Paradoxically, while sub-Saharan migration represents only a small fraction of total migrant flows and stocks in Europe, it has become a major topic of public debate. This is perhaps because quantitative data on African migration are quite scarce, so the phenomenon is poorly understood. In response to this knowledge gap, the MAFE project (Migration between Africa and Europe) was set up to collect original quantitative data with the aim of shedding new light on African migration. This article presents the methodology of the MAFE project, whose goal was to produce data for analysis of migration trends, their causes and consequences at micro level. The MAFE project is based on comparable, longitudinal (retrospective), multilevel and multi-topic data concerning three African migration flows (migrants from DR Congo, Ghana and Senegal). A multisite approach is adopted, based on surveys conducted in countries of both origin and destination. Beyond the MAFE project, this article highlights the classic problems encountered by surveys on international migration, and recommends a more (self-)critical approach in this field of research.