The fickle P value generates irreproducible results

By L. Halsey, Douglas Curran-Everett, S. Vowler, G. Drummond. Nature Methods, 2015

739 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Inferential Statistics as Descriptive Statistics: There Is No Replication Crisis if We Don’t Expect Replication (2018, background): “…direct presentation of precise P -values—including P - values for alternative hypotheses (Poole 1987a, 1987b; Cohen 1994; Ziliak and McCloskey 2008; Hurlbert and Lombardi 2009; Amrhein, Korner-Nievergelt, and Roth 2017; Greenland 2017; Amrhein and Greenland 2018; Trafimow et al. 2018; Greenland…”
A protocol for conducting and presenting results of regression‐type analyses (2016, background): “An alternative is to present 95% confidence intervals for the regression parameters and effect size estimates and their precision (Halsey et al. 2015).”
An Introduction to Deep Reinforcement Learning (2018, neutral): “Across various fields, investigations in reproducibility have found problems in numerous publications, resulting in several works providing experimental guidelines in proper scientific practices (Sandve et al., 2013; Baker, 2016; Halsey et al., 2015; Casadevall and Fang, 2010).”
Experimental design and analysis and their reporting: new guidance for publication in BJP (2015, methodology): “…investigators have suggested that using ‘ P ’ to denote ‘statistical signiﬁcance’ as a way to denote detection of an ‘effect’ is inappropriate, and offer other solutions such as provision of effect size estimates and their precision from conﬁdence intervals (Colquhoun, 2014; Halsey et al ., 2015).”