Estimating the Completeness of Death Registration in a Closed Population

Preston and his colleagues have proposed a method to estimate the completeness of adult death registration in populations that are approximately stable. In this paper we have developed a logical extension of that method which by eliminating the restrictive assumption of stability allows 1 to determine the extent of underregistration relative to the completeness of enumeration in populations which may be far from stable. Another application of this technique is in its use with populations that are known to have rigorous registration systems but relatively poor enumeration. The impracticability of the method of extinct generations for use with data from current populations can now be circumvented. Thus we must no longer wait for cohorts to become extinct. Rather we can now simply adjust cross-sectional death distribution data and subsequently obtain estimates of underenumeration by age. (Authors modified)