In this paper, we discuss generalized regression (GREG) estimation and model-calibration estimation for population subgroups or domains under unequal probability sampling. The classical GREG estimator of Särndal, Swensson and Wretman (1992) uses a fixed-effects linear assisting model. A multinomial logistic model is introduced as an assisting model for GREG in Lehtonen and Veijanen (1998). Logistic GREG has been examined further for domain estimation in Lehtonen, Särndal and Veijanen (2003, 2005), Lehtonen and Veijanen (2008) and Myrskylä (2007). Model calibration (MC), which also provides a design-based method, was introduced by Wu and Sitter (2001) and was further discussed in Wu (2003), Lehtonen, Myrskylä, Särndal and Veijanen (2007), Särndal (2007) and Lehtonen, Särndal and Veijanen (2008). A key property of MC is that the weights are calibrated to the population total of the predictions derived via an assumed model. For comparability with the GREG approach, we use a logistic model. Under this model, GREG and MC require an access to unit-level auxiliary information. Both GREG and MC provide nearly design unbiased methods. We extend in this paper the model-calibration method to domain estimation. We present results on the accuracy of logistic GREG and MC estimators of domain totals of a binary response variable. The results are based on Monte Carlo experiments where repeated probability proportional-to-size samples were drawn from an artificially generated finite population.
|Title of host publication||Proceedings of the Baltic-Nordic-Ukrainian Workshop on Survey Sampling Theory and Methodology 2008|
|Number of pages||5|
|Place of Publication||Tallinn|
|Publication status||Published - 2008|
|MoE publication type||A4 Article in conference proceedings|
Fields of Science
- 112 Statistics and probability
Lehtonen, R., Särndal, C-E., & Veijanen, A. (2008). Generalized regression and model-calibration estimation for domains: Accuracy comparison. In Proceedings of the Baltic-Nordic-Ukrainian Workshop on Survey Sampling Theory and Methodology 2008 Tallinn: Statistics Estonia.