The abstract of Psacharopoulos’ (1987) paper claims that his purpose is to ‘present evidence on differences between public and private schools in Columbia and Tanzania on a number of indicators’. He observes that ‘the findings are mixed . . .‘, and concludes with a relatively anodyne reference to a ‘possible explanation’ for the statistical results. If the paper reflected this academic and careful line of argument, it would be an interesting cont~bution to a general debate about the relative importance of school and family background in determining educational achievement. (Psacharopoulos cites Coleman (1981) for the U.S.A., and Heyneman and Loxley (1983), but the debate is, of course, far wider.) In fact, the paper starts by referring to the debate on ‘tuition tax credit’ in the United States and then says that a comparison of private and public schools is important because ‘the encouragement of private schooling may (a) provide incentives towards a better utilisation of social resources allocated to schools, and (b) relax the limited public capacity to support an educational system’ (pp. I-Z). To claim that private schools are more efficient than public schools is a very different proposition and much more controversial than the abstract would suggest, so that the status of the evidence used to make that claim is crucial. The data provided is, in fact, based on student achievement scores. Whilst recognising that ‘school quality is a multifaceted concept’ Psacharopoulos then says ‘it can be measured by either input or output’ so that he treats the educational system as a black box whose only function and connection with the rest of society is to process inputs (capital, labour and new students) and produce outcomes (qualifiedor unqualified students). Any notion that there is limited utility to the production of quali~cations (Hirsch, 1977; Dore, 1978) or that the educational process is, of itself, a valued experience is ignored. Moreover, schooling has other functions than the production of (un-)qualified students and interacts in far more complex ways with society. In particular, selectivity into private and public schools has impo~ant implications for elite formation and reproduction in developing countries (Foster, 1965, Bibby and Peil, 1974; Nyerere, 1983). This is not just a sterile argument from a non-economist-it crucially affects the interpretation of the results.