The attention to
small area estimation has increase along with the increased of government and private sector demand to provide accurate information quickly, not only for national level but also for small
domain such as district or village. However, problem which often emerges is that data in small areas are sparse which result in a difficulty to produce
reliable estimates because the
sample size from these areas is not large enough to support the specified accuracy. For these reason, to provide more reliable
estimate in a small area, it is very convenient to use information (borrowing strength) from other related areas. The procedure to borrow information from other small domain (areas) depends on the estimator and it usually involves using a class of regression methods. This paper will discuss and review several small area estimation methods and comparisons will be made with respect to a case-study in Indonesia. We will introduce some methods like synthetic estimator, empirical Bayes, and (empirical) best linear unbiased prediction.
More abstracts about the Small Area Estimation: A Review and Comparison on Various Methods