An experimental design that compares real interventions against dummy versions to isolate true effects from confounds.