A change to Title IX has spurred new single-sex public schooling in the US. Until recently, nearly all gender-segregated schools were private, and I therefore address potential selection bias in the effects on educational and labor market outcomes using within private sector comparisons, an index comparing expectations to outcomes, quantile regressions, and other techniques. Descriptive statistics suggest significant benefits, but more consideration of selection bias reveals less consistency. Girls' school alumnae are more likely than their coed peers to receive scholarships, but they are not more likely to pursue college degrees, and both genders are less likely to meet their own educational expectations. Moreover, single-sex schooling is not universally superior in supporting gender equity, as coeducational public schools yield the least segregated college major choices. On the other hand, I find 15-20% higher starting salaries among single-sex school graduates, but only persistently for men of median ability. Much of the benefit from single-sex schooling accrues to students already likely to succeed, but favorable selection is an insufficient explanation for all gains. Most notably, there are clear returns for both African-Americans and low income students.