Thread by @tedfrank

Saw this in @SCOTUSblog. Wait, what? Affirmative action doctors *double* the chances of black infant survival?!?

Sure enough, that’s what the brief says. Let’s look at the study in the footnote. https://t.co/O7YjighTyysupremecourt.gov/DocketPDF/20/2…

The lawyers are lying (or, more likely, mathematically illiterate). The study says no such thing.

The study purports to find black-doctor/black-infant have survival rate of 99.8% vs. 99.6% for white-doctor/black-infant.

pnas.org/doi/10.1073/pn…

So right away that’s not a doubling of anything.

But the ceteris isn’t paribus. The white docs aren’t seeing the same infants as the black docs. They’re more likely to get the NICU cases where all infants are less likely to survive, and study doesn’t control for that.

So the study is confusing correlation with causation: if you have a black doctor, your baby is more likely to survive, but that’s because that means you’re less likely to be in the NICU, where there are fewer black doctors. It has nothing to do with the race of the doctor.

Anyway, anyone want to place a bet whether the game of telephone works and takes a bad legal writeup of a bad study and the entirely fictional (but striking!) claim in the brief ends up in a SCOTUS opinion?

I read the wrong chart, in part because the study’s meaningful data is in an appendix. The difference is 99.96% vs 99.91%. And the difference isn’t even statistically significant.

https://twitter.com/tedfrank/status/1586786826219827200

Haven’t listened to this @VPrasadMDMPH podcast yet. Public policy community refuted it contemporaneously when the study came out. Knew the study would be pushed to SCOTUS (I’m sure it’s mentioned in other briefs); just didn’t think the dishonesty of the study would be multiplied.

https://twitter.com/SudarRajagopal/status/1586908090431053824