A new National Bureau of Economic Research study examines how gender composition in schools and classrooms impacts student performance. Analysts study middle school classrooms in Seoul, South Korea, where the students are randomly assigned to single sex schools, co-ed schools with single sex classes, and co-ed schools with mixed gender classes. In Seoul, under the “Equalization Policy,” the prior achievement of students is used to “balance” the performance in the classroom. In other words, they dole out kids of varying performance levels across classes—so essentially the average quality of peers is held constant. Students are not allowed to submit schooling preferences, which means that compliance to random assignment is high. Other variables, like curriculum and school funding, are more or less held constant because Korea adheres to a national curriculum and schools are centrally financed. The sample comprises 76 percent of middle schools in Seoul; no differences were found in student background across school types. There are two key findings. First, classroom or school gender composition does not impact outcomes for females. Second, the impact of single sex education on male achievement varies by school gender composition, with single sex schools increasing achievement and single  sex classes within mixed gender schools decreasing achievement. Yet interestingly, they find that the impact of single sex schools on male achievement is driven by increases in student effort and study time—not gender. Males report spending more time on homework and tutoring. Analysts hypothesize that single sex schools are able to specialize teaching techniques when they teach only males—but in single sex classes within co-ed schools, they teach both girls and boys, so educators are less likely to change their teaching style. One possibility: Without girls around, the boys had nothing better to do than their homework.

SOURCE: Soohyung Lee, Lesley J. Turner, Seokjin Woo, and Kyunghee Kim, “All or Nothing? The Impact of School and Classroom Gender Compositi...,” National Bureau of Economic Research, Working Paper No. 20722 (December 2014).