A two-stage sampling method with statified secondary units

Abstract

Two-stage sampling is frequently used to estimate the density of insects living in a hierarchical habitat. In an ordinary two-stage sampling, secondary sampling units (SSUs) are selected at random from each primary sampling unit (PSU) on the assumption that the SSUs are homogeneous. In some situations, however, SSUs consist of different components which have different levels of insect densities. To obtain an efficient estimate of density in such a heterogeneous environment, we proposed a two-stage sampling method with stratified SSUs. As a special case, we discuss the incorporation of a binomial sampling to the two-stage sampling method. The proportion of SSUs infested by one or more insects (i.e. the proportion of occurrence) is first estimated through a binomial sampling for each PSU. If this can be estimated with a high precision, we can approximately apply the two-stage sampling method to the PSUs with two strata: infested SSUs and uninfested ones.