Key points are not available for this paper at this time.
Abstract Specialized samples of very rare populations require expensive cluster search procedures. Most of the effort is wasted since no or very few members of the rare population are located. Bayesian optimum sampling procedures exclude such clusters, but the critical problem is identifying them. The ideas of Wald on sequential sampling are applied and prove to be highly efficient.
Seymour Sudman (Thu,) studied this question.