ANALYSIS OF FACTORS EXPLAINING SENIOR HIGH SCHOOL DROPOUT RATE USING GEOGRAPHICALLY WEIGHTED REGRESSION

Abstract

East Nusa Tenggara (NTT) Province is one of the areas experiencing the problem of dropping out of high school education. Even though NTT Province has adequate educational facilities and teaching staff, the high school dropout rate in NTT Province always ranks top 9 in Indonesia for the 2019/2020 academic year to 2021/2022. Dropping out of school can be influenced by region (spatial) and does not occur at one time, so research is needed using panel-structured spatial data that accommodates spatial effects over time. Geographically Weighted Panel Regression (GWPR) is a local regression analysis method that considers the effect of spatial heterogeneity on panel-structured spatial data. This study aims to analyse the factors that explain the high school dropout rate in NTT Province in 2019-2021 using GWPR. The results showed that the GWPR model with the Fixed Exponential weighting function was the best model compared to other weighting functions based on  and AIC. Population density, student-teacher ratio, regional minimum wage, open unemployment rate, student-to-school ratio, average length of schooling, and Smart Indonesia Program budget have a significant effect on explaining high school dropout rates in at least 21 regencies/cities in NTT Province. Grouping districts/cities based on variable significance using k-modes clustering produces 4 groups.