Nate Silver

Nate Silver @NateSilver538

Biden's national lead has also ticked up to 7.8 points, as our polling averages are starting to see more evidence that Trump got a small bump from the RNC which has since faded (or is still fading?). https://t.co/cy51vc5isJ https://t.co/ZSWsg8yeLl — PolitiTweet.org

Nate Silver @NateSilver538

Other good news for Biden: Marist's poll showed him doing 1 point *better* among likely voters than registered voters. Same was true in their FL poll. Pollsters often switch from RV to LV after Labor Day, which sometimes helps Republicans, but doesn't seem to be helping Trump. — PolitiTweet.org

Nate Silver @NateSilver538

Since PA is the tipping point state in nearly one-third of our simulations, any polling movement in the state can have an impact on the model's bottom line. Between Marist and some national polls, Biden's Electoral College chances have risen to 74%. https://t.co/ajG88SznSA — PolitiTweet.org

Nate Silver @NateSilver538

Lotta polling out this morning and it's mostly been in a direction our model likes for Biden. The most influential number is in Pennsylvania, where Marist has Biden up 9 among likely voters. That gets Biden up to +5.1 in Pennsylvania from +4.1 before. https://t.co/wKbW02AGO5 https://t.co/hDZtxcMiEg — PolitiTweet.org

Nate Silver @NateSilver538

@jonathanchait Maybe we need to add more categories (Pretty Favored, More Favored Than Not, Favored To Some Extent) — PolitiTweet.org

Nate Silver @NateSilver538

We ran 12 updates to our model today. Here's how Biden's odds changed over the course of the day. Riveting stuff. 71% (overnight) 72% (6 AM) 72% (7 AM) 72% (10 AM) 72% (10:30 AM) 72% (10:50 AM) 71% (1 PM) 71% (1:20 PM) 71% (2 PM) 71% (3 PM) 72% (4 PM) 71% (5 PM) 71% (6 PM) — PolitiTweet.org

Nate Silver @NateSilver538

The phenomenon of polls recently underestimating Republican support in heavily WWC areas is well-known. But Democrats have often beat their polls in the past couple of cycles in Hispanic-heavy states like AZ, NV, TX and CA (though *not* FL, notably). — PolitiTweet.org

Nate Silver @NateSilver538

There are some parallels between Biden's apparent gains in polls with white noncollege voters and Trump's apparent gains with Hispanics. Both are probably real, and in general you should trust the polls IMO, but these groups have been sources of polling error in the past. — PolitiTweet.org

Nate Silver @NateSilver538

RT @mcimaps: And if he fails to take #NE02, its a tie! What happens then? Well you can read right here. https://t.co/AW5PR3bFnv — PolitiTweet.org

Nate Silver @NateSilver538

With PA + AZ, he could also lose Wisconsin *and* Minnesota if he holds onto the others. Those 20 electoral votes in Pennsylvania matter a lot, which is why it's showing up as the tipping-point state almost 1/3 of the time. https://t.co/ajG88SznSA — PolitiTweet.org

Nate Silver @NateSilver538

Conversely, winning PA but losing WI would be a lot more robust, because that puts Biden at 280 EV with Arizona. He could lose any 2 out of 3 of the NH/NV/ME-2 group and still have enough electoral votes to win. https://t.co/98DDfDOx7w — PolitiTweet.org

Nate Silver @NateSilver538

Biden's winning 270 electoral votes there. But that assumes he wins NE-2 in that map (where he's been well ahead in D partisan polls but which hasn't had any nonpartisan polling) plus hold on in NV and NH, plus hope there are no faithless electors. — PolitiTweet.org

Nate Silver @NateSilver538

If you take the polls at face value right now—with Biden gaining/holding steady in AZ, WI, MI but losing ground in PA & FL, then Arizona is the tipping-point state right now, which could lead to the map shown here. This is a winning map for Biden, but it's precarious. https://t.co/bLTzvk2E3l — PolitiTweet.org

Nate Silver @NateSilver538

If you look at this map, there's a lot of attendance from rural counties in the Upper Midwest/Great Plains/Rocky Mountains. These places generally hadn't been hit hard by COVID before and for various reasons may therefore be more prone to case growth now, especially on a % basis. https://t.co/su560n5OIk — PolitiTweet.org

Nate Silver @NateSilver538

I think this Sturgis study deserves some skepticism. Another problem is that I'm not sure they're really adequately controlling for where the attendees are coming from, i.e. it may be that cases happen to be increasing in those areas, but not because of Sturgis. — PolitiTweet.org

Nate Silver @NateSilver538

So FL is Biden's worst state in terms of the current polling trend, but TX and AZ (which are the most demographically similar to Florida on this list) are his best ones? He's losing ground in PA, OH and MI but gaining in WI and MN? A lot of this looks like noise. — PolitiTweet.org

Nate Silver @NateSilver538

In order of worst to best for Biden: Florida -3.4 Pennsylvania -2.2 Ohio -1.4 Michigan -1.1 {{{National -0.9}}} North Carolina +0.4 Wisconsin +0.6 Minnesota +0.6 Texas +0.9 Arizona +0.9 — PolitiTweet.org

Nate Silver @NateSilver538

Not super easy to find patterns in where Biden is gaining or losing ground. Here's a comparison of the 538 polling average in swing states that have been polled at least twice since the conventions began, between Aug. 17 (first night of the DNC) and today. https://t.co/aa0d5Sv2RH — PolitiTweet.org

Nate Silver @NateSilver538

RT @gercohen: @LPDonovan @daveweigel I have more updated info on the 30 “duplicates” and “spoiled” in the NC absentee file. Many are colleg… — PolitiTweet.org

Nate Silver @NateSilver538

Still, this polling movement in Florida—even if it exaggerates the situation a bit—makes Biden's strategy less robust. It had the potential to deliver a knockout blow. Now, it's one of a number of Plan B states (along with AZ, NC) if Biden doesn't hold his leads in the Midwest. — PolitiTweet.org

Nate Silver @NateSilver538

Florida's likelihood of being the tipping-point state has declined a fair bit as it has moved away from the national averages. Until fairly recently, it was basically tied with PA for No. 1. Now it's more in a group with WI, MN, MI, AZ and NC. https://t.co/viTrjbCt4b — PolitiTweet.org

Nate Silver @NateSilver538

So rather than either unskew this poll (sure, there are crosstabs you can critique... but that's true in every poll!) *or* treat it as the new normal, it's probably best to toss it into the model, which can figure out how much to hedge it against other polls and other data. — PolitiTweet.org

Nate Silver @NateSilver538

At the same time, when you see the polls move strongly in one swing state when they aren't doing so in other swing states (Biden's polling has been good in WI, AZ lately, etc.) and there's no particular reason to expect the state to behave differently, it is often a quirk. — PolitiTweet.org

Nate Silver @NateSilver538

Biden's lead is down to 1.9 points in Florida after this poll. Was 2.7 before we added this poll. Maybe more troublingly for Biden, had been a 5 or 6 points lead before the conventions. https://t.co/nWUo9X5F4y https://t.co/DNE8WRZXpp — PolitiTweet.org

Nate Silver @NateSilver538

This continues a pretty good run of Florida polls for Trump. — PolitiTweet.org

Nate Silver @NateSilver538

@SenhorRaposa 😬 https://t.co/vPtsrKnx07 — PolitiTweet.org

Nate Silver @NateSilver538

@SenhorRaposa Wait is it literally McLaughlin? — PolitiTweet.org

Nate Silver @NateSilver538

RT @ryanabest: By popular demand, we've added some extra info to our 🐍 snake 🐍 chart!! Hover over a state segment to see exactly how many… — PolitiTweet.org

Nate Silver @NateSilver538

@gercohen @WinWithJMC @NCSBE Ahh, thank you -- that's great context. — PolitiTweet.org

Nate Silver @NateSilver538

RT @gercohen: @LPDonovan @daveweigel sample here are all military or overseas. “Spoiled” in NC election lingo is either person voluntarily… — PolitiTweet.org