Description of node <>

Subject Predicate Object <p>From 2011, in metropolitan France, the sample consists of dwellings from the housing tax register.<br /> The sample is &quot;areal&quot;: it is not derived directly from a drawing of dwellings, but from a drawing of groups of twenty contiguous dwellings.<br /> These groups or &quot;clusters&quot; have been automatically constructed using a bottom-up logic : clusters are defined among all dwellings in France, then sets of 6 clusters are put together to form sectors. Sampled sectors are then drawn in each region from this database in a balanced way  and a serial number is randomly assigned for the clusters from each sector. A cluster is surveyed for six consecutive quarters and then is replaced in the sample by the next cluster of the same sector (the one with immediately-higher order number).</p> <p>The sampling was done using a stratification by region and degree of urbanization. Henceforth, each quarter, approximately 73 000 dwellings are identified as main homes and interviewed. Also, each quarter, a sixth of them are replaced.</p> <p>Ultimately, there are survey files for about 108,000 people aged 15 or over who are respondents every quarter, in 58,000 households.</p> <p>The overseas sample is made up of dwellings drawn in census annual survey files; as in metropolitan France, these units are grouped in clusters, a cluster being surveyed for six consecutive quarters and then replaced by a new cluster located in the same sector.</p> <p>The complementary module is asked to the outgoing clusters.</p>