ABSTRACT
This chapter describes methods for determining how large a sample is needed to obtain precise estimates of process capability. The methods described are based on two approaches: limiting the size of the estimation error, and obtaining adequate power in a test of hypotheses. The capability measures considered include the proportion of nonconforming items, the rate of nonconformities, capability indices such as C p and C pk , and statistical tolerance limits. For nonnormal distributions, Monte Carlo methods are used to determine the required sample size.
