Please use this identifier to cite or link to this item: http://hdl.handle.net/10419/209564
Authors: 
Lapierre, David
Lefebvre, Pierre
Merrigan, Philip
Year of Publication: 
2017
Series/Report no.: 
Research Group on Human Capital - Working Papers Series 16-02
Abstract: 
Very few studies analyze the long-term educational effects of private secondary school attendance while controlling for socioeconomic status. In Québec, the second most populous Canadian province, twenty percent of students at this level are enrolled in private schools subsidized by the government that however sets a relatively low ceiling for the fees in exchange for subsidies. Selection bias arising from a host of factors, preclude simplistic comparisons of their educational results with those of their public sector peers. This study uses the first four longitudinal waves of the two cohorts from Statistics Canada?s Youth in Transition Survey (YITS) to estimate the average treatment on the treated effect of private school on the high school graduation rate within the expected number of years after starting high school (5), enrolment in postsecondary institutions at age 19, university enrolment at age 21 or more, university graduation at age 24 or more, and enrolment in a professional degree program. The econometric estimation of treatment effects is based on a particular entropy balancing algorithm with a large set of key balancing covariates. Results are validated by a simulation-based sensitivity analysis for matching estimators. We find large, positive, robust, and statistically significant effects of private schooling on almost all outcomes analyzed. Most results are not sensitive to simulations of omitted variable bias.
Subjects: 
YITS
high school graduation
postsecondary education and professional programs enrolment and graduation
longitudinal data
treatment effect
entropy balancing
JEL: 
I20
I21
I28
Document Type: 
Working Paper

Files in This Item:
File
Size
1.23 MB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.