NHST and Related Problems

adopt the author’s assertions. In particular, the purpose of scientific research articles is to devise a certain law or hypothesis on the basis of the researcher’s observations or collected data and to persuade with the law or the hypothesis. For effective persuasion, the expression as well as the assertion should be logical. A method which is used extensively in a number of scientific research articles to prove a certain hypothesis on the basis of collected data is null hypothesis significance testing (NHST). For example, to observe a difference among reactions under a specific experimental condition, under the null hypothesis that the means of reactions do not differ between two groups and an alternative hypothesis that the difference in the means of reactions between the two groups is not zero, a parametric statistical test is performed to compare the means, after which whether to accept or reject the alternative hypothesis is determined on the basis of the P value, which is the result of the parametric statistical test. However, the NHST, used by many researchers without any doubts as to its validity, has many problems. The NHST is often abused or misused, even seriously at times, without accurate knowledge, leading to incorrect interpretations of research results. This article briefly introduces the NHST, discusses the problems related to it, presents examples of its misuse, and suggests alternatives to solve or offset the problems.