Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search
44 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Nate Silver's 538 Puts Nevada Caucus Odds at 50/50 for Hillary & Bernie (Original Post) JudyM Feb 2016 OP
Nice. Jefferson23 Feb 2016 #1
Hah! Trajan Feb 2016 #2
That explains a lot. Hassin Bin Sober Feb 2016 #3
Sure does... nt Purveyor Feb 2016 #10
boy howdy! frylock Feb 2016 #36
Oooo that's a horse of a different color... nt Purveyor Feb 2016 #4
Considering that a suspect Republican poll has swung the average KingFlorez Feb 2016 #5
538 weights the polls according to their methodological soundness. JudyM Feb 2016 #8
The suspect poll still swung the average KingFlorez Feb 2016 #11
These guys are legitimate. Statisticians, not Jonathan Capehart. JudyM Feb 2016 #13
That poll is weighted at 100% CajunBlazer Feb 2016 #29
It's not even an average of polls - the "tie" is only based one the suspect poll CajunBlazer Feb 2016 #18
52 Bernie 47 Hillary the rest go O'Malley and Vermin Supreme. Kalidurga Feb 2016 #6
Oopsie pintobean Feb 2016 #23
To the alerter: Reading is fundamental. PonyUp Feb 2016 #25
LOL Kalidurga Feb 2016 #32
! frylock Feb 2016 #37
Another stupid hide request to shut down discourse--all because someone can't comprehend? TheBlackAdder Feb 2016 #39
Uh oh...heads are gonna roll!! n/t PonyUp Feb 2016 #7
this is very interesting.... kgnu_fan Feb 2016 #9
I wouldn't get too too excited SheenaR Feb 2016 #12
it's just one state 6chars Feb 2016 #35
Check back at Silver's site after the next Nevada poll comes in. CajunBlazer Feb 2016 #14
It was based on 7 polls. I was posting the update, but I don't think it's gospel. Polls are JudyM Feb 2016 #15
No, sorry, you're reading that wrong CajunBlazer Feb 2016 #16
Yes, I saw that before. As he notes, he used 7 polls, I'm assuming he's not misstating. JudyM Feb 2016 #17
Sorry, you're simply wrong! CajunBlazer Feb 2016 #22
For the record I still think Nate is fucking it up. joshcryer Feb 2016 #33
We are funding a local group to go to NV from S. Oregon bkkyosemite Feb 2016 #19
Message auto-removed Name removed Feb 2016 #20
Absolutely not! At least not this Clinton supporter CajunBlazer Feb 2016 #26
Nobody, even the admired N. Silver can sadoldgirl Feb 2016 #21
+1 eom Arazi Feb 2016 #28
No, not hedging his bets - it's based only on one poll CajunBlazer Feb 2016 #30
Not enough polling TSIAS Feb 2016 #24
Agreed, not enough polling yet CajunBlazer Feb 2016 #31
k & r (nt) pat_k Feb 2016 #27
I'm not surprised. kiva Feb 2016 #34
Lucky you. Wish I was there. 840high Feb 2016 #38
It was a great event, kiva Feb 2016 #42
I can't see how Hillary can pull off a win. delrem Feb 2016 #40
Where is that little emoticon that just shakes his head when you need him CajunBlazer Feb 2016 #41
538's got a cushy gig: take the poll numbers, subtract them from 50%, MisterP Feb 2016 #43
Have a recent example? JudyM Feb 2016 #44

KingFlorez

(12,689 posts)
5. Considering that a suspect Republican poll has swung the average
Fri Feb 12, 2016, 11:24 PM
Feb 2016

I don't think the odds are reliable. We need more polling that is actually not full of push questions from the GOP.

KingFlorez

(12,689 posts)
11. The suspect poll still swung the average
Fri Feb 12, 2016, 11:30 PM
Feb 2016

The graph clearly shows that. We need more reliable polling, period. This pollster worked for Mitt Romney of all people, so I think we need to be careful about trusting it. The race could be close, but Republican pollster are sketchy.

JudyM

(29,251 posts)
13. These guys are legitimate. Statisticians, not Jonathan Capehart.
Sat Feb 13, 2016, 12:03 AM
Feb 2016

Yes, that poll is the most recent so it's weighted more, but they take the source's validity into consideration:

"The weights account for the quality of each poll based on its track record and its methodological standards. They also account for sample size and how recently it was conducted; recent polls are weighted much more heavily than older ones."

CajunBlazer

(5,648 posts)
29. That poll is weighted at 100%
Sat Feb 13, 2016, 01:08 AM
Feb 2016

Meaning that it was the only one used. That makes sense since the last Nevada poll before the present one - one that resulted in a tie - was taken in late December - much too old.

CajunBlazer

(5,648 posts)
18. It's not even an average of polls - the "tie" is only based one the suspect poll
Sat Feb 13, 2016, 12:39 AM
Feb 2016

Last edited Sat Feb 13, 2016, 01:10 AM - Edit history (1)

Silver does not use older polls. The last poll before the current "tie" poll was back around Christmas - much to old to use. (See my reply to the OP below.)

With the Nevada caucuses coming up soon, we should get additional Nevada polls fairly quickly and then we well have a better idea what is going on.

 

pintobean

(18,101 posts)
23. Oopsie
Sat Feb 13, 2016, 12:56 AM
Feb 2016

On Sat Feb 13, 2016, 04:42 AM an alert was sent on the following post:

52 Bernie 47 Hillary the rest go O'Malley and Vermin Supreme.
http://www.democraticunderground.com/?com=view_post&forum=1251&pid=1218718

REASON FOR ALERT

This post is disruptive, hurtful, rude, insensitive, over-the-top, or otherwise inappropriate.

ALERTER'S COMMENTS

Calling Hillary Clinton, a Democratic candidate for President, on a Democratic discussion board "Vermin Supreme" is ground for hide if not expulsion.

You served on a randomly-selected Jury of DU members which reviewed this post. The review was completed at Sat Feb 13, 2016, 04:53 AM, and the Jury voted 0-7 to LEAVE IT.

Juror #1 voted to LEAVE IT ALONE
Explanation: http://www.cbsnews.com/news/vermin-supreme-finishes-fourth-in-new-hampshire-democratic-primary/
Juror #2 voted to LEAVE IT ALONE
Explanation: I don't think this poster was calling Mrs. Clinton Vermin Supreme. Not sure who the poster was referring to. Not good - yes. Hideable - no.
Juror #3 voted to LEAVE IT ALONE
Explanation: No explanation given
Juror #4 voted to LEAVE IT ALONE
Explanation: How embarrassing for you. Google "Vermin Supreme" and get over your daily dose of outrage.
Juror #5 voted to LEAVE IT ALONE
Explanation: No explanation given
Juror #6 voted to LEAVE IT ALONE
Explanation: No explanation given
Juror #7 voted to LEAVE IT ALONE
Explanation: No explanation given

Thank you very much for participating in our Jury system, and we hope you will be able to participate again in the future.

SheenaR

(2,052 posts)
12. I wouldn't get too too excited
Fri Feb 12, 2016, 11:30 PM
Feb 2016

The writing on the wall seems to me that Hillary is writing off Nevada to emphasize the March 1 primaries.

That way if she wins its a plus. If she loses, she plays it off as it were no big deal.

Just my theory.

6chars

(3,967 posts)
35. it's just one state
Sat Feb 13, 2016, 02:33 AM
Feb 2016

just like NH.

Hillary goes into SC with a Victory from Iowa, and after winning SC she will go into Super Tuesday tied on states and way ahead on delegates (counting supers).

CajunBlazer

(5,648 posts)
14. Check back at Silver's site after the next Nevada poll comes in.
Sat Feb 13, 2016, 12:12 AM
Feb 2016

I thought Sanders supporters considered Nate Silver is a loser who made all kinds of mistakes in his predictions. Oh, that's right, that was when Silver was predicting Bernie would lose a state.

You should understand that according to Silver's his current prediction of a tie is based the only poll in Nevada which has been taken lately. According to Nate's ratings of organization which conducted the poll, it is not known for producing accurate polling data. If you don't believe me check it out for yourself using the poll information available on the link and Silver's ratings of polling organizations on his site.

If you dig deeper you will find that polling organization is a right wing political group and if you look at the questions on the poll itself you will find that they written to try to lead the people answering the questions in a particular direction. Not good.

Many Bernie supporters have claimed that the polls are rigged, well this might serve as their best example. Now ask yourself, why would a right wing organization try to get people to believe that Clinton is not as far ahead in Nevada as they thought? You might also what to ask yourself why Carl Rowe's super PAC spent $1.5 million in NH right before the primary on ads attacking Clinton using Bernie's talking points?

Here's the bottom line, I would wait until you see the next Nevada poll to get excited.
I don't know what the next poll is going to show, but it will probably be much more accurate than the current one. I am going to wait until the next two or three Nevada polls come out before I get excited one way or the other.

JudyM

(29,251 posts)
15. It was based on 7 polls. I was posting the update, but I don't think it's gospel. Polls are
Sat Feb 13, 2016, 12:14 AM
Feb 2016

Last edited Sat Feb 13, 2016, 01:06 PM - Edit history (1)

approximations.

CajunBlazer

(5,648 posts)
16. No, sorry, you're reading that wrong
Sat Feb 13, 2016, 12:31 AM
Feb 2016

I am familiar with Silver's techniques - he uses only recent polls because the public's opinion can change a great deal over time.

Note that the second to last poll - Gravis Marketing poll was taken why back between Dec. 23rd and 27th. and it had Clinton ahead by +23%. That was a long time ago in polling terms. Look at the polls before that one where Clinton was ahead by 29% 16%, 27% etc. Those couldn't possibly be combined with a poll which showed a tie to come out with combined score of a tie. No, only the last poll - with the tie result - was used.

Also look at the graph of Clinton's and Sander's poll results over time where the two line intersect at the present date. Note that those lines correspond exactly to the results of the other six polls and ultimately to tie poll.

You can see that now can't you, right JudyM?

JudyM

(29,251 posts)
17. Yes, I saw that before. As he notes, he used 7 polls, I'm assuming he's not misstating.
Sat Feb 13, 2016, 12:39 AM
Feb 2016

Also I noted that his comment about the most recent polls being heavily weighted. No argument. But he determines the validity of the poll and I trust that if he's counting it he made an independent determination how to weight it.

Have a good night, signing off. I'm directing a canvassing team in the morning.

CajunBlazer

(5,648 posts)
22. Sorry, you're simply wrong!
Sat Feb 13, 2016, 12:51 AM
Feb 2016

Try averaging 6 polls where Clinton was way ahead with one poll where she and Sanders tied and see if you come up with a tie. (please show your work)

Mathematically, the only way you can do that, even with weighting the polls, is to put all of the weight on the last poll and none on the previous six. That is the mathematical equivalent of using only the last poll - the was one that resulted in a tie.

I have taken far too many statistics and math courses to doubt myself on this one. Even a Sanders supporter who knows his/her math will support me on this.

joshcryer

(62,276 posts)
33. For the record I still think Nate is fucking it up.
Sat Feb 13, 2016, 02:25 AM
Feb 2016

The reason he gets general elections right is because there are dozens and dozens of polls running up to the election. Tens of thousands of people polled. Of course if you aggregate them you're going to have fantastic results. The primaries, however, are all over the place.

Response to JudyM (Original post)

CajunBlazer

(5,648 posts)
26. Absolutely not! At least not this Clinton supporter
Sat Feb 13, 2016, 01:03 AM
Feb 2016

Nate Silver is an excellent statistician. His formulas and techniques are very sound.

But as they say in the computer trade: Garbage in, garbage out. Silver does not use old polling data because public opinion can change quickly over time - witness the Sanders surge. The only recent Nevada poll of the Democratic race came out of a polling organization which Silver has rated as "poor". But it is the only recent Nevada poll he has; the last poll before this one was taken back around Christmas - much to old to use.

I suggest that Bernie supporters waiting until new Nevada polls come out before getting excited. With the primary coming up soon, that shouldn't take too long.

However, base on the posts above, but now that you mention it, I did notice that some Sanders fans are suddenly taking a liking to Silver.

sadoldgirl

(3,431 posts)
21. Nobody, even the admired N. Silver can
Sat Feb 13, 2016, 12:46 AM
Feb 2016

predict a caucus state, especially one with same
day registration.

Thus this is just hedging his bets,imo.

CajunBlazer

(5,648 posts)
30. No, not hedging his bets - it's based only on one poll
Sat Feb 13, 2016, 01:12 AM
Feb 2016

Silver knows that other (better) Nevada polls will be out soon.

TSIAS

(14,689 posts)
24. Not enough polling
Sat Feb 13, 2016, 12:57 AM
Feb 2016

I like Sanders' chances, but there hasn't been much reliable polling recently. Reid has been neutral so far. Wonder if he will send out his vaunted machine to buttress the chances of Clinton.

kiva

(4,373 posts)
34. I'm not surprised.
Sat Feb 13, 2016, 02:30 AM
Feb 2016

I just went to a panel tonight in Las Vegas that included Cornel West, Donna Murch, Christopher Whitt and James Small who all talked about why they were supporting Bernie. Given the reaction of the audience, I'm thinking Hillary's firewall is crumbling.

kiva

(4,373 posts)
42. It was a great event,
Sat Feb 13, 2016, 11:25 AM
Feb 2016

lots of discussion...and we get Bernie tomorrow night. It's an amazing weekend in Vegas.

delrem

(9,688 posts)
40. I can't see how Hillary can pull off a win.
Sat Feb 13, 2016, 05:41 AM
Feb 2016

I think she's dead in the water.

Just my opinion, tho' - so don't mind me.

MisterP

(23,730 posts)
43. 538's got a cushy gig: take the poll numbers, subtract them from 50%,
Sat Feb 13, 2016, 03:34 PM
Feb 2016

square the result, and subtract that from the poll number

you'll predict "right" every single time!

Latest Discussions»Retired Forums»2016 Postmortem»Nate Silver's 538 Puts Ne...