Comparison of small area estimation methods: simulation study in EURAREA project

  • Kaja Sõstra Statistical Office of Estonia
Keywords: small area estimation, linear mixed model, spatial correlated area effects, correlated time effects


In literature several techniques have been introduced for small area estimation. It has not been proved theoretically that some estimator is superior to others. The focus of this paper is on comparison of some estimators (direct, GREG and EBLUP estimators with correlated area and time effects) in a simulation study. The individual-level models are used. The target of estimation is the average disposable income in small geographical regions of Finland. The results given here are developed in the framework of the EURAREA project, where the author of this paper was responsible for the simulation study.


Download data is not yet available.