Wondering why all the polls were wrong about the Iowa Caucus winner.

[sagesnow]

Don't remember any poll showing Buttigieg winning. Monmouth was most accurate- they reported the Caucus was up for grabs 5 days before the Caucus.

http://www.borntorunthenumbers.com/2020/02/btrtn-iowa-caucuses-prediction-bernie.html

https://www.democraticunderground.com/1287446363

https://www.democraticunderground.com/1287455724

[Mike Nelson]

1. Hmm..

... there was also a top tier poll cancelled when they found out if wasn't counting Pete... I'm not one for "tin foil hats," but I have to say they have a narrative, if they want one. I should add: I'm not one who thinks there are ultra-sinister forces at work, other than candidates want to win and use the advantages which are there. I think the "caucus" system is un-Democratic, but I don't hold it against candidates that were able to take advantage of the system in place.

[Loki Liesmith]

2. The cancelled DMR poll had Buttigieg in third

And it was cancelled because one caller had forgotten to include Buttigieg in at least one interview. It wasn’t cancelled because it went against the narrative. It was cancelled because Ann Selzer, the pollster, couldn’t assure its quality.

That said it was very close on all other candidates.

[riversedge]

4. glad the DM register kept its credibility.

[Mike Nelson]

7. Thank you for...

... explaining why the poll was cancelled. In all the reports, I had not seen any reasons... To restate, I am not one that believes most expository narratives... I think most foul-ups are just that... I know some have sinister origins, but most do not and I don't think this Iowa story has a "manipulative" narrative. If there does turn out to be mischief, I don't think it's from the Democrats involved.

[Scurrilous]

12. That 'top tier' poll was the one that gave Bernie his first lead in a previous iteration (Jan 10).

Iowa poll shows a tight four-way race in final weeks before the Democratic caucuses

https://www.cnn.com/2020/01/10/politics/iowa-poll-democrats-2020/index.html

Prior to that poll Bernie never led. Warren, Biden, and Buttigieg did, but never Bernie. It placed previous leader Buttigieg third.

And now we have a race. Yay!


x =


Eve of the endlessly hyped Iowa caucus, CNN (DMR's partner in this thimblerig), schedules an entire hour to striptease the newest poll results a la Survivor Finale et al.

On the eve! The night before the fucking fuckers go to their fucked-up square dance for delegates. That particular poll. The now proven ass surge precipitator.



Need more? Trusty NH, down w/ surge since 2016, will get its snowy primary results stripteased on a hour long live version of Roseanne's old show.

Narrative ? Manipulation?

[Mike Nelson]

15. Now, that I agree with...

... the Media is at work, yes. I do think the media would prefer a Trump v Bernie race over a Trump v Biden race. And, they want conflict... it means ratings. They are making much more of the Iowa voting confusion, too... with the race so close between Pete and Bernie, it might have gone one for days, anyway. They were correct, perhaps, not to call it right away.

[Scurrilous]

16. They do it because they can, a public flexing of what they can do.

Goes over big in meetings with ad execs who admire and value that sort of thing.

Trump vs Bernie would be ratings gold for them. The stuff of fables. Raging capitalistic monstrosity slaying the socialist peter pan promisebot, laying waste to his youthful crusade and cruelly crushing its every aspiration.

[MasonDreams]

25. I'm sure you know, after 37,707 posts, that calling Iowa caucus voters fucking fuckers,

and referring to the Iowa caucus as a fucked up square dance is disrespectful and inappropriate.

[Recursion]

3. And the caucus was indeed up for grabs

Buttigieg won a majority of SDEs, and Sanders won a majority of first-round alignments, and the two of them tied on delegates. The undecideds broke against Biden.

[apcalc]

5. Caucuses generally represent party activists, well organized groups,

Or those people who have time, are well abled etc.

Good polling requires good sampling. If your sample consists of Democrats in general, it likely will not show the same result.

[MasonDreams]

6. Do we have the results yet? Last I saw with 96% in Pete's lead was only 0.1%.

Congratulations to Mayor Pete for coming in first or second.

[Donkees]

14. A dowser will be reporting shortly ...

[bottomofthehill]

8. Because it is incredibly difficult to poll caucuses

Tou don’t know who will be viable where you are and where the votes may go. It is just a shitty system. Vote with paper back up

[hlthe2b]

10. Exactly. Anyone who had ever participated in a caucus would realize what folly it is poll on them

Beyond being nearly impossible to collect a representative sample on who will actually show up in a pre-poll, all things can change when one arrives and attendees come under the collective pressure of their peers.

No, the polls aren't reflecting some damned conspiracy. They are just reflecting the fact that without one hell of a lot of "luck," they can never accurately measure caucus behavior as they would anonymous primary vote behavior.

[FBaggins]

9. They weren't

The results were entirely consistent with how most analysts were interpreting the polls.

It’s well understood that standard polling methodology is not well matched to caucus environments... with the normal expectation that candidates with strong local organizations and more-energized supporters tend to outperform.

It was anticipated (for months) that Bernie/Pete were likely to outperform while Biden would underperform in IA. I suspect that most analysts would say that if you ran the state over again as a primary, the results would match up pretty well to the polls.

[Laelth]

32. +1

The Iowa results that we have so far do not shock me.

-Laelth

[BlueMTexpat]

11. This is one reason why

Warren doesn't rely wholly on polls.

But, as explained by some respondents, these polls really weren't THAT far off AND polling caucus results is literally a fool's game.

[Iliyah]

28. At this point in time I don't blame her . . .

Actually I think a few other candidate don't reply on poll except for their internal one(s).

[BlueMTexpat]

33. That makes more sense. eom

[bearsfootball516]

13. It's extremely difficult to poll a caucus, much harder than a normal election.

It's tough to determine who's going to show up, and then trying to nail down someone's 2nd or 3rd choice if their original choice isn't viable? Good luck.

[mgcgulfcoast]

17. Sanders got the most votes

The party is headed for another 2016 if he is screwed over again.

[sagesnow]

18. Do you mean he won the popular vote but lost the Caucus?

It is very difficult to see the results on the IDP page. I cannot get my computer to scroll over. Is there a better way to see the results
https://results.thecaucuses.org/

[R B Garr]

19. He lost by millions -- hardly getting "screwed" over. nt

[thesquanderer]

20. I think he's referring to Iowa, which was also controversial in 2016. (n/t)

[thesquanderer]

21. re: "Don't remember any poll showing Buttigieg winning. "

One of the last polls was the Jan 28-20 poll from David Binder Research, which showed Buttigieg first with 19%, Sanders a pretty close second at 17%, Warren and Biden tied for third at 15%.

[WhiskeyGrinder]

22. Because polls measure individual people, and caucuses measure sentiment across a group that is

working both with and against itself internally.

[HarlanPepper]

23. Bernie was supposed to win big.

That was the Twitter narrative from his followers and media surrogates.

[ibegurpard]

24. polls don't do well with caucuses

when you have to factor in that the winner of a caucus is the result of several rounds of voting that include discarding "non-viables", realignments and second and third choices.
The only thing you should compare to any pre-caucus Iowa polls are the available first round numbers if possible.

[Hortensis]

26. Deja vu all over again. Suspect serious foreign and RW interference.

[JudyM]

29. There's a paper trail...

[lunasun]

27. Pete overperformed Biden underperformed and Bernie did not break away from others to win

So many candidates it’s hard to poll their second choice which counts in IA imo

[Tweedy]

30. The polls were right

The top two are neck and neck.

[Drunken Irishman]

31. Unless a candidate has a very sizable lead, caucus polling is flawed.

What the polling told us was that Bernie had a lot of strong support but the gap between him and others was so minimal that it essentially became impossible to predict the outcome. Remember, caucuses aren't like going to the polls and voting. Candidates have to be viable at a specific level and if they're not, those candidates get nothing.

For that reason, caucuses are extremely difficult to poll. Even in 2016, the polling was off - as Hillary was supposed to win by a wider margin. In 2008, it was a toss-up and Obama ended up out-performing most polls.