Jiménez-Paneque Rosa
Subeditora y revisora estadística, Medwave. Address: Villaseca 21 Of. 702, Ñuñoa, Santiago, Chile. Email:
Medwave. 2016 Sep 9;16(8):e6534. doi: 10.5867/medwave.2016.08.6534.
The use of p-value and statistical significance have been questioned since the early 80s in the last century until today. Much has been discussed about it in the field of statistics and its applications, especially in Epidemiology and Public Health. As a matter of fact, the p-value and its equivalent, statistical significance, are difficult concepts to grasp for the many health professionals some way involved in research applied to their work areas. However, its meaning should be clear in intuitive terms although it is based on theoretical concepts of the field of Statistics. This paper attempts to present the p-value as a concept that applies to everyday life and therefore intuitively simple but whose proper use cannot be separated from theoretical and methodological elements of inherent complexity. The reasons behind the criticism received by the p-value and its isolated use are intuitively explained, mainly the need to demarcate statistical significance from clinical significance and some of the recommended remedies for these problems are approached as well. It finally refers to the current trend to vindicate the p-value appealing to the convenience of its use in certain situations and the recent statement of the American Statistical Association in this regard.
自上世纪80年代初至今，p值和统计学显著性的使用一直备受质疑。统计学领域及其应用，尤其是流行病学和公共卫生领域，对此展开了诸多讨论。事实上，对于许多或多或少参与到应用于其工作领域的研究中的卫生专业人员而言，p值及其等价物统计学显著性是难以理解的概念。然而，尽管它基于统计学领域的理论概念，但其含义在直观层面上应该是清晰的。本文试图将p值呈现为一个适用于日常生活的概念，因此直观上很简单，但它的正确使用离不开内在复杂的理论和方法要素。直观地解释了p值受到批评及其单独使用的背后原因，主要是需要区分统计学显著性和临床显著性，同时也探讨了针对这些问题的一些推荐补救措施。最后提到了当前为p值辩护的趋势，这一趋势基于其在某些情况下使用的便利性以及美国统计协会近期在这方面的声明。
