Bias may be inadvertently introduced to a drug effect study by restricting the study sample or controlling for a variable that is a “collider”.