New on the blog: Using Bayesian tools to be a better frequentist
Turns out that for neg. bin. regression with small samples, standard frequentist tools fail to achieve their stated goals. Bayesian computation ends up providing better frequentist guarantees. www.martinmodrak.cz/2025/07/09/u...
Turns out that for neg. bin. regression with small samples, standard frequentist tools fail to achieve their stated goals. Bayesian computation ends up providing better frequentist guarantees. www.martinmodrak.cz/2025/07/09/u...
Using Bayesian tools to be a better frequentist
www.martinmodrak.cz
July 11, 2025 at 5:49 AM
New on the blog: Using Bayesian tools to be a better frequentist
Turns out that for neg. bin. regression with small samples, standard frequentist tools fail to achieve their stated goals. Bayesian computation ends up providing better frequentist guarantees. www.martinmodrak.cz/2025/07/09/u...
Turns out that for neg. bin. regression with small samples, standard frequentist tools fail to achieve their stated goals. Bayesian computation ends up providing better frequentist guarantees. www.martinmodrak.cz/2025/07/09/u...
So much statistics discourse orbits around Bayes verses frequentist debates, but in practice Bayes verses frequentist is often a false dichotomy. Just as important as the inferential strategy is the underlying model, and how that model is specified. 🧵
December 3, 2025 at 5:02 AM
So much statistics discourse orbits around Bayes verses frequentist debates, but in practice Bayes verses frequentist is often a false dichotomy. Just as important as the inferential strategy is the underlying model, and how that model is specified. 🧵
I’m a frequentist, when I give a movie 7/10 it means if I watch it 10 times I enjoy 7 of them
i'm a bayesian. when i give a movie a 7/10 it means i'm 70% sure that it's good
December 25, 2024 at 4:24 PM
I’m a frequentist, when I give a movie 7/10 it means if I watch it 10 times I enjoy 7 of them
I’m economically Bayesian but socially frequentist
October 9, 2025 at 10:50 PM
I’m economically Bayesian but socially frequentist
i find it very odd that overleaf's native spellchecker does not recognise words like "frequentist" or "covariate"
May 22, 2026 at 9:39 AM
i find it very odd that overleaf's native spellchecker does not recognise words like "frequentist" or "covariate"
Are you a Frequentist/Bayesian? Or do you just happen to use Frequentist/Bayesian methods.
Were you a Frequentist/Bayesian? Or did you just happen to use Frequentist/Bayesian methods.
Were you a Frequentist/Bayesian? Or did you just happen to use Frequentist/Bayesian methods.
September 25, 2026 at 1:53 PM
Are you a Frequentist/Bayesian? Or do you just happen to use Frequentist/Bayesian methods.
Were you a Frequentist/Bayesian? Or did you just happen to use Frequentist/Bayesian methods.
Were you a Frequentist/Bayesian? Or did you just happen to use Frequentist/Bayesian methods.
To Be a Frequentist or Bayesian? Five Positions in a Spectrum hdsr.mitpress.mit.edu/pub/axvcupj4
#philsci
#philsci
To Be a Frequentist or Bayesian? Five Positions in a Spectrum
hdsr.mitpress.mit.edu
December 4, 2024 at 11:32 AM
To Be a Frequentist or Bayesian? Five Positions in a Spectrum hdsr.mitpress.mit.edu/pub/axvcupj4
#philsci
#philsci
In 2011 Bayesians argued science would improve if we switch from frequentist to Bayesian statistics.
In 2025 we have realized science improves if we have scientists with smarter ideas about improving science than proposing to switch from Bayesian to frequentist statistics.
In 2025 we have realized science improves if we have scientists with smarter ideas about improving science than proposing to switch from Bayesian to frequentist statistics.
December 4, 2025 at 12:21 PM
In 2011 Bayesians argued science would improve if we switch from frequentist to Bayesian statistics.
In 2025 we have realized science improves if we have scientists with smarter ideas about improving science than proposing to switch from Bayesian to frequentist statistics.
In 2025 we have realized science improves if we have scientists with smarter ideas about improving science than proposing to switch from Bayesian to frequentist statistics.
If your frequentist analysis gives the same result as a Bayesian analysis, that means nothing for either paradigm. No frequentist is bothered that Bayes "works" - they use Bayesian calculations all the time. No Bayesian is bothered that a freq estimator agrees - it's interpretation that is doubted.
June 4, 2025 at 6:30 AM
If your frequentist analysis gives the same result as a Bayesian analysis, that means nothing for either paradigm. No frequentist is bothered that Bayes "works" - they use Bayesian calculations all the time. No Bayesian is bothered that a freq estimator agrees - it's interpretation that is doubted.
A frequentist showing their analysis to a Bayesian
A social scientist showing their Stata code to a #Rstats person
July 26, 2024 at 12:27 PM
A frequentist showing their analysis to a Bayesian
Instead of the Kinsey scale we need the Lindley scale for statistical orientation:
0 - exclusively frequentist
1 - occasionally tempted by a posterior
2 - frequentist in practice but relaxed about it
...
5 - Bayesian in practice but not subjectivist
6 - Savage Bayesian
X - data scientist
0 - exclusively frequentist
1 - occasionally tempted by a posterior
2 - frequentist in practice but relaxed about it
...
5 - Bayesian in practice but not subjectivist
6 - Savage Bayesian
X - data scientist
September 17, 2026 at 8:23 PM
Instead of the Kinsey scale we need the Lindley scale for statistical orientation:
0 - exclusively frequentist
1 - occasionally tempted by a posterior
2 - frequentist in practice but relaxed about it
...
5 - Bayesian in practice but not subjectivist
6 - Savage Bayesian
X - data scientist
0 - exclusively frequentist
1 - occasionally tempted by a posterior
2 - frequentist in practice but relaxed about it
...
5 - Bayesian in practice but not subjectivist
6 - Savage Bayesian
X - data scientist
📢 Frequentist != NHST 📢
February 13, 2025 at 4:21 AM
📢 Frequentist != NHST 📢
i am teaching the undergrads the behrens-fisher problem so all my frequentist colleagues are on notice
April 5, 2025 at 7:58 AM
i am teaching the undergrads the behrens-fisher problem so all my frequentist colleagues are on notice
this is why i was extra proud that (1) we assessed the sensitivtiy of my bayesian work alongside the frequentist examples in our AMIP / data dropping paper, and (2) it did garbage just like in the frequentist examples
Crazy coincidence how often [thing that will fix science of all of its ailments] coincides with [thing the speaker considers themselves an expert in]. Makes you think
July 15, 2025 at 1:31 PM
this is why i was extra proud that (1) we assessed the sensitivtiy of my bayesian work alongside the frequentist examples in our AMIP / data dropping paper, and (2) it did garbage just like in the frequentist examples
has anyone written a nice review or summary of Bayesian approaches to causal inference? most of what I'm aware of is focused on frequentist approaches, or implicitly frequentist (caring about bias, ignoring variance); and methods which may not translate easily to the Bayes world (e.g. weighting)
June 28, 2025 at 7:12 AM
has anyone written a nice review or summary of Bayesian approaches to causal inference? most of what I'm aware of is focused on frequentist approaches, or implicitly frequentist (caring about bias, ignoring variance); and methods which may not translate easily to the Bayes world (e.g. weighting)
I wonder about that. Were they raised Frequentist? Or did they just learn Statistics and Probability theory that happened to be Frequentist?
September 25, 2026 at 3:28 PM
I wonder about that. Were they raised Frequentist? Or did they just learn Statistics and Probability theory that happened to be Frequentist?
They will be *scandalised* to find that the frequentist foundation of statistics is openly debated and that Objective Bayesianism as a statistical epistemology is actually quite popular
These fools are about to be absolutely fuckass mad, just infuriated and offended beyond all belie, when they find out that truth does not consist in a proposition corresponding to facts but rather is a disquotational device used for blind ascriptions.
April 20, 2026 at 9:10 AM
They will be *scandalised* to find that the frequentist foundation of statistics is openly debated and that Objective Bayesianism as a statistical epistemology is actually quite popular
Bayesian vs frequentist is definitely older than rationalism. The issue with attempting to be explicitly Bayesian about everything is actually assigning likelihoods to things when unknown unknowns are common generates spurious precision
It's also an insane amount of work that they ~never do.
It's also an insane amount of work that they ~never do.
September 25, 2026 at 2:53 PM
Bayesian vs frequentist is definitely older than rationalism. The issue with attempting to be explicitly Bayesian about everything is actually assigning likelihoods to things when unknown unknowns are common generates spurious precision
It's also an insane amount of work that they ~never do.
It's also an insane amount of work that they ~never do.
Before we know it, Canada will have a Bayesian and a Frequentist party.
"I'm going to govern in econometrics" is a winning campaign slogan for the ages.
April 29, 2025 at 4:46 AM
Before we know it, Canada will have a Bayesian and a Frequentist party.
It’s also very funny that, unless I’m misunderstanding, his “Bayesian priors” are purely frequentist.
September 4, 2026 at 12:19 AM
It’s also very funny that, unless I’m misunderstanding, his “Bayesian priors” are purely frequentist.
better than having a frequentist pope tbqh
woke pope this, woke pope that, bro we've got bigger problems he's BAYESIAN
May 9, 2025 at 5:36 PM
better than having a frequentist pope tbqh
If you're a frequentist it's called "Fewer Wrong."
June 16, 2024 at 4:50 PM
If you're a frequentist it's called "Fewer Wrong."
frequentist statistics is about long run frequency behavior because the conversations about it go on and on forever
February 27, 2026 at 5:51 PM
frequentist statistics is about long run frequency behavior because the conversations about it go on and on forever
was looking at a paper, got confused about its statistical methods, went down a rabbit hole on Bayesian stats, and now I both have a headache and don't trust frequentist methods
January 3, 2026 at 1:17 PM
was looking at a paper, got confused about its statistical methods, went down a rabbit hole on Bayesian stats, and now I both have a headache and don't trust frequentist methods
Another front just opened in the frequentist-Bayesian war
September 17, 2026 at 6:23 PM
Another front just opened in the frequentist-Bayesian war