Silver: About 50-60% of national poll leaders at this point in the campaign won.

[highplainsdem]

Nate Silver thread responding to a complaint from Dan Pfeiffer, who apparently isn't very happy about all the national polls that just came out, saying he "will never understand why media organizations continue to do so many national primary polls."

This is the first tweet - full text below:

Link to tweet

One good reason is that they're reasonably predictive of who wins the primary. About 50-60% of national poll leaders at this point in the campaign won.


Another good reason is that they're more reflective of Democratic voters overall, who are 40% nonwhite, than Iowa and New Hampshire, which are very white. If the media truly gave SC/NV coequal billing with IA/NH, that would obviate the need somewhat, but they don't.


p.s. If you look at *all 4* early states as a group, they look quite similar to national polls, especially if you weight by the number of voters or delegates they have. Buttigieg maybe a hair stronger in the early states (although weak in SC) but pretty minor differences overall.

Btw, Pfeiffer is apparently a Warren supporter, judging by the tweets of his that I can recall mentioning her. Which might be why he isn't happy with today's national polls.

[crazytown]

1. cf Dean 2004, HRC 2008.

[Fiendish Thingy]

6. Not to mention the change in primary rules re: delegate assignment

No supers this time until the convention, so these national polls seem to have same odds as a coin toss.

[BeyondGeography]

2. So about half the time they lose

Ok.

[33taw]

3. Exactly.

[George II]

4. I like that thought, but it also means that 40-50% of them lose, too.

[comradebillyboy]

5. I'm pretty much convinced it's going to be Biden or Warren...

[Bradshaw3]

7. That includes incumbents which we don't have this time

The proposition above is even worse when there is not an incumbent. Since 1972, in a year in which a non-incumbent POTUS or vice-POTUS is not running in the Democratic primary - as is the case in 2020 - the leader in the early polls is NOT the nominee in the majority of the cases. The candidate who is sometimes way behind comes out on top. That goes for McGovern in '72, Carter in '76, Clinton in 1992 and Obama in 2008. That's four out of seven that dispute Silver's hypothesis, when incumbency is taken into account.

https://fivethirtyeight.com/features/what-more-than-40-years-of-early-primary-polls-tell-us-about-2020-part-1/



Sorry to burst your bubble.

[Scurrilous]

8. ...

[Stuy]

9. So almost half don't win it

Not an impressive number if you ask me.