Current NIH leadership want you to think they are using rigorous, consistent & scientific processes to screen studies to align them with agency priorities.
But the process that they have put down on paper is a sham.
It’s important to know NIH is not following its own guidance. Here’s why:
🧵1/
i second oli's question - which movie?
Manuscript peer review is not
1) journal club
2) dissertation committee meeting
3) zero sum grant award
4) mortal kombat
When bullshiters tell you they will soon be able to make "digital twins" of patients, remember articles like this about how hard it is to model a single cell.
Here is a link to a video of the speech.
xcancel.com/netflix/stat...
Today should be your 38th birthday.. When you lose your son you lose more than a child you lose a piece of your heart and your joy. Life is forever altered, and nothing can ever fill that space. Yet in my soul, he remains, my precious boy, forever loved. His laughter, his smile, and his love are woven into who I am. My love for him will never fade. Yor are my pride and my joy - you've made a change that cannot be broken..
Alex's mom on his 38th birthday.
We’re hiring!🔬The MAClab at Texas A&M is looking for a lab manager to join our team in College Station. If you’re organized, passionate about research, and interested in emotion and the brain, we want to hear from you!
🔗Apply at the link in our bio!
#PsychSky #PsychSciSky #neuroskyence #AcademicSky
That's right. Worse than Pointless
www.nytimes.com/2026/02/23/o...
Technological and logistical tour de force, *but* what is the scientific or practical significance of showing that you can reliably classify/decode spontaneous pain in n=1, given hours of scanner data
While we sit around to find out how bad exactly it all is, a bit of cheer: the best baguette of Paris competition was won by a Sri Lankan who appears to have started as a dish washer in restaurants and eventually trained as a Boulanger.
www.paris.fr/pages/la-mei...
“Dolly Parton Children’s Hospital did not share how much Parton donated as part of the naming announcement. But Matt Schaefer, its president and CEO, said her support would ensure ‘every child who walks through our doors receives the treatment they deserve.’”
GOD BLESS DOLLY PARTON
Another day of repeating the mantra, there are no banned words, just words that will prevent an otherwise fundable application from being funded…
blood boiling
Jasp is pretty good for this
origins of MADRS-SR
Back in business! @foxdotdrew.bsky.social @sgrogans.bsky.social (note hard to see menu on the far left). wanna plot data OLS vs Spearman/Rank vs Robust/Tukey? Dynamically censor points (for thinking about leverage, not p-hacking)? Try Scatterize 3!
Slightly apropos of this: In honor of Drew Fox's birthday, this weekend I (finally!) resurrected scatterize as a browser-only setup, in the hope that it may live as HTML and javascript. It probably still has a few rough edges, but give it a try!
njvack.github.io/scatterize/
give em hell 75 !
Kurt Vonnegut stop being so applicable to all time periods of American life, you can’t do that Kurt Vonnegut, your insights are too evergreen Kurt Vonnegut
now walking the poor dog!
6/n. PPS: An unexpected bonus: 13-y.o. asks what I'm doing and why I'm grinning so much. Lower right panel by Hannah Shackman. Neg exp framing: 'weed smoking (NumTimes) among 8th graders'
5/n. PS: One of the best parts of spending 10! years working on my PhD is that I regularly had the pleasure and the luxury of engaging in these sorts of stats rabbit holes and sharing the fruits of my labor w/ others. (Screen cap is archived Wisco stats page circa 2012)
4/n. sort, and which is routinely applied to similar real-world data (teen drinking, insurance claims, hospital stay length, etc.)
3/n. do a little Sat morning geek-out detective work...when I should be walking the dog barking downstairs. I think what we actually want is a Tweedie regression, which is fun to say aloud, which is designed to capture mixed distributions of this
2/n. We ended on an equally inelegant 2-stage "hurdle" approach: logistic regression of 0-vs-1+ . If signif, then followed by OLS linear on the GT 0 values (censor 0s). Blech. Skimming the primer paper, which features several 0-inflated distributions led me to
