Experimentally Estimated Impacts of School Vouchers on Educational Attainments of Moderately and Severely Disadvantaged Students

• In Event: Moscone Center West Roundtable Session 9
  In Roundtable Session: Making Sense out of Recent Mixed Results From Private School Voucher Evaluations

Abstract

Objective. Prior research on private-school vouchers has generally not reported long-term impacts on educational attainment separately for students in moderate and extreme poverty. We conduct an analysis on the heterogeneous effects of a private-school voucher program in New York City on college graduation.
Theoretical Framework. College enrollment rates vary substantially for students from moderately low-income and extremely low-income households (Bailey and Dynarski 2011). Theoretical and qualitative empirical research suggest that the ability of families to exploit opportunities for social mobility differ sharply between those who have modest material and cultural resources and those who are extremely deprived (Bourdieu, 1986). Wilson (1991; 2012) and others have drawn clear distinctions between the moderately and “truly” disadvantaged. The extent to which private-school vouchers can benefit all students is debatable. Indeed, critiques of private school vouchers have cautioned that choice schools cream the relatively advantaged within communities from lower socioeconomic backgrounds (Wells and Scott, 2001). By evaluating the voucher program separately for students from moderately and extremely impoverished backgrounds, we assess the effectiveness of vouchers for students across different levels of socioeconomic disadvantage.
Methods. In 1997, a program in New York City awarded private-school vouchers via random lottery to students from low-income backgrounds. We conduct an experimental analysis of the effect of being offered and using the voucher on college degree completion. Essentially, we compare postsecondary degree attainment rates between students who were awarded a voucher and students who were not awarded a voucher. We use multiple regression techniques to estimate the effect of receiving a voucher and conduct a mediation analysis to analyze effect heterogeneity across students with differing levels of poverty.
Data. Data come from the administrative records of the private school voucher program and are matched with National Student Clearinghouse data to track two- and four-year college degree attainment as of fall 2017. We have baseline and outcome data for 99 percent of our sample for a total of 2,634 observations.
Results. Considering students from severely and moderately disadvantaged backgrounds, we only find significant effects for the latter. Students from moderately disadvantaged backgrounds are 6 to 8 percentage points more likely to complete college. The offer of a voucher to students from extremely disadvantaged backgrounds yielded no effect on college completion.
Scientific or scholarly significance. The effects of this voucher program solely on moderately disadvantaged families suggest that cultural as well as pecuniary factors may be at work. Social and cultural capital are crucial elements for improving educational attainment and a private-school voucher that does not provide those elements may not deliver the desired, positive outcomes to students from the most severely disadvantaged backgrounds (Bourdieu, 1986). Although students from moderately disadvantaged backgrounds benefited from the voucher, the presence gaps in social and cultural capital among the severely disadvantaged might preclude them from realizing the same benefits. These findings have implications for the promise and potential of private-school voucher programs.

Authors

• Albert Cheng, University of Arkansas at Fayetteville

• Matthew Chingos, Urban Institute

• Paul E. Peterson, Harvard University