Spatial Analysis of Gross Enrollment Ratio in Higher Education Using the Geographically Weighted Regression Approach

Abstract

This study analyzes the factors influencing the Gross Enrollment Rate of Higher Education (GER-HE) in Indonesia by emphasizing regional differences that cannot be adequately captured by global Ordinary Least Squares (OLS) or panel data models. The analysis is based on secondary data for 2023 obtained from Badan Pusat Statistik (BPS) and the Ministry of Education, including Gross Regional Domestic Product (GRDP), education funding, lecturer–student ratio, number of students, per capita expenditure, total population, and the poverty depth index. To capture spatial heterogeneity in these relationships, the study applies Geographically Weighted Regression (GWR) with a Gaussian kernel weighting function. The results indicate that the GWR model explains 50.57% of the variation in GER-HE, reflecting improved model performance after accounting for spatial variation across regions and providing a better fit than the OLS regression. The effects of explanatory variables on GER-HE vary across provinces, allowing regions to be classified into five groups based on combinations of statistically significant factors, particularly the number of students, per capita expenditure, and the poverty depth index. These findings suggest that higher education policies should be tailored to the specific characteristics of each regional group to enhance GER-HE and reduce interprovincial disparities.