Sample Size Calculator
Calculate the minimum number of survey respondents needed for statistically valid results. Includes finite population correction for small communities.
Typical surveys: 3–5%
Use 50% for maximum sample size
| Confidence / Margin of Error | Required Sample Size | Typical Use | Who Uses This |
|---|---|---|---|
| 95% / ±5% | 385 | Standard survey | Academic research, company surveys |
| 95% / ±3% | 1,068 | National poll | YouGov UK, Gallup US, Morgan Stanley AU |
| 95% / ±2% | 2,401 | High-accuracy poll | Major election polls |
| 99% / ±5% | 664 | High-confidence survey | Medical surveys, FDA studies |
| 99% / ±1% | 16,588 | Very precise estimate | Government census supplements |
| 90% / ±5% | 271 | Exploratory research | Business pilots, quick polls |
Frequently Asked Questions
Why does population size barely affect the required sample size?
For large populations (N > 10,000), the sample size formula n = (z²pq)/e² is independent of population size. This is counterintuitive but correct: accuracy depends on the absolute sample size, not the proportion of the population sampled. A sample of 400 gives ±5% accuracy whether the population is 10,000 or 1 billion. The finite population correction only makes a meaningful difference when you're sampling more than ~10% of the total population.
How do UK and US polling companies determine sample size?
YouGov UK typically uses weighted samples of 1,000–2,000 for national surveys (targeting ±3% MoE at 95% CI). Ipsos UK and MORI use similar sizes. US Gallup polls use 1,000–1,500 adults. Australian Essential Media uses 1,000 respondents. All target 95% confidence. The difference from a strict statistical sample is that all these organisations use quota sampling and weighting — not purely random sampling — which affects the effective sample size.