About these polls: I think CBS/YouGov has it figured out.

[Buzz Clik]

Last edited Sun Mar 13, 2016, 03:05 PM - Edit history (1)

Not everyone likes polls, but a well designed poll can be highly informative. Michigan last week on the Dem side did not have good polling numbers, for a bunch of reasons. After standing back, digging in, and discussing this, it was clear that Michigan was NOT an anomaly but a warning about poorly constructed polls.

With about 1/3 of the states' primaries in our rear view mirror and lots of polling data, the pollsters are starting to figure it out. Specifically, CBS/YouGov seems to be nailing it: they have Sanders leading Clinton by 2% in Illinois, a state that fits the Michigan model: an open primary with similar demographics.

I'm going by the assumption that CBS/YouGov has good numbers. So, let's take a look at their poll numbers:

Illinois: Clinton 46%, Sanders 48%

Florida: Clinton 62%, Sanders 34%

Ohio: Clinton 52%, Sanders 43%
(I cannot find any other YouGov polls for this Tuesday)

Resulting delegate split: Clinton 279, Sanders 234
(I'm giving Sanders ALL the undecideds, so this is best case scenario for Sanders with these poll numbers)

===========

Giving Missouri to Sanders by 10% and NC to Clinton by 10%, those delegates get split Clinton 91, Sanders 88

My projected total for Tuesday only: Clinton 370; Sanders 321

Running total (pledged only): Clinton 1,145; Sanders 871

[idea5]

1. How did CBS/YouGov do on MI?

Were they accurate there?

[Buzz Clik]

2. As explained in the OP...

... all the polls missed. CBS/YouGov seems to have it figured out.

[Barack_America]

12. Based on....???

[Buzz Clik]

21. Based on the fact the the CBS/YouGov made a huge correction

They initially had HRC with a double digit lead, just as all the polls did before Michigan. The new results have Sanders in the lead due to a more appropriate polling method.

Time will tell.

[Arazi]

4. No, they had Clinton up by 11

http://www.realclearpolitics.com/epolls/2016/president/mi/michigan_democratic_presidential_primary-5224.html#polls

I'm not sure why CBS is suddenly the credible one

[idea5]

6. I see.

If were interested in trends and movements, let's wait to see if a pollster can pull together a string of correct predictions .

[Buzz Clik]

8. CBS/YouGov is the poll be hailed and celebrated by the Berners all over DU.

They also seem to have the concept correct. Tuesday will tell.

[Arazi]

9. Not this "berner". I don't think anything is predictable this primary

[Buzz Clik]

23. The polls were decent up until Michigan.

[Garrett78]

3. If Clinton increases her margin by 'just' 51 on Tuesday...

...that will be considered a very good day for Sanders. That exemplifies just how much of an uphill battle he faces.

[Dem2]

5. CBS/YouGov shows 48-46

here: http://elections.huffingtonpost.com/pollster/polls/cbs-yougov-24062

But it's interesting that the polls are now attempting to reflect the reality on the ground, we'll see soon enough.

[Buzz Clik]

7. I updated and made the correction. One delegate less for Clinton...

The reality on the ground is that polling "Democrats likely to vote" was a miss. I believe the qualifier is now "registered voter likely to vote in the Democratic primary." A very important difference in an open primary.

[Dem2]

10. Seems to make a huge difference in open primaries

Glad they learned their lesson, we'll see in 2 days if this is accurate.

[Buzz Clik]

13. Pollster.com welcomed the change -- the canned the rest of the polls to go with CBS/YouGov

Good move.

[Barack_America]

11. I'll wait until one of them actually gets a result right...

...before proclaims any poll has "got it".

[Buzz Clik]

15. Most of the polls get it close. MI was interesting because they were all wrong.

[Barack_America]

16. Actual most of the polls have severely underestimated Sanders' support.

By a factor of 10-15%.

[Buzz Clik]

17. Not, not really.

In the states that Sanders has won, the polls underestimated his margin of victory by an average of 15%. Removing Minnesota from that equation, and the underestimation is only 7%.

In the states that Hillary has won, the polls have underestimated her margin by 12%. That is, Sanders performed worse than predicted.

So, the polls are not skewed toward either candidate; they simply underestimate the margin of victory.

[DCBob]

14. Well.. even if it ends up that way the Hillary team would be happy.

As long as Bernie doesnt get any big wins he cant win.

[november3rd]

19. Did you hear about what happened in Iowa?

A bunch of Hillary's county delegates didn't show up to vote on the state delegation for the national convention. Bernie came out on top.

[Joe the Revelator]

20. If she goes to the convention without having enough pledged delegates, she's done

She will be unviable as a candidate. I don't know if Bernie gets the nomination, or someone like Biden, but I do know that if Hillary can't close out Bernie, she won't be the nominee.

[Buzz Clik]

22. How many is "enough"?

[Garrett78]

24. One of them will get a majority of the pledged delegates, given that there are only 2 candidates.

Clinton might even get to 2383 without superdelegates, though that's a stretch. I think much will depend on just how well she does on Tuesday, as a sweep of all 5 states (unlikely at this point) would devastate the Sanders campaign.

If it ends up being really close (e.g., 2030-2021), then the superdelegates will play a role. If Sanders wins, it's more likely that it'll be a close finish. If Clinton wins, it's more likely that she'll win by a pretty big margin. I say that because of her 200+ lead as I type this, and also because she's still favored to win the big states going forward. Whereas it'll probably take until June for Sanders to surpass Clinton, if that even happens, which is doubtful. But Tuesday will tell us a lot.

[oasis]

18. Kick

[Buzz Clik]

25. Poll updates:

PPP has submitted a poll that shows Clinton with a slight lead in Illinois. I think we can assume Illinois will be tight.

Pollster.com added PPP to their polls, but have not changed their assessment of Illinois -- the continue to give the nod to Sanders by 3%.

Using PPP's polling data would give Clinton another few delegates, but I'm not changing my numbers right now.

Keep in mind that the OP is the best case scenario using the polling numbers and giving all the undecided voters to Sanders.