New @pewresearch poll: Biden leads 69%-27% among those who plan to vote by mail, Trump leads 63%-31% among those who plan to vote on Election Day. pewresearch.org/politics/2020/…
Also interesting: @pewresearch finds that of the 5% of RVs who support Jorgensen/Hawkins, they would lean Trump over Biden 54%-42% if forced to choose - suggesting their presence might actually aid Biden?
Sorry, should have noted Biden also leading 55%-40% among early in person, per @pewresearch.
Another big finding: Biden leads Trump 49%-26% among 2016 Johnson/Stein voters, per @pewresearch, as well as 54%-38% among those who didn’t cast ballots in 2016.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
My new @nytimes op-ed: meet the 10 counties in America that hold the key to President Trump's fate - and illustrate why he's in such deep trouble. nytimes.com/2020/10/06/opi…
Today, I'd rate all ten of these top bellwether counties as either "toss ups" or "lean Biden:"
1. Maricopa, AZ 2. Pinellas, FL 3. Peach, GA 4. Marshall, IA 5. Kent, MI 6. New Hanover, NC 7. Wood, OH 8. Erie, PA 9. Collin, TX 10. Sauk, WI
These places don't fit one stereotype: they run the gamut from meatpacking hubs to military bases and peach orchards to yacht-dense retiree havens.
But one thing they have in common: their votes will matter a whole lot more than most Americans' votes.
The key to the stability of Biden's lead to date? He's basically stayed out of the way and ceded the spotlight to Trump. It's October, and this is still a referendum (vs. choice) election.
Hearing a few people ask: then why is Trump's approval up to 44.4% at @FiveThirtyEight? My answer: we're pretty close to Nov. and if the election were today, I wouldn't be surprised if Trump got ~44.4% of the vote.
This might be a good opportunity to explain why I will *not* be saying "I've seen enough" on Election Night 11/3 (besides the main reason, which is a longstanding contractual obligation w/ a network desk)...
My general approach to "seeing enough" on Twitter is to make that declaration at a ~98% personal confidence interval, whereas a media organization might prudently wait to project a race at a ~99%+ confidence interval.
In other words, I'd expect that for every 98 times "I've seen enough" ends up being correct, there might be two times it ends up being incorrect.
I'm comfortable w/ that level of risk on most nights. But on 11/3, the stakes are way too high to mischaracterize results.
If you were a voter who went in w/ concerns about Biden's age, he didn't do much to dispel them. If you went in w/ concerns about Trump's temperament, his tantrums sent them into overdrive.
If you were a casually politically-interested voter, you probably changed the channel.
Sorry, but last night was not Biden at his sharpest. He's almost 78 and looked/sounded like it. Fortunately for him, Trump, 74, behaved/sounded like he was 3.
p.s. Not sure how anyone can read this as "both sides" or equating their performances. Trump lost. Especially b/c it's Trump who needs to change the trajectory of the race, and last night didn't.
If you were only to go by September's live-interview national polls (averaging demographic crosstabs, ignoring state polls), here's how that might roughly translate in the Electoral College: cookpolitical.com/swingometer?CE…
As you might glean, it's not *that* far off from what state polls are showing! The notable exceptions might be:
1) FL, where Trump's gains w/ Hispanics have outpaced his national gains 2) GA, where non-college white voters are less elastic than they are in Upper Midwest
Also, in the above map, I'm assuming every demographic group sees a 5 point turnout bump vs. '16, but being extra-generous to Trump by assuming a 10 point bump in non-college white turnout.