Nate Cohn

Nate Cohn @Nate_Cohn

RT @davidshor: @xenocryptsite @Nate_Cohn This is also a fun one! https://t.co/tg1ozYYWtX — PolitiTweet.org

Nate Cohn @Nate_Cohn

the upshot? astrology is fake. don't @ me — PolitiTweet.org

Nate Cohn @Nate_Cohn

They're also quite old, which confirms to my satisfaction that it's just a place where incomplete birthdates are going https://t.co/lTGT90bSHN — PolitiTweet.org

Nate Cohn @Nate_Cohn

The 1/1 folks are also very old, which also adds up https://t.co/JnIiBbJNfi — PolitiTweet.org

Nate Cohn @Nate_Cohn

After some examination, I've determined that the seemingly notable Dem-lean of Capircorn is exclusively a product of people born on 1/1, who are disproportionately nonwhite and Democratic. I'd guess its an artifact of the registration data (default blanks to 1/1) — PolitiTweet.org

Nate Cohn @Nate_Cohn

I won't respond to any questions about why this chart exists. But, now that it does... https://t.co/lyfbC27E19 — PolitiTweet.org

Nate Cohn @Nate_Cohn

The new state polls are of, shall we say, uneven quality. But they're quite consistent and so is the trend in Trump approval. It'll be interesting to see whether this shows up in the net wave of national polls — PolitiTweet.org

Nate Cohn @Nate_Cohn

Well, Biden was up 2 among RVs and 1 among LVs in states Trump won by 1.6. So, that's basically the same picture from the national polls today and not far from state polls conducted until about one week ago, which have since lurched toward Biden https://t.co/WCk82UFThC — PolitiTweet.org

Nate Cohn @Nate_Cohn

And this is pre-coronavirus, of course. Can confirm the party registration distributions on older voters are unbiased over all for '19, too (can't for 16, since it includes data from RDD polls) — PolitiTweet.org

Nate Cohn @Nate_Cohn

Here's what that looks like, but for age--the first column being voters by 2016 age, the second column being the age of voters in 2019 (so a 62 year old in 2016 counts as 65) https://t.co/A0Amw4ctWQ — PolitiTweet.org

Nate Cohn @Nate_Cohn

You may recall that the Times/Siena polls back in October had this accompanying chart showing the shift in results by demographic group https://t.co/0yNXE9EUdo — PolitiTweet.org

Nate Cohn @Nate_Cohn

And the state pollsters, as a whole, are extremely heterogeneous. NBC/WSJ, CBS/NYT, ABC/Post polls are all more or less the same commodity; NBC/Marist, NYT/Siena, CBS/YouGov are nothing alike. And then there's the huge pool of largely unknown state poll mush — PolitiTweet.org

Nate Cohn @Nate_Cohn

And look, *in general* the state polling is lower in quality than the national polling right now. There are state pollsters who do crazy things, imo, like weighting on self-reported party reg or not even asking about education. State online polling is hard, due to few panelists — PolitiTweet.org

Nate Cohn @Nate_Cohn

That said, there are many things that have not changed. A very substantial share of state polls still aren't adjusting on self-reported education. Virtually none ensure they represent sporadic/low turnout voters, who are stronger for Trump controlling for other demographics. — PolitiTweet.org

Nate Cohn @Nate_Cohn

Another thing has definitely changed for the better, at least in some cases: more pollsters weighting on self-reported education; failing to do so was considered major driver of the polling error. Nowhere near all do this, to be sure. But, still more than '16. — PolitiTweet.org

Nate Cohn @Nate_Cohn

One thing definitely changed for the better: fewer undecided voters, esp in the Midwest. So, even if the polls are just as bad, they'd still be more accurate. Indeed, the Midwestern polls today are better for Trump than '16. Many have found him up in WI, for ex., vs. 0 in 16' — PolitiTweet.org

Nate Cohn @Nate_Cohn

It's a good question, and there isn't an overwhelmingly clear answer, but here are a few thoughts https://t.co/yISvB1Q8mS — PolitiTweet.org

Nate Cohn @Nate_Cohn

And, importantly as far as I'm concerned, last week of polls is beginning to go farther than just undoing Trump's initial covid/stimulus approval bump — PolitiTweet.org

Nate Cohn @Nate_Cohn

And, importantly as far as I'm concerned, the polls over the last week of polls is beginning to further than just undoing what was left of Trump's initial stimulus/coronavirus bump — PolitiTweet.org

Nate Cohn @Nate_Cohn

Very clear negative trendline for Trump on these approval polls over the last two-to-three weeks, with the Navigator Tracker only this morning's newest example https://t.co/sXaPaPM7bQ — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @billscher: Kamala Harris feels like the Joe Biden of the VP race. A little too obvious, imperfect on the campaign trail, veering betwee… — PolitiTweet.org

Nate Cohn @Nate_Cohn

And that's before you consider the uncertainty in the forecast itself, which can't possibly be particularly confident about what's going to happen by Q3 given all of the external factors — PolitiTweet.org

Nate Cohn @Nate_Cohn

But look, I'm not at all contending that this is sure to be good for Trump! I'm just saying this is a highly sensitive and difficult to evaluate situation if you have a quick V-type recovery right ahead of an election. It could be a disaster for Trump; it might not be so bad. — PolitiTweet.org

Nate Cohn @Nate_Cohn

It's not simply GDP matters and unemployment doesn't. It's that the rate of change matters. You can feel optimistic if you're convinced things are getting better, even if they're bad in absolute terms (FDR '36, as just one ex) https://t.co/DSqXNaAm5i — PolitiTweet.org

Nate Cohn @Nate_Cohn

Now, I don't think you'd find anyone who argues that the only thing that matters is the rate of growth over the last month. But it is a tough case if you have a surging economy, over a short period, that's still bad in absolute terms. It's hard to know what that would feel like — PolitiTweet.org

Nate Cohn @Nate_Cohn

This would be so tough because it really pushes the limits of two general assumptions about the role of the economy : --Voters are fairly myopic, and focused on the most recent events --Rate of change matters far more than the absolute state of the economy — PolitiTweet.org

Nate Cohn @Nate_Cohn

@ForecasterEnten If I were president, I would actually make sure the economy did as bad as possible in all other quarters of my term so I could maximize annual growth in Q2 — PolitiTweet.org

Nate Cohn @Nate_Cohn

This is the exactly the sort of reason why it's so hard to figure out what the national political environment is going to look like in November https://t.co/hfLd2ervMR — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @brahmresnik: One month ago... https://t.co/G3w1QkbJOU — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @TheOnion: Man Just Buying One Of Every Cleaning Product In Case Trump Announces It’s Coronavirus Cure https://t.co/3bKatQGz04 https://t… — PolitiTweet.org