Not “50 grand I get this on one take”
He said: 🤷🏼
Happening now: anti-war protestors are gathereing in Chicago's Congress Plaza for a demonstration against ongoing U.S. and Israeli attacks on Iran.
Tonight's rally follows a similar protest in Chicago on Saturday; other anti-war actions are being planned across the city over the week.
What good can you do today? What trouble can you cause? What suffering can you ease? What needs to exist, and can you help build it?
We’ve kidnapped one head of state and assassinated another, and then we wonder why …
People are out grilling in Chicago … I tell you what …
These tech companies know everything about me … except my cookie preferences 🤔
Again and again …
Just catching the tribute to D’Angelo and Roberta Flack … phew
Lauryn Hill from NJ, btw …
And Bilal is a bad man …
So maybe I should be asking, if theory predicts b1>b2, what test *should* I be running?
Oh, I see what you mean, standardizing the regressors. But after that, does that help me with testing these inequalities?
In other words what does it mean to fail to reject the null in a positive one-sided test when the coefficient is so negative it would have failed a two-sided test?
Embarrassing follow-up question. Philosophically, a one-sided test is supposed to be for when I “know” the sign of the coefficient. What if theory says what the sign should be, but I don’t “know” if the theory is correct, am I still supposed to do a one-sided test?
In my case, I have inequalities between the betas.
I guess I could write a regression that estimates b1, b2-b1, and b3-b2, my null is then b1-b2=0 and b3-b2 =0, and then what? Impose b1-b2=b2-b3 and test that in a one sided test?
No, do mean divide by SD?
Q: There’s a regression with three coefficients: b1, b2, b3
A theory says b1 > b2 > b3
What statistical test should I perform to test this joint prediction?
The point estimates are such that b3 > b1 > b2, if that matters
Q: There’s a regression with three coefficients: b1, b2, b3
A theory says b1 > b2 > b3
What statistical test should I perform to test this joint prediction?
The point estimates are such that b3 > b1 > b2, if that matters
ICE also shot a woman here in Chicago, whom they made similar claims against. She didn't die, despite being shot multiple times, and the charges against her were eventually dropped, because the stories they tell to justify their violence are always bullshit and never hold up under scrutiny.
The Yo! MTV Raps channel is cool …
This Kansas game is wild …
Hmm, but WE both seem to agree that Walter was trash for that … sooo, then we can move on to analyzing Carol’s latest choices, no?
Two words Carol Sturka: con-sent … 😬
Ok, 5 minutes later, Carol is digging an even deeper hole … she is lost in the sauce … “you are mine?”
I’ve been defending Carol all season … halfway through the finale and I’m ’bout done … maybe humanity won’t make it, but her behavior is unacceptable …
🤲🏾
🧵1/3 J-PAL North America is thrilled to announce our 2025 #EvidenceChampion awardees! Both @nomadj1s.bsky.social and Dr. Owen Garrick have dedicated their careers to advancing racial equity and evidence-based policymaking. Learn more 👇
There’s a logic to not inviting them, but it has backfired … hence a misstep …
The first time 11M people die, she didn’t know it was going to happen, the second time? Probably want a do over in that spot …
