In the past decade there has been a remarkable surge of interest in using impact evaluation to establish the effectiveness of development interventions. The conventional evaluation methods used by development consultants typically rely on simplistic before-after or with-without comparisons. Impact evaluation, by contrast, involves comparing actual outcomes with a formal counterfactual. Any differences between the two can be attributed to the intervention if, and only if, the counterfactual is credible. A conventional before-after comparison fails this test since the “before” situation is obviously not a credible counterfactual: outcomes could have changed over time for reasons unrelated to the intervention.
Randomized controlled trials (RCTs) are usually considered the preferred design for impact evaluation, but there are also regression-based techniques such as regression discontinuity designs and regressions in first differences (double differencing).
Gunning, J-W. "How Can Development NGOs Be Evaluated?". Ferdi, Document de travail P51, octobre 2012
Codes JEL : L31