Nate Cohn

Nate Cohn @Nate_Cohn

I don't know exactly what this means (and let's not kid ourselves, they might not either, if this happens at all), but ~ one-third of doctors are foreign born and that's only the first and most obvious example of the problems you run into here https://t.co/eH7UpccXY2 — PolitiTweet.org

Nate Cohn @Nate_Cohn

@joshtpm i mean gwen graham is the one who doesn't fit and deserves the question-mark — PolitiTweet.org

Nate Cohn @Nate_Cohn

But I do not regret taking the opportunity to note that Graham's defeat in the primary would loom awfully large right now if she (or Gillium) had won — PolitiTweet.org

Nate Cohn @Nate_Cohn

the only valid critique of this list is that I snubbed donna shalala for gwen graham — PolitiTweet.org

Nate Cohn @Nate_Cohn

ok who is your top choice for VP — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @Redistrict: Kinda skeptical of arguments that certain former D primary contenders would add a "much-needed jolt of energy/enthusiasm/tu… — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @Neil_Irwin: What a world. — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @clcrozier: I spent a humiliating amount of time learning how to make animated graphs, just to illustrate a fairly obvious point. “Fo… — PolitiTweet.org

Nate Cohn @Nate_Cohn

Anyway, we'll see what this all looks like when we return to the field, whenever that may be — PolitiTweet.org

Nate Cohn @Nate_Cohn

Another interesting cut: the Sanders-not-Biden vote in our poll was Warren 49, Trump 36 in a Warren v. Trump general election matchup. So, you know, you could probably appeal to some of these folks by moving left, but I'm going to guess that's not going to do it for most — PolitiTweet.org

Nate Cohn @Nate_Cohn

Of the 16% of Sanders voters who didn't back Biden against Trump--about 2% of the electorate--43% said they approved of the president's performance and 23% opposed Medicare For All, so this was not obviously a group of like DSA-type Bernie supporters — PolitiTweet.org

Nate Cohn @Nate_Cohn

Way back in October, we did a poll of the top six battleground states and found Biden+2 over Trump and Biden leading Warren, 27 to 19 with Sanders at 16. This pretty core group of Sanders supporters said they favored Biden over Trump by an 84-6 margin. (Warren vote was 93-2) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@ArmenianBacon true, though weighting on 18 to 34 year old men is an order of magnitude different than weighting on 18 to 21 year old black men! — PolitiTweet.org

Nate Cohn @Nate_Cohn

@sfcpoll @PollsterPatrick depends on your purpose, i suppose. totally fine for a topline. probably a problem for 59% of 18 to 29 year olds to be male, like your March poll, if you want to an unbiased account of that group — PolitiTweet.org

Nate Cohn @Nate_Cohn

There's essentially no difference on the topline for RVs, but it does make somewhat more of a difference in terms of likely voters. Biden's lead among those who are "almost certain" to vote shrinks by 0.5 pts if you weight by the interaction — PolitiTweet.org

Nate Cohn @Nate_Cohn

@jbarro i have no serious theories — PolitiTweet.org

Nate Cohn @Nate_Cohn

(Indeed, Trump leads among seniors this way) — PolitiTweet.org

Nate Cohn @Nate_Cohn

The difference on the topline was non-existent (a net-.03% shift). But the tabs do shift a bit. Biden's lead grows by nearly 4 points among voters under age 35 if you use the interaction between age x sex, while Trump's lead grows by over 2 points among voters over age 65 — PolitiTweet.org

Nate Cohn @Nate_Cohn

To test this out, I went and pulled the most recent Monmouth data and tested two very simple sets weights: race, edu, age, sex vs. race, edu, age-x-sex. (cc @PollsterPatrick) — PolitiTweet.org

Nate Cohn @Nate_Cohn

@ThinkUpstream they do, just separately. they make sure they have the right number of young people over all, and right number of women over all, but not necessarily the right number of young women — PolitiTweet.org

Nate Cohn @Nate_Cohn

There are a few polls that take steps to help this--Selzer, Times/Siena, at least some polls fielded by SSRS. But most others, including very good polls like Monmouth or ABC/Post, don't weight on the interaction between age and gender, and you get this bias in the age crosstabs — PolitiTweet.org

Nate Cohn @Nate_Cohn

TLDR: without any special steps to address it, older respondents to telephone surveys are wayyyy too female, helping Biden, and young voters are overwhelmingly male, helping Trump — PolitiTweet.org

Nate Cohn @Nate_Cohn

One interesting trend in recent polls is that Biden seems to be doing a bit better among over 65, and worse among under 30. I think some of it is probably attributable to this: https://t.co/f38QjLa0u9 — PolitiTweet.org

Nate Cohn @Nate_Cohn

@polipi17 @magi_jay exits are wrong https://t.co/xH7subXCRV — PolitiTweet.org

Nate Cohn @Nate_Cohn

@magi_jay they're highly correlated and they're small enough that it's not clear that there are changes at all. and i complied that data myself — PolitiTweet.org

Nate Cohn @Nate_Cohn

@magi_jay it hasn't really dropped much tbh https://t.co/xYu9cn1jxf https://t.co/rcS1XsNdzQ — PolitiTweet.org

Nate Cohn @Nate_Cohn

One misconception in my replies over the last few days: "Trump is faring much worse in well-educated areas than '16, since Dems flipped burbs." What's really happened: there's not much change in attitudes about Trump, but now attitudes about Trump drive down ballot vote choice — PolitiTweet.org

Nate Cohn @Nate_Cohn

RT @Toure: The 7pm clap ritual is loud today! — PolitiTweet.org

Nate Cohn @Nate_Cohn

@BrianStryker @mattyglesias @nbeaudrot @davidshor @spinons @SpecialPuppy1 @beyerstein In Oct/Nov RDD national polls, I've got a tie between Clinton/Trump among LVs in WI/OH/MI/PA/MN/IN/MO/IA, v. Trump+6 IRL. Same polls were Clinton+6 outside of those states v. Clinton+4 IRL — PolitiTweet.org

Nate Cohn @Nate_Cohn

@henryolsenEPPC given that whites and maricopa are such a large share of the electorate, is that really any different than saying don't trust this poll because sinema won by 2 not 9? — PolitiTweet.org