We use several data sets to consider the effect of teaching practices on student beliefs, as well as on organization of firms and institutions. In cross-country data, we show that teaching practices (such as copying from the board versus working on projects together) are strongly related to various dimensions of social capital, from beliefs in cooperation to institutional outcomes. We then use micro-data to investigate the influence of teaching practices on student beliefs about cooperation and students' involvement in civic life. A two-stage least square strategy provides evidence that teaching practices have an independent sizeable effect on student social capital. The relationship between teaching practices and student test performance is nonlinear. The evidence supports the idea that progressive education promotes social capital.