New Documents
Evaluating
the Impact of Development Aid Program: The Role of Randomized
Evaluations Esther Duflo Massachusetts Institute of
Technology Department of Economics and Poverty Action Lab Paper
prepared for the AFD Conference, November 25, Paris, 2004 "We do not
propose that all projects be subject to randomized evaluations. But we
argue that there is currently a tremendous imbalance in evaluation
methodology, and that increasing the share of projects subject to
randomized evaluation from near-zero to even a small fraction could
have a tremendous impact on knowledge about what works in development.
All too often development policy is based on fads, and randomized
evaluations could allow it to be based on evidence. 3 The paper
proceeds as follows: Section 1 discusses the methodology of randomized
evaluations: we present the impact evaluation problem, review why other
current evaluation methods may often be unable to adequately control
for selection bias, and discuss why randomized evaluations can be
useful in addressing the problems encountered by other evaluation
practices. Section 2 reviews recent randomized evaluations in
developing countries. Section 3 extracts lessons from the evaluations
described in Section 2, and Section 4 reviews an example of current
practice, offers political economy explanations for why randomized
evaluations are so rare, and discusses the role the international
agencies can play in promoting and financing rigorous evaluations,
including randomized evaluations. Section 5 concludes" (posted
25/06/05).
Email
the Editor: To tell him if a link is not working
Launch a new search
Return to MandE News homepage