Scientists rise up against statistical significance

By Valentin Amrhein, Sander Greenland, Blakeley B. McShane. Nature, 2019

2,466 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/bd933600365e9fe235425e932ac9c76652b59ab3

How researchers see this paper

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

In Statistics and Probability

Citations in the corpus

Statistical significance: p value, 0.05 threshold, and applications to radiomics—reasons for a conservative approach (2020, background): “Nature 520:612 https://doi.org/10.1038/520612a 10.”
Statistical Significance Testing for Natural Language Processing (2020, background): “However, experienced readers could still beneﬁt from reading this chapter, as concepts such as p-values are widely misunderstood and misused (Amrhein, Greenland, and McShane 2019).”
Factorial Designs, Model Selection, and (Incorrect) Inference in Randomized Experiments (2019, methodology): “Finally, we contribute to the literature on the pitfalls of focusing on statistical significance in applied work (e.g., Brodeur et al., 2016; Wasserstein & Lazar, 2016; Amrhein et al., 2019; Wasserstein et al., 2019; Brodeur et al., 2020).”
Deep Reinforcement Learning at the Edge of the Statistical Precipice (2021, background): “[2] Valentin Amrhein, Sander Greenland, and Blake McShane.”
The consequences of checking for zero‐inflation and overdispersion in the analysis of count data (2019, background): “2019) (and in other fields (Amrhein et al. 2019)), hypothesis testing remains prevalent.”
Analysis of compositions of microbiomes with bias correction (2020, background): “This is particularly important in the present climate when researchers are increasingly skeptical about making decisions based on p values (alone) 15 .”
GEDI launches a new era of biomass inference from space (2022, methodology): “While formal hypothesis tests based on a specific confidence level could be performed, the value of such tests has been questioned [38] and in our casemay be overly constrainingwhere the goal is to use observed departures from the validation data as a guide towards discovering potential biases.”
Recent Trends in the Use of Statistical Tests for Comparing Swarm and Evolutionary Computing Algorithms: Practical Guidelines and a Critical Review (2020, background): “A recent Nature paper [95] warns about the common mistakes in the interpretation about the meaning of the p-value, especially the statements about the alleged “no difference” between the groups when the null hypothesis is not rejected.”
Do energy efficiency policies save energy? A new approach based on energy policy indicators (in the EU Member States) (2020, neutral): “…and business or policy decisions should not be based only on whether a p-value passes a specific threshold ” (see also McShane and Gal (2017), Amrhein et al. (2019), and most of the contributions to the 2019 special issue of The American Statistician “ Statistical Inference in the 21st…”
Enhancing Rigor in Quantitative Entrepreneurship Research (2019, background): “…significance level thresholds to categorize findings either as significant or not significant has recently raised serious concerns across disciplines, leading to calls to ban the term “statistical significance” altogether (e.g., Amrhein et al., 2019; McShane et al., 2019; Wasserstein et al., 2019).”
Safe Testing (2019, background): “the 800 co-signatories of the recent Amrhein et al. (2019), or the call to abandon significance by McShane et al. (2019)).”
Evaluating the relationship between circulating lipoprotein lipids and apolipoproteins with risk of coronary heart disease: A multivariable Mendelian randomisation analysis (2020, methodology): “Interpretation of findings Whilst we desisted from dichotomising results of analyses purely on the basis of a P-value into being ‘significant’ or not [49,50], as a means of grading the strength of evidence against the null hypothesis, in both the univariable and multivariable MR analyses, we used a ”