Résumé
The growing popularity of experimentation in economics has widened the scope for economic experiments. In this paper we question the relevance of experimental methods for economic impact assessment. The major issue of impact evaluation is answering a counterfactual question. We show that the economists' experimental toolbox can provide the appropriate method to give the right answer, especially by relying on randomized field experiments (RFEs). We contrast RFEs to other types of experiments, and discuss the limitations of experiments for evaluation purposes, by presenting 3 case studies that relied on economic experiments at the individual, local and national levels.