Prediction theory for finite populations (Q1189566)

This book is a systematic treatment of prediction theory for finite populations based on superpopulation models, and brings together material scattered in papers during the last two decades. After a brief outline of ideas and principles, the following topics are dealt with in separate chapters: Optimal predictors of population quantities; Bayes and minimax predictors; maximum likelihood predictors; classical and Bayes prediction intervals; the effects of model misspecification; conditions for robustness and Bayesian modelling; models with measurement errors; asymptotic properties in finite populations, design characteristics of predictors. The emphasis is on the analysis of the data rather than on the design.