This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
To the Editor:—
The interpretation of statistical statements of significance in medical articles was recently clarified by Schor1 who pointed out that the expression "P < 0.05" means that "... if the author were to draw conclusions from different research projects day in and day out, he would come to the wrong conclusion less than 5% of the time." This statement reflects the usefulness of statistical tests in controlling error, but there are some additional implications of great practical importance.First, the statement "P < 0.05" expresses the probability of the author being wrong for certain types of conclusions only. These conclusions are the kind that we would ordinarily label as "discoveries" or "demonstrations" of some phenomenon. For example, concluding that two treatment conditions differ in their effects or that two variables are related in some way. "P < 0.05" specifies our risk in making these positive assertions, but it
The Mystic Statistic P < 0.05. JAMA. 1966;195(7):594. doi:10.1001/jama.1966.03100070138051