SOME STATISTICAL ASPECTS OF THE CONTINUOUS MORTALITY INVESTIGATION BUREAU'S MORTALITY INVESTIGATIONS

Abstract

This article discusses the formulae for the select and ultimate exposed-to-risk in the CMIB's mortality studies, and certain statistical aspects of these investigations. It is shown that there are difficulties in the traditional binomial approach to the distribution of deaths, particularly for select rates, and the use of the Poisson distribution, rather than the binomial, is advocated.