Multivariate processing of subpopulations by social survey: the construction of social zones

Abstract

When one carries out a Social Survey in a wide population, one would like to carry out the same analysis with a small population, using, of course, the same data of the Survey. The aim of this paper is to show each stratum of a stratified sample can be a sample population which allow an analysis with the reliable, or at least, calculable and acceptable guarantees without increasing the sample number of the general Survey.