It’s still Australian tho and can eat a sealed wooden deck bc fuck you that’s why.

Cheer up, BMJ. You can’t be responsible for the ludicrous overconfidence and innumeracy of some authors.

Anyway. Trust no one.

A big practical point I don’t see reflected elsewhere is: the sample size of the resultant groups is super important. Smaller potential groups are *noticeably* easier to threshold hack with. Surprised how much it helped.

It is quite easy to simulate this as a general case and only slightly trickier to do it on a real dataset.

I think it took about twenty minutes with NHANES data until we had one that was not only a good spurious result but also funny.

How do we know they’re playing silly buggers?

1) the cutoffs change over time / author
2) the cutoffs aren’t just binary delineations they get creative (e.g. the effect of an artificial A/B on the top two quintiles of C!)
3) physiology can go screw - we got numbers, we don’t need no stinkin bodies

Remarkable Debacle is a top tier indie band name.

ooh please leave seventeen crotchety reasons why nothing can ever be done anywhere ever

i can’t wait to hear what’s impossible this week

“As the physicians are extremely busy, lacking sufficient experimental facilities and time to complete the entire project independently, foreign professors will be responsible for experimental execution and paper.”

A genius wrinkle on the regular mill flavor. The hot fudge sundae to mill vanilla.

Who is this shifty vagrant.

Not sure what you mean.

if you weren't supposed to invest in them then why are they named financiers