Open Conference Systems, ITACOSM 2019 - Survey and Data Science

Font Size: 
Spatial sampling for the French Master Sample
Thomas Merly Alpa

Building: Learning Center Morgagni
Room: Aula 210
Date: 2019-06-06 05:00 PM – 06:30 PM
Last modified: 2019-05-06

Abstract


The French National Institute of Statistics (INSEE) conducts face-to-face surveys on a variety of subjects: wealth, well-being, housing... These surveys are based on a two-phase sampling design, starting with the drawing of the same Master Sample for around ten years. This Master Sample is a collection of geographical areas which are meant to represent the national population in many ways.

In this presentation, we'll discuss the construction of the primary units and then the methods considered to draw the Master Sample : Doubly Balanced Spatial Sampling (Grafström, Tillé), pivotal sampling... These methods are compared using sociological and economical data on territories and in regard to the coverage of France by Insee's surveyer network. The method we chose to use is doubly balanced spatial sampling, because on one hand it is based on a well-known method at the Insee, and on the other hand because it allowed us to coordinate our Master Sample with the sample of the LFS (Labor Force Survey), which is drawn in a different way. This coordination ease the data collection work.


Full Text: SLIDES