Guillaume Rousselet

Guillaume Rousselet

@robustgar

Followers3.5K
Following364

data parasite | queue jumper | statistics | visualisation | neuroscience | MEEG | #rstats #matlab #barbarplots EJN & BNA editor @CCNi_UofG

Glasgow, Scotland
Joined on September 05, 2013
Statistics

We looked inside some of the tweets by @robustgar and here's what we found interesting.

Inside 100 Tweets

Time between tweets:
9 days
Average replies
1
Average retweets
9
Average likes
25
Tweets with photos
32 / 100
Tweets with videos
0 / 100
Tweets with links
0 / 100

Fantastic Q&A about the roles of journals, peer review, data sharing, and the path forward for open science -- with @robustgar, @JohnnyFoxe, @JuanLerma1, and Jeff Dalley (of @CambPsych) They shared great practical advice too! Thinking of adding some sections to my CV… πŸ€”

New blog post: Interpreting regression models: a reading list https://t.co/xcOZisTbhr

Here is the cure to all these unwarranted "A predicts B" statements: cross-validation #rstat tutorial https://t.co/jPtovgbSdC to read in conjunction with https://t.co/9f4uCAcXLO

Out in @Meta_Psy Reaction Times and other Skewed Distributions: Problems with the Mean and the Median https://t.co/vMfXwgPfz0

I'm happy to say we're now starting to publish the backlog of our "in press" papers. Two highly technical papers published today, (@robustgar and @R__INDEX), and goal is for 10 more to be published within a month. https://t.co/a7MfpYES9b

[NEW POST] Mean or median reaction time? An open review of Miller (2020) https://t.co/0FZPm2Y0ZE Miller (2020) https://t.co/be4tJC15yd is a reply to Rousselet & Wilcox (2020) https://t.co/Zzm6bKB2do which builds upon Miller (1988) https://t.co/iWfvi7fqjL

Rigor Mortis: How Sloppy Science Creates Worthless Cures, Crushes Hope, and Wastes Billions

https://t.co/b6uZXXNHtf https://t.co/NF1W443fv1

Rigor Mortis: How Sloppy Science Creates Worthless Cures, Crushes Hope, and Wastes Billions https://t.co/b6uZXXNHtf https://t.co/NF1W443fv1

using robust measures of association (thanks @CyrilRPernet, @robustgar, and Rand Wilcox), we found both positive and negative associations of synegy with age, in localized clusters https://t.co/2wCulxAKq4

using robust measures of association (thanks @CyrilRPernet, @robustgar, and Rand Wilcox), we found both positive and negative associations of synegy with age, in localized clusters https://t.co/2wCulxAKq4

Does anyone know of a good, accessible introduction to defining outliers and different approaches to dealing with them? (The "accessible" bit of that sentence is really important!)

Quoted @Social__AI

We're really excited about the range of amazing SOCIAL PhD projects for September 2020 entry. See https://t.co/XTWnGIgcgW for the full list, including a project on using e-skin to generate emotive features, and another on optimising interactions with virtual environments.

Our @UKRI_News Centre for Doctoral Training in Socially Intelligent Artificial Agents @Social__AI is now accepting PhD applications. We will fund up to 15 SOCIAL AI PHD projects at the intersection of @UofGPsychology and @GlasgowCS from September 2020. https://t.co/Ceyose5zgX

Should scientists who average across trials before modelling be reported for fraud?

https://t.co/XxlHxXOVdB https://t.co/kcFRAS6GM8

Should scientists who average across trials before modelling be reported for fraud? https://t.co/XxlHxXOVdB https://t.co/kcFRAS6GM8

brms and multi-level models are πŸ’ͺ! πŸš€ Another: https://t.co/Vw21TgMxf6 https://t.co/ScQd9REel2

Quoted @jonmummolo

Johnson and Cesario have stated their analysis is still informative because it controls for county crime rates by race and other shooting attributes. But as Bayes’ rule shows, the addition of these control variables, X, does not solve the fundamental conceptual problem. 11/n https://t.co/6PclW6Y0Pb

Johnson and Cesario have stated their analysis is still informative because it controls for county crime rates by race and other shooting attributes. But as Bayes’ rule shows, the addition of these control variables, X, does not solve the fundamental conceptual problem. 11/n https://t.co/6PclW6Y0Pb

Re study of police shootings, scroll up for full flavor. Another example of social science using β€œcontrol” like a magic verb that manufactures truth. If only there were an axiomatic framework for deciding causal queries... https://t.co/aUxYahuXJg

New blog post: Baby steps in Bayes: Accounting for measurement error on a control variable #rstats

https://t.co/thfy1twBY1 https://t.co/dDSDCzrC5L

New blog post: Baby steps in Bayes: Accounting for measurement error on a control variable #rstats https://t.co/thfy1twBY1 https://t.co/dDSDCzrC5L

To look at individual differences, it is essential to consider test-retest reliability: https://t.co/IReAyyPNzR However, standard methods such as ICC can be very misleading: https://t.co/XxlHxXOVdB

This brilliant blog post by @Nate__Haines demonstrates the huge benefits of estimating test-retest reliability using hierarchical models instead of standard ICC on averaged data:

https://t.co/XxlHxXOVdB https://t.co/aAxnZOwUa9

This brilliant blog post by @Nate__Haines demonstrates the huge benefits of estimating test-retest reliability using hierarchical models instead of standard ICC on averaged data: https://t.co/XxlHxXOVdB https://t.co/aAxnZOwUa9

More good reasons to move away from idolising P from single studies. And towards cumulative science. https://t.co/2e3OrMdqoH https://t.co/UZ4PSJpNLE

More good reasons to move away from idolising P from single studies. And towards cumulative science. https://t.co/2e3OrMdqoH https://t.co/UZ4PSJpNLE

Next Page