Nate Silver

Nate Silver @NateSilver538

Yeah. Religious white Christians are the most reliably Republican voters. Black voters are the most reliably Democratic ones (other nonwhite groups to a slightly lesser degree). So states with a lot of irreligious whites tend to be swingy. — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias https://t.co/EXvEP5u5Yw — PolitiTweet.org

Nate Silver @NateSilver538

Elasticity ratings for 2020! Elasticity scores reflect how much a state’s polls would be expected to change based on a change in national. e.g. if Idaho’s elasticity score is 1.1, a 5-point swing in national polls would be expected to produce a 1.1 x 5 = 5.5-point swing in Idaho. https://t.co/9BIk53SsSw — PolitiTweet.org

Nate Silver @NateSilver538

Maybe also noteworthy that Trump has lost ground more than Biden has gained it. There may be some R-leaning voters who are just fed up with things right now but could come back to Trump later. Then again, Biden is >=50% right now, whereas Clinton peaked at 46% in national polls. — PolitiTweet.org

Nate Silver @NateSilver538

So it's not like the race has actually swung *that* much. But 4 points vs. 8 points is sort of an inflection point where you go from "whoa, the Electoral College could still be close" to "whoa, Biden up everywhere". But things could easily enough swing back in Trump's direction. — PolitiTweet.org

Nate Silver @NateSilver538

Our (soon-to-be-published) national polling average over time: March 1: Biden 49.7, Trump 45.6 (+4.1) Apri 1: Biden 49.2, Trump 43.1 (+6.1) May 1: Biden 49.5, Trump 43.4 (+6.1) June 1: Biden 49.2, Trump 42.9 (+6.3) June 15 (today): Biden 50.2, Trump 42.1 (+8.1) — PolitiTweet.org

Nate Silver @NateSilver538

RT @FiveThirtyEight: Here's how to read polls like a pro: https://t.co/Me2bceTvLh — PolitiTweet.org

Nate Silver @NateSilver538

I understand the impulse to post pictures of people not adhering to social distancing—and at this point there are a wide variety of examples to choose from. But I've long worried that shaming burns people out in the long run…and now it's the long run and people seem burned out. — PolitiTweet.org

Nate Silver @NateSilver538

The other measures don't look so hot, though. New cases have been at a plateau, and the positive test rate has been stuck between 4-5% for a while now. — PolitiTweet.org

Nate Silver @NateSilver538

Sundays have featured really low numbers recently, especially since New York (which does *not* have strong week vs. weekend effects) has become a less important part of the picture. Still, 358 deaths is the fewest reported in a day since March 26. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 358 Yesterday: 695 One week ago (6/7): 460 Newly reported cases T: 21K Y: 25K 6/7: 19K Newly reported tests T: 485K Y: 519K 6/7: 445K Positive test rate T: 4.4% Y: 4.9% 6/7: 4.3% — PolitiTweet.org

Nate Silver @NateSilver538

The thing about there being so much actual, really important *news* is that there's no need to waste your time with all the poll unskewing. But the poll unskewing is going to be dumber than ever. — PolitiTweet.org

Nate Silver @NateSilver538

One other thing to consider when infections are low in a certain area and testing is high is the role of false positives. There don't seem to be a lot of false positives in PCR tests (more in antibody tests) but there probably are *some*: https://t.co/SrOqYIWU5R — PolitiTweet.org

Nate Silver @NateSilver538

NYC hit another new low positive test rate (1.4%) after having bobbed around a bit for the past few days. So that's good. For the week as a whole, NYC had a 1.6% positive rate this week as compared to 2.1% the previous week. https://t.co/1w2VrzC4FG — PolitiTweet.org

Nate Silver @NateSilver538

In addition to COVID+ people, NYC has also been offering free hotel rooms to healthcare workers who are at risk of catching the coronavirus. https://t.co/qJcsMLyOLe — PolitiTweet.org

Nate Silver @NateSilver538

For instance, would Republican governors in red states be even quicker to open up if it was seen as defying Clinton? Would mask-wearing be even more of a partisan signifier? I don't know, but those things seem possible, at least. And the CARES Act might have been weaker. — PolitiTweet.org

Nate Silver @NateSilver538

Seeing some threads about how a Clinton White House would have handled COVID-19. While the answer almost certainly has to be "better" given the low bar the actual White House set, it's also worth thinking about negative partisanship, i.e. how Republicans would react to Clinton. — PolitiTweet.org

Nate Silver @NateSilver538

This feels like a little bit of a canary in the coal mine. Ernst has been considered a rising star; Iowa has trended red; she hasn't had any particular scandals or gotten in any particular trouble. If this race is competitive, the GOP is in a lot of trouble in a lot of places. — PolitiTweet.org

Nate Silver @NateSilver538

The good news is that deaths remain well below their peak. It's not clear why that is: 1. Could be because deaths lag infections (i.e. they'll begin increasing later) 2. Could be that new cases are hitting mostly younger populations 3. Could be that better treatments are helping — PolitiTweet.org

Nate Silver @NateSilver538

This is a *lot* of new cases. The most since May 14. It's *also* a lot of tests. Seems likely though that infections are growing in a number of Southern, Western and Southwestern states even once you control for test volume. Other states still improving. https://t.co/oAQ2y08bJ5 — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking Newly reported deaths Today: 695 Yesterday: 751 One week ago (6/6): 717 Newly reported cases T: 26K Y: 24K 6/6: 23K Newly reported tests T: 523K Y: 584K 6/6: 478K Positive test rate T: 5.0% Y: 4.1% 6/6: 4.7% — PolitiTweet.org

Nate Silver @NateSilver538

@jbarro @mattyglesias These averages contain a trendline adjustment (i.e. it adjusts old polls based on what's happening in national polls and polls of similar states). So stale isn't an issue per se. It's more just that there isn't a ton of recent polling in PA, so the estimate is a bit noisy. — PolitiTweet.org

Nate Silver @NateSilver538

I guess one thing that really stands out here is that it doesn't take that much to go from a massive Biden Electoral College landslide (if he wins everywhere on the list below, that's 413 EV) to a narrow Biden loss (he's only ahead 3.7 points in the tipping-point state). — PolitiTweet.org

Nate Silver @NateSilver538

The tipping-point state here would be PA, where there's been a real lack of high-quality polling for the past couple months. I doubt that Biden is actually ahead by 8 in MI and 6 in WI but only 3-4 in PA. But, that's what the polls show if you don't blend them with priors. — PolitiTweet.org

Nate Silver @NateSilver538

National: Biden +7.6 CO: Biden +15.6 ME: Biden +12.4 VA: Biden +9.5 MI: Biden +8.2 NE-2: Biden +7.4 NH: Biden +6.6 WI: Biden +5.9 NV: Biden +5.7 MN: Biden +5.5 PA: Biden +3.7 AZ: Biden +3.6 FL: Biden +3.5 NC: Biden +2.0 OH: Biden +1.6 GA: TIE TX: Trump +1.4 IA: Trump +1.6 — PolitiTweet.org

Nate Silver @NateSilver538

We'll have our national and state polling averages coming out soon. Note that these are polling averages only; they're NOT yet blended with any sort of demographic priors. Here's what the draft versions are currently showing.... — PolitiTweet.org

Nate Silver @NateSilver538

Here are states that added the most cases PER CAPITA over the past week. Arizona: 11.8 newly-reported cases per 10,000 people Arkansas: 9.6 Alabama: 8.8 Mississippi: 7.8 North Carolina: 7.6 Utah: 7.3 Iowa: 6.9 Louisiana: 6.5 D.C.: 6.4 Maryland: 6.4 Nebraska: 6.2 — PolitiTweet.org

Nate Silver @NateSilver538

Not so bad, all in all. A fairly high number of cases BUT a record number of tests. So this was actually the second lowest positive test rate of all time. Still, there are some fairly large states where cases are growing faster than test volume. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 751 Yesterday: 936 One week ago (6/5): 848 Newly reported cases T: 24K Y: 22K 6/5: 23K Newly reported tests T: 584K Y: 447K 6/5: 540K Positive test rate T: 4.1% Y: 5.0% 6/5: 4.3% — PolitiTweet.org

Nate Silver @NateSilver538

As for "spikes", the word means a sharp, dramatic increase. While those can happen, what's both more common AND the bigger concern IMO is sustained gradual increases. If cases increase by 10% per week for 6 weeks, well, there's never a spike, but you can have really big problems. — PolitiTweet.org