An Introduction to Impact Evaluations with Randomized Designs

This note discusses impact evaluations of development interventions with randomized controlled trials (RCTs).

The most difficult part of evaluating the impact of an intervention is to separate its causal role. To estimate impact, researchers have to find ways to approximate the counterfactual, which is a prediction of what would have happened without the intervention. In practice, the counterfactual is estimated by measuring impacts for individuals who do not participate in the intervention, but are similar to those who do in as many respects as possible.

The paper presents studies from the Philippines, South Africa, India and Sri Lanka to demonstrate use of RCTs to measure impacts of microfinance. It states that RCTs:

• Are increasingly popular in impact evaluations of development interventions;
• Help to address the selection and reverse causation biases that are common to nearly all statistical evaluations;
• Can provide clear, transparent and credible evidence in complicated contexts;
• Have been used in studies on the impact of microfinance loans, savings, education and health interventions;
• Are not always feasible, representative, or focused on larger questions of interest.