Who Benefits Most From Shadow Education? The Heterogeneous Effects of SAT Preparation Activities on SAT Scores

• In Event: Promising Scholarship in Education: Dissertation Fellows and Their Research

Abstract

This study asks the question, what are the causal effects of shadow education on SAT scores? Shadow education for the SAT includes various commercial preparation activities such as coaching, cram school, and even 1-on-1 tutoring. This study uses data from the Educational Longitudinal Survey of 2002 (ELS02). I particularly focus on understanding how the effects of shadow education differ according to socioeconomic and educational background. To do this, I examine variation in the effects on SAT scores depending on the individual likelihood of receiving shadow education, and I summarize the systematic trend in effects. My research design can answer such questions as who is most likely to utilize shadow education and who benefits most from shadow education. Does shadow education primarily benefit already advantaged students and thereby further increase the education gap (i.e., Positive selection; the enhancement function), or does it benefit students from less privileged backgrounds by helping them to close the education gap (i.e., Negative selection; the remedial function)? By examining this, I want to show how shadow education affects educational inequality and opportunity in the U.S.
In this study, I use the propensity score matching methods (hereafter PSM) to examine causal relationships between shadow education and SAT scores. Considering the fact that preexisting differences between those who receive treatment and those who do not may confound the effects of shadow education, I employ PSM to better conceptualize problems of selection bias to estimate the causal effect of shadow education. I also propose to extend the PSM framework to attend to treatment effect heterogeneity. The treatment effect heterogeneity approach assumes that individual responses to certain treatments are not identical and examines variation in the effects of treatment on outcomes for those with different probabilities of selection into treatment. To examine treatment effect heterogeneity, I use the stratification-multilevel (SM) method (Brand and Xie 2010). The SM method is based on propensity score matching estimation, but it particularly examines the variation of treatment effects as a function of the propensity strata and summarizes the systematic trend using a hierarchical linear model (Brand and Xie 2010:281; Brand and Davis 2011).
If shadow education does make a difference in academic achievement, it carries important implications concerning educational opportunity and stratification in society. It is obviously a mechanism for maintaining and increasing social stratification by conferring educational advantages on students who are already advantaged in terms of their rich economic, social and cultural capital. Given the increasing trend of shadow education in the U.S, I pose two key questions as to its effects: does shadow education have a causal effect on academic performance, and how do the effects of shadow education differ depending on an individual’s social and economic backgrounds? By examining these two questions, this research has important implications, not only for understanding how more advantaged families utilize their economic resources to acquire the educational advantages of shadow education, but also for suggesting how policymakers might approach the issue of shadow education to diminish educational stratification in the U. S.

Author

• Yool Choi, University of California - Los Angeles