OK, for those of you who haven't had enough of Rev ranting about journalism, let's turn the page to this story, which is...fishy.https://twitter.com/washingtonpost/status/1001463917145620481 …
-
Show this thread
-
Several things jumped out at me in this paragraph. First, if 14.3/1k is a 62% increase, then the expected is baseline is 8.8/1k for that period. That corresponds to about 32 deaths per thousand per year. That implies life expectancy in PR is about 31 years. 2/pic.twitter.com/AVAIC4tgA6
1 reply 3 retweets 11 likesShow this thread -
Ah, this says the *yearly* mortality rate is 8.6/1000. https://www.indexmundi.com/puerto_rico/death_rate.html … If the population were steady-state that would imply an expected lifespan at birth of 116 years, but...eh, close enough to the reported value of 80ish https://www.indexmundi.com/puerto_rico/life_expectancy_at_birth.html … 3/
1 reply 0 retweets 5 likesShow this thread -
But that's quibbling. Here's the real problem. -- They surveyed about 3,300 households. Let's generously call that 10,000 people. -- The baseline mortality rate is 8.6. That would be 86 deaths in this group. -- They claim elevated mortality of 62%. 53 excess deaths... 4/
1 reply 1 retweet 4 likesShow this thread -
...Except, the excess mortality is only happening over the period from September 20 to Dec. 31, which is about 103 days. So the survey must have found about 53 x 103 / 365 = 15 extra deaths (and 39 total deaths by the same math). 5/
1 reply 1 retweet 5 likesShow this thread -
It would be easy (easier than it is for me, at least) for WaPo to present this reasoning -- or at least link to the fucking paper -- but then they'd have to mention the error bars. And the error bars are going to be pretty big. 6/
1 reply 1 retweet 10 likesShow this thread -
I've lost track of my point and I'm tired and today was generally infuriating, so I'm going to stop, except to note that as usual, WaPo used some rickety numbers to push a particular emotional narrative, because that's what they do. The end. 7/7
1 reply 1 retweet 13 likesShow this thread -
Ah, NPR is slightly more honest. They lead with "near 5,000", but it turns out the 95% confidence interval is [800, 8500]! https://www.npr.org/sections/health-shots/2018/05/29/615120123/study-puts-puerto-rico-death-toll-at-5-000-from-hurricane-maria-in-2017 …
1 reply 3 retweets 20 likesShow this thread -
Also: "The research team randomly selected 3,299 households in Puerto Rico. Local scientists surveyed them over the course of three weeks in January. People in those homes reported a total of 38 deaths." Rev was right again!
2 replies 2 retweets 19 likesShow this thread -
Replying to @St_Rev
Also, it's not like they used this sampling method pre-Maria,&were comparing the
#s, right? They were comparing their sampling method to published statistics. If it's just a biased method(for example, people are in multiple 'households' so they get overcounted) we wouldn't know?1 reply 0 retweets 3 likes
Presumably they counted # per household but who knows what other adjustments they used, let alone their validity. It's in AMA, I don't have access to that.
-
-
Replying to @St_Rev
I was converted to PR statehood yesterday solely on the basis of learning that vital statistics data for PR and territories aren't in CDC Wonder so we can't make easy inferences about whether their methods are any good.
0 replies 1 retweet 8 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
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.