To the Editor.— Jeoffrey K. Stross, MD, and William R. Harlan, MD, stated that they assessed knowledge of the findings of the Hypertension Detection and Follow-up Program (HDFP) in family physicians and internists at two months and six months, respectively (1981;246:360). Because of this, any comparison between the two groups as groups of physicians that differ in fields of specialization is inappropriate. However, they publish in table form, specifically Table 2, a comparison of such with a P value as listed. Listing a P value represents a misuse of statistical inference. It is rather like posing a mathematical problem to left-handed students and right-handed students and then listing a P value referring to only the hand dominance without mentioning that one group of students was in the third grade while the other was in the fourth grade. Statistical comparison should only be made between two groups that differ in only