The international trend towards the use of randomised trials in economics, particularly on topics pertaining to development, is methodologically suspect. In South Africa it is being used to advance a set of quasi-ideological positions under rubrics like ‘evidence-based policy’ that constitute a new missionary approach to basic education. Besides the methodological limitations, the factual premises of this work are also flawed; these compound the harms of an inappropriately hierarchical policymaking process and are likely to hinder, rather than advance, long-term improvement of the education system.