Nate Cohn

Nate Cohn @Nate_Cohn

Monmouth in PA at 11AM will be our first live interview poll in a while. It's also nice because PA has been sort of sparsely polled this year, at least by higher quality firms, and the poll averages there have often looked ok for Trump. OTOH, NYT/Siena was Biden+10 — PolitiTweet.org

Nate Cohn @Nate_Cohn

Monmouth in PA at 11AM will be our first live interview poll in a while. It's also nice because PA has been sort of sparsely polled this year, at least by higher quality firms, and the poll averages there have often looked ok for Trump OTOH, NYT/Siena was Biden+10 — PolitiTweet.org

Nate Cohn @Nate_Cohn

Missed these changes! I'd guess @Redistrict won't be too far behind https://t.co/LabNFb0ucJ — PolitiTweet.org

Nate Cohn @Nate_Cohn

As a result, I have no real doubts about whether Biden *could* win a district like TX03, which was Trump+14 last time. That's not to say he will. But I don't see how that's even controversial at this point, and if so it would pbly mean no better than a low single digit House race — PolitiTweet.org

Nate Cohn @Nate_Cohn

These districts don't strike me as very different from an MO02 and IN05, if you factor in the uncertain D upside of rapid demographic change. TX is changing but these districts are transforming. What's going to happen with high turnout and registration in a competitive race? — PolitiTweet.org

Nate Cohn @Nate_Cohn

All of these races were single digits in the House last time, despite no serious contest. All were Cruz+5 or less (TX10 was a Beto win). TX3, last I looked, has the highest college graduate percentage of any GOP held district remaining in the land — PolitiTweet.org

Nate Cohn @Nate_Cohn

With the conclusion of the primaries in Texas last night, I want to flag a tier of districts that haven't gotten a ton of attention from the race raters (perhaps for good reason, I haven't been following the candidates) but seem ripe in this environment: TX3, 6, 10, 25 — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @nickgourevitch: Before diving deep into the schools data, a quick note on the President's standing. His job approval is 39%/59% in our… — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @maggieNYT: The White House isn’t going after Fauci, you see, they’re just letting their trade adviser write an op-ed devoted to critici… — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @AaronBlake: Kansas Congressman Steve Watkins (R) charged with 3 felonies, including: -Voting without being qualified -Unlawful advance… — PolitiTweet.org

Nate Cohn @Nate_Cohn

@nataliemj10 oh i was thinking narrower swaths of online life, perhaps even as narrow as NYT employees on a day like after the Cotton oped — PolitiTweet.org

Nate Cohn @Nate_Cohn

My prior is that the people who choose to participate in these sort of debates are largely unrepresentative of broader opinion, for a lot of reasons. They could easily be vastly more representative than Twitter was in the primary. But maybe they're not. We just don't know — PolitiTweet.org

Nate Cohn @Nate_Cohn

There are many high-profile cultural debates going on right now, including some in journalism. We see the views of the debaters, whether they participate because they're active in the salient forum or because they feel most strongly. Are the debaters representative? Idk. — PolitiTweet.org

Nate Cohn @Nate_Cohn

I think one obvious lesson of the Democratic primary was the extent that Twitter is not representative of the real Democratic electorate. I often find myself wondering whether this lesson extends beyond this website and the Democratic primary https://t.co/kBGZMYSwyW — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @mhugolopez: NEW from @pewresearch: The share of U.S. adults who say illegal immigration is a "very big" problem for the nation is down… — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @ForecasterEnten: I see something similar... But wanna see more data for making too big a thing of it... https://t.co/hLqfwmwdgU https:… — PolitiTweet.org

Nate Cohn @Nate_Cohn

We haven't had very many live interview polls yet in July, but it does seem to me that almost all of the online data is a couple of points better for Trump on approval/ballot than before the holiday break. Will be interesting to see if that holds up — PolitiTweet.org

Nate Cohn @Nate_Cohn

Well we polled NC. We had Cooper 50-39 in Gov, Cunningham 42-39 in the Senate. They had Cooper 53-44 in the Gov and Cunningham 50-41. Now I don't want to go declare we're right, they're wrong. There are possible explanations too. But the pattern's there. https://t.co/AKrAuSZfRD — PolitiTweet.org

Nate Cohn @Nate_Cohn

Anyway, I don't want to be too tough here. This is a challenge for all pollsters and especially state online pollsters working in a tiny state. I think their work is about as good as we're getting from panels nowadays. But it's something that I look at and figured I'd share — PolitiTweet.org

Nate Cohn @Nate_Cohn

I should note that they have had polls--KS, for instance--where they do show healthy ticket splitting. But this isn't the first time I've thought they were on the low-end of that range. An earlier NC poll, with Cooper and Cunningham both +9, has lingered in my mind — PolitiTweet.org

Nate Cohn @Nate_Cohn

I'm pointing this out because of the Civiqs poll in MT this AM. I think the Civiqs folks are doing about as well as anyone to meet the challenge of modeling out bias in online polling. I think it's good stuff. But nonexistent undecideds/ticket splitting is a warning to me — PolitiTweet.org

Nate Cohn @Nate_Cohn

It's totally possible to still produce compelling topline results, even if you have too many highly engaged activists in your polls. Your average error might be still be low, since maybe you have a decent partisan balance even though you have too many activists, etc. — PolitiTweet.org

Nate Cohn @Nate_Cohn

If you show those things, I take it as a sign that you have some normal humans in your poll. Here on political twitter, there are no undecideds and there's no ticket splitting because you're mainly a bunch of ideological consistent partisans. But that's not most of the electorate — PolitiTweet.org

Nate Cohn @Nate_Cohn

In general, I take it as a good sign if a pollster shows a meaty number of undecided voters, particularly as you head down the ballot, and if you show some real gaps between top and bottom of the ticket — PolitiTweet.org

Nate Cohn @Nate_Cohn

I think you can learn a lot about a pollster by comparing races up and down the ballot. Does the pollster have more undecided voters as you head down ballot? Does the pollster show differences in the horse race between different races? — PolitiTweet.org

Nate Cohn @Nate_Cohn

Yeah, the content of the ad does cut against making that appeal at this point, even if they intend to do it in time https://t.co/ObjNDswxdl — PolitiTweet.org

Nate Cohn @Nate_Cohn

@SaveGarrett Yeah, the content of the ad does cut against making the partisan appeal at this point https://t.co/ObjNDswxdl — PolitiTweet.org

Nate Cohn @Nate_Cohn

I thought Biden campaign would go on the air in Texas, but I'm somewhat surprised there isn't an accompanying and explicit 'help us turn tx blue' kind of fundraising appeal after all the money people threw at Beto — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @MediumBuying: Biden wades into Texas with first general election TV ad in state https://t.co/1koqK9kOfO — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @MachinePix: "Stairway to Heaven" ladder in the Salzkammergut resort area of Austria. https://t.co/Wg7BzWeoDp — PolitiTweet.org