A Study of Disability Across Districts of the Province of Wielkopolska Using Small Area Estimation Methods

Abstract

Surveys and censuses conducted by the Central Statistical Office in Poland are the main sources of information about disability for official statistics. Because sample sizes for relevant cross-classification domains are too small to employ classical estimation methods, results are usually published at a relatively high level of aggregation (at country or province level) or for very broadly defined domains. To meet the growing demand for detailed information about disability, the authors present an attempt of applying the methodology of small area estimation to estimate the percentage of disabled people, in the legal and biological sense, across districts (NUTS 4/LAU 1 units) of the province of Wielkopolska crossclassified by the level of education. This methodological exercise is based on data from the 2011 census and employs selected techniques of indirect estimation. Estimates obtained in the study provide an indication of the spatial variation of disability in the target domains with greater precision. It is worth noting that this level of aggregation has not been considered for purposes of official statistical outputs because of unacceptably high estimation errors of the direct estimator.