Ten ironic rules for non-statistical reviewers

By Karl J. Friston. NeuroImage, 2012

362 citations recorded by the source corpus.

Subfields: Sociology and Political Science, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/8a8db07fae1d07826f3b0d959552453c8b4a55eb

How researchers see this paper

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

In Sociology and Political Science

In Statistics and Probability

Citations in the corpus

The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences (2017, background): “…optimal to pursue large sample sizes in neuroimaging because effects that require large samples are not sufficiently large to be of practical or theoretical importance (Friston, 2012, though see commentaries; Button et al., 2013a; Friston, 2013; Ingre, 2013; Lindquist, Caffo, & Crainiceanu, 2013).”
Ten simple rules for predictive modeling of individual differences in neuroimaging (2019, neutral): “For a discussion of the pros of explanatory models, we point the interested reader to a series of comments and replies (Friston, 2012, 2013; Lindquist et al., 2013; Reiss, 2015) on this topic.”
The relation between statistical power and inference in fMRI (2017, background): “The primary interest in large effects, and consequently, the sufficiency of small sample sizes to provide reasonable statistical power, has more recently been restated [7], but also criticized [8,40].”
A statistical framework for neuroimaging data analysis based on mutual information estimated via a gaussian copula (2016, background): “MI measures dependence on a common scale (bits), which provides a meaningful effect size [Friston, 2012] and allows direct comparisons across different responses, experimental modalities or with behavior.”
Mind The Hype: A Critical Evaluation and Prescriptive Agenda for Research on Mindfulness and Meditation (2017, neutral): “Yet calculating valid estimates of effect sizes in neuroimaging data is extremely difficult (Fox et al., 2016; Fox et al., 2014; Friston, 2012; Hupé, 2015).”
Evaluation of a decided sample size in machine learning applications (2023, methodology): “Indeed, it is commonly accepted that large sample sizes are not a substitute for good hypothesis testing [10]; Friston [10] suggested that a minimal sample size for producing statistically significant results is 16 subjects with a good effect size.”