An application of sequential sampling to analytical surveys.
J. Sedransk · PubMed · 1966
SUMMARY The nature of an analytical sample surveyis discussed in general, and a sequential sampling scheme is proposed for certain situations. It is assumed that the subgroups of the population which are to be compared form a one-way classification, and are not identifiable in advance of drawing a simple random sample from the population. With sequential sampling, one continues to sample from the general population until the variance constraints are satisfied. It is required that, for each pair of group means, the variance of the difference between them be no larger than a specified constant. Knowledge of the probability that the constraints are met by the time n observations have been taken can be used to answer questions about the possible cost of the contemplated sampling scheme. Therefore, emphasis is placed on obtaining exact and approximate expressions for this probability and relating these probability statements to the cost question. The approximate procedures are checked (for the three group case) using a series of numerical examples.