Smallest number of entities or observations required in each test sample to ensure that any variation between the characteristics of the test sample and those of the entire associated group (batch, lot, population, or universe) are due only to chance.