Aubrey Clayton
@aubreyclaytonAuthor of Bernoulli's Fallacy | @NautilusMag @GlobeIdeas | Opinions and Bayesian priors are my own. He/him
Similar User
@StatModeling
@LearnBayesStats
@alex_andorra
@rlmcelreath
@LearnedLeague
@mcmc_stan
@paulbuerkner
@willkurt
@AllenDowney
@f2harrell
@eliasbareinboim
@NicholasChiuMD
@AnkaReuel
@aloctavodia
@VanishingData
"If Ed Jaynes were to rise from the grave to lambast frequentist statistics one last time, Bernoulli’s Fallacy is the kind of book he would want to write." 😍😍😍 tandfonline.com/doi/full/10.10…
Deep learning algorithms are revolutionizing weather forecasting. Learn more about it as @photayy reads "How to Predict Extreme Weather" by @aubreyclayton in a new Nautilus Narration. YouTube: bit.ly/3CiJeIk Spotify: bit.ly/3NXPecb
Stats people love to talk about how ML/AI is "just stats" and yet don't like to chat about the co-development of early statistics with the eugenics movement. Trust me, AI is not the first time mathematics was used for questionable ends.
My latest on “The New Artificial Intelligentsia,” part of the Legacies of Eugenics series @LAReviewofBooks Thanks to all who read + share!
My personal belief is you should just have basic human decency to not to talk to people this way, especially if you’re a public figure with 1M+ followers and the person you’re yelling at is a grad student with ~400 followers
Next Tuesday (9/17) at 4pm ET I'll be joining an online panel hosted by the University of Virginia on the subjects of statistics, AI, and society. Naturally I'll use my time to talk about how racist Karl Pearson was. Registration is open now! karshinstitute.virginia.edu/events/co-opti…
A fascinating example of statistical brain poisoning. What would it mean for a die to be "unbiased" and how could you ever tell? Unbiased dice, fair coins, etc., are like isosceles right triangles. Useful to think about in theory but never encountered in the real world.
Mathematics. You toss a dice, and it lands "6." You repeat the toss. It lands "6" again. How many repeats of 6 are required for you to declare that the dice is biased.
Messier relationships, you say? I hope the Eagle Nebula finally gets together with Ptolemy's Cluster
“Emily in Paris” has returned to Netflix for a fourth season with its charming absurdity, maximalist accessories and a Parisian love triangle. Get an inside look at baby pink (red) carpet premiere. nyti.ms/3SMymHR
I got one of my favorite emails this morning: "The analysis of the Monty Hall problem in your book is wrong !" Well, my friend, your brain is in for a long ride -- let's get started...
Then there are some who, upon learning that frequentist statistics is entangled with the eugenics movement, will find it a huge point in its favor.
@aubreyclayton Bernoulli's Fallacy 'the best single-volume work describing and contributing to the debates in modern statistics on the shelves today.' Klyve 2022 tandfonline.com/doi/full/10.10…
The interrelated histories of the eugenics movement and statistics, particularly as they overlap in the concept of correlation, should caution us against being taken in by the seeming neutrality of mathematical formulae. @LAReviewofBooks lareviewofbooks.org/article/the-co…
On Reddit's statistics forum, the most common question is "What test should I use?" My answer, from 2011, is "There is only one test" allendowney.blogspot.com/2011/05/there-…
Think I'll take to the sky for a bit. See you up there? bsky.app/profile/aubrey…
Yes, it's true (except for the way in which it's false).
In intro stats we should just teach that a 95% confidence interval means that there's a 95% probability the true value is in the interval. It's fine. It's true (before the data are collected).
Think of all the ways correlation is ruling our world: the discussion of technology and young people; algos' decisions to target content & ads; polling... It is a means of trapping people into categories to manage them: social eugenics.
I know nothing of this research, but I think the idea that a trial with n=2 can never be meaningful is a product of frequentist-statistical brain poisoning. Many important experiments have had small n. To quote Oppenheimer: "Now I am become death, the destroyer of worlds (n=1)."
In intro stats we should just teach that a 95% confidence interval means that there's a 95% probability the true value is in the interval. It's fine. It's true (before the data are collected).
"Eugenics was never about passive best-estimate prediction but rather about interventions." @aubreyclayton excavates the troubling correlation between the birth of statistical methods and the history of eugenics. lareviewofbooks.org/article/the-co…
![Aspiring philosopher; tolerable human; "amusing combination of sardonic detachment & literally all the feelings felt entirely unironically all at once" [he/his]](https://pbs.twimg.com/profile_images/1064478691068137473/NnNlroSU.jpg)
United States Trends
- 1. Tyson 517 B posts
- 2. Syracuse 21,1 B posts
- 3. #wompwomp 6.652 posts
- 4. Karoline Leavitt 14,2 B posts
- 5. Kiyan 25,3 B posts
- 6. Pence 60,8 B posts
- 7. Paige 7.489 posts
- 8. Kash 105 B posts
- 9. The FBI 270 B posts
- 10. Debbie 35,2 B posts
- 11. Whoopi 115 B posts
- 12. #LetsBONK 16,4 B posts
- 13. Villanova 1.732 posts
- 14. White House Press Secretary 18,5 B posts
- 15. Dora 24,5 B posts
- 16. Shu Shu 27,2 B posts
- 17. Bucks County 94,7 B posts
- 18. Mike Rogers 21,9 B posts
- 19. Kyle Neptune N/A
- 20. Boiled 4.866 posts
Who to follow
-
Andrew Gelman et al.
@StatModeling -
Pierre-Simon Laplace
@LearnBayesStats -
Alexandre Andorra
@alex_andorra -
Richard McElreath 🦔
@rlmcelreath -
LearnedLeague
@LearnedLeague -
Stan
@mcmc_stan -
Paul Bürkner
@paulbuerkner -
Will Kurt
@willkurt -
Allen Downey
@AllenDowney -
Frank Harrell
@f2harrell -
Elias Bareinboim
@eliasbareinboim -
Nicholas Chiu, MD MPH
@NicholasChiuMD -
Anka Reuel
@AnkaReuel -
[email protected]
@aloctavodia -
Vanishing Gradients Podcast
@VanishingData
Something went wrong.
Something went wrong.