Frame Correction Modelling with Applications to the German Register-Assisted Census 2011

摘要

Frame corrections have been studied in census applications for a long time. One very promising method is dual system estimation, which is based on capture-recapture models. These methods have been applied recently in the USA, England, Israel and Switzerland. In order to gain information on subgroups of the population, structure preserving estimators can be applied [i.e. structure preserving estimation (SPREE) and generalized SPREE]. The present paper extends the SPREE approach with an alternative distance function, the chi-square. The new method has shown improved estimates in our application with very small domains. A comparative study based on a large-scale Monte Carlo simulation elaborates on advantages and disadvantages of the estimators in the context of the German register-assisted Census 2011.