Using prospective data from an 18-year longitudinal study of a birth cohort of 657 New Zealand children, this paper examines the effects of single-sex and coeducational secondary schooling on children's academic achievement. This analysis showed a pervasive tendency for children attending single-sex schools to have greater success in the School Certificate examinations, higher Burt reading scores, greater school retention, less likelihood of leaving school without qualifications, and less exposure to unemployment than children attending coeducational schools. However, a substantial amount of this association was explained by pre-entry differences in children's academic, behavioural, social, and family functioning. None the less, even after control for selection processes, children attending single-sex schools tended to perform better than their coeducated peers across a number of educational outcomes. Possible explanations for these differences are considered, including school gender composition, school climate and traditions, and possible inadequate control for selection and confounding factors.