An examination of aspects of probability sampling and application of the sampling method to cost analysis. The relative merits of the sample as a means of data collection, sampling theory, and the more basic techniques of sample design and estimation are discussed. Sampling offers an approach to the data quality problem that is usually cheaper, faster, and more flexible than attempts to modify existing massive data collection systems. Basic tools that the analyst has at his disposal are stratification, clustering, subsampling, systematic sampling, ratio and regression estimators, and sampling with unequal probabilities. The study was prepared to provide an introduction to sampling methods. There is sufficient coverage to guide simple survey investigations, and attention is given throughout to the use of cost-effectiveness criteria in choosing among alternative sampling plans. 92 pp. Bibliog.