Nate Cohn

Nate Cohn @Nate_Cohn

I think the needle's wait-and-see move right now is entirely the right call — PolitiTweet.org

Nate Cohn @Nate_Cohn

That parenthetical is worth taking very seriously: there just isn't much that's unequivocally *done* outside of FL (and to a lesser extent Indiana), so early vote contamination risk remains very real — PolitiTweet.org

Nate Cohn @Nate_Cohn

So far, Democrats are running ahead of our expectations outside of Florida (in places with meaningful vote, which may still be contaminated with early votes); they're way running behind in Florida. It's almost entirely cancelling out at the moment — PolitiTweet.org

Nate Cohn @Nate_Cohn

And unfortunately these validations are... very necessary. https://t.co/Bmq7l9deR9 — PolitiTweet.org

Nate Cohn @Nate_Cohn

This is how much the needle cares https://t.co/I6qUrYUsPk — PolitiTweet.org

Nate Cohn @Nate_Cohn

You should basically ignore the early votes in Ohio, Pennsylvania and North Carolina, which will be extremely unrerpesentative. Wait for completed votes — PolitiTweet.org

Nate Cohn @Nate_Cohn

@tcb_dna taco — PolitiTweet.org

Nate Cohn @Nate_Cohn

Over time, additional data may be added to the needle. For the first time, we will attempt to integrate multiple data sources into a single integrated national model. We're going to iteratively add and validate -- elements of this are essentially untestable in advance — PolitiTweet.org

Nate Cohn @Nate_Cohn

Over time, additional needle may be added to the needle. For the first time, we will attempt to integrate multiple data sources into a single integrated national model. We're going to iteratively add and validate -- elements of this are essentially untestable in advance — PolitiTweet.org

Nate Cohn @Nate_Cohn

'It doesn't know much' has consequences for how its initial expectations compare other forecasts. The House is more uncertain than 538 (despite R>4 pop vote and Dems at 203 seats). Dems slightly favored in Sen (more uncertainty = asymmetrically makes WI/OH/NC in play) — PolitiTweet.org

Nate Cohn @Nate_Cohn

At the moment, the needle is based on county data (like our 2018/2016 needles that modeled the whole country). The most important thing to know about the county needle: it cares a lot about completed votes, and not much else. Right now, it doesn't know much. — PolitiTweet.org

Nate Cohn @Nate_Cohn

The needle is on its way to the internet, after a last typo is fixed. Not surprisingly, it'll open with a tossup in the Senate and the GOP favored in the House — PolitiTweet.org

Nate Cohn @Nate_Cohn

Or put differently: I don't think we will know the Senate tonight if the final results are close to the averages (and this isn't just about the needle) — PolitiTweet.org

Nate Cohn @Nate_Cohn

GA/PA release the detail necessary, but GA could go to a runoff while PA may have so many outstanding votes that we still can't be sure — PolitiTweet.org

Nate Cohn @Nate_Cohn

The late votes in Nevada and Arizona will be numerous and systematically different, but neither state releases the data in a good enough way to make an especially clear estimate using the data alone. — PolitiTweet.org

Nate Cohn @Nate_Cohn

This is the first election year where my best ex ante guess is that the needle will not be able to offer an especially clear answer on the most important question posed to it: Senate control — PolitiTweet.org

Nate Cohn @Nate_Cohn

@KevinQ taco — PolitiTweet.org

Nate Cohn @Nate_Cohn

it is back https://t.co/9HYh3lxK8T — PolitiTweet.org

Nate Cohn @Nate_Cohn

(I should caveat here that I did not get around to formally model the Arizona early vote as it would stand on the first batch, so if Kelly doesn't lead double-digits i wouldn't read much into it either) https://t.co/fybZWhThpf — PolitiTweet.org

Nate Cohn @Nate_Cohn

When early vote is second -- PA -- the mirage may be a little more muted than '20. In many states, Democrats may fare better on Election Day than '20 (with Election Day a larger share, so doesn't matter), meaning a smaller GOP lead before early counted — PolitiTweet.org

Nate Cohn @Nate_Cohn

This is especially true where the early vote is first. It would not be at all surprising to me if Kelly has a double digit lead in Arizona in the first batch of mail votes -- even more extreme than '20. — PolitiTweet.org

Nate Cohn @Nate_Cohn

The Dem early vote hopium has important consequences for 'blue mirage' or 'red mirage' tonight. In many states, the early vote is more Democratic than '20 (but likely a smaller share of the electorate, so don't think anything of it). As a result, 'mirages' can be more extreme — PolitiTweet.org

Nate Cohn @Nate_Cohn

@SpecialPuppy1 https://t.co/FaOXVzgAJB — PolitiTweet.org

Nate Cohn @Nate_Cohn

And then there's the interesting task of evaluating whether any of this offers a credible path toward improving the low-incentive survey data — PolitiTweet.org

Nate Cohn @Nate_Cohn

There's a lot of interesting analysis for us in the months ahead. This data isn't final - people can still mail back envelopes. We don't have the Ipsos/KP data. We don't have the ABS survey joined to the voter file, which may be where the most fun begins — PolitiTweet.org

Nate Cohn @Nate_Cohn

They could both be off! It's pretty sobering that going from 1-->30% response rate still yields a D+7 survey in the raw data (v D+9 in NYT/Siena). https://t.co/ukesF2NPKh — PolitiTweet.org

Nate Cohn @Nate_Cohn

We have one last final bit of polling for you: the preliminary findings from the NYT/Ipsos high-incentive mail survey from Wisconsin, compared to a typical NYT/Siena poll: --A little more GOP --Much more moderate A lot more to come in months ahead https://t.co/856X4UTKBF — PolitiTweet.org

Nate Cohn @Nate_Cohn

@jon_m_rob tho given the pattern in the rest of our polling, probably would make it worse in 2020 #modeeffect — PolitiTweet.org

Nate Cohn @Nate_Cohn

Half the reason the Democrats spent the last two years flagellating themselves over going too far left and so on is because the expectations in '20 were set so high. Similarly, the GOP felt no need to moderate or anything in no small part bc they got to feel great after losing — PolitiTweet.org

Nate Cohn @Nate_Cohn

One interesting consequence of the GOP polls 'flooding the zone' is that the expectations for the Senate is very different than it would be otherwise https://t.co/eR02p8LFHT — PolitiTweet.org