Sampling

All households that received information on the HFCS personally addressed to them are part of the survey sample. They were chosen by random selection from among the frame population (= all households resident in Austria) subject to complex selection rules (“the sample design”) that ensured that the selected households were representative of the target population.

Technically speaking, the survey was based on stratified multi-stage cluster random sampling. We used a probabilistic sampling procedure: all households resident in Austria had a strictly positive probability of being drawn from the frame population. At the same time, all units of the sample (= all households invited for an interview) were selected by chance (“random sampling”).

“Stratified sampling” ensured that the sample contained households from all over Austria and adequately reflected the size of the individual regions of Austria. In the HFCS survey, the sample was stratified first on a regional scale (based on the “NUTS-3” regions into which Austria is divided statistically; NUTS = Nomenclature des unités territoriales statistiques) and second by municipality size (based on eight different categories of size, ranging from municipalities with up to 2,000 inhabitants to municipalities with over 1 million inhabitants). In Austria, each household selected for the survey sample ultimately represented about 900 households.

The number of addresses to be contacted were determined in line with the shares that the individual provinces (i.e. NUTS-3 regions and municipality size categories) had in the entire population based on Statistics Austria’s Micro census (labor force survey) of 2009. The gross sample eventually consisted of 4,436 households.

For detailed information on the sampling design used in the first wave of the HFCS in Austria, see the methodological notes.