Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We randomly assigned 115 primary schools in Bangladesh to one of two settings: children studying in groups with friends and children studying in groups with peers. The groups consisted of four people with similar average cognitive abilities and household characteristics. While the achievement of male students was not affected by the group assignment, low-ability females with friends outperformed low-ability females working with peers by roughly 0.4 standard deviations of the test score distribution. This is not due to the fact that friends tend to be of the same gender or to a higher frequency of interactions among friends.