Effect size – one size doesn't fit all

• Published in: Multiple sclerosis and related disorders Vol. 46; p. 102587
• Main Author: Cutter, Gary
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.11.2020
• Subjects: effect size; Neurology; p-values; reporting standards; effect size; reporting standards; p-values
• ISSN: 2211-0348; 2211-0356; 2211-0356
• DOI: 10.1016/j.msard.2020.102587

•P-values are misused•Use of p-values is still valid, but should be accompanied by effect sizes•The context and impact should be considered with results There remains a pendulum swing to avoid p-values, but the binary, use or don't use p-values may be replacing one problem with another. This paper elaborates on the use of p-values and effect sizes, points out their utility and reminds that the context of the question remains critically important. There is no single measure that provides the sole value of a study. The over or under interpretation of various measures is cautioned. Researchers need to keep in mind there is bio in biostatistics and statistics in biology and medicine. Both are needed to understand results in context.