Max Rubinstein

[Max Rubinstein @max_rubinstein]

@paulgp Something related to prob causation, which is P(Y0=0| Y1=1)

[Max Rubinstein @max_rubinstein]

@apaullo95 Jfc glad you're still here!!

[Max Rubinstein @max_rubinstein]

@jennifer__holz Hmm ok well I'd be interested to hear more about that sometime on a non online forum lol

[Max Rubinstein @max_rubinstein]

@seancollierpgh What was the show?!

[Max Rubinstein @max_rubinstein]

almost engaged in an lpm v logit debate. but then i stopped and asked myself, is this really what the world needs right now?

[Max Rubinstein @max_rubinstein]

@jennifer__holz i could keep up with that pace for like maybe 1/4 of a mile 10 years ago

[Max Rubinstein @max_rubinstein]

@jennifer__holz tell that to suburban wine parents

[Max Rubinstein @max_rubinstein]

google history: "taj mahal"; "taj mahal, mcknight road"

[Max Rubinstein @max_rubinstein]

@nileshshah196 any idea why that looks so different from this? https://t.co/GDd1yACRRs

[Max Rubinstein @max_rubinstein]

@apaullo95 semi-relatedly, annoying that improv seems to get characterized in ppl's minds by everyone's worst memory from a bad show they saw, while stand-up by everyone's best. and yet most improv is way less cringe than most stand up

[Max Rubinstein @max_rubinstein]

@apaullo95 (gotta look up what game changer is tho)

[Max Rubinstein @max_rubinstein]

@jennifer__holz Is that what he's looking at here??

[Max Rubinstein @max_rubinstein]

@jennifer__holz But who's seeing them?

[Max Rubinstein @max_rubinstein]

@carlislerainey If the hypotheses are theory driven, so that the final tests are meant to avoid finding evidence contradicting the theory, this is bad. Re inference tho, as long as the test sample is independent from the pilot study the CIs on the hypothesis tests should be fine

[Max Rubinstein @max_rubinstein]

@EricNMoody Role models

[Max Rubinstein @max_rubinstein]

@tcarpenter216 in addition to what ppl have said below, sometimes it's reasonable to think the treatment assignment process is easier to model than the outcome

[Max Rubinstein @max_rubinstein]

@instrumenthull Take-up rates?

[Max Rubinstein @max_rubinstein]

@cosmicfibretion very interested to know what any of this means!

[Max Rubinstein @max_rubinstein]

@thedumbestrobot @jennifer__holz Sometimes it takes me years. Eg I never disliked det crashmore, but I have consistently liked that one more and more w time