Risto Lehtonen, Graham Kalton
2016 Statistics in Transition New Series  
The second part of this Joint Issue of Statistics in Transition and Survey Methodology includes seven articles. These two issues have been split according to which guest editors have been looking after the articles. They are not necessarily sequenced according to the themes that appeared in the original Conference programme. The first paper, by Erciulescu and Fuller, presents a small area procedure where the mean and variance of an auxiliary variable are subject to estimation error. They
more » ... error. They consider fixed and random specifications for these auxiliary variables. Their study was motivated by a situation where the sample used for small area estimation was a subsample of a larger survey. The larger survey furnished estimates of the distribution of the auxiliary variables. They demonstrate that efficiency gains associated with the random specification for the auxiliary variable measured with an error can be obtained. They propose a parametric bootstrap procedure for the mean squared error of the predictor based on a logit model. The resulting bootstrap procedure has a smaller bootstrap error than a classical double bootstrap procedure with the same number of samples. The second paper, by Münnich, Burgard, Gabler, Ganninger and Kolb, develops a sampling design that can support accurate estimation for the 2011 German Census. In contrast to carrying out a classical census, a register-assisted census, using population register data and an additional sample, was implemented. The main objective of the census was to produce the total population counts at fairly low levels of geography. Ralf Münnich et al. provide an overview of how the sampling design recommendations were set up to fulfill legal requirements and to guarantee an optimal, yet flexible, source of information. Small area methods, as well as traditional methods, were used to produce these counts. Empirical results of the small area estimation are presented. The next three papers present developments in small area estimation methodology and practical application in various fields of empirical research and statistics production, including poverty research and fisheries statistics. The first paper, by Guadarrama, Molina and J. N. K. Rao, provides a review on methods for the estimation of poverty indicators for small areas, including design-based direct estimation and a number of model-based small area estimation methods: the Fay-Herriot area level model, the World Bank poverty mapping method (the ELL method) and three Bayesian variants previously published by the authors. These are the empirical best/Bayes (EB) and hierarchical Bayes (HB) methods and a Census EB method providing an extension of the EB method. While the
doi:10.21307/stattrans-2016-002 fatcat:6c2akmngfrhqffbf3eua4v37m4