PENDUGAAN ANGKA PUTUS SEKOLAH DI KABUPATEN SEMARANG DENGAN METODE PREDIKSI TAK BIAS LINIER TERBAIK EMPIRIK PADA MODEL PENDUGAAN AREA KECIL

Abstract

Nowadays, small area information that has a small sample size is needed. A direct estimation in the small area will produce a large variance of values. In order of that, another alternative is needed that can be used is the indirect estimation. Small area estimation is an indirect estimation method that can be used to estimate parameters in a small area by utilizing information from outside the area, from the area itself, and from outside the survey. One of the methods that can be used is the empirical best linear unbiased prediction (EBLUP). EBLUP will be used to estimate the dropout rate for each village in the district of Semarang. Additional information used in this EBLUP method are the number of educational facilities, population, average expenditure per capita and distance from village to district. The results of EBLUP estimation showed that the lowest dropout rate village is Beji village and the highest is Pledokan village. Indirect estimation with EBLUP methods for the case of dropout rate in the district of Semarang has a coefficient variance 0,598% smaller than the coefficient variance that obtained from direct estimation