A note on sampling to locate rare defectives with strong prior evidence
Tommy Wright · Biometrika · 1992
Inference about the number of defectives in a finite population is considered. Using a Bayesian model for computing the probability of unobserved defectives given the results of the sample, a criterion for sample size determination is introduced for two cases: (i) when there is a uniform prior, and (ii) when there is strong prior information. Depending on the value of the population size N, the savings in terms of a sampling effort rather than a census can be significant. When there is strong prior information, an explicit decision rule is given for determining whether a sample is needed or if the prior information alone is sufficient.