How could the addition of a control group improve the scientific accuracy of an experiment on the effectiveness of a new medication?