Abstract:
|
The properties of dual system estimation when applied to census underenumeration are well-developed in the U.S., but there is no such history in the UK. As part of the 2001 Census of the UK, a fresh approach has been taken, based on an estimation strategy, that differs from the U.S. model. This employs unweighted dual system estimation at a low level of aggregation and combines it with a robust approach to ratio estimation to produce population estimates by age and sex for subnational areas of the UK. These estimates are quality assured by comparing them with demographic estimates and other administrative sources. Historically, census overenumeration has not been a significant issue in the UK. However, the estimation approach is easily adapted to cope with overenumeration, and, as part of the quality control in the 2001 Census, data have been collected to allow such an adjustment. This talk will provide insight into the UK Census estimation methodology and our experience with its implementation.
|