The paper approaches the issue of school choice in an indirect manner by investigating the effectiveness of public, private government-dependent and private independent schools in 19 Organisation for Economic Co-Operation and Development countries selected from the PISA 2000 survey for this purpose. In a multilevel approach we estimate these sector effects, controlling for sociological characteristics of students and parents, school composition, teaching and learning conditions of schools and students', and principals' perception of the climate of their schools. The main explanation of the gross differences in mathematical achievement is the better social composition of private schools, both government-dependent and independent, which is a clear consequence of school choice. But our analysis also reveals that private independent schools are less effective than public schools with the same students, parents, and social composition, while private dependent schools are more effective than comparable public schools. The explanation of these remaining net differences in mathematical achievement seems to be the better school climate of private dependent schools. The comparison concludes that these net differences in mathematical achievement between public and private school sectors are equal across nations, despite the historical and legal variations in their educational systems and school choice approaches. (Contains 4 tables and 36 notes.)