The EU-SILC sample design variables: critical review and recommendations

The EU Statistics on Income and Living Conditions (EU-SILC) are the principal data source for analysing the social situation in Europe. Given that EU-SILC is based on a representative sample in each participating country, estimates based on EU-SILC are subject to sampling variance. One of the principal determinants of the sampling variance is the sample design that has been used for drawing the sample. Therefore, standard errors, significance tests and confidence intervals should be computed taking the sample design as much as possible into account. For doing so, good sample design variables are an indispensable starting point. In this paper, I review the quality of sample design information in the EU-SILC dataset and formulate recommendations for data producers about how to improve the quality of sample design variables and for data users about how to make optimal use of the information that is already available in the EU-SILC UDB.