Good fact-check -- in the Washington Post! -- about that paper on Puerto Rico hurricane deaths. Rev was right again! https://www.washingtonpost.com/news/fact-checker/wp/2018/06/02/did-4645-people-die-in-hurricane-maria-nope/ …
-
Show this thread
-
Replying to @jaylateofthepnw
He's directly quoting from the paper! To be fair, the definition of CI is one of those things I try to get into my head every couple of years, and it falls right back out! Also true of some people much smarter than me. (cc:
@Meaningness)2 replies 0 retweets 2 likes -
Replying to @St_Rev @Meaningness
I'm statistically challenged, but I thought disaster-related death tolls were estimated through deviation from an expected mortality baseline -- like when they estimate deaths from heat/cold. Surprised the range of "excess deaths" would be so wide for an island population.
1 reply 0 retweets 0 likes -
Replying to @NineBandedBooks @Meaningness
It's just because the study was underpowered. If mortality was at baseline, they would expect 23 deaths in their sample (literally the people they got to answer the survey). Instead, they found 38. Scaling those 15 extra deaths to the whole island gives you 4,645.
2 replies 0 retweets 1 like -
The point is, though, that simply scaling up is Problematic, in the correct sense of the word. It's not really that much information at all, which is why the confidence interval is so wide. There are a bunch of other studies that estimated around 1000 extra deaths.
1 reply 0 retweets 3 likes -
Given that, and the iffiness of this estimate, the smart money is on something like 'probably in the 1,000-1,500 range, and this sample was mostly a fluke'.
1 reply 0 retweets 1 like -
But newspapers thrive on scary nonsense, so they're going to jettison the consensus in favor of the biggest number.
1 reply 0 retweets 1 like
Also note: "three died directly from medical complications, illness or trauma because of the hurricane, and 12 died because of the interruption of medical services in the aftermath". Some/a lot of those people were probably near death already. Maybe this moved it up a few months.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
. Banned in Sweden. SubGenius, Zhuangist, white-hat troll. Defrocked mathematician. Brain problems.