Is the Kish household sampling method better than the birthday method?

Abstract

Practically all studies on household survey sampling have given some space to the different dwelling selection processes. Since most surveys are administered to ONE person within each household, reselection is necessary within those households where there are two or more people. This study compares two within-household selection methods: the last-birthday method and the Kish method. The hypothesis is that the last-birthday method represents the population better than Kish method. It complements the ‘classic’ representation of sex and age distribution with the representation of educational attainment, labor force participation rates, employment and unemployment by sex. The data from the European Social Survey (8th wave) shown point toward accepting this hypothesis. In spite of the last-birthday method producing a greater selection of women, the differences in educational level and labor force participation are smaller than with the Kish method.