I can see clearly now: Reinterpreting statistical significance

By Jonathan Dushoff, Morgan P. Kain, Benjamin M. Bolker. Methods in Ecology and Evolution, 2018

156 citations recorded by the source corpus.

Null hypothesis significance testing (NHST) remains popular despite decades of concern about misuse and misinterpretation. There are many recent suggestions for mitigating problems arising from NHST, including calls for abandoning NHST in favour of Bayesian or information‐theoretic approaches. We believe that NHST will continue to be widely used, and can be most usefully interpreted as a guide to whether a certain effect can be seen clearly in a particular context (e.g. whether we can clearly see that a correlation or between‐group difference is positive or negative). We believe that much misinterpretation of NHST is due to language: significance testing has little to do with other meanings of the word ‘significance’. We therefore suggest that researchers describe the conclusions of null‐hypothesis tests in terms of statistical ‘clarity’ rather than ‘significance’. We illustrate our point by rewriting common misinterpretations of the meaning of statistical tests found in the literature using the language of ‘clarity’. The meaning of statistical tests become easier to interpret and explain when viewed through the lens of ‘statistical clarity’. Our suggestion is mild, but practical: this simple semantic change could enhance clarity in statistical communication.

Subfields: Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/5c0f4774adf7e407f71b7a72ec7eadfe840eadea

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

What drives study‐dependent differences in distance–decay relationships of microbial communities? (2021, result): “We report the results of all null hypothesis tests in terms of statistical “clarity” rather than “significance”, in line with recommendations from Dushoff et al. (2019).”
Machine learning and deep learning—A review for ecologists (2022, unknown): “…of ML and DL and where traditional statistical methods are preferable and highlight current and emerging applications for ecological problems. dominated data analysis in E&E to this day (but see Dushoff et al., 2019 ; Gelman & Loken, 2014; Hartig & Barraquand, 2022 ; Muff et al., 2022 ).”
Parsimoniously Fitting Large Multivariate Random Effects in glmmTMB (2024, unknown): “After controlling for other covariates, Oxsimpa densities were clearly larger in the North zone than in the Wind Farm zone (i.e., the contrast is clearly positive, Dushoff, Kain, and Bolker 2019), while the South-Wind Farm contrast is negative and close to zero.”
The consequences of checking for zero‐inflation and overdispersion in the analysis of count data (2019, background): “Despite the ongoing debate surrounding the use (and misuse) of significance testing in ecology (Dushoff et al., 2019; Murtaugh, 2014) and in other fields (Amrhein et al., 2019), hypothesis testing remains prevalent.”