Small Area Estimation Using Estimated Population Level Auxiliary Data

Abstract

Unit level linear mixed models are often used in small area estimation (SAE), and the empirical best linear unbiased prediction (EBLUP) is widely used for the estimation of small area means under such models. However, EBLUP requires population level auxiliary data, atleast area specific aggregated values. Sometimes population level auxiliary data is either not available or not consistent with the survey data. We describe a SAE method that uses estimated population auxiliary information. Empirical results show that proposed method for SAE produces an efficient set of small area estimates.

Keywords:

• Auxiliary data
• Linear mixed model
• Small area
• Survey weights

Mathematics Subject Classification:

• 62D05

Acknowledgment

The authors would like to acknowledge the valuable comments and suggestions of the Editor and the referee. These led to a considerable improvement in the article.