Updating Small Area Population Estimates in England and Wales

Abstract

"Population estimates have important implications for resource allocation within government and commerce, and are often assumed to be without error. Currently, central government provides annual population estimates for all the local and health authority districts in Britain, but estimates are needed for smaller areas, typically for electoral wards and postal sectors. Small area estimates are provided by some local authorities and commercial organizations, using different methods; the accuracy of these estimates is modelled here within a multilevel framework. Certain characteristics of the small area and of the method of estimation are included as explanatory variables. Results show that the method of estimation used is of great importance."

Differences in Population Estimates Between an Administrative System and Census: The Case of Israel

“…The next step is to estimate population by both methods until the next population census in 2008. After that census, the results of the census will be comparable with the results obtained under each of the two systems, using a method similar to that of Simpson, Diamond, Tonkin and Tye (1996) for England and Wales. This comparison would enable us to compare the quality of the two methods with the actual estimates.…”

Childhood leukaemia and non-Hodgkin's lymphoma in relation to proximity to railways

“…Although the population estimates in the two studies were derived from different censuses, it is possible that both results may have been influenced by differential underestimation of the population in Inner London (Simpson et al, 1991;Champion, 1995), an area that also had high exposure to pollution from road and rail traffic. Census underenumeration is known to be higher for urban, deprived wards (Simpson et al, 1996); the influential wards in our study showed these characteristics. Hence, the apparent association in Inner London between childhood leukaemia and exposure to both benzene and railways may be because of underestimation of both the underlying population and the effect of population mixing.…”

Small-area population estimates: a review of methods used in Britain in the 1990s

“…All population changes Areas with less than 5% pop'n change, Areas with 5% or greater 1981-1991 pop'n change, 1981-1991 All sizes < 2000 2500 + All sizes < 2500 2500 + All sizes < 2500 2500 + Simpson et al (1996) found that about half the variation in the accuracy of estimates of smallarea total population is due to the type of area being estimated. Apart from small and rapidly changing areas, those with many students or armed forces, and urban areas, are particularly difficult to estimate.…”

“…Large inaccuracies are those of 10% or more, the figure that most users feel would be damaging to applications (Tye, 1994). The preparation of the data-set is described in Lunn et al (1996) and in Simpson et al (1996). Table 5 reflects the inaccuracy of population estimates made for 1991 before the results of the 1991 Census were known.…”

“…The project sought documentation of estimation methods as well as the population estimates themselves. Early reports on the accuracy of estimation methods have been published elsewhere (see Simpson, 1995 for an overview; Simpson et al, 1996 andLunn et al, 1998 for methodological issues and statistical modelling). The next section estimates the level of demographic estimation activity in Britain.…”