Cochran’s formula gives the sample size needed to estimate a proportion with a chosen margin of error and confidence level. If the population is small, the finite population correction reduces it.
n₀ = z² p(1 − p) ÷ e² · n = n₀ ÷ (1 + (n₀ − 1) ÷ N)
With 95% confidence, p = 0.5 and e = 5%, n₀ = 384.16, rounded up to 385.
Effect of the inputs
- A smaller margin of error needs a much larger sample: halving e quadruples n₀.
- Higher confidence needs more: 99% instead of 95% multiplies n₀ by about 1.73.
- p = 0.5 gives the largest sample. If you expect around 20%, n₀ falls by 36%.