Inverse, Multiple and Sequential Sample Censuses

Abstract

Various approaches to the tag and sample method of enumerating populations are considered. In multiple censuses there are several stages of tagging and sampling. In inverse censuses one fixes the number of tagged (or untagged) individuals to be recovered rather than the sample size. Sequential censuses are like multiple censuses except that the number of stages is not fixed in advance. The author discusses all approaches singly and in combination with others. He concludes that sequential sample censuses are optimum in most cases. Both estimation and hypothesis testing are considered. Although tables are given which aid in designing such censuses, no illustrative examples are given which aid in designing such censuses, no illustrative examples are included.