Boys will be boys

Some new research on gender and education, via Tim Harford:

A new working paper from economists Victor Lavy of Hebrew University and Analía Schlosser of Princeton attempts to unpick the peer effects associated with gender, using data on nearly half a million students passing through Israel's school system in the 1990s. They compared consecutive year groups passing through the same school, figuring that if one year's group was 55 percent boys and the next year's was 55 percent girls, that difference was very likely to be random and thus susceptible to meaningful number crunching. Their answer chimes perfectly with the conventional wisdom: Boys benefit from being in a classroom with girls, but girls do not benefit from being in a classroom with boys.