Nate Silver

Nate Silver @NateSilver538

@thecity2 They're live caller polls with cellphones, should be pretty decent quality. — PolitiTweet.org

Nate Silver @NateSilver538

Bunch of live-caller polls in this morning, conducted on behalf of the AARP. Generally a mediocre set of polls for Biden and D Senate candidates relative to polling averages, although with exceptions. Note the timing—not super recent—and big sample sizes. https://t.co/9AeTKX5o0O https://t.co/TChzBwh4XB — PolitiTweet.org

Nate Silver @NateSilver538

@SenhorRaposa Agree that *now* would be a bad time. But it's kind of an interesting state! It's sort of white-working class. Portland is a national story. Is it becoming bluer like WA or CA or moving in a different direction? — PolitiTweet.org

Nate Silver @NateSilver538

@TrackerDebate It did vote for Obama! — PolitiTweet.org

Nate Silver @NateSilver538

Like, a firm that was like "we're going to focus on the Southwest, including Texas and House races in California" and develop expertise in issues that are pertinent in the region, such as in-depth polling of Hispanics" ... there's a big payoff there, I think. — PolitiTweet.org

Nate Silver @NateSilver538

There are also a handful of states where we could really use a "pollster of record", along the lines of Selzer in Iowa or Marquette in Wisconsin. Georgia, Nevada, Arizona and Minnesota come to mind. Lots of competitive races, few pollsters that focus on these states. — PolitiTweet.org

Nate Silver @NateSilver538

@NilesGApol @jacobfhsmith @kilometerbryman Yeah GA has quickly gone from a bit overpolled to underpolled. — PolitiTweet.org

Nate Silver @NateSilver538

@NickRiccardi I'm not sure it's competitive, but it sort of fits the pattern of a state where like "oops, nobody polled the race in October and somehow Gardner won!", kind of like the Maryland gubernatorial race a few years ago. — PolitiTweet.org

Nate Silver @NateSilver538

@TrackerDebate That is a very hot take. — PolitiTweet.org

Nate Silver @NateSilver538

I certainly wouldn't object to more polling in AK, MT and SC, which have interesting Senate races and could be competitive for POTUS in a Biden landslide. Oregon is the closest thing to a presidential swing state that hasn't been polled *at all* so one poll there would be nice. — PolitiTweet.org

Nate Silver @NateSilver538

Where could we use more polling? I think it's some of the smaller electoral vote states, which tend to get bypassed by both the online panels (hard to get a good sample size there) and the big-live caller pollsters. So: NV, NH, NE-2, Maine (especially with the Senate race there). — PolitiTweet.org

Nate Silver @NateSilver538

@conorsen Although arguably Michigan does not belong with the other three, in which case the remaining states give you some, um, interesting options. — PolitiTweet.org

Nate Silver @NateSilver538

@conorsen WAMP? — PolitiTweet.org

Nate Silver @NateSilver538

Here's Biden's lead in our polling average and forecast, respectively, in the "Big 4" Midwestern / Rust Belt states: MI: +7.5 / +7.6 MN: +6.5 / +5.9 WI: +7.0 / +5.6 PA: +5.1 / +4.3 If Biden wins all 4 of these he wins the election, assuming nothing else too weird happens. — PolitiTweet.org

Nate Silver @NateSilver538

Still no live-caller polls of Minnesota since Fox News polled the state in July (Biden +13). But SurveyUSA is the highest-rated firm to poll the state since then and has Biden +9. His lead has inched up to 6.5 points in our polling average there. https://t.co/WfszY2xOvB https://t.co/VMmJKvUkua — PolitiTweet.org

Nate Silver @NateSilver538

Oh shit somebody caught on. — PolitiTweet.org

Nate Silver @NateSilver538

@foxjust @conorsen @StaceyShick I hope they make outdoor seating year-round and let restaurants get creative with tents, heat-lamps, etc. It probably doesn't work most days in Dec.-Feb. but you can claw back November and March that way. — PolitiTweet.org

Nate Silver @NateSilver538

Typically likely voter polls help Republicans but hurt incumbents. Since Trump is a Republican but an incumbent, our prior was that Trump would only gain 1 point or so from RV>LV. And that's actually been pretty close so far; he's gaining maybe 0.5 points or something on average. — PolitiTweet.org

Nate Silver @NateSilver538

@Redistrict @FiveThirtyEight Our *polling average* has Arizona as the tipping-point state now, where Biden is +5.1, versus +7.6 nationally. Our *forecast* has Pennsylvania as the tipping-point state at Biden +4.3, versus +6.6 nationally. — PolitiTweet.org

Nate Silver @NateSilver538

@conorsen @foxjust @StaceyShick Some if of the *very* high-end restaurants in NYC are actually fairly large and/or well-spaced out. That's part of the luxury you're paying for. It's the bistros that will have more issues, although many can use an outdoor/indoor combo for a few weeks. — PolitiTweet.org

Nate Silver @NateSilver538

If we have ~6 states that are polled constantly, another ~6 polled somewhat frequently, plus national polls, and we can also slice and dice things along other dimensions (e.g. live-caller vs. online) that's a lot of opportunities to find patterns, but many will be spurious. — PolitiTweet.org

Nate Silver @NateSilver538

I talk a lot about how there's a lot of noise in polling averages and today was a good example of that. There were questions about why Biden was polling better in WI and PA. But today he got a great poll in PA and a perfectly fine but sort of meh one in WI. These things even out. — PolitiTweet.org

Nate Silver @NateSilver538

Marquette has Biden +4 in Wisconsin (that's without leaners; we'll use the version with leaners once it's available) which is ~the same as their last poll of the race. It does knock Biden down a bit from +7.5 to +7.0 in our Wisconsin average. https://t.co/C8GYebMykJ https://t.co/ln9vzswB5Q — PolitiTweet.org

Nate Silver @NateSilver538

Oops, there's an 11% chance that Trump wins the *Electoral College* but not the popular vote, not the other way around. (Got that revered in my tweet so reposting, h/t to @greenfield64 & others.) Just one of the many "fun" scenarios you can now explore. https://t.co/ajG88SznSA https://t.co/xmdS23LOCw — PolitiTweet.org

Nate Silver @NateSilver538

We've added our old friend the scenario analysis chart to our forecast interactive. There's an 11 percent chance Trump wins the popular vote but not the Electoral College (<1% chance of the other way around). https://t.co/ajG88SznSA https://t.co/m2oH2XMH3p — PolitiTweet.org

Nate Silver @NateSilver538

RT @ryanabest: 🚨🚨 We've added another module to the forecast!! 🚨🚨 What are the chances that some wild (or not so wild) scenarios actually… — PolitiTweet.org

Nate Silver @NateSilver538

As mail voting becomes a bigger factor, real-time election night predictions are going to get harder to the point where needle-like calculations at the very least require a **lot** more work and may not be prudent or viable. But at least in 2016, it was very successful. — PolitiTweet.org

Nate Silver @NateSilver538

I understand why liberals have negative memories of the NYT needle, but as a feat of reporting, it was awesome. It figured out quickly that Trump was winning on election night, overcoming its initial skepticism far sooner than the networks, etc did. It basically scooped everyone. — PolitiTweet.org

Nate Silver @NateSilver538

FWIW, here's a list of all nonpartisan live caller polls conducted since the conventions. https://t.co/QPMmvai2DT — PolitiTweet.org

Nate Silver @NateSilver538

I've pointed out this contradiction before, but there's a type of person who believes that outward, highly public signs of Trump enthusiasm (yard signs, rally attendance, etc.) are more reliable than polls BUT that Trump voters are too "shy" to tell the truth to pollsters. — PolitiTweet.org