2 appalling aspects of discourse:
1. Deaths statistic is being treated as factual even when we know it's a WILD undercount. This is deeply irresponsible. EVERY sentence referencing the stats MUST include the fact it is SERIOUSLY too low.
2. MANY "models" project only 1 wave.
1. Deaths statistic is being treated as factual even when we know it's a WILD undercount. This is deeply irresponsible. EVERY sentence referencing the stats MUST include the fact it is SERIOUSLY too low.
2. MANY "models" project only 1 wave.
SOME LINKS (but there are many out there):
https://www.nytimes.com/2020/04/10/nyregion/new-york-coronavirus-death-count.html
https://www.nytimes.com/2020/04/05/us/coronavirus-deaths-undercount.html
I *strongly* suspect red state/rural counts are WAY off, given inadequacy of testing & political incentives.
https://www.nytimes.com/2020/04/10/nyregion/new-york-coronavirus-death-count.html
https://www.nytimes.com/2020/04/05/us/coronavirus-deaths-undercount.html
I *strongly* suspect red state/rural counts are WAY off, given inadequacy of testing & political incentives.
Why does this matter? By underplaying the human cost of our irresponsibly delayed social distancing, "stats based" people are facilitating a premature return to normalcy.
Statistics *seem objective,* and so limitations must be highlighted.
Statistics *seem objective,* and so limitations must be highlighted.
(btw, we were last in office physically 3/10 at @revolvingdoorDC/ @ceprdc)
One sports context: Utilizing "deaths due to COVID-19" without adjusting for the undercount is like reporting "home runs hit by the 1990s Colorado Rockies" without noting Coors Field's unique impact on baseball.
It's tendentiously correct and wildly wrong at the same time!
It's tendentiously correct and wildly wrong at the same time!
As I keep yelling about... there is major variability in how many COVID-19 deaths are being undercounted and the right wing UK government is behaving as expected. https://www.politico.com/newsletters/politico-nightly-coronavirus-special-edition/2020/04/13/open-and-shut-488897
cc: @mattyglesias @NateSilver538
cc: @mattyglesias @NateSilver538
Our nursing home count is a mess: https://www.nbcnews.com/news/us-news/more-2-200-coronavirus-deaths-nursing-homes-federal-government-isn-n1181026
Good that New York City is belatedly *beginning* the incomplete process of updating their statistics; crazy that just recently death certificates would mention "COVID" & yet the deaths had been excluded from the stats "data analysts" trumpet.
https://www.politico.com/states/new-york/albany/story/2020/04/14/new-york-city-coronavirus-death-toll-jumps-by-3-700-after-uncounted-fatalities-are-added-1275931
https://www.politico.com/states/new-york/albany/story/2020/04/14/new-york-city-coronavirus-death-toll-jumps-by-3-700-after-uncounted-fatalities-are-added-1275931
I bet Florida under criminally negligent Governor @GovRonDeSantis is VASTLY worse in undercounts. If you see a stat analyst taking Florida stats literally or seriously... adjust your priors on their analysis.
Florida is going to be such a catastrophe and I am skeptical the public stats will reflect that for a while.... https://www.miamiherald.com/news/health-care/article241996231.html
Here's @propublica on the unattributed death issue. By failing to account for deaths accurately we're encouraging unwise public policy and distorting public opinion.
Too many striking visual representations with far too little humility about data quality. https://www.propublica.org/article/theres-been-a-spike-in-people-dying-at-home-in-several-cities-that-suggests-coronavirus-deaths-are-higher-than-reported
Too many striking visual representations with far too little humility about data quality. https://www.propublica.org/article/theres-been-a-spike-in-people-dying-at-home-in-several-cities-that-suggests-coronavirus-deaths-are-higher-than-reported
Federalism in America, man. https://www.propublica.org/article/theres-been-a-spike-in-people-dying-at-home-in-several-cities-that-suggests-coronavirus-deaths-are-higher-than-reported
Yeah, between penal incarceration and the incarceration of migrants... I bet thousands of human beings are dying without being counted. https://twitter.com/TomJawetz/status/1250518674689097731
WHEN WILL THESE HANDLES AND WEBSITES quit misleading people? https://twitter.com/jeffhauser/status/1251123746624151553
Do I think that nursing home TRAGEDIES like this are limited to NYC? https://twitter.com/rontkim/status/1251121549643522048
Do we think the situation in Florida is going okay???? https://www.miamiherald.com/news/state/florida/article241966186.html
Do we think the situation in Florida is going okay???? https://www.miamiherald.com/news/state/florida/article241966186.html
Read this thread and recognize that there is a political movement which wants FEWER RULES ON HOW "PRIVATE SECTOR" spend taxpayer money in their nursing homes/assisted living facilities: https://www.nytimes.com/2020/04/16/nyregion/new-york-nj-nursing-homes-coronavirus-deaths.html
There are no "2 sides" to the argument for Big Government Regulation.
There are no "2 sides" to the argument for Big Government Regulation.
Here I was on April 1st telling everyone that the leaders of the "narrative" were pretending statistics that were being juked were being objective.
I bet a ton of the pushback on social distancing is due to the stat juking understating COVID-19's impact: https://twitter.com/jeffhauser/status/1245347547880796161
I bet a ton of the pushback on social distancing is due to the stat juking understating COVID-19's impact: https://twitter.com/jeffhauser/status/1245347547880796161
More like this headline from the @ajc, please. https://www.ajc.com/news/undercount-covid-deaths-means-full-effects-georgia-unknown/Cg1na060vIiOH6VM3ecztO/
I mean, how can you justify limiting "statistics" purported to be complete on COVID-19 deaths to positive tests when American testing capacity is tragically inadequate?
My 3+ weeks of obsession on this point is because we're going to see premature re-openings (and inadequate provision of free food, shelter, and health care to the vulnerable) due to the sense the pandemic is overrated... because of juked "statistics."
That Boris Johnson's United Kingdom is juking the stats is... unsurprising, although there may be hopes for improvement there. https://www.telegraph.co.uk/news/2020/04/17/number-coronavirus-deaths-care-homes-could-high-7500-according/
GREAT TO HAVE The Economist on the case: https://twitter.com/J_CD_T/status/1250815341044924425
Pennsylvania undercount: https://whyy.org/articles/pa-underreported-18-of-covid-deaths-this-week-officials-blame-computer-glitches/
Good to have @thedailybeast on the case: https://www.thedailybeast.com/the-covid-19-death-undercount-is-scarier-than-you-think?ref=home
The "re-open fast" case is basically based on juked stats, and everyone who navel gazes about stats on Twitter without emphasizing the omissions is encouraging reliance on inherently unreliable analysis. https://www.thedailybeast.com/the-covid-19-death-undercount-is-scarier-than-you-think?ref=home
All of America is Puerto Rico now, and #COVIDー19 is #HurricaneMaria, as many of us have suggested it would be from the jump.
Hurricane Maria is too often ignored; here we were on it last fall as a possible basis for impeachment, https://washingtonmonthly.com/2019/09/19/the-kitchen-table-case-for-impeaching-trump/
As I surmised upthread, UK & USA way ahead of rest of developed world in... covering-up COVID-19 deaths. Some of the tragic truth is getting out there now.
England & Wales: https://www.theguardian.com/world/2020/apr/21/coronavirus-deaths-in-care-homes-in-england-and-wales-more-than-quadruple-in-a-week
England & Wales: https://www.theguardian.com/world/2020/apr/21/coronavirus-deaths-in-care-homes-in-england-and-wales-more-than-quadruple-in-a-week
Good to see the @nytimes joining in my obsession about representing the true cost of the COVID-19 crisis -- far too much ostensibly objective elevation of CRAP statistics on deaths leading people to downplay immense losses. https://www.nytimes.com/2020/04/21/world/coronavirus-live-world-cases-global.html?type=styln-live-updates&label=global&index=1&action=click&module=Spotlight&pgtype=Homepage#link-567f32d9
Important! Also: Every reason to fear "immunity" may not last that long, so people who fended off COVID-19 last winter might be susceptible to re-infection as soon as next winter or spring. https://www.sfchronicle.com/health/article/First-known-U-S-coronavirus-death-occurred-on-15217316.php
Sadly influential with me, but I have seen all of these claims elsewhere as well: https://twitter.com/statesdj/status/1252698777296797698
If COVID-19 not going away anytime soon, then clear-eyed truth about known and LIKELY deaths is better than false objectivity via clear undercount "statistics."
If COVID-19 not going away anytime soon, then clear-eyed truth about known and LIKELY deaths is better than false objectivity via clear undercount "statistics."
I mean *this disease is really bad* -- see, e.g., https://www.washingtonpost.com/health/2020/04/22/coronavirus-blood-clots/
United Kingdom is likely underestimating fatalities by more than a factor of 2, per the Financial Times: https://www.ft.com/content/67e6a4ee-3d05-43bc-ba03-e239799fa6ab
Strokes are likely an occasional outcome of COVID-19, and probably a cause of many at home sudden deaths being omitted from ostensibly objective statistics: https://www.washingtonpost.com/health/2020/04/24/strokes-coronavirus-young-patients/
The FT, which had been exacerbating unwarranted belief in the accuracy of COVID-19 death totals... now has a substantial corrective out. https://www.ft.com/content/6bd88b7d-3386-4543-b2e9-0d5c6fac846c
Now the @washingtonpost belatedly joins this beat: https://www.washingtonpost.com/investigations/2020/04/27/covid-19-death-toll-undercounted/?arc404=true
As Washington Post notes, the miscommunication about undercounting has been deleterious to discussing policy options/cost-benefit analysis.
Statistics tends to be perceived as "objective," & so all good faith analysts should YELL repeatedly about systemic biases within numbers.
Statistics tends to be perceived as "objective," & so all good faith analysts should YELL repeatedly about systemic biases within numbers.
BTW, for all the ghoulish anti-mass transit people out there... not only has Seoul South Korea proven an early hit from an epidemic isn't inherently bad for mass transit... cars, which are A MASSIVE public health scourge, being off the road is a huge positive. #CarsKill #climate
Read https://www.washingtonpost.com/investigations/2020/05/02/excess-deaths-during-covid-19/?arc404=true on undercounts through April 11th.
Fatality undercounting is part of why people have stopped wearing masls & taking distancing seriously.
Undercounting is killing people.
Fatality undercounting is part of why people have stopped wearing masls & taking distancing seriously.
Undercounting is killing people.
Enormous stat juking going on with respect to the imprisoned. Here's a key thread on Texas -- but I wouldn't be surprised if the same is going on in FL, AL, MS, GA, AZ, and many other states. https://twitter.com/TexasObserver/status/1258512836101206016
Yup, the Trump Administration is still trying to hide nursing home deaths. https://www.nbcnews.com/health/health-news/government-still-doesn-t-know-how-many-nursing-homes-have-n1202281
I am totally curious about Florida, where SOME @GovRonDeSantis scams are known publicly -- how many are still hidden???
I am totally curious about Florida, where SOME @GovRonDeSantis scams are known publicly -- how many are still hidden???
Here @FAIRmediawatch provides analysis of "undercounts" -- basically, a ton of states don't provide enough data to analyze undercounts yet.
Which makes them look artificially good.
But the NYT persisted, nonetheless. https://fair.org/home/nyts-real-coronavirus-death-toll-is-anything-but/ by @JNaureckas
Which makes them look artificially good.
But the NYT persisted, nonetheless. https://fair.org/home/nyts-real-coronavirus-death-toll-is-anything-but/ by @JNaureckas
Here's @FiveThirtyEight's take on a topic I have been obsessed with: Undercounting COVID-19 death. (I'm also obsessed with the fact that COVID-19 outcomes < death can often be much worser than from a normal "flu"). https://fivethirtyeight.com/features/coronavirus-deaths/