Abstract
In this study, we examine the effect of religiosity as measured by attendance at religious services on religious school choice. Particular attention is given to the possibly endogenous relationship between school choice and religiosity. We find that religiosity has an important causal effect on the demand for parochial schools. It is also shown that religiosity is substantially biased downward in OLS and probit estimates of parochial school choice. Finally, we discuss the implications of our findings for estimating the treatment effect of private school attendance on student outcomes.
Acknowledgements
We are grateful to two anonymous reviewers, the editors, Todd Elder, Naomi Feldman, and Oren Rigbi for their valuable comments and suggestions.
Notes
In a related study on Catholic schools, it could not be shown that religiosity was endogenous with the demand for Catholic schooling (Sander Citation2005).
The Amemiya–Lee–Newey test results for over identification of instruments were generated using Baum et al.'s (Citation2006) overid.ado program for Stata.9.
We are grateful to Todd Elder for providing us with the Stata program that calculates the bias in addition to many useful discussions on this procedure.
The size of the bias was found to be 0.08 and its standard error was 0.18.