Replication Data for: Local Context and Mobilization in Poor Communities: Lessons from Kenya, Malawi and Zambia

Abstract

Why are some actors better able than others to mobilize the poor? Some scholars approaching this question focus on leaders, suggesting that different types of actors and their corresponding roles shape their abilities to encourage participation in community initiatives or support for specific candidates. Others emphasize actors’ connections to their followers, positing that individuals closely connected within a community mobilize more effectively. Together, these approaches suggest that the mobilizers’ success varies by community, yet there remains limited understanding of where and why certain actors are more influential. In this paper, we turn attention to how the social context moderates actors’ abilities to influence the poor. To do so, we analyze data from a factorial vignette experiment with over 14,117 poor respondents in 631 communities embedded in an original survey fielded in Kenya, Malawi, and Zambia. We find more local actors, including those without formal leadership positions, have greater influence than more distant ones, and their comparative advantage is magnified in communities with dense social ties. Results also indicate that the fear of community sanctions and desire to cooperate with others drive mobilization, while fear of mobilizer sanctioning does not. We further rule out the possibility that these findings are driven by the homogeneity of socially dense communities or their levels of social inequality or ethnolinguistic fractionalization. These findings turn attention away from the focus on leaders in official positions, whether customary or state. They also suggest that social development programs can be more effective when engaging locally embedded actors, particularly in socially dense communities and when aimed at promoting community social goods. This approach may help international donors and NGOs achieve more sustainable outcomes and foster community resilience.