One-sentence summary : Randomized controlled trials by behavioural economists pretend to be pragmatic and only interested in what really works to solve practical problems but in reality they have notorious normative and ideological aspects. Key points: Behavioural RCTs ignore contexts and composition effects and reflect the biases of those who perform assessments. Behavioural randomizers presume without demonstrating that market exchanges are the most effective form of regulation for societies in all situations of social life. The positive or negative incentives ("nudges") offered by behavioural economics aim to normalize the behaviour of consumers, users, employees or small/independent producers. They are part of a set of power devices by which individual behaviours are shaped and forced, without their knowledge, to conform to dominant class interests.