We are now recruiting participants for a series of upcoming focus groups that will discuss the impacts of AI on health and well-being. If you or any friends/colleagues are interested, please visit the study site:
umhealthresearch.org/study-team/s...
All focus groups will be 100% online (via Zoom).
This is how my sister raises her kids and while it’s obviously not causally identified, her kids are way more well adjusted and mature than I was at their age.
Live and lapply()
Ah, yes this is my problem in that I am almost always making more than two eggs. Maybe I just need more pans! 😂
Ok what is your secret to getting such a perfect fried egg?
How was I entirely unaware of this package?!?
The sapply() who loved me
You know what I'm gonna give this one to her. 10/10
To each their own I suppose.
Watching S1E1 of The Pitt. It’s gotta get better right? Is this just the usual crappy pilot after which they bring in a showrunner that knows what they’re doing?
One of those legendary pieces of online statistical lore that everyone needs be aware of.
You might check out this paper by Cernat and @daob.nl that seems along the lines of what you’re thinking. Not substantively but as a method for doing it.
With very large numbers of n’s you don’t need randomization, and with LLM’s we can generate very large numbers of n’s, so I think all of science is solved by now. I don’t see any problems with this.
Nobody can post unless they’re connected over dialup at speeds not to exceed 14.4 kbps.
Can. Not. Wait.
If they don’t know it THAT’S 50 DKP MINUS!!!
Woke up and chose violence, eh?
Would pay good money for this.
This chapter from a volume edited by Cernat might be of interest.
academic.oup.com/book/39454/c...
This is a good question for @daob.nl. This is very much in his area of expertise. Alexandru Cernat also but I don’t think he’s on here.
I see. That makes sense. As @cameronpat.bsky.social said elsewhere, I guess it depends on the estimand. Correlated measurement error is a giant morass.
E.g. if people who got the treatment remember their pre status as better than it really was, modeling change as a multivariate outcome seems like it would have the same bias as if your model was post ~ pre + treatment.
It seems like this would only be necessary if the treatment would plausibly affect the way the pre condition is reported. Although in that case, an estimand measuring change since baseline seems of limited utility (although maybe it does in context).
“average monkey asks 3 questions a day" factoid actualy just statistical error. average monkey asks 0 questions per year. Curious georg, who lives in cave & asks over 10,000 each day, is an outlier adn should not have been counted
lol, I wouldn’t know!
I’m probably going to buy it, play it for about an hour, get freaked out by something creepy and never touch again. This is my standard MO for Resident Evil games.
Always
Yeah, I really can't stress enough that AI code assistants royally fuck up statistical analyses. And they do it with absolute confidence.
data.table - for when there aren’t enough brackets in your life.
I have wanted to attend since I first heard about it!
