Nate Cohn

Nate Cohn @Nate_Cohn

RT @PatrickRuffini: Remember folks: We are one standard error away from being one standard error away from Trump eking it out in the Electo… — PolitiTweet.org

Nate Cohn @Nate_Cohn

@IChotiner you’re going to get hired at this rate — PolitiTweet.org

Nate Cohn @Nate_Cohn

@RossRichendrfer @meredimw ! when — PolitiTweet.org

Nate Cohn @Nate_Cohn

not *that* far off. it's adjacent. but still, seems odd since the right LD should just be on the voter file — PolitiTweet.org

Nate Cohn @Nate_Cohn

they've got me in the wrong state legislative district somehow, based on the questionnaire — PolitiTweet.org

Nate Cohn @Nate_Cohn

Now there’s a voter file I probably shouldn’t be on anymore https://t.co/52ibVzWKRA — PolitiTweet.org

Nate Cohn @Nate_Cohn

@julia_azari don't we all stand to lose if the only lesson of history is that everyone was evil until college grads 20 years ago, or if the standard for 'good' isn't progress but perfection by the standard of the future, making us just carbon emitting monsters? — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @maggieNYT: Wowww — PolitiTweet.org

Nate Cohn @Nate_Cohn

Here, they choose to weight to a Utah exit poll (which, itself is not weighted by education). That appears to be less educated than their own poll, but still far better educated than the census-based estimates that most state pollsters (us, Marquette, Monmouth) use — PolitiTweet.org

Nate Cohn @Nate_Cohn

An inside look at a state pollster (Y2, Utah) evaluating whether to weight by education They switch in the end, finding a 3-7 pt shift toward Trump when they add education weights (though they don't go as most edu-weighted state polls) https://t.co/Q5fLt2J0QX — PolitiTweet.org

Nate Cohn @Nate_Cohn

And if you look at the wording, Fox doesn't ask 'how' he's doing the job as president' but if you approve of 'the' job he's doing. In the extreme, a respondent could approve of the job he's doing (TRYING to fix covid, i approve!) but think he's doing terribly at it — PolitiTweet.org

Nate Cohn @Nate_Cohn

My working theory is that 'handling' evokes conduct and behavior, while 'doing' evokes actions, policies and the task you're trying to solve. https://t.co/KZdsf5fYvW — PolitiTweet.org

Nate Cohn @Nate_Cohn

@PollsterPatrick It is. There’s no question it lines up, but other things could conceivably explain the difference — PolitiTweet.org

Nate Cohn @Nate_Cohn

One thing I believe--but haven't proven in an experiment--is that the Fox News/Monmouth/NYT question of whether you approve of how the president is 'doing' his job yields far better results for Trump than how he's 'handling' his job (Gallup/most others) https://t.co/jHipGh0bbT — PolitiTweet.org

Nate Cohn @Nate_Cohn

i mean the fox poll definitely wasn't better for trump than the quinnipiac poll, so this tweet stands strong https://t.co/eK0IwMyAsn — PolitiTweet.org

Nate Cohn @Nate_Cohn

I could argue this is Trump's best poll in three weeks, given that they've tended to have a pretty Dem house effect https://t.co/MIVcoPhjJW — PolitiTweet.org

Nate Cohn @Nate_Cohn

Now, what works on paper doesn't may not work IRL. For instance, the 'low profile in Senate' weakness (which is probably a strength IMO) could be reinterpreted as 'probably not as great of a politician as her bio/resume.' But when you're up 10, "do no harm" weighs heavily — PolitiTweet.org

Nate Cohn @Nate_Cohn

When I read through the @kkondik 'pros' and 'cons' for each Biden VP, it seems easy to imagine the Biden folks reaching the end of a long VP hunt and simply asking "why not Duckworth?" and struggling to come up with an answer https://t.co/2R9WGNmNKP — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @MULawPoll: Heads up. New @MULawPoll will be out next Wednesday, June 24th. Video at 12:15, results posted shortly after 12:30. I’ll p… — PolitiTweet.org

Nate Cohn @Nate_Cohn

@DanRosenheck it has potential to induce bias if, for instance, voters are less likely to report voting for the loser, more likely to say they voted for some other candidate, more likely to report voting, and so on. there are other practical challenges — PolitiTweet.org

Nate Cohn @Nate_Cohn

@QuickOne76 it's the poll in the article — PolitiTweet.org

Nate Cohn @Nate_Cohn

This could easily be a big deal again if we wind up back in a spot where Biden's up 6 in national RV polls and up 5 in a bunch of battleground state polls unweighted by education. But that's not where we're at right now — PolitiTweet.org

Nate Cohn @Nate_Cohn

I'm pretty sympathetic to many of the points in this article--many of which I've argued before. But Biden's lead is big enough that it's not terribly material. Like, ok, maybe Biden's up 12 instead of 16 in Michigan once you weight by education. https://t.co/zDQp3H6kU0 — PolitiTweet.org

Nate Cohn @Nate_Cohn

@b_schaffner @DataProgress true, though fwiw most polls don't have biden gaining so much up to 54 (if he was at 50 i think you could more reasonably speculate about a simple trump--> undec pattern) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@BenjySarlin i don't think there's clear evidence that the economy is hurting him, exactly, though on the other hand it is probably mitigating his best advantage — PolitiTweet.org

Nate Cohn @Nate_Cohn

In this case, there's an argument that mean-reversion would involve unemployment dropping 10 pts, COVID going away, and BLM net-fav dropping 10 or 15 pts https://t.co/cmLbVZ6U3T — PolitiTweet.org

Nate Cohn @Nate_Cohn

My own view is it's not social desirability. Look at the Ipsos poll this morning. It's a good poll that's done online. I don't think Trump voters would be embarrassed to merely 'click' Trump. You're not talking to some interviewer judging you. But Trump's at 35 and down 13! — PolitiTweet.org

Nate Cohn @Nate_Cohn

Given that Trump's support has dropped, I'd guess there are probably Trump voters who currently say they're undecided. Whether that's because of social desirability or because, in this moment, they don't want to vote for Trump is another question https://t.co/M8SHD0b9mD — PolitiTweet.org

Nate Cohn @Nate_Cohn

The last national poll to show Trump within a mere 7 points was the Emerson College poll fielded 6/1-6/2. Since then, the average national poll (n=17) is Biden +10.6 — PolitiTweet.org

Nate Cohn @Nate_Cohn

But it is noteworthy if a pollster that claims to be weighting on self-reported vote and yet winds up with a sample that's 4 pts off the target/their last poll on self-reported vote, while the topline shifts by a similar amount — PolitiTweet.org