Towards a Social Statistical Database and Unified Estimates at Statistics Netherlands
Statistics Netherlands aims at improving the accuracy and reliability of estimates by using data from registers and surveys in an optimal way. To this end, Statistics Netherlands is constructing a Social Statistical Database, in which several registers are via a unique key linked to each other, as well as to data from sample surveys. All estimates related to social statistics will be obtained from this database. Many estimates can simply be counted from the (combined) registers. Moreover, the presence of ample register data offers far better opportunities for nonresponse correction of estimators from the surveys. Furthermore, by combining data from surveys having variables in common, the accuracy of estimators involving these variables can be improved. In addition, Statistics Netherlands prefers to publish a single figure for each statistical concept. Numerical consistency between estimates may be achieved by using the calibration properties of the regression estimator. In this article, we explain how the social statistical database is constructed, and how reliable, accurate, and numerically consistent tables can be estimated from it. We also mention some theoretical and practical problems, and discuss possible solutions.
Combining registers and surveys, consistent estimates, record linkage, regression estimator, repeated weighting