Documenting Research in Scientific Articles: Guidelines for Authors: 2. Reporting Hypothesis Tests

摘要

Proposed by Sir Ronald Fisher in 1920 as anmeasure of the strength of evidence, p values arenpart of an area of statistics called the frequentistnapproach to statistics. Also a part of the frequentistnapproach is a method of choosing between hypoth-neses, called hypothesis testing, which was developednby mathematicians Jerzy Neyman and Egon Pearsonnin the 1930s. Probability values and hypothesis test-ning are actually quite different concepts, but they arenwidely, if mistakenly, seen as parts of a coherentnapproach to statistical inference.n1 In fact, the fre-nquentist approach is widely used in biomedical re-nsearch. Although the logic behind it is elegant, it isnnot intuitively obvious, which is why it is so oftennmisunderstood. The guidelines here should help tonmake reports of hypothesis testing more complete.nThe guidelines here have been condensed fromnthose presented in How To Report Statistics InnMedicine.