Steve Haroz
@steveharoz.com
Visual perception and cognition scientist
(he/him)
My site: http://steveharoz.com
R guide: https://r-guide.steveharoz.com
StatCheck Simple: http://statcheck.steveharoz.com
(he/him)
My site: http://steveharoz.com
R guide: https://r-guide.steveharoz.com
StatCheck Simple: http://statcheck.steveharoz.com
Pinned
Steve Haroz
@steveharoz.com
· Mar 14
Invalid Conclusions Built on Statistical Errors
When you see the statistical result of an experiment like p = 0.003 or a 95% confidence interval of , you might assume a certain clarity and definitiveness in those results. A null effect is very unli...
steveharoz.com
BLOG POST: Invalid Conclusions Built on Statistical errors
If you see a small p-value or a large separation in confidence intervals, you may assume that an effect is reliable. But overly simple reporting may be hiding serious conclusion-altering errors.
Here are some example scenarios:
1/🧵 #stats
If you see a small p-value or a large separation in confidence intervals, you may assume that an effect is reliable. But overly simple reporting may be hiding serious conclusion-altering errors.
Here are some example scenarios:
1/🧵 #stats
It's frustrating that so many development tools and platform instructions now start with "Step 1: Run a python script to launch a web server."
February 12, 2026 at 5:31 AM
It's frustrating that so many development tools and platform instructions now start with "Step 1: Run a python script to launch a web server."
Reposted by Steve Haroz
Were wealthy donors key to Trump’s campaigns in 2016 and 2020?
I'm thrilled to announce a new paper in which Sean Kates, Eric Manning, Tali Mendelberg and I analyzed data on 108 million (!) homeowner-voters.
See “Plutopopulism: Wealth and Trump’s Financial Base.” Open access: cup.org/4cfm0Az 🧵 1/
I'm thrilled to announce a new paper in which Sean Kates, Eric Manning, Tali Mendelberg and I analyzed data on 108 million (!) homeowner-voters.
See “Plutopopulism: Wealth and Trump’s Financial Base.” Open access: cup.org/4cfm0Az 🧵 1/
April 16, 2025 at 4:37 PM
Were wealthy donors key to Trump’s campaigns in 2016 and 2020?
I'm thrilled to announce a new paper in which Sean Kates, Eric Manning, Tali Mendelberg and I analyzed data on 108 million (!) homeowner-voters.
See “Plutopopulism: Wealth and Trump’s Financial Base.” Open access: cup.org/4cfm0Az 🧵 1/
I'm thrilled to announce a new paper in which Sean Kates, Eric Manning, Tali Mendelberg and I analyzed data on 108 million (!) homeowner-voters.
See “Plutopopulism: Wealth and Trump’s Financial Base.” Open access: cup.org/4cfm0Az 🧵 1/
Reposted by Steve Haroz
LLMs good performance on medical exams does not translate to accurate performance in real-world settings (preregistered n~1,300 study). This can't be explained by current standard benchmarks for medical knowledge & simulated patient interactions.
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Reliability of LLMs as medical assistants for the general public: a randomized preregistered study - Nature Medicine
In a randomized controlled study involving 1,298 participants from a general sample, performance of humans when assisted by a large language model (LLM) was sensibly inferior to that of the LLM alone ...
www.nature.com
February 11, 2026 at 10:30 AM
LLMs good performance on medical exams does not translate to accurate performance in real-world settings (preregistered n~1,300 study). This can't be explained by current standard benchmarks for medical knowledge & simulated patient interactions.
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Reposted by Steve Haroz
Anyone want to do a PhD with me at the Sunny Coast? I'm recruiting, and I wanna do some fun psychophysics (but the possibilities for the PhD are very broad). Domestic students only, sadly.
In case y'all happen to know someone:
@nataliepeluso.com
@reubenrideaux.bsky.social
@visnerd.bsky.social
In case y'all happen to know someone:
@nataliepeluso.com
@reubenrideaux.bsky.social
@visnerd.bsky.social
February 10, 2026 at 9:12 PM
Anyone want to do a PhD with me at the Sunny Coast? I'm recruiting, and I wanna do some fun psychophysics (but the possibilities for the PhD are very broad). Domestic students only, sadly.
In case y'all happen to know someone:
@nataliepeluso.com
@reubenrideaux.bsky.social
@visnerd.bsky.social
In case y'all happen to know someone:
@nataliepeluso.com
@reubenrideaux.bsky.social
@visnerd.bsky.social
Reposted by Steve Haroz
When I was putting my "deep dive" AI series on how LLMs impact human learning this was my very strong anecdotal impression -- just tons and tons of low-quality crap pushing the "AI is great" agenda, and a few high-quality things with negative or nuanced takes
February 10, 2026 at 9:15 PM
When I was putting my "deep dive" AI series on how LLMs impact human learning this was my very strong anecdotal impression -- just tons and tons of low-quality crap pushing the "AI is great" agenda, and a few high-quality things with negative or nuanced takes
Reposted by Steve Haroz
Hats off to the great Anne "I don't do fancy" Treisman!
I was curious about the fact that there were seemingly no women scientists mentioned, even of equivalent caliber, and came across what is, seemingly, beef between Epstein and Brockman debating whether "the women are all weak" or not
February 1, 2026 at 9:55 PM
Hats off to the great Anne "I don't do fancy" Treisman!
Reposted by Steve Haroz
💀
From the comments: "[T]o be mentioned TWICE as a thorn in the side of pedos. That’s a badge of honor I’d wear everyday.
RebeccaWatsonRules"
To be called "a rather nasty young woman," by Richard Dawkins is another badge of honor.
@skepchick.org, always right about these heinous creatures.
RebeccaWatsonRules"
To be called "a rather nasty young woman," by Richard Dawkins is another badge of honor.
@skepchick.org, always right about these heinous creatures.
From the SGU community on Reddit: In the new Epstein Files release there's a previously unpublished email form Richard Dawkins talking about Rebecca Watson
Explore this post and more from the SGU community
www.reddit.com
January 31, 2026 at 4:41 PM
💀
Reposted by Steve Haroz
I beg you please stop calling p-hacking exploratory analysis. Call it data dredging, fishing expedition, cherry picking, significance chasing. Labeling your non-pre-reg'd analyses clearly as exploratory isn't a license to p-hack. Stop bastardizing scientific exploration and stop hiding behind it.
May 23, 2025 at 4:56 PM
I beg you please stop calling p-hacking exploratory analysis. Call it data dredging, fishing expedition, cherry picking, significance chasing. Labeling your non-pre-reg'd analyses clearly as exploratory isn't a license to p-hack. Stop bastardizing scientific exploration and stop hiding behind it.
Friends don't let friends ask open-ended uninterpretable survey questions.
January 30, 2026 at 11:48 PM
Friends don't let friends ask open-ended uninterpretable survey questions.
Reposted by Steve Haroz
You are being misled about renewable energy technology.
YouTube video by Technology Connections
www.youtube.com
January 30, 2026 at 5:27 PM
I like that this scale doesn't have a neutral "Unsure" or "I don't know" option. Having a neutral options conflates confidence with the decision.
Amid ramped-up immigration enforcement efforts around the country, Americans’ opinions on the actions of ordinary people and federal immigration officers vary.
Read more about the different outlooks. 🧵⬇️
Read more about the different outlooks. 🧵⬇️
January 29, 2026 at 7:27 PM
I like that this scale doesn't have a neutral "Unsure" or "I don't know" option. Having a neutral options conflates confidence with the decision.
January 29, 2026 at 1:02 PM
I was thinking about how to improve the Scoville scale (a measure of spiciness) without relying on chromatography.
The scale works by taking a sample of food and adding sugar water until it doesn't taste spicy. That ratio of sugar water to food is the Scoville number.
The scale works by taking a sample of food and adding sugar water until it doesn't taste spicy. That ratio of sugar water to food is the Scoville number.
January 28, 2026 at 6:01 PM
I was thinking about how to improve the Scoville scale (a measure of spiciness) without relying on chromatography.
The scale works by taking a sample of food and adding sugar water until it doesn't taste spicy. That ratio of sugar water to food is the Scoville number.
The scale works by taking a sample of food and adding sugar water until it doesn't taste spicy. That ratio of sugar water to food is the Scoville number.
"[ML researchers] don’t consider it a priority to get the facts right"
Yup.
Yup.
"Machine learning research is not serious research & therefore hallucinated references are not necessarily a big deal, agrees prestigious group of ML researchers"
I ❤️ when the titles write themselves
But seriously, I don't think there's much to debate statmodeling.stat.columbia.edu/2026/01/26/m...
I ❤️ when the titles write themselves
But seriously, I don't think there's much to debate statmodeling.stat.columbia.edu/2026/01/26/m...
Machine learning research is not serious research and therefore hallucinated references are not necessarily a big deal, agrees a prestigious group of machine learning researchers | Statistical Modeli...
statmodeling.stat.columbia.edu
January 27, 2026 at 12:46 AM
"[ML researchers] don’t consider it a priority to get the facts right"
Yup.
Yup.
Reposted by Steve Haroz
How do people compute a sense of confidence? This question is usually addressed using very simple images because we don't know how complex stimuli are represented internally. In a new paper, we addressed this question using artificial neural networks (ANNs).
journals.plos.org/ploscompbiol...
journals.plos.org/ploscompbiol...
Using artificial neural networks to reveal the human confidence computation
Author summary Human decisions are accompanied by a sense of confidence which reflects the decision accuracy. Conventionally, human confidence has been studied using two-choice tasks with simple stimu...
journals.plos.org
January 26, 2026 at 7:18 PM
How do people compute a sense of confidence? This question is usually addressed using very simple images because we don't know how complex stimuli are represented internally. In a new paper, we addressed this question using artificial neural networks (ANNs).
journals.plos.org/ploscompbiol...
journals.plos.org/ploscompbiol...
Idea: A list of movies and series where an MRI can easily be turned on or off without a massive cloud of helium.
The magnet is ALWAYS on.
The magnet is ALWAYS on.
January 26, 2026 at 6:17 AM
Idea: A list of movies and series where an MRI can easily be turned on or off without a massive cloud of helium.
The magnet is ALWAYS on.
The magnet is ALWAYS on.
Reposted by Steve Haroz
We present a typology of traps to avoid:
1. Believing that AI systems are minds
2. Believing that AI systems are theories
3. Believing that cognitive science can be automated.
Learn to recognise and avoid these traps. Failure to avoid leads to numerous problems.
3/🧵
1. Believing that AI systems are minds
2. Believing that AI systems are theories
3. Believing that cognitive science can be automated.
Learn to recognise and avoid these traps. Failure to avoid leads to numerous problems.
3/🧵
January 6, 2026 at 5:54 PM
We present a typology of traps to avoid:
1. Believing that AI systems are minds
2. Believing that AI systems are theories
3. Believing that cognitive science can be automated.
Learn to recognise and avoid these traps. Failure to avoid leads to numerous problems.
3/🧵
1. Believing that AI systems are minds
2. Believing that AI systems are theories
3. Believing that cognitive science can be automated.
Learn to recognise and avoid these traps. Failure to avoid leads to numerous problems.
3/🧵
Reposted by Steve Haroz
beyond words to express how much i hate this... looks like "AI summary" not the author's written abstract is displayed by default for all ACM digital library publications. i, for one, didn't ask for it and hate it to my core
January 23, 2026 at 2:50 PM
beyond words to express how much i hate this... looks like "AI summary" not the author's written abstract is displayed by default for all ACM digital library publications. i, for one, didn't ask for it and hate it to my core
Reposted by Steve Haroz
Possums in your yard? News got you down? Fear not, I have made playlists for the A and B sections of my ongoing Statistical Rethinking course. Click the section of your choice, sit back, and forget the possums and decay of the international order while your brain updates. github.com/rmcelreath/s...
January 24, 2026 at 7:09 AM
Possums in your yard? News got you down? Fear not, I have made playlists for the A and B sections of my ongoing Statistical Rethinking course. Click the section of your choice, sit back, and forget the possums and decay of the international order while your brain updates. github.com/rmcelreath/s...
Reposted by Steve Haroz
In 1916, The United States bought the Little Saint James island from Denmark. It is now known as Epstein Island.
I don't know what to do with that information, so I'm posting it.
en.wikipedia.org/wiki/Little_...
I don't know what to do with that information, so I'm posting it.
en.wikipedia.org/wiki/Little_...
Little Saint James - Wikipedia
en.wikipedia.org
January 20, 2026 at 11:03 PM
In 1916, The United States bought the Little Saint James island from Denmark. It is now known as Epstein Island.
I don't know what to do with that information, so I'm posting it.
en.wikipedia.org/wiki/Little_...
I don't know what to do with that information, so I'm posting it.
en.wikipedia.org/wiki/Little_...
Reposted by Steve Haroz
ICE’s actions tear families apart, intimidate communities, & undermine the conditions that make education, research, healthcare, & public trust possible.
We’re calling on scientists to join the general anti-ICE strike in Minnesota on January 23rd.
We’re calling on scientists to join the general anti-ICE strike in Minnesota on January 23rd.
January 19, 2026 at 10:22 PM
ICE’s actions tear families apart, intimidate communities, & undermine the conditions that make education, research, healthcare, & public trust possible.
We’re calling on scientists to join the general anti-ICE strike in Minnesota on January 23rd.
We’re calling on scientists to join the general anti-ICE strike in Minnesota on January 23rd.
In the late 90s, many companies did nothing but chase the latest tech fad.
What does your company do?
We have a website!
What does the website do?
It uses the latest tech!
But what do you do with that?
We have a website!
The result was the dotcom bust.
All these AI startups look the same to me.
What does your company do?
We have a website!
What does the website do?
It uses the latest tech!
But what do you do with that?
We have a website!
The result was the dotcom bust.
All these AI startups look the same to me.
January 19, 2026 at 4:22 AM
In the late 90s, many companies did nothing but chase the latest tech fad.
What does your company do?
We have a website!
What does the website do?
It uses the latest tech!
But what do you do with that?
We have a website!
The result was the dotcom bust.
All these AI startups look the same to me.
What does your company do?
We have a website!
What does the website do?
It uses the latest tech!
But what do you do with that?
We have a website!
The result was the dotcom bust.
All these AI startups look the same to me.
Could be useful for journalists
There are many terms to refer to a person’s sex: sex, birth sex, natal sex — and more contentiously, assigned sex at birth and biological sex.
Here's our expanded guidance for journalists interested in reporting accurately about trans and intersex people:
www.transjournalists.org/why-biologic...
Here's our expanded guidance for journalists interested in reporting accurately about trans and intersex people:
www.transjournalists.org/why-biologic...
Why “biological sex” may not mean what you think
There are a lot of terms to refer to a person’s sex: sex, birth sex, natal sex — and more contentiously, assigned sex at birth and biological sex.
The nuances and differences aren’t always important....
www.transjournalists.org
January 16, 2026 at 7:30 PM
Could be useful for journalists