THE ESTIMATION OF POPULATION MICRODATA BY USING DATA FROM SMALL-AREA STATISTICS AND SAMPLES OF ANONYMISED RECORDS

Citation
P. Williamson et al., THE ESTIMATION OF POPULATION MICRODATA BY USING DATA FROM SMALL-AREA STATISTICS AND SAMPLES OF ANONYMISED RECORDS, Environment & planning A, 30(5), 1998, pp. 785-816
Citations number
35
Language
INGLESE
art.tipo
Article
Categorie Soggetti
Environmental Studies",Geografhy
Journal title
ISSN journal
0308-518X
Volume
30
Issue
5
Year of publication
1998
Pages
785 - 816
Database
ISI
SICI code
0308-518X(1998)30:5<785:TEOPMB>2.0.ZU;2-U
Abstract
Census data can be represented both as lists and as tabulations of hou sehold/individual attributes. List representation of Census data offer s greater flexibility, as the exploration of interrelationships betwee n population characteristics is limited only by the quality and scope of the data collected. Unfortunately, the released lists of household/ individual attributes (Samples of Anonymised Records, SARs) are spatia lly referenced only to areas (single or merged districts) with populat ions of 120 000 or more, whereas released tabulations are available fo r units as small its single enumeration districts (Small Area Statisti cs, SAS). Intuitively, it should be possible to derive list-based esti mates of enumeration district populations by combining information con tained in the SAR and the SAS. In this paper we explore the range of s olutions that could be adapted to this problem which, ultimately, is p resented as a complex combinatorial optimisation problem. Various tech niques of combinatorial optimisation are tested, and preliminary resul ts from the best performing algorithm are evaluated. Through this proc ess, the lack of suitable test statistics for the comparison of observ ed and expected tabulations of population data is highlighted.