Evaluating Social Programs: The Process of Impact Evaluations Described and Applied

Abstract

There is an increasing demand for quality evaluations of aid projects to find out what works for development and what does not. The effect of specific aid projects should be measured through impact evaluations using robust methodologies. We believe that evidence from proper impact evaluations can help move the world towards better policy-making and poverty reduction. This thesis provides a framework for the process of conducting an impact evaluation from beginning to end. Our proposal is that the best way to design an aid project is to (1) include the steps of careful resultbased monitoring in the impact evaluation process, and (2) make the impact evaluation a randomized controlled experiment if the circumstances allow for it. Randomization should ensure that the measured effects can be attributed to the project in question. To demonstrate the practical application of our framework, the framework is applied to an agricultural development project by Norwegian Church Aid aimed at smallholder farmers in Malawi.