Given a confidence level, margin of error and population standard deviation, estimate the minimum sample size needed to estimate a population mean, for experiment design.
| Item | Formula / Note | Source / Standard |
|---|---|---|
| Mean sample-size formula | n = (z × σ ÷ E)² with E the margin of error; z=1.96 for 95% | Cochran 1977, Sampling Techniques |
| Finite population correction | n' = n ÷ (1 + (n−1) ÷ N) for finite population N | Cochran 1977 FPC section |
| Design effect | Multiply by deff (1.5–3 for clusters) for complex designs | Kish 1965, Survey Sampling |
| Nonresponse allowance | n_final = n ÷ expected response rate (e.g. 80%) to secure valid n | Survey-methods practice |