Many participatory projects in rural Africa aim indirectly to enhance development by promoting different dimensions of social capital: cooperation in networks (formal or informal), trust, and norms of behavior that encourage mutually beneficial action. However, it is unclear whether these development initiatives can actually influence social capital, especially in the short term. To address this question, we used semi-experimental data to investigate the effects of agricultural research and development (ARD) on various indicators of social capital in the border region of Rwanda, Uganda, and the DRC. Specifically, we focused on the effects of the “Integrated Agricultural Research for Development Approach” (IAR4D) and compared it to conventional ARD efforts. We show that IAR4D has influenced the level of social capital, although not in all dimensions and not consistently for all countries. In the Democratic Republic of Congo (DRC) and Uganda, for example, IAR4D strengthened the networks that link villages to the outside world (bridging social capital), but not in other countries. We also find indications that IAR4D resulted in higher levels of intra-village networks (bonding social capital) in Rwanda and improved trust and norms of cooperation (cognitive social capital) in the DRC. Finally, we showed that traditional agricultural extension (ARD) has been less successful than IAR4D in increasing the level of social capital.