Nothing So Practical
@nsp-blog.bsky.social
We write about a range of topics from a behavioral science perspective, mostly related to quantitative methods. This includes causal inference, Bayesian statistics, study design, and measurement.
The fact is we have been overproducing PhDs for decades while the number of tenure track positions has failed to keep pace. I think we can afford to award fewer doctorates for a while. www.nytimes.com/2026/07/06/u...
Decline of Ph.D. Admissions Could Imperil a ‘Generation of New Talent’
nytimes.com
I'd love to believe this but it looks like nothing but a selection effect. Real effects are never this large. Also, what is a plausible biological pathway that could produce such "broad protections"? www.washingtonpost.com/health/2026/...
Covid vaccine linked to broad protections against heart conditions, study finds
The vaccine reduced cardiovascular conditions, hospitalizations and deaths of all causes, including those not linked to covid, researchers said.
washingtonpost.com
LET THIS SINK IN—the U.S. now has no confirmed CDC director, no FDA commissioner, no Surgeon General, an NIH director with no medical license, and the last Acting CDC Director was a George Bush speechwriter. In fact—the only doctor/scientist we have in the White House is… Dr Oz.
It's amazing what branding does. A neural network is just billions of logistic regressions. RLHF is just GANs. Perplexity is just entropy. Simulated language is AGI. smh
Stephen Wild@stephenjwild.bsky.social · 5mo agoAI is just billions of logistic regressions in a trench coat, but saying so means being a party pooper
New paper— everyone is collecting intensive longitudinal data but recent review studies report that less than half of studies consider measurement/psychometrics. The paper covers a few foundational psychometric methods for ILD and provides a shiny app to apply them link.springer.com/article/10.3...
A primer on intensive longitudinal psychometrics - Behavior Research Methods
Many intensive longitudinal studies are interested in topics that are not always amenable to direct physical measurement and instead are often theorized as latent constructs (e.g., affect, emotion, mo...
link.springer.com
Just gave a seminar on causal inference in which students had to draw graphs which involved the observation that nausea during pregnancy seems to be a positive sign of a healthy pregnancy. Afterwards a student thanked me for the seminar and wished me lots of nausea going forward 😭😭😭
A peer reviewed study shows a promising link between flu and shingles vaccines and a reduced risk of dementia. A heroin addict who cut off a raccoon’s penis for “further study” says these vaccines cause autism. For busy Americans, it can be hard to know who to trust.
Just want to be clear that when RFK Jr. said in the hearing today that he is “overhauling” the U.S. Preventative Services Task Force, it is because this administration wants to reduce the number of services that health insurance companies are required to cover.
I would have preferred to have the "draw the rest of the owl" meme on the cover, but this will do. Seems like it is on schedule, and we'll leave some typos so you know we didn't write it with AI.
The category error we all make about the Monty Hall problem is that we think it is a probability problem, when in fact it was created in a lab to make us all feel stupid.
Kevin M. King@kevinmking.bsky.social · 6mo agoQuick reminder that the Monte Hall problem is a lot easier to understand if you think of a variation with 100 doors instead of 3. Here you're presented with 100 doors, you choose 1, and the host opens 98 others. Should you switch?
Design > Model, always. Another good one from Frank Harrell www.fharrell.com/post/causal/
Causal by Design – Statistical Thinking
Well-controlled randomized experiments when analyzed in a randomization-respecting way are causal by design and need no causal calculus to infer causation. If an observational study has any hope of pr...
fharrell.com
"But the mystery of replication still seems too deep; the predictions from A.I. aren’t entirely random, but they’re far from perfect, the SCORE team found." www.nytimes.com/2026/04/01/s...
Can Science Predict When a Study Won’t Hold Up?
nytimes.com
This is my favorite climate change chart. Japanese monks, aristocrats, and emperors kept meticulous records of cherry blossom festivals for 1,200 years and accidentally built the world's longest climate dataset.
A few people have replied to my Artemis posts with links to Gil Scott-Heron's "Whitey on the Moon." Here's a story from Space.com about Artemis II pilot Victor Glover, currently en route to fly around the Moon, and his connection to that piece. www.space.com/space-explor...
NASA's Artemis 2 pilot Victor Glover listens to 'Whitey on the Moon' every Monday. This is why.
"It ain't about racism. It's about the human condition."
space.com
View the presentation recording on YouTube: youtu.be/Q8A5NEmRxyM?...
2026 Methods Series with Ellie Murray
YouTube video by CAUSALab at Harvard T.H. Chan
youtu.be
Is CUPED just regression adjustment in disguise? You be the judge: nothing-so-practical.com/post/cuped-v... #ABtest #experiment #CausalSky #Causalinference
CUPED: Old Wine in a New Bottle?
CUPED has become a popular feature in online experimentation platforms. Regression works fine too.
nothing-so-practical.com
The best way to prevent being blocked by Daniel Lakens. IYKYK
Statistical Rethinking 2026 is done: 20 new lectures emphasizing logical and critical statistical workflow, from basics of probability theory to causal inference to reliable computation to sensitivity. It's all free, made just for you. Lecture list and links: github.com/rmcelreath/s...
DAGs and potential outcomes are useful frameworks but they miss some things that are helpful to understand in practice. #CausalInference is Easy: nothing-so-practical.com/post/causal-... #cause #causal #PotentialOutcomes #DAG
Causal Inference Is Easy
An accessible approach to thinking about causal inference
nothing-so-practical.com
Is CUPED just regression adjustment in disguise? You be the judge: nothing-so-practical.com/post/cuped-v... #ABtest #CUPED #experiment
CUPED: Old Wine in a New Bottle?
CUPED has become a popular feature in online experimentation platforms. Regression works fine too.
nothing-so-practical.com
