CNN's Harry Enten: Why Twitter underestimates Joe Biden

[highplainsdem]

https://www.cnn.com/2019/12/08/politics/twitter-joe-biden-analysis/index.html

According to a 2018 Pew Research study, only 24% of potential Democratic primary voters say they use Twitter. An even smaller 7% of potential Democratic primary voters say they are on Twitter and have used hashtags about a political or social issue. (I'm going to call this group "politically vocal on Twitter." ) If you have ever been on Twitter, this 7% seems to dominate the political conversation. Yet they are only a tiny fraction of Democrats in the real world.

Biden seems to get the ire of a lot of these hashtag users. It shouldn't be too surprising when looking at who Biden's base is.

We'll start with the most obvious divide among Democrats: ideology. Biden cruises with moderate and conservative Democrats and struggles more with liberals. In our last two CNN/SSRS polls, moderate and conservative Democrats were a majority (54%). Among the politically vocal on Twitter in that 2018 Pew poll, they were just 35%. No wonder Medicare for All sells a lot more on Twitter than among all primary voters.

Age has been another big predictor of support in this primary. Biden does his best with voters age 50 and older. Among all potential Democratic voters nationwide, this group has been 50% in our last two CNN polls. In that 2018 Pew poll, those age 50 and older were only 31% of those politically vocal on Twitter. The real world is much more OK with Boomers than Twitter.

Biden also wouldn't be doing well in this primary if not for black voters. They're about 21% of all potential primary voters in our last two polls. They're 15% of those politically vocal on Twitter. Digging down deeper, you could argue Biden's most loyal base is black voters age 50 and older. This group is about 12% of all potential Democratic primary voters nationwide, but are a mere 4% of those politically vocal on Twitter.

Finally, one of Biden's weakest groups is whites with a college degree. This is a group that South Bend, Indiana, Mayor Pete Buttigieg and Warren have been fighting over. They're 28% of all potential primary voters. Among those politically vocal on Twitter, they're 39%.

Add it all together and you can see how Biden can be doing poorly among the politically vocal on Twitter and well in the real world.

-snip-

[redqueen]

1. "potential Democratic primary voters"

I'm always curious how this is determined.

[TwilightZone]

5. They ask people questions and compile the answers.

[redqueen]

9. So does that mean they ask the general public at random,

And then filter our which ones said they're likely to vote in the primary?

[MH1]

14. One way (not necessarily used in the op article)

is to look at historical voting records, which are publicly available.

The simplest is to say that someone who is a 90%+ voter in primaries, historically, is a very likely primary voter. A less, but still likely primary voter would be those over 50% and less than 90%. I just picked those cutoffs, a real analyst in this area would probably have more refined numbers.

That's the easy way that has been available for a long time. Nowadays, there are all sorts of interesting things one can do by correlating other "big data". For example, local trends (maybe stuff indicating how meaningful this election might be). I don't work in this area but I know these things exist and I bet the professionals are cooking up ever better ways to predict "likely voters".

Or maybe they just ask, as people said. If it is a properly determined statistically relevant sample, it's as good as any other properly done poll.

[LakeArenal]

2. I must be the one that proved this rule.?.

“Biden..... struggles more with liberals. In our last two CNN/SSRS polls.”


I am white, old and college educated.

I consider my self very liberal. I want free hot lunch for everybody.

This is why depending on polls bothers me.

Even tho, High, you post the best ones.

Off this topic: I also wonder when who collects the most money became a thing.
Although not my candidate, think constantly reminding folks Harris wasn’t getting $$ became a self- fulfilling event.

Why give money to a loser, no matter how great her ideas and plans are? To me she was the 2nd or third most qualified.

[Thekaspervote]

3. Great breakdown!! Agree completely

[highplainsdem]

7. There are always exceptions, different preferences, among blocs of voters identified as favoring

particular candidates.

Like you, I'd probably fit with the voters most likely to support one of Biden's more leftwing rivals.

But I also consider myself very much a realist, which is why I favor Biden, in addition to admiring him much more than I admire any of his rivals.

As for "when who collects the most money became a thing" -- it's always been, to some extent, for as long as I can recall.

But Nate Silver pointed out that polls and endorsements tend to be more predictive of who'll win.

Drat, can't find what I posted about that earlier.

But here's a thread of Silver's from two days ago about endorsements and fundraising affecting polls, with endorsements being more predictive than money:

Link to tweet

Full text of that thread:

FWIW, I've been working on how to model the impact of money and endorsements on the primary—historically, endorsements are more predictive of polling movement but raising money doesn't hurt—and most of candidates are about where they "should" be in polls given those inputs.


Here's where the model thinks the candidates "should" be in national polls given their fundraising and endorsements:

Biden: 25%
Warren: 13%
Sanders: 11%
Buttigieg: 9%
Klobuchar: 5%
Castro: 5%
Booker: 4%
Yang: 4%
Bloomberg: 4%
Bennet: 4%

Not too far from where they actually are.


Handling the endorsement data is tricky—you need to weight recent endorsements more heavily but also account for how many people are sitting on the sidelines. Anyway, a bit too esoteric for Twitter, but the point is that there isn't a huge gap between polls and "fundamentals".

[TwilightZone]

4. I think pretty much everyone outside of his supporters underestimates him.

He's proven rather resilient in the primary season thus far. Still a few months until the voting starts, of course, but his consistency in the polls has been a significant story this cycle. His demise is continually predicted, especially on places like DU, but so far, it's been mostly wishful thinking, it seems.

"In our last two CNN/SSRS polls, moderate and conservative Democrats were a majority (54%)"

This is in line with other polls the past few years. About half of Democrats ID as moderate or conservative on a fairly consistent basis. Warren and Sanders will have to tap into that if they're going to outlast Biden and, possibly, Buttigieg.

[highplainsdem]

6. +1,000,000

[Hoyt]

8. One can be for MFA, college tuition, etc., but accept the fastest way forward is to beat GOPers,

even if we don't nominate the most "progressive" candidate.

There is nothing progressive about losing the Presidential election or House and/or Senate. That's why I detest the term "progressive" as used nowadays. Just for the heck of it, I don't like "populist" or "populism," either. That helped give us trump.

[Wawannabe]

10. 49

Not on twitter

Voting Biden!

[Gothmog]

11. Why Twitter underestimates Joe Biden - CNNPolitics

I like this article

Link to tweet

[Cha]

12. This tells me what I need to.. they're the smartest Voting Bloc, imv..

Biden also wouldn't be doing well in this primary if not for black voters. They're about 21% of all potential primary voters in our last two polls. They're 15% of those politically vocal on Twitter. Digging down deeper, you could argue Biden's most loyal base is black voters age 50 and older. This group is about 12% of all potential Democratic primary voters nationwide, but are a mere 4% of those politically vocal on Twitter.

Thank you, hpd!

[Gothmog]

13. This makes me smile

Link to tweet