Margin of error by sample size (95% confidence)
| Responses | Margin of error |
|---|---|
| 100 | ±9.8% |
| 250 | ±6.2% |
| 400 | ±4.9% |
| 600 | ±4.0% |
| 1,000 | ±3.1% |
| 2,500 | ±2.0% |
z-scores for common confidence levels
| 80% | 1.282 |
| 90% | 1.645 |
| 95% | 1.960 |
| 98% | 2.326 |
| 99% | 2.576 |
Small populations
When you survey a large share of a small group, say 400 of 1,000 employees, the finite population correction shrinks the margin: multiply by √((N − n) ÷ (N − 1)). For that example ±4.9% becomes about ±3.8%. Enter the population size above to apply it.
What it doesn't cover
Margin of error only measures random sampling error. It says nothing about a biased sample, leading questions or people who chose not to answer, which often matter more.
How to calculate margin of error by hand
- Pick a confidence level and find its z-score (1.96 for 95%).
- Use the proportion from your results, or 0.5 if you don't have one yet.
- Compute √(p × (1 − p) ÷ n). For 400 responses at p = 0.5: √(0.25 ÷ 400) = 0.025.
- Multiply by z: 1.96 × 0.025 = 0.049, or ±4.9%.
Looking for profit margins instead? That's a different kind of margin: try the margin calculator or the profit margin calculator.