Most of the aid effectiveness literature has focused on potential growth effects of aggregate aid, with inconclusive results. Given that donors have repeatedly stressed the multidimensionality of their objectives, a more disaggregated view on aid effectiveness is warranted. Specifically, the impact of aid on education is analyzed empirically for almost 100 countries over the period 1970-2004. The effectiveness of sector-specific aid is assessed within the framework of social production functions. The Millennium Development Goals related to education, notably primary school enrolment, are considered as outcome variables. The main finding is that higher per-capita aid given for education significantly increases primary school enrolment, while domestic government spending on education does not. This result is robust to the method of estimation, employing instruments to control for the endogeneity of aid, and the set of control variables included in the estimations.