Modeling and estimation methods for household size in the presence of nonresponse

This paper considers the problem of estimating the number of private households of various sizes and the total number of households in Norway. The approach is model-based with a population model for household size given registered family size while taking into account possible nonresponse biases by modeling the response mechanism conditional on household size. Various models are considered together with regression estimation and imputation-based poststratification. Comparisons are made with the estimation methods used in official statistics for The Norwegian Consumer Expenditure Survey. We conclude that poststratification, response modeling and imputation should be used instead of the current method in official statistics, a modified Horvitz-Thompson method.