Attendance at Catholic high schools is estimated to improve math test scores and to increase high school graduation rates and enrolment in 4-year college. Propensity score matching methods are used to obtain these estimated effects, based on data from the National Educational Longitudinal Study. Since selection into Catholic schools is non-random, matching methods help to overcome the problem of choosing instruments for identifying the Catholic school effect on educational outcomes. The difference-in-differences approach is used on test score data in order to control for fixed unobservable influences on outcomes.