School sector and cognitive performance: When is a little a little?

Coleman, Hoffer, and Kilgore's claims regarding the effects of Catholic schools on cognitive achievement have, evoked much controversy. Critics have argued that Catholic schools enroll students of superior academic competency, and that Coleman et al., using cross-sectional testing data, could not distinguish differential sector effectiveness from this selection effect. The first follow-up (1982) of the High School and Beyond base-year sophomore cohort allows a stronger design for studying this issue. We use sophomore test performance to control for input-level differences in competency while predicting senior test performance in several cognitive domains. The omission of such input controls leads to a substantial upward bias in the estimate of Catholic-school effects on achievement. We also show that the so-called common-school effect found by Coleman and his colleagues disappears when appropriate input-level test controls are applied. Our best estimate of the Catholic-school effect on cognitive growth from the sophomore to senior year, using aggregate sophomore-to-senior change in performance as a yardstick, is about two thirds of a year's growth. We judge differences of this magnitude to be substantively trivial because they correspond to less than 0.1 standard deviation in test performance. We conclude that sector differences in test performance are too small to warrant the attention they have received.