Medical authors are not apt to submit for publication manuscripts which contradict facts of human anatomy, or flaunt laws of physiology. Extensive training in the basic sciences and corrective influences of friendly experts caution them against perpetrating these offenses. Yet, authors of the same manuscripts are more likely than not to contravene "anatomy" and "physiology" of experimental design and statistical "dissection." In a study of 149 articles selected at random from ten widely read and highly regarded medical periodicals, Schor and Karten (p 1123) found that less than 28% have sufficient statistical support for drawn conclusions. None of the remaining 72% could pass muster in terms of experimental design, applicability of statistical tests, and the type of analysis performed.
There is no mystery attached to causes of this delinquency on the part of a medical author. The meager biostatistical fare offered by the medical curriculum does not prepare him for