Toward an Ethical Experiment

Abstract

Randomized Controlled Trials (RCTs) enroll hundreds of millions of subjects and involve many human lives. To improve subjects’ welfare, I propose an alternative design of RCTs that I call Experiment-as-Market (EXAM). EXAM Pareto optimally randomly assigns each treatment to subjects predicted to experience better treatment effects or to subjects with stronger preferences for the treatment. EXAM is also asymptotically incentive compatible for preference elicitation. Finally, EXAM unbiasedly estimates any causal effect estimable with standard RCTs. I quantify the welfare, incentive, and information properties by applying EXAM to a water cleaning experiment in Kenya (Kremer et al., 2011). Compared to standard RCTs, EXAM substantially improves subjects’ predicted well-being while reaching similar treatment effect estimates with similar precision.

Keywords: Research Ethics, Clinical Trial, Social Experiment, A/B Test, Market Design, Causal Inference, Development Economics, Spring Protection, Discrete Choice