Predictors of educational attainment in Indonesia
Abstract
The current study applied quantile regression analysis to estimate the relationship between educational attainment and its predictors, and compared the results to parameter estimates using OLS regression. OLS regression is a regression technique that uses conditional mean as a solution to minimize the error variance. Predictors of educational attainment used were socioeconomic status represented by parent’s education, parent’s occupation, and family income, hours of study at school, intelligence, and students’ employment. Results from the quantile regression analysis showed that for several variables parameter estimates were significant only for certain quantiles. Parameters for hours of study at school and family income were significant only for lower quantiles, while intelligence and managerial/professional were significant for higher quantiles. There were variables that had significant parameters on OLS but not on all quantile from quantile regression. Significance tests of difference between quantiles showed non-significant values. Therefore, an analysis to estimate scale and skewness shift were not reasonable to be conducted.