An empirical study of the possible relation of treatment differences to quality scores in controlled randomized clinical trials