Nate Cohn

Nate Cohn @Nate_Cohn

Even the black turnout of 2008 involves black turnout going from... what 60 to 65 percent. That's a big deal that could totally bias a poll of NC. It's also on a vastly different scale than doubling turnout in TX23 or something — PolitiTweet.org

Nate Cohn @Nate_Cohn

That said, I do think there's a lot less reason to be worried about this particular kind of bias in 2020. 2018 turnout surge was truly huge; 2014 was rock bottom, 2020 was a record. It's a real worst case scenario. Presidential elections don't have that kind of shift. — PolitiTweet.org

Nate Cohn @Nate_Cohn

For all we know, we'll be biased toward Dems in these same spots in 2020. I'd argue the two AZ polls we've done so far were unusually strong for Biden, v other firms, at those points of the race. What if we have a Dem bias there obscured by high turnout in '18? IDK — PolitiTweet.org

Nate Cohn @Nate_Cohn

If it's sampling, then we'd be persistently biased against Democrats in the Southwest. Turnout, on the other hand, can go either way from race to race. We could overestimate Latino turnout in 2022 if Biden wins and be biased toward Dems in the exact same spots. — PolitiTweet.org

Nate Cohn @Nate_Cohn

Now, I'm not saying we should get credit for showing a Heller defeat. But for a pollster, there's a big difference between getting a group wrong because you're sampling them poorly or because your electorate is wrong. — PolitiTweet.org

Nate Cohn @Nate_Cohn

It turns out, maybe not surprisingly, that our polls in all of these places would have been very good, let alone merely better, if we got the turnout right. We would have had Heller behind in NV with the 2018 electorate, for ex. — PolitiTweet.org

Nate Cohn @Nate_Cohn

Truth is, it didn't really happen--particularly among Latinos. Even weeks before the election, low turnout Latino voters just weren't indicating their intention to vote in the kind of numbers that would move our estimates. They were also likelier to be undecided — PolitiTweet.org

Nate Cohn @Nate_Cohn

But as I wrote in the next graf, the hope was that voters would nudge us out of the issue https://t.co/8AuJWyRoR9 — PolitiTweet.org

Nate Cohn @Nate_Cohn

This was fairly predictable, tbh, as I wrote before live polling again https://t.co/Ge6MENDqK1 — PolitiTweet.org

Nate Cohn @Nate_Cohn

2018 was really tough on the 'track record of voting' stuff, since midterm turnout was wayyy higher than ever before, but especially in CA/TX, where midterms are usually totally noncompetitive and turnout increased by wild amounts. We baked some increase in, but not nearly enough — PolitiTweet.org

Nate Cohn @Nate_Cohn

Our turnout estimates have two components, basically: whether voters have a track record of voting, and whether they say they'll vote. Neither got the job done, at least in 2018 — PolitiTweet.org

Nate Cohn @Nate_Cohn

In the post-election analysis, the explanation is quite straightforward: we systematically underestimated turnout, and particularly Hispanic turnout. And by a *lot* in some districts. — PolitiTweet.org

Nate Cohn @Nate_Cohn

As a general rule, our polls fared quite well in 2018 but our results in the heavily Hispanic southwest were not nearly so good. Not all of them. But the record there was plainly worse than in the rest of the country, where the average bias was ~0 and average error < 3 — PolitiTweet.org

Nate Cohn @Nate_Cohn

I wouldn't call it a flub; it's a ~6 pt error more than 3 weeks out! But it is an interesting question https://t.co/VR6b2HaAZq — PolitiTweet.org

Nate Cohn @Nate_Cohn

@jon_m_rob i fear that this is going to need to change a lot in response to mail and i'm a little concerned about it. for now, in practice, VF winds up dominating over self-report if you're a high turnout voter, but self-report can move it if you're a low turnout voter — PolitiTweet.org

Nate Cohn @Nate_Cohn

@jon_m_rob for now, we're sampling the whole RV universe, raking to a modeled electorate with the turnout score as a first stage weight, estimating turnout as a function of both self report and VF, and then the final weight is rake weight * estimated turnout/modeled — PolitiTweet.org

Nate Cohn @Nate_Cohn

As an aside, you really do learn a lot about polling a state after you've actually done it! They're not all the same and it does make a bit of a difference to have some prior data and experience there — PolitiTweet.org

Nate Cohn @Nate_Cohn

Honestly it's the same as last time, but a few notes: --we're switching to LVs --this was our first time polling NH (and I should note we've polled basically everywhere!) and it's a much rougher voter file than I had assumed! low productivity; missing age. The Michigan of the NE — PolitiTweet.org

Nate Cohn @Nate_Cohn

Well they're not all done yet so I don't really have an opportunity this time around https://t.co/rfZE4DXtct — PolitiTweet.org

Nate Cohn @Nate_Cohn

@B_M_Finnigan it's the very last state on the list — PolitiTweet.org

Nate Cohn @Nate_Cohn

@micahcohen who said they were all done? — PolitiTweet.org

Nate Cohn @Nate_Cohn

We'll be in the field more-or-less continuously between now and the election, so there's a lot more to come — PolitiTweet.org

Nate Cohn @Nate_Cohn

Tomorrow AM: NYT/Siena in Wisconsin, Minnesota, Nevada and New Hampshire — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @afreedma: I didn't think the satellite images of the West Coast fires could get more jaw-dropping and alarming. I was wrong. The smoke… — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @AP: BREAKING: Authorities say about 500,000 have been forced to evacuate due to Oregon wildfires — roughly 10% of state's population.… — PolitiTweet.org

Nate Cohn @Nate_Cohn

This seems fairly testable by the folks at Pew or USC, though even then panel attrition is a real and expected thing and I'm not sure whether we could read something more generalizable into anything they saw. It would still be an interesting data point imo — PolitiTweet.org

Nate Cohn @Nate_Cohn

How do you get enough Trump vote in '16, but too much Trump-Biden vote in 20? It would need to be an awfully targeted kind of bias, though I could make up a story where committed Trump voters become disillusioned with polling and drop out while weaker Trump support remains. — PolitiTweet.org

Nate Cohn @Nate_Cohn

It is pretty tough, IMO, to come up with an explanation for how a panel like the Pew ATP or USC (weighting issues aside) would be pretty good in '16, retain many of the same panelists, and be badly biased in '20. — PolitiTweet.org

Nate Cohn @Nate_Cohn

One thing that's clear: an apples-to-apples comparison, today, unmistakably shows Biden doing better than Clinton, including among the kind of white, working class voters where we might expect problems. Any issues would almost certainly need to have gotten much worse since '16 — PolitiTweet.org

Nate Cohn @Nate_Cohn

Now for that same reason, there's just not a ton of evidence that this is a real problem, either. Education + undecided voters + turnout covers a whole lot of what went wrong in '16 and 2018 was good enough for most purposes. But I don't see how you can disprove it, either — PolitiTweet.org