p value variability and subgroup testing
Open Access
- 13 February 2021
- journal article
- research article
- Published by Springer Science and Business Media LLC in European Journal of Nutrition
- Vol. 60 (8), 4203-4206
- https://doi.org/10.1007/s00394-021-02498-z
Abstract
This article discusses the variability and randomness of p values, the most widely used currency of evidence in nutritional and health studies. One implication of this, the importance of always testing interaction terms when subgroups are examined and presented separately is also discussed.Funding Information
- Rural and Environment Science and Analytical Services Division
This publication has 12 references indexed in Scilit:
- Scientists rise up against statistical significanceNature, 2019
- A manifesto for reproducible scienceNature Human Behaviour, 2017
- Common pitfalls in statistical analysis: Clinical versus statistical significancePerspectives in Clinical Research, 2015
- WhyPValues Are Not a Useful Measure of Evidence in Statistical Significance TestingTheory & Psychology, 2008
- Statistics in Medicine — Reporting of Subgroup Analyses in Clinical TrialsThe New England Journal of Medicine, 2007
- Data dredging, bias, or confoundingBMJ, 2002
- Statistics notes: Interaction 3: How to examine heterogeneityBMJ, 1996
- Statistics Notes: Interaction 2: compare effect sizes not P valuesBMJ, 1996
- Statistics Notes: Interaction 1: heterogeneity of effectsBMJ, 1996