Our study investigates the impact of different social contexts (family, peer groups, school) on student outcomes of 5,110 students in Dutch secondary schools. Configuration is used to construct typologies from the intersection of these 3 social contexts of learning. Four empirically based configurations were found. The relative and additional impact of these configurations on cognitive and social outcomes of secondary school students have been analysed. Multilevel results show that certain elements of effective families and effective peer groups contribute to the math performance and school perception of secondary school students. Finally, our assumptions about additional effects of the configurations have been tested. (Contains 5 tables.)