A method of statistical sampling in which the population is divided into intervals, and a single selection is sampled from each interval. For example, suppose an auditor wishes to sample 50 invoices from a population of 1,000 invoices arranged in chronological order. Dividing the population of 1,000 by 50 yields 20 intervals. Thus, each 20th invoice is chosen, beginning with a random starting point between 1 and 20. If 3 is chosen as the random starting point, invoices selected will be 3, 23, 43, 63, and so forth.