Work out audit sample size instantly with clear inputs, formula shown and shareable results.
Attribute sampling for control testing sizes the sample from the confidence level and the precision, which is the gap between the tolerable and expected error rates. A narrow gap demands a large sample, which is why an auditor expecting 4 percent errors against a 5 percent tolerance needs far more items than one expecting zero. With no expected errors, zero-acceptance sampling gives a much smaller sample.
Attribute sampling
precision = tolerable - expected; n = z^2 p(1-p) / precision^2; with zero expected errors n = ln(1 - confidence) / ln(1 - tolerable); apply the finite-population correction
Sample size scales with the inverse square of precision. Halving the tolerable rate roughly quadruples the items to test.
The control is assessed as ineffective for the period. Expanding the sample after the fact is generally not acceptable, since it biases the conclusion.