The authors develop an evaluation method that combines qualitative evidence with quantitative survey data analyzed with propensity score methods on matched samples to study the impact of a participatory community-driven social fund on preference targeting, collective action, and community decision-making. The data come from a case study of five pairs of communities in Jamaica where one community in the pair has received funds from the Jamaica social investment fund (JSIF) while the other has not-but has been picked to match the funded community in its social and economic characteristics. The qualitative data reveal that the social fund process is elite-driven, and decision-making tends to be dominated by a small group of motivated individuals. But by the end of the project there was broad-based satisfaction with the outcome. The quantitative data from 500 households mirror these findings by showing that ex-ante the social fund does not address the expressed needs of the majority of individuals in the majority of communities. By the end of the construction process, however, 80 percent of the community expressed satisfaction with the outcome. An analysis of the determinants of participation shows that better educated and better networked individuals dominate the process. Propensity score analysis reveals that the JSIF has had a causal impact on improvements in trust and the capacity for collective action, but these gains are greater for elites within the community. Both JSIF and non-JSIF communities are more likely now to make decisions that affect their lives which indicates a broad-based effort to promote participatory development in the country, but JSIF communities do not show higher levels of community-driven decisions than non-JSIF communities. The authors shed light on the complex ways in which community-driven development works inside communities-a process that is deeply imbedded within Jamaica's socio-cultural and political context.