Many different Small Area Estimation (SAE) methods have been proposed to overcome the challenge of findingreliable estimates for small domains. Often, the required data for various research purposes are available at differentlevels of aggregation. Based on the available data, individual-level or aggregated-level models are used in SAE.However, parameter estimates obtained from individual and aggregated level analysis may be different, in practice.This may happen due to some substantial contextual or area-level effects in the covariates which may be misspecifiedin individual-level analysis. If small area models are going to be interpretable in practice, possible contextualeffects should be included. Ignoring these effects leads to misleading results. In this paper, synthetic estimators andEmpirical Best Linear Unbiased Predictors (EBLUPs) are evaluated in SAE based on different levels of linear mixedmodels. Using a numerical simulation study, the key role of contextual effects is examined for model selection inSAE.Key words: Contextual Effect; EBLUP; Small Area Estimation; Synthetic Estimator.
Namazi-Rad, M. & Steel, D. (2011). Contextual effects in modeling for small domain estimation. In E. Beh, L. Park & K. Russell (Eds.), Proceedings of the 4th Applied Statistics Education and Research Collaboration (ASEARC) Conference (pp. 12-14). Wollongong: University of Wollongong.