Abstract:
|
Model-based small area estimation methods can lead to reliable estimators. provided the underlying model assumptions hold at least approximately. In this talk I will address practical issues related to assumed models, focusing on basic area level and unit level models. In particular, I will address measurement errors in covariates, misspecification of linking models, benchmarking to ensure agreement with a reliable direct estimator at an aggregate level, informative sampling and robust estimation. Methods for measuring variability of area estimators will also be examined.
|