Stephen Wild
@stephenjwild.bsky.social
I try to put straight lines through things but usually fail. Try to be Bayesian when I can. Views my own. RT/like != endorsement.
Social media bad
A good friend had a baby a couple of years ago so I briefly signed up for fb with a burner email, told it I was a middle aged dude and got served a torrent of utterly hateful content, much of it targeting Trans people.
the minute the fb algorithm clocks you as a straight man it immediately sets about trying to destroy your life & endanger those around you
February 13, 2026 at 11:08 PM
Social media bad
Always retaliate first
The only thing necessary for the triumph of evil is for good players to not play tit-for-tat
February 13, 2026 at 10:48 PM
Always retaliate first
Reposted by Stephen Wild
Fair. I'm never quite sure how to take these conversations as I'm genuinely not sure (and not everyone is on the same page about) what journal articles are for. Nail that down and a Kasy-style decision theory exercise will sort out the rest, fine, but until then...
February 13, 2026 at 4:40 PM
Fair. I'm never quite sure how to take these conversations as I'm genuinely not sure (and not everyone is on the same page about) what journal articles are for. Nail that down and a Kasy-style decision theory exercise will sort out the rest, fine, but until then...
This a great case study in statistics, economics, and business. Why is their match rate is so high now, and can they maintain it as they expand?
"Our matches convert to actual dates at about 10x the rate of Tinder," Weng told TechCrunch.
A Stanford grad student created an algorithm to help his classmates find love; now, Date Drop is the basis of his new startup | TechCrunch
"Our matches convert to actual dates at about 10x the rate of Tinder," Weng told TechCrunch.
techcrunch.com
February 13, 2026 at 4:40 PM
This a great case study in statistics, economics, and business. Why is their match rate is so high now, and can they maintain it as they expand?
Hehehe
"Proper understanding of the answers of the above questions should in most cases make you at best ambivalent about DiD. If you still think you have a DiD problem, expect me to try to help you figure out what else you should do"
"Proper understanding of the answers of the above questions should in most cases make you at best ambivalent about DiD. If you still think you have a DiD problem, expect me to try to help you figure out what else you should do"
For Spring semester, I'm bringing back free weekly open office hours for anyone in the world with Econometrics questions.
Tuesdays 4-6 PM Central European Time (US EST 10AM-12PM) or by appointment; sign up and drop by!
Details and signup at: donskerclass.github.io/OfficeHours....
Tuesdays 4-6 PM Central European Time (US EST 10AM-12PM) or by appointment; sign up and drop by!
Details and signup at: donskerclass.github.io/OfficeHours....
February 13, 2026 at 1:10 PM
Hehehe
"Proper understanding of the answers of the above questions should in most cases make you at best ambivalent about DiD. If you still think you have a DiD problem, expect me to try to help you figure out what else you should do"
"Proper understanding of the answers of the above questions should in most cases make you at best ambivalent about DiD. If you still think you have a DiD problem, expect me to try to help you figure out what else you should do"
Reposted by Stephen Wild
Stats enthusiasts: I like to weave narratives using insights cross-tabbed from the latest poll
Stats experts: I only estimate conditional expectations and I keep a gun nearby so I can shoot them if they're from a distribution I don't recognize
Stats experts: I only estimate conditional expectations and I keep a gun nearby so I can shoot them if they're from a distribution I don't recognize
February 12, 2026 at 10:20 PM
Stats enthusiasts: I like to weave narratives using insights cross-tabbed from the latest poll
Stats experts: I only estimate conditional expectations and I keep a gun nearby so I can shoot them if they're from a distribution I don't recognize
Stats experts: I only estimate conditional expectations and I keep a gun nearby so I can shoot them if they're from a distribution I don't recognize
Stan bat signal
February 12, 2026 at 9:55 PM
Stan bat signal
Me, a genius: Q is at its profit-maximizing level when marginal revenue equals marginal cost
nothinghuman.substack.com/p/the-tyrann...
Being popular causes all kinds of bad context collapse phenomena
Being popular causes all kinds of bad context collapse phenomena
The Tyranny of the Marginal User
why consumer software gets worse, not better, over time
nothinghuman.substack.com
February 12, 2026 at 8:56 PM
Me, a genius: Q is at its profit-maximizing level when marginal revenue equals marginal cost
Having thought carefully* about this for 24 hours, my 2 cents as a non-academic about the discussion around this point.
1) there is a huge blurring of the is-ought distinction
2) the system is what the system does
*not really. Also I am thinking out loud right now
1/N
1) there is a huge blurring of the is-ought distinction
2) the system is what the system does
*not really. Also I am thinking out loud right now
1/N
Overall, I think our results are:
1. Believable. I think deep down we all know we publish "too many" significant results.
2. Disastrous. Selecting on significance in this way biases results across our whole literature away from zero & it stops us from learning what doesn't work.
1. Believable. I think deep down we all know we publish "too many" significant results.
2. Disastrous. Selecting on significance in this way biases results across our whole literature away from zero & it stops us from learning what doesn't work.
February 12, 2026 at 2:45 PM
Having thought carefully* about this for 24 hours, my 2 cents as a non-academic about the discussion around this point.
1) there is a huge blurring of the is-ought distinction
2) the system is what the system does
*not really. Also I am thinking out loud right now
1/N
1) there is a huge blurring of the is-ought distinction
2) the system is what the system does
*not really. Also I am thinking out loud right now
1/N
The Schitt's Creek gifs will continue until morale improves
February 12, 2026 at 11:42 AM
The Schitt's Creek gifs will continue until morale improves
*Wearing my Übergangsjacke*
9°C, misery. 10.5°C, bliss.
9°C, misery. 10.5°C, bliss.
look I don’t make the rules but Übergangsjacke has a clearly defined narrow temperature range.
February 11, 2026 at 3:29 PM
*Wearing my Übergangsjacke*
9°C, misery. 10.5°C, bliss.
9°C, misery. 10.5°C, bliss.
Reposted by Stephen Wild
That seems fine, but the Cox PH model can also be written instead in terms of the cumulative hazard, H(t), which doesn't condition on survival till time t anymore, so (1) doesn't hold and the rest of the argument falls apart
February 11, 2026 at 2:39 PM
That seems fine, but the Cox PH model can also be written instead in terms of the cumulative hazard, H(t), which doesn't condition on survival till time t anymore, so (1) doesn't hold and the rest of the argument falls apart
Two things (among many) that seem to pop out from this thread, the replies, and the QTs:
1) people do a lot out of spite
2) people who "aren't good at math" can learn math when either math becomes interesting to them or math is grounded in something that interests them.
No lessons here, I'm sure.
1) people do a lot out of spite
2) people who "aren't good at math" can learn math when either math becomes interesting to them or math is grounded in something that interests them.
No lessons here, I'm sure.
Funny story: my career in drones exists because I have a arithmetic learning disability, tested awfully in math and couldn’t study STEM in college, and became so angered by how normal it was to look down upon non-tech people in the 2010s that I decided to master a hard tech field out of pure spite.
February 11, 2026 at 2:39 PM
Two things (among many) that seem to pop out from this thread, the replies, and the QTs:
1) people do a lot out of spite
2) people who "aren't good at math" can learn math when either math becomes interesting to them or math is grounded in something that interests them.
No lessons here, I'm sure.
1) people do a lot out of spite
2) people who "aren't good at math" can learn math when either math becomes interesting to them or math is grounded in something that interests them.
No lessons here, I'm sure.
Always the Nurses' Health Study
February 11, 2026 at 11:42 AM
Always the Nurses' Health Study
THANK YOU FOR YOUR ATTENTION TO THIS MATTER
This means smartphones have probably expanded the sample of people who respond to surveys, even if they harm knowledge-based measures.
February 10, 2026 at 9:56 PM
THANK YOU FOR YOUR ATTENTION TO THIS MATTER
Reposted by Stephen Wild
I have plenty of reservations about the reliability of funnel-plots and z-curves and such, as well as their interpretation.....
But holy shit look at that.
But holy shit look at that.
February 10, 2026 at 3:17 PM
I have plenty of reservations about the reliability of funnel-plots and z-curves and such, as well as their interpretation.....
But holy shit look at that.
But holy shit look at that.
10/10 joke
The latest from YouGov (n ~ 7k) gives me some methodological concerns. They allowed people to rate their pun making as "outstanding" but, statistically, most are sitting indoors at the time
today.yougov.com/topics/socie...
today.yougov.com/topics/socie...
February 10, 2026 at 3:15 PM
10/10 joke
February 10, 2026 at 2:56 PM
Genuine question for people follow this sort of thing more closely than I do: why is NCI looking st Ivermectin as s cancer treatment?
(I know there are snarky and cynical answers, but I am not after those)
(I know there are snarky and cynical answers, but I am not after those)
The National Cancer Institute will begin studying ivermectin's ability to kill cancer cells, @rachanadpradhan.bsky.social reports
www.statnews.com/2026/02/10/i...
www.statnews.com/2026/02/10/i...
National Cancer Institute studying ivermectin’s ‘ability to kill cancer cells,’ alarming career scientists
The National Cancer Institute is studying ivermectin as a potential cancer treatment, according to its top official, alarming career scientists.
www.statnews.com
February 10, 2026 at 2:33 PM
Genuine question for people follow this sort of thing more closely than I do: why is NCI looking st Ivermectin as s cancer treatment?
(I know there are snarky and cynical answers, but I am not after those)
(I know there are snarky and cynical answers, but I am not after those)
*whistles*
Good paper.
Good paper.
Has financial literacy been declining or are smartphones changing the way we answer surveys? I was on Marketplace yesterday to talk about how smartphones are making it more challenging to gauge knowledge on surveys. Longer thread coming soon!
Check out our full study here:
lnkd.in/guiqHmKW
Check out our full study here:
lnkd.in/guiqHmKW
February 10, 2026 at 1:14 PM
*whistles*
Good paper.
Good paper.
Incredible.
February 9, 2026 at 11:59 PM
Incredible.
Still funny seeing someone discover Gladwell for the first time
This leads into one of my chief annoyances with this book, there has been a lot, and I mean a lot of ink spilled on strategic bombing in WWII, and the amount of scholarship that Gladwell either intentionally or unintentionally leaves out is borderline negligence
Just flipping through the endnotes, how could possibly write about American strategic bombing in WWII and not include any work by Richard overy, Michael Sherrill, literally any study of the air tactical school, or the official AAF histories lol?
February 9, 2026 at 7:29 PM
Still funny seeing someone discover Gladwell for the first time
The moralization of food and cooking is *fascinating*
tl;dr - it's the nutrient content, not the processing
For a healthier diet:
- less meat, less sugar, less salt
- more veggies, more legumes
Also makes the under-appreciated point that cooking from scratch requires more labour (typically female & unpaid) that isn't an option for many households.
For a healthier diet:
- less meat, less sugar, less salt
- more veggies, more legumes
Also makes the under-appreciated point that cooking from scratch requires more labour (typically female & unpaid) that isn't an option for many households.
ICYMI @gnrosenberg.bsky.social and I wrote about why you should not want to eat like your great-great-grandparents and why modern, industrially-produced food is "real food."
www.nytimes.com/2026/02/08/o...
www.nytimes.com/2026/02/08/o...
February 9, 2026 at 4:22 PM
The moralization of food and cooking is *fascinating*
Reposted by Stephen Wild
My social media echo chamber is dozens of science, academic, and tech people agreeing that the Bad Bunny halftime show was a-maze-ing
February 9, 2026 at 2:07 AM
My social media echo chamber is dozens of science, academic, and tech people agreeing that the Bad Bunny halftime show was a-maze-ing