We're not actually measuring educational attainment (whatever that is) but admission into a 4 year college program. This is like measuring the effectiveness of

We're not actually measuring educational attainment (whatever that is) but admission into a 4 year college program. This is like measuring the effectiveness of the speed of your network by how many feet of cable you bought. Measuring graduation from a 4-year college program would at least be a measure of attainment of credentials, but this can be gamed simply by lowering bars. (The campaign against high school drop-out rates has largely consisted of making it easier to stay and harder to drop out (rather than actually improve any useful outcomes); the campaign to improve girls' science scores in Australian schools has largely comprised removing science from the science curriculum. Educational research is always plagued by these issues.) Single sex schools clearly helped improve the lot of women in the US (and other countries). But it's not clear that it was because single sex schools are better than co-ed (which is at issue) or that they are better than no schools at all (which appears blindingly obvious).