ABSTRACT
This chapter argues that interaction terms between Catholic and urban were good candidates for identifying variables because they were related to the probability of attending a Catholic school and they were unrelated to test scores. The corrected estimates indicate that Catholic school has no effect on test scores in four of five cases and a significant negative effect in the case of reading. Two sets of uncorrected estimates of test scores were undertaken. These estimates were undertaken to show how Catholic school variables are correlated with test scores if selection bias is not taken into account. The best Catholic high school in Chicago is ranked third while the best public high school in Chicago is ranked sixtieth. Twenty-three of sixty-two high schools in the Chicago public school system have an average score on the ACT exam that places them in the bottom one percentile nationally.
