SMALL AREA ESTIMATION PADA TINGKAT KEMISKINAN DI PROVINSI JAWA TENGAH DENGAN PENDEKATAN EMPIRICAL BEST LINIER UNBIASED PREDICTION

Abstract

Poverty is a complex problem for every country, similar to Indonesia. Poverty is
one of the important measures to determine the level of welfare of a household.
Factors that cause poverty include low income, the number of family
dependents, health, and education levels that characterize poor families in
Indonesia. The purpose of this research is to know the level of impact at
districts level in Central Java Province by using Small Area Estimation (SAE)
method with Empirical Best Linier Prediction (EBLUP) approach. The data
used in this research are poverty data obtained from SUSENAS of Central Java
Province with the response variable that is the number of poor population,
while as the participant variable is selected gross enrollment rate (X1), school
participation rate (X2), health insurance (X3), goods per capita (X4) and life
expectancy (X5). The results of the MSE study of the SAE model were smaller
than the direct predicted MSE, indicating the SAE model was better than the
direct estimates in the estimated number of poor people in each district in
Central Java Province.

Keywords : Poverty Rate, SAE and EBLUP.