Randomized controlled trials (RCTs) are now widely used in development economics. However, their use is often resisted by non-governmental development organizations. The objections they raise differ between the three types of activities of such non-governmental organizations (NGOs): capacity building, advocacy, and service delivery. This paper discusses the objections and alternatives to RCTs for each type. RCTs might not be appropriate even for service delivery, the activity which would appear to be best suited to their use. This is because typically local NGO staff can use their discretion in selecting communities or individuals for participation in a service-delivery programme. A standard RCT does not mimic the use of private knowledge of local circumstances and can therefore be misleading.
Chris Elbers and Jan Willem Gunning
WIDER Working Paper 2014/026