Can Participants Estimate Program Impacts ? Evidence from a Labor Market Intervention

We test whether participants in two labor market interventions hold accurate beliefs about program impacts, exploiting a randomized evaluation to compare participants’ beliefs to intent-to-treat estimates. We provide a framework for comparing methods of belief elicitation and identifying potential biases. We find the program participants do a poor job of estimating their own counterfactual (probabilistic) outcomes. However, we extend the existing set of best practices by demonstrating that participants are quite good at estimating average treatment impacts on the population once behavioral biases are taken into account.