What strategy can be used to reduce bias for a study in which participants are assigned to one of two treatment interventions?