The purpose of this study is to analyze the reality of the causal effect of application plan to
special high schools on private tutoring expenditure using data from 4th to 6th year of Seoul
Education Longitudinal Study(SELS). In particular, we use a penalty regression model to find out
the pre-covariates which affects the application plan to special high schools, math and English
private tutoring expenditure, and estimate the propensity score considering the selected covariates.
After adjusting the selection bias using the inverse probability weighting(IPWT) method, we
estimated the average treatment effect(ATE) of the application plan to special high schools on
private tutoring expenditure.
The result of the study shows that there is no causal effect of the application plan to special
high schools on private tutoring expenditure for English after balancing pre-covariates between
special high school planning group and general high school planning group. But there is a causal
relationship on the private tutoring expenditure for mathematics. In addition, it was confirmed that
parents' economic and educational support was confirmed as a direct variable to the student's
high school application plan and participation in private tutoring, as well as working as a mediator
variable. We also discussed the importance of balancing of pre-covariates in the analysis of
causality effect of the application plan to special high schools