3 - Randomised Trials As A Dead-End For African Development

Abstract

Randomised Control Trials (RCTs) have been presented as a method for identifying interventions that will improve the lives of residents of economically poorer countries. I argue that RCT-driven development policy is unlikely to achieve even its narrower objectives and, instead, is likely to hinder the attainment of long-term economic and social development objectives in African countries. The article examines two fundamental methodological problems: first, that proselytisers of this approach, despite claims to the contrary, have not addressed the challenge of extrapolating from experimental results to policy interventions; second, that while the use of RCTs is framed as an objective scientific approach to policymaking, it systematically smuggles in the ideological and other biases of the researchers involved. These claims are illustrated with three sets of examples: controversial studies that involved cutting off water to poor households in Nairobi and randomising exposure to a Christian missionary programme in the Philippines; studies on civil servant absenteeism in Kenya and India cited in the 2019 Nobel award; and two case studies from South Africa pertaining to labour market and educational interventions where RCTs distorted the policy process with negative consequences. The conclusion is that the RCT approach is a dead end for African development.