Don't freak out over Nate Silver. He's plugging in shit polls like Gravis into his crappy machine

[Cali_Democrat]

and it spits out bullshit!

Garbage in, garbage out.

[TroyD]

1. I agree that Nate Silver should not be relying on Gravis & ARG

They are both Republican-leaning, as is Rasmussen (although they have improved over the past week).

But that doesn't change the fact that Obama collapsed nationally, and we have to get those national numbers back up.

[1StrongBlackMan]

10. Of course they have ...

as is Rasmussen (although they have improved over the past week).

It's rasmussen's business model ... and a brilliant one at that!

Rasmussen serves as a consistent conservative-leaning polling out-fit for 40 weeks out of any given election year. Then, beginning about week 441-48 they slowly pull back to the pack, producing polling results that are closer to what is really going on. By has week 49-50, they are actually pretty accurate, but still err to the right. By election day, they are spot on and actually out perform most other polling outfits.

For about 50 weeks out of the election year, you service a a propaganda shop drawing empirical support for your gop clients' talking points (remember: "A majority of Americans hate ObamaCare" being conflated with "A majority of Americans want Obamacare ended?&quot Then, by election day, they point to their spot on polling results to sell their next round of polling.

That really is a brilliant business model ... Wish I had dreamed it up ...

[Maximumnegro]

11. OMG Troy your are just all over these doom threads!

Fancy meeting YOU here!

So. What's goin' on.

[TroyD]

12. Malarkey!

[progressivebydesign]

18. (R)Assmussen sent out an email PUSHING the 2016 DVD today!!!

They are NOT to be trusted. And please "Collapsed nationally?" BULLSHIT. The polls have shifted to Likely Voters, and in the past they were not reported as such, and LV is NOT people who say they'll likely vote, it's who the pollsters think will vote. Big diff Debbie Downer.

[speedoo]

2. Has anyone informed Nate of Grantcart's expose re. Gravis?

Probably not, right?

[woolldog]

3. This is a big problem I have with him.

And I had the same problem with him when he had Obama winning big.

[MjolnirTime]

4. He was never our friend. In the end, he is on a payroll.

He doesn't conduct polls for the betterment of the United States.

[Codeine]

21. He doesn't conduct polls at all.

And nobody was bitching about him when he had the President up.

[Blaukraut]

5. He was better when he was poblano and just a numbers cruncher

Now he has been hired by the NYT and has to push out a column every day.

[cheriemedium59]

6. I am not living in denial but...

I have to say I am curious how Romney can be doing so well in the polls?
I admit and accept President Obama did not have a stellar performance during
his first debate BUT it just smells fishy that out of no where Romney gains?
I mean the unemployment numbers are improving, consumer confidence is improving,
the housing market is improving and Romney continues with his gaffes almost daily
in the media.

I am not saying the polls are wrong, but I am saying it just seems bizarre.

Does anyone else think it seems odd?

[Proud Liberal Dem]

28. President Obama's approval numbers are up into the 50's (per Gallup)

So, yeah, how/why are the polls tightening? All of the indicators are not consistent with an incumbent President about to get defeated for re-election.

[CreekDog]

34. Obama was never as strong as Nate said and isn't as weak as Nate is saying

all his numbers and predictions are just insights into a race that moves just barely in one direction or another in recent weeks.

[Panasonic]

7. Nate needs a few days to get the new numbers post-debate

and I know it'll be the death knell of the Rmoney/Rlyin campaign. Three weeks before elections.

Obama will destroy Rmoney in the town hall debate. And finish him off in the foreign policy debate (By that time, the Libya issue would be wrapped up)

[JackN415]

8. Don't worry

Nate Silver is a political scientist and he has to be neutral by including everything, including bad polls because in principle you should never ever discount data, unless the measurements are proven erroneous beyond a reasonable doubt.

All polls, slightly biased or not have a way to average themselves out. And if the reality is that, we should accept it, remember, the probability is still highly favorable to Obama because of electoral math.

My view is the follow: the post conventional bounce for Obama is a tad too good a story (probably higher than reality), and the post-debate bounce of Romney is likely the same. At the end of the day, Obama is still ahead.

[woolldog]

9. well what used to separate him from RCP and the like

was the way he evaluated pollsters. He's gotten away from that it seem like since he went from Poblano to MSM guy. Maybe not enough time anymore, I dunno.

[Cali_Democrat]

14. Nate now has a blog on the NYT website and he needs to drive traffic there.

He's now essentially a corporate media whore. He does some good work, but he's not above criticism.

[roguevalley]

15. when he does better numbers in our book he won't be a whore. this is cyclical. I will hold judgment.

this is the long haul here.

[Maximumnegro]

13. Nate is Nate. Go for his numbers, not analysis

He is terrible at the punditry part. Always has been. I won't dispute his number crunching except to state the obvious: 10 days ago Obama was at what 85% to win. So the numbers go up just as quickly. The important thing to note is the SENATE numbers are holding, that means Romney is not truly surging - if he was he would be taking his party with him. He has soft support and people want Obama to close the deal for them. No way it makes sense that the Senate holds but Romney rockets up. But no question the narrative needs to change in the next week. Personally I think if Romney is going to get a boost better now than later. 3 1/2 weeks is a LOT of time in politics.

[geek tragedy]

20. Senate numbers not changing is a bad sign for Obama.

It means that the fall in his numbers isn't due to Democratic enthusiasm falling, but rather to people switching to Romney but splitting tickets by voting Democratic locally.

[Maximumnegro]

24. Disagree. If this were regional, sure

but it doesn't appear that way. To be across the board that doesn't hold up as an explanation.

[geek tragedy]

26. I disagree. What we're seeing is Obama losing 1-4 % across the board.

That the trend isn't hitting other Democrats means that this trend is all about people defecting from Obama to Romney.

If it were about Democratic enthusiasm/turnout we'd see Sherrod Brown and Bill Nelson losing support--but they're holding steady or even expanding their leads.

[allrevvedup]

16. Nate does what he's paid to do

just like the rest of the liars. K/R.

[gkhouston]

17. This is what happens when you plug and chug without thinking about your calcuation.

He's not skeptical enough about his input data, IMO.

[geek tragedy]

19. The polls are skewed! Obama trailing nationally is a great thing!

He meant to lose the first debate. 11 dimensional chess.

[Democat]

29. It's part of the secret plan

The good news is that Biden's performance will hopefully remotivate some Democrats. The question is whether we can get back the voters that Obama gave away at the first debate.

[geek tragedy]

32. My guess is no, but we'll have to see. PPP will show OH a dead heat tomorrow. nt

[former9thward]

22. Just a week ago Nate was praised on DU.

Now he is "a corporate whore" and "just another liar". It seems many here know all there is to know about professional polling. They should go into that line of work. I'm sure they would do well.

[geek tragedy]

23. That's what he gets for going by the numbers.

I imagine we'll have skewed poll conspiracy theories around here pretty soon.

[Lex]

25. Was he using those same polls when it was showing Obama was way UP?

[geek tragedy]

27. His model hasn't changed.

I've played with HuffPo's model where they let you pick the pollsters you include. Picture is still ugly, ugly, ugly.

[HooptieWagon]

30. Oh good lord...

Thanks for your concern, chicken little.

[ChimpersMcSmirkers]

31. Nate's model is the gold standard for me. His model and Intrade are all you need. And it looks...

crappier by the day with both. Obama is now below 60% on intrade.

Obama screwed the pooch on his first debate and he needs to hit it out of park on the next one. I'm nervous and you should be too.

PS: unlike Greek_Trollergy, I'm not enjoying this.

[politicman]

33. Majority of swing state polls are from

People, you all need to chill out and look at this in context.

Majority of the swing state polls that have been released after the debate are from right-leaning pollsters.

Sure there is a trend going in Romney's direction BUT we have no idea of knowing just how big or small Romney's bounce was.

There are certain polls which shows that Romney has closed a gap of 18 points among women, There are certain polls which show that Romney is getting 50% of the black vote, There are certain polls which virtually don't include any Hispanic numbers (like Pew's latest poll), There are polls that have typically been favorable to Romney that are now showing even more favorability.

In the end, we can conclude that there is a trend towards Romney and that the race has tightened BUT poll averages at the moment tend to be skewed by some certain polls that have unrealistic numbers in them (like a certain poll that shows Romney way up because he is getting half the black vote, like a certain poll that shows Romney up because he has pulled even with Hispanics, etc)..

Nate is factoring these obviously bad polls into his model THUS using them as further evidence that Romney got a huge bounce, BUT this is a flaw in his methology, because a bad poll should be discounted entirely as it cannot give a picture of the landscape if it is using flawed numbers.

We KNOW that Obama will get 90% or so of the black vote, we know that he will get close to 70% Hispanic vote and we know that he should get more women than Romney, so any poll that shows figures that are not in allignment with these should be discounted.