Thread by @charles_gaba

🧵 Scores of people have called my attention to the new NBER study by @paulgp, @jwswallace and @jasonlschwartz confirming that COVID deaths have been MUCH higher among Republicans than Democrats since the moment the vaccines became widely available. 1/
acasignups.net/22/10/03/new-n…

It's worth noting that one of the stories
@paulgp
linked to in his original thread a few days ago, from the NY Times last November, uses data from, well...me: 2/

https://twitter.com/paulgp/status/1576899940525084672

The NY Times story links to this post of mine from last fall: acasignups.net/21/10/14/weekl…

I'm not saying this to criticize the new study--theirs was done in a FAR more in-depth, methodical way which also addresses some of the concerns folks have brought up in an attempt to poke holes in the data. I'm only writing this to reassure folks that yes, I know about it. 4/

And of course there are plenty of others who have tracked/analyzed the Red/Blue divide, from @pbump at the Washington Post to @greg_travis, @JonathanCOnP and others. It's still always good to get additional confirmation, especially via a respected outlet such as @nberpubs. 5/

Anyway, here's an *animated* look at how the red/blue vaxx divide played out for the first year of availability (Feb. 2021 - Feb. 2022). Notice how there was almost no gap for the first few months...but it exploded starting in May 2021? 6/

What happened in May 2021? Well, at first vaxxes were limited to seniors, healthcare workers & some immunocompromised groups. In late March/early April it was opened up to ALL adults (& 12+ adolescents shortly after that).

It then took 3-4 weeks for 2nd doses...or early May. 7/

Here's how the Red/Blue gap R^2 (correlation) has increased over time from February 2021 through September 2022 (last month): 8/
acasignups.net/22/09/11/septe…

Here's how the *slope* of the Red/Blue gap has grown steeper over the past 19 months.

Notice how in both cases it increased *dramatically* once the floodgates were open, then leveled off...until 5-11 yr olds were authorized to get vaxxed, at which point it jumped again. 9/

The reason for the second jump around Thanksgiving 2021 seems to be that Dems got their 5-11 yr old children vaxxed while GOP parents refused to.

Since April, the newly-vaxxed rate has dropped off across the board. The 6mo-4yr authorization hasn't caused much of a 3rd jump. 10/

OK, that's the vaccination side. What has this looked like on the *death rate* side? For that, I've mostly used a bar graph. Here's what it looks like from the beginning of the pandemic (Feb. 2020) through November 2021, animated: 11/
acasignups.net/21/11/17/red-s…

The animation above spans the first 20 months of the pandemic.

The graph below shows what cumulative COVID death rates looked like at each partisan bracket for the first 16 months (thru 5/01/21)...and for the 16 months after (thru 9/01/22): 12/

The cumulative COVID death rate in the bluest tenth of the U.S. has been 55% *lower* in the post-vaccine 16 months vs. the pre-vaccine 16 months.

The COVID death rate in the reddest tenth has been 18% *higher* in the post-vaccine 16 months vs. the pre-vaccine 16 months. 13/

So, how is all of this likely to impact the 2022 midterms? Here's my final updated attempt to analyze that question from mid-September: 14/
acasignups.net/22/09/17/eleph…

NOTE: This ISN'T about how *family/friends* of those who've died are likely to vote. This is just looking at how many of those who've died of COVID since the 2020 election voted for Trump vs. Biden. That doesn't necessarily overlap 100% w/the midterms but it should be close. 15/

Here's COVID death rates along the Red/Blue partisan line *since 11/04/20 only,* as of mid-September 2022.

The death rate since then in the reddest decile is over 2.5x higher than in the bluest.

If you look at the reddest/bluest 20%, it's still over twice as high. 16/

NATIONALLY, *since 11/04/20,* ~815,000 Americans have died of COVID (going by the official data; this doesn't take into account misdiagnoses, etc, which I'll go into later).

Of those, I figure perhaps 568,000 actually voted for Biden or Trump (~70% of all COVID deaths). 17/

Of those ~568,000, if I *only* adjust them based on age bracket using 2020 state level exit polling (which I realize is flawed but it's all I have to work with), I estimate perhaps 299K were Trump voters vs. 269K Biden voters, a difference of ~30K or so.

HOWEVER... 18/

...as noted above, there's also the "excess deaths" factor. A pre-print study last spring found that while excess deaths NOT attributed to COVID were fairly similar along partisan lines in 2020... 19/
acasignups.net/22/05/09/exclu…

...in *2021* there was a SHOCKING discrepancy in these "non-COVID excess deaths."

How shocking? They ran a stunning 21x higher in the reddest tenth of the U.S. than in the bluest. 20/

W/this in mind, I've adjusted my Trump/Biden voter COVID death estimates to shift towards Trump voters by ~20% overall, giving raw estimated numbers of ~359K Trump voters vs. ~210K Biden voters, or roughly 150K more Trump voters having died of COVID since the 2020 election. 21/

While that's a shockingly large raw number, keep in mind that around 113 million Americans voted in the 2018 midterms. Assuming similar turnout nationally this year, that 150K is just 0.13% of total voter turnout. 22/

Beyond that, remember that most of those deaths happened in either fairly solid red or solid blue states. Here's what the gap looks like in the 12 closest states in 2020, compared to either Biden or Trump's margin: 23/

In the 6 Biden-won states, similar R/D turnout in 2022 would just pad the Dem win slightly.

In the 6 Trump-won states, it wouldn't shift the results to the Dem nominee by more than ~8% of the final GOP margin (in NC, it would've shaved Trump's margin from 74.5K to 68.1K). 24/

It's also important to remember that ALL of these COVID deaths should be baked into polling anyway, since you can't answer a survey if you're dead.

In short, I wouldn't expect the Red/Blue COVID death gap to impact state-level races unless they're razor-thin to begin with. 25/

How about at the district level, though?

Unfortunately, most Congressional/ legislative districts don't overlap nicely with full county borders, so this isn't something that I was able to do. Plus, there's already a LOT of assumptions/caveats I'm making at the state level. 26/

Having said that, it could make a difference in some otherwise-razor thin House & state legislative races in certain districts, as well as county-level/etc races.

Anyway, #GetVaxxed, #GetBoosted and continue to #WearAFuckingMask when indoors at crowded public locations. /END

P.S. Here's 500 ways to help Democrats win up & down the ballot this year: AmericaBlueIn22.com/final-500

P.P.S. If you find my work useful & want to help support it, you can do so here, thanks! Acasignups.net/support

Addendum: I should note that my 150K estimate assumes only the GOP COVID death rate post-2020 election is only 71% higher than the Dem death rate, vs the NBER study’s 153%. However, their 153% starts in spring 2021, while my 71% starts in November 2020.

In order to equal 153% higher, the gap would have to be more like 406K Trump voters vs. 163K Biden voters, or a 243K difference nationally.