Dean “Unskewed Polls” Chambers: I Was Wrong, Nate Silver Was Right
I was wondering whether we’d hear from the guy behind UskewedPolls.com given last night’s results. His website hasn’t been updated since yesterday and he hasn’t posted anything at his Examiner.com site. However, he did give an interview to BBrett LoGiurato at Business Insider:
Dean Chambers, the man who garnered praise from the right and notoriety on the left for his “Unskewed Polling” site, admitted today that his method was flawed.
“Nate Silver was right, and I was wrong,” Chambers said in a phone interview.
Chambers’ method of “unskewing” polls involved re-weighting the sample to match what he believed the electorate would look like, in terms of party identification. He thought the electorate would lean more Republican when mainstream pollsters routinely found samples that leaned Democratic.
But as it turned out, the pollsters were right — self-identified Democrats outnumbered Republicans by 6% in election exit polls.
“I think it was much more in the Democratic direction than most people predicted,” Chambers said. “But those assumptions — my assumptions — were wrong.”
(…)
He actually thinks conservative-leaning pollsters like Scott Rasmussen have a lot more explaining to do.
“He has lost a lot of credibility, as far as I’m concerned,” Chambers said. “He did a lot of surveys. A lot of those surveys were wrong.”
I have to admit that I’ll give the guy some credit for admitting an error and not trying to paper it over, although it is worth noting that he made a major change in his Electoral College projection over the course of just four days. The comments about Rasmussen are also interesting. I wonder if others on the right will feel the same way. Rasmussen’s polling was way off, we’ll have to wait until Nate Silver does his Pollster Rankings to see how far off, but he was clearly didn’t reach the level of accuracy he achieved in 2008. But that what happens when you tailor the presumptions behind your polls to please your customers.
Notice he didn’t explain his “effeminate man” quips.
Obviously “the very small statured thin effeminate man with a soft- sounding voice” is far more credible than the fat blowhard hack…
Where is Jan????
Yeah, Rasmussen is an absolutely joke.
Their party ID registration in October had GOP +5, which is so far off. There was no chance in hell the electorate was going to be that old and white.
I give him credit for admitting he was wrong and for calling our Rasmussen.
Anyone seen Smooth Jazz?
Part of the problem was that rasmussen had that zombie lie that showed them as the best pollster in 08 when really they were like the 6th best after all the votes had been counted. Maybe now, that zombie lie will die.
@Console:
Yeah, they weren’t that good in 2008.
If you actually look at the state polling, they were pretty off (ex. they had OH tied, but Obama won it by 5). They were even worse in 2010 and they may be the worst pollster in 2012 after Silver does his analysis on accuracy.
Kudos for him admitting he was wrong. Better than Doug and Benghazi…
@Steven L. Taylor:
Been thinking the exact same thing.
I will give one more thing to Chamber’s, his explanation was far more genuine and honest than the one that good ‘ol Dick Morris issued (which included blaming Chris Christie for making him wrong). I’m still wrapping my head around Michael Barone’s explanation.
That said, while he might get credit for saying Silver was right, Chambers has still not apologized for all the personal attacks on Silver he made as part of that odious editorial.
Way to stay classy Chambers.
Steven,
This is true. I’m not sure the reporter asked him about that
Rasmussen, Mason-Dixon, and Suffolk were the three biggest losers.
Rasmussen’s methodology of weighting by party was shown to be a farce, as were all of his polls. PPP beat him in every single state, despite both being partisan robopollers. If you’re a Republican and want to know the actual leanings of a state, you’d hire PPP over Rasmussen.
Mason-Dixon had Romney leading Obama in Florida by twice (+6) what Obama was leading Romney in Minnesota (+3). Final results were Obama +1 in Florida and Obama +8 in Minnesota.
In early October, Suffolk pulled the plug in North Carolina, Florida, and Virginia–claiming that Romney had them locked up because Obama was too far below 50%–even though those polls showed him LEADING in Florida and Virginia.
Mason-Dixon missing FL is pretty surprising.
We all know Rasmussen sucks, but Mason-Dixon generally polls FL pretty well. To miss it by 5 points is pretty remarkable.
@Steven L. Taylor: Even Dick Morris admitted he was wrong. He said he didn’t believe that the electorate would look like 2008, that Obama’s first victory was a once-in-a-lifetime thing, and that America would revert to it’s old self again, as it seemed to in 2010. But now, he concedes, this is the new face of the American electorate and the Republicans better find a way to adjust to it. Sounded like the first shot in a civil war to me — but surprisingly mature and self-aware for Morris.
And speaking of mature and self-aware, I sure do wonder why we haven’t heard from Jan, Bit, Drew and Smoovie today. At least Jaynos showed up to spew some argle-bargle about Obummer using Fast and Furious to smuggle guns to kill Americans in Benghazi, or whatever he’s going to live off of for the next four years…
@wr: Good for Morris, although he inability to understand that one cannot compare turnout in a mid-term to turnout in a presidential year does not speak well to his understanding of some pretty basic stuff.
@Jr:
Mason-Dixon has trouble polling racial minorities.
In 2004, not such a problem. In 2012, fatal.
@Jr:
Also, M/D didn’t miss FL by 5. They missed it by 7. They had Romney +6, it was Obama +1.
Their mid-October poll had Romney +7.
Why were they off? They showed a roughly 50/50 split in the Latino vote. Which instantly discredited it.
@Steven L. Taylor:
Well, he predicted a Romney “landslide” of 323 EVs but then downplayed Obama’s winning of the same margins as W Bush won in 2004 and 332 EVs as a “squeaker.”
Interesting. This mea culpa suggests that he actually believed what he wrote.
@Steven L. Taylor: Well, yeah. I’m not claiming that the author of the famous Hillary Condi Smackdown book is suddenly brilliant, but there must have been a time before he became a Fox whore that he had some political insights or ability. Maybe he had a little flashback yesterday…
I’d looked through Rasmussen’s state polling in 2008 and 2010 and found that in every close race, Rasmussen was 2-4 points too GOP. Every single one. I haven’t double checked this year, but it looked like they were likely to maintain that 2-4 point GOP bias. I still love Rasmussen, because he really does give an excellent picture of how things look – 3 points better for the Democrats than he claims.
@Steven L. Taylor:
Sorry. No bully for Morris.
His explanation was largely a dodge filled with gobbledy gook and weasel words. He essentially said that he disbelieved the polls simply because he was sure that (a) 2008 was an anomaly, (b) that all posters are biased towards liberals, and (c) he refused to believe that many brown people would actually show up and vote, and finally (d) it was also the fault of Chris Christie and Hurricane Sandy.
@mattb: Gotcha. And yes, Sandy (and Christie) will be receiving a lot of blame.
@Steven L. Taylor:
Yup… btw, here’s the actual quote from Dick Morris on this (it’s pretty snort worthy):
@mattb: Snort-worthy, indeed.
BTW, Chamber’s mea culpa article for Examiner.com should be nominated from an “understatement of the year” award:
I really have to remember “depending on your point of view” line for the next time I run into trouble on something.
@mattb: Heh. Depending on your point of view, Hurricane Sandy was a bad storm.
The thing about Morris is that any apology he makes now is too little, too late. He’s been wrong so many times throughout his career that Andrew Sullivan was right to rename his bad-predictions award after him (who the hell is Von Hoffmann?), and if Morris possessed an ounce of integrity he’d refrain from ever trying to predict the future again, because he just plain sucks at it. It’s just that the bad predictions of conservative punditry in this race was such a high-profile blunder they had little choice but to acknowledge their error. They rarely are forced into that situation, which is why they usually get away with being proven wrong again and again and again–nobody bothers to call them out on it. (Nobody they have to answer to, anyway.) Perhaps Nate Silver had a hand in this, by shining a light on the subject while inspiring them to foolishly wage war against him.
Lots of our dear friends are suddenly MIA. What about Florack? I’m sure he’s eager to explain what he said on 10/21:
kari:
I looked pretty carefully at some of his numbers for 2012 and 2008 and found pretty much the same thing. Link.
“…but I don’t take back any of that effeminate stuff. I mean, just look at him. Girly. That’s the one call I made right this whole cycle.”
@Steven L. Taylor: “Good for Morris, although he inability to understand that one cannot compare turnout in a mid-term to turnout in a presidential year does not speak well to his understanding of some pretty basic stuff. ”
He wasn’t unable, he was lying. He could not physically have participated in national politics as he did and not understood that. And he’s lived that, going through 1992-4-6.
And I’ll believe his apology if and only if it doesn’t evaporate over the next few weeks. Which I think that it will, or he’ll find himself looking for honest work.
@Barry: I was trying to be charitable, but that is a bad idea where Morris is concerned. Having read the quote above and having seen clips of his “apology” I must confess he deserves no credit.
Anyone who thinks Nate Silver is an “effeminate man with a soft- sounding voice” needs to visit more gay bars.
You are so right. Not that there’s anything wrong with that.