Work out community survey cluster size instantly with clear inputs, formula shown and shareable results.
Calculates the simple random sample size for a proportion, inflates it by the design effect and divides by the cluster size to give the number of clusters.
Cluster sample size
n = 1.96^2 x p(1-p) / d^2, multiplied by the design effect; clusters = n / cluster size
The variance inflation caused by clustering; 1.5 to 2 is typical for household surveys of health indicators.
The classic EPI cluster survey used 30 clusters of 7, which balances precision against field logistics.