering a population size shrinks the required sample when that population is small relative to the uncorrected sample size, and has little effect when the population is large.
Before you rely on the result
- You've used 50% for the estimated proportion unless you have a genuine prior estimate.
- Your margin of error and confidence level match what your study actually needs to claim.
- You've only entered a population size if you have a real, known total population to correct for.
- You understand a tighter margin of error can dramatically increase the required sample size.
Who this is for
- Researchers and survey designers planning how many responses to collect.
- Analysts checking how sample size requirements scale with margin of error.
- Anyone surveying a small, known population who needs a finite-population correction.
- People comparing required sample sizes across different confidence levels before fielding a study.
Real-world examples
Planning a survey before fielding it
A researcher decides on a target margin of error and confidence level, then calculates how many respondents they need to recruit.
Budgeting around precision tradeoffs
An analyst compares the sample size needed at a 5% versus a 3% margin of error to weigh the cost of a more precise estimate.
Surveying a small known population
Someone surveying a small membership list enters the known population size to get a finite-population-corrected sample size instead of over-sampling.
Limitations
- This estimates sample size for a proportion (e.g. "what percent of people prefer X"), not for a mean or other kind of estimate.
Frequently asked questions
How do I calculate sample size for a survey?
Pick a confidence level (95% is typical), enter your margin of error (for example 5%) and an expected proportion (use 50% if unsure), and click Calculate. For a 95% confidence level, 5% margin of error and 50% proportion, the answer is 385 respondents.
What is a good margin of error and confidence level?
95% confidence with a 5% margin of error is the most common choice for surveys. A tighter margin or a higher confidence level needs a larger sample.
What does population size do in the sample size formula?
It applies a finite-population correction. For a large or unknown population, leave it blank. For a small, known population (for example 500 members), the required sample shrinks noticeably.
Can I use this to find a sample size for a mean?
No — this is the sample size for estimating a proportion. A mean needs a different formula that depends on the standard deviation of what you're measuring.
Why does the calculator default to a 50% estimated proportion?
50% produces the largest possible required sample size for a given margin of error and confidence level, so it's the safe default when you don't have a prior estimate of the true proportion.
What happens if I leave population size blank?
The calculation assumes an effectively infinite population — appropriate unless you're sampling from a small, fully known population.
Why did the required sample size jump so much when I tightened the margin of error?
Margin of error is squared in the formula's denominator, so even a small reduction in margin of error can substantially increase the required sample size.
Is this sample size for estimating a mean or a proportion?
This is specifically for estimating a proportion (like a percentage of respondents who answer a certain way). For a mean, a different formula based on the outcome's standard deviation would be needed.
🔒 Runs in your browser — nothing is uploaded. Results are for reference only; see the disclaimer. Last updated: 2026-10-03
Subscribe for updates
New tools and tips, straight to your inbox. No spam, unsubscribe anytime.