Racial Segregation Among Public and Private Schools

Coleman, Hoffer and Kilgore (1982) analyze racial enrollment in public and private high schools and conclude that private schooling does not affect the overall racial segregation in American schools. Four serious flaws in their argument render their analysis meaningless and their conclusion improper. (a) The segregation index they use in their empirical analysis is a function of the racial composition of the set of schools to which it is applied. Because public and private sectors differ in proportion black, their index exaggerates the difference between public and private sectors and misspecifies the key comparison in the final stage of their argument. (b) To permit evaulation of their argument, they assume that private school pupils are reallocated to public schools according to the existing racial composition of public schools. This assumption is implausible, and its use contributes tautologically to their conclusion. (c) Their claim that within-sector and between-sector components of segregation are counterbalancing is based on a very insensitive test of their argument, and the test itself is statistically incorrect-variance components are all positive. (d) Their treatment of the nation's public schools as one school district and the nation's private schools as another has no real-world applicability. Segregation should be examined within specific administrative and geographic settings. The only unflawed portion of their analysis is their confirmation of the fact that private schooling, nationwide, tends to separate white students from the racially more diverse public sector. The rest of their analysis and conclusions should be recognized as irrelevant to the policy debate they seek to influence.