Public and Catholic Schools: A Reanalysis of "Public and Private Schools"

In this study we reanalyze the impact of Catholic school attendance on the reading and mathematics achievement of senior and sophomore pupils using the same data base employed by Coleman, Hoffer and Kilgore in their report, Public and Private Schools. We identify four variables they did not control for-sex, handicap status, region of residence, and eighth grade college attendance expectations-that could be associated with self-selection into Catholic schools, and hence responsible for the apparent significant impact of Catholic school attendance on cognitive outcomes reported by Coleman et al. Using a variety of specifications and techniques for handling missing data, we find that we cannot confirm Coleman et al.'s findings. Except for a statistically significant but small advantage on sophomore reading tests, Catholic school pupils are found to do no better-or worse-than public school pupils.