Journal of Official Statistics, Vol.19, No.4, 2003. pp. 421–447
Estimating Census Undercount by Demographic Analysis: New Approaches to the Emigrant Component
Abstract:This is a method of jointly measuring net undercount in a census of population and the stock of emigrants as at census date (that is, the stock of persons born here and resident abroad). It is a variant of demographic analysis but does not require the data on migratory flows that are an essential component of the conventional analysis. The method is applied here to the 1991 Census of England and Wales (E&W).The calculations start from statistics of births and deaths. Part of the stock of emigrants is measured by data from contemporaneous censuses in a few foreign countries - perhaps 3 or 4 foreign countries or, to avoid the delay in assembling foreign data, even 0 foreign countries. The rest of the emigrant stock is estimated from data extracted from the E&W Census on persons who have returned from abroad, supplemented (in some versions of the method) by historical data from foreign censuses. Each of the uncertain elements in the calculations is given an a priori error distribution. Empirical constraints, based on evidence from comparable countries, are imposed on the sex-age profile of percentage net undercount.The method is inexpensive. A Monte Carlo process generates results in the form of frequency distributions. Most importantly, the near congruence of the results from different versions of the model demonstrates the robustness of the methodology.
Keywords:Census of population; demographic analysis; emigration; international migration; overcount; population statistics; post-enumeration survey; undercount; under-enumeration.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group