Estimates of Mortality and Population from Survey-Removal Records

Abstract

If a population consists of 2 subclasses and if these are subject to differential removal, then knowledge of the relative population composition before and after this removal can be used to estimate population parameters. The statistical theory of this, the ''Kelker index'' has been studied primarily for the case where natural mortality is assumed to be zero. In this paper this restrictive assumption is removed. Two situations are considered: where the removal is ''instantaneous'' and where there is a continuous uniform removal process. A table is provided to facilitate estimation of the parameters in this 2nd case. Variances of the estimates are given under assumptions which are believed to be more realistic than those of previous studies.