Summing up the Polling Debate
Dean Chambers of UnSkewed fame, puts the polling debate into sharp perspective.
Dean Chambers, the fellow behind the whole UnSkewedPolls bit, has a column at the Washington Examiner that well sums up the true nature of the ongoing debate over polls. Chambers starts with the following:
While many conservatives look to former Clinton political consultant Dick Morris to understand the polls and political surveys on the elections, or even a site like UnSkewedPolls.com, those on the left look to New York Times blogger Nate Silver.
First, I think that is noteworthy that the entire conversation is cast in terms not of right/wrong, accurate/inaccurate, etc., but it rather a contest of left and right.
Second, consider what he is saying: his claim is that “right” looks to a fellow (Morris) whose most famous manifestation of predictive skills can be found in his 2006 book entitled Condi vs. Hillary: The Next Great Presidential Race. Beyond that, he gets paid to boost the GOP on Fox News and his columns are mostly (if not exclusively) published in clearly partisan outlets. Indeed, the notion that in 2012 the best identifier for Morris is as a “former Clinton political consultant” is amusing. While it is true, it also ignores Morris’ career trajectory since those days. The other place that Chambers says that those on the right go is his website, which blatantly recalculates poll results to make them more palatable to Republicans.
This is to be compared to Silver, who uses complex and meticulous statistical models to analyze reams and reams of polling data. And yes, I know that many think that Silver works for a left-wing propaganda machine, but the fact that one cannot see that NYT is not the leftward equivalent of Fox News is part of the problem. And yes, I am sure that Silver will vote for Obama. The problem is, the numbers are the numbers and the methodologies are the methodologies.
The bottom line here is that Chambers is claiming that the “right” like to get their numbers and analysis from a blatant partisan commentator and/or from a source that re-calculates the polling results while the “left” likes to go to a fellow who analyzes copious amounts of data via sophisticated modeling. This strikes me as damning the “right” quite frankly. It certainly puts Chambers’ POV into perspective (not that this is a surprise).
The quality of Chambers’ analysis get worse, as it only takes a handful of paragraphs to get very personal:
Nate Silver is a man of very small stature, a thin and effeminate man with a soft-sounding voice that sounds almost exactly like the “Mr. New Castrati” voice used by Rush Limbaugh on his program. In fact, Silver could easily be the poster child for the New Castrati in both image and sound. Nate Silver, like most liberal and leftist celebrities and favorites, might be of average intelligence but is surely not the genius he’s made out to be.
First: what is the point of this? Is Silver not manly enough to do statistical analysis? Huh?
Second: the Limbaugh reference is telling because Chambers is using Limbaughesque techniques here: ridicule as “analysis.”
Third: the issue should be the number and the methodology and nothing more.
This column is an excellent microcosm of the current debate about polling from the fellow at the epicenter of the conversation. It demonstrates the dominance of poor reasoning, personal attacks, and partisanship over, well, the numbers.
I love this little tidbit
Did no one think to check this? Nate Silver was accurate. I almost feel like the article is suggesting some sort of Orwellian cover-up. Something along the lines of Silver being horrendously wrong in 08, but no one remembers because the “fact checkers” are rewriting history or something.
The current conservative right wing mode is to attack as biased or skewed any organization or institution that presents a view, opinion or facts that are contrary to right wing ideology.
This has been going on for a while now. It started with the ‘liberal media’ betrayal, and it’s widened to include: the education establishment, the science establishment, unions, ‘Hollywood,’ “liberal judges,’ and now …. polling organizations.
It all coincides with the dumbing down and decline of America. Basically, the modern right wing movement is largely responsible for the dumbing down of America.
@A: It is shame that it so difficult to do research on such things. If only we have some sort of electronic and easily searchable portal that could take us back to what was written in 2008….
@al-Ameda:
It’s actually more cynical than that — it’s the belief that everything is fundamentally driven by political ideology. There’s no possibility of a finding or a statement of fact that isn’t political in this model. In many respects it’s taking the worst excesses of (dreaded) postmodernism and academic critique and turning the on itself.
Without a doubt Silver is liberal — as are many scientists as well — but this doesn’t mean his statistical work (versus his short and long form writing) must inherently have a skewed liberal bias.
What this misses is that Silver is a trained statistician. And part of that training is learning to minimized bias in his models (just as should be done in good science and arguably any other practice).
But this issue of rigor is not well understood. And people like a Limbaugh, who views everything through a partisan lens themselves (at least on their program), are not remotely interested in understanding this.
To be fair — all conservatives should not be painted with this anti-science or anti-polling brush (and I’m not suggesting you’re doing that Steven… I’m trying to head off some of the comments that will most likely follow).
In particular, the bloggers at the American Conservative are doing some amazingly smart and brave writing about a number of complex science, social science, political and cultural issues right now. If people are not following most of those blogs, they should be.
@mattb:
Indeed–although it has dominated most of the conversations about “skewed” polls. And, unfortunately, it seems to dominate huge swaths of conservative media.
Mr. Taylor, we’ll need some more information before we can properly evaluate your article. How tall are you, what’s your waistline and are you a tenor?
“…..but the fact that one cannot see that NYT is not the leftward equivalent of Fox News is part of the problem.”
The NYT used to be the media standard for excellence. This is no longer the case. More and more it is derided, not only for it’s left-sided opinion/editorial bias, but for it’s tendency of omission –selectively placing news items on it’s front and interior pages, not for timely relevance or worthiness, but more for how such news might either positively or adversely effect the democratic party. F & F and now Benghazi are prime examples of muting stories by their publication dismissal of them . The NYTs has basically become the “print twin” of MSNBC, a notorious cable news shill of the left with a shrinking viewing audience.
“The bottom line here is that Chambers is claiming that the “right” like to get their numbers and analysis from a blatant partisan commentator.” …that partisan commentator being Dick Morris.
While Morris is a favorite on Fox talk shows, he, by no means, is the polling guru which seems to be insinuated here. As for right-sided blogs, Morris is considered a ‘joke’ in most of those quarters. Sure, his numbers are eye-candy of encouragement, for those who want Romney to win. But, his POV and secret internal polls are not taken seriously. Rasmussen, criticized here as being slanted towards the R’s, is who is looked at with credibility, especially regarding his screens, sampling size and weighting of those samples.
As far as Nate Silver goes, people on the right look at his models with a mixture of respect and taking his projections with a grain of salt. He is well known for his left leaning politics, as well as a side-show kind of competition with Rasmussen. He would love to take Rasmussen down. In this election, in fact, Silver seems to be betting his reputation by going diametrically in the opposite direction of Rasmussen. One of these pollsters, IMO, will come out of 2012 as damaged goods, because of how badly they missed the mood of the public, consequently the projected outcome of the election.
@Steven L. Taylor:
Correct and this speaks to a greater problem with Conservative Media (or rather, what Sean Scallon has correctly dubbed “Conservative Inc”). So long as the commercial aspect of aspect of the conservative enterprise is of equal (if not greater importance) than the political/governing side, there’s no interest in having honest debate.
Sadly, (as Dick Morris continually proves) there’s little profit to be made in rendering honest appraisals (because, more often than not, it requires making the audience uncomfortable).
The polling issue is a symptom of the disease that riddles the Republican Party. The Chairman of the House Science Commitee thinks Evolution is a Hoax. Another Republican member of the Commitee thinks you cant get pregnant if raped. The keynote speaker at the RNC Convention speaks to an imaginary person. Romney economic plan is mathematically impossible. He takes every position on every issue. As I heard somewhere…Picasso couldn’t paint all the sides of his face.
Seriously…how does an intelligent person take the Republican Party seriously?
“…One of these pollsters, IMO, will come out of 2012 as damaged goods…”
But not in your eyes if it’s the notoriously right leaning Rasmussen…you’ll simply find a way to rationalize it in terms friendly to your team. Your record, as well as the records of Morris and Rasmussen, is crystal clear…as is Silvers.
In 2010, Nate Silver predicted:
Senate: 51.6 – 48.3 – 0.1 (D – R – I)
Actual outcome (51 – 47 – 2)
So, according to his prediction, Republicans would have gotten one more Senator.
House: 202.8 – 232.2 – 0 (D – R – I)
Actual outcome (193 – 242 – 0)
He predicted 10 more Democrats in the House than the election results.
Governors: 19.3 – 29.9 – 0.8 (D – R – I)
Actual outcome (29 – 20 – 1)
He predicted one more Republican governor.
Anyone arguing that Silver is a partisan Democrat should really look at his record. Then they can compare it to Dick Morris, Dean Chambers, etc.
The model where the “right” prefers analysis that gives the results they want, while the “left” prefers analysis that is empirically sound could be applied to much, much more than the polling debate (social mobility analysis, inequality analysis, global warming analysis, etc).
Damning to the “right” indeed.
Bob Krumm delivers a complex and compelling analysis of this presidential race, including a perspective on Nate Silver’s polling model.
Krumm continues to view the 2008 election, though, as more of an outlier, and therefore cannot be utilized as being totally representative of what is going on currently in 2012. He denotes, as an example of this, how poorly Silver’s predictions turned out to be in the outcome of House seats taken by the R’s in 2010.
Fundamentally, Krumm asserts the following about Silver’s model:
Mr. Chambers’s missive: …
Nate Silver is a man of very small stature, a thin and effeminate man with a soft-sounding voice that sounds almost exactly like the “Mr. New Castrati” voice used by Rush Limbaugh on his program. In fact, Silver could easily be the poster child for the New Castrati in both image and sound. Nate Silver, like most liberal and leftist celebrities and favorites, might be of average intelligence but is surely not the genius he’s made out to be. His political analyses are average at best and his projections, at least this year, are extremely biased in favor of the Democrats.
…
Mr. Chambers is heavy on the vicious invective – and there’s some revolting stuff in there such as the quoted bit – but comes up short on the analysis. Might need to deal with the mote in his own eye before finding fault with others.
Nate Silver is currently predicting Obama winning 295.4 – 242.6.
Dean Chambers is currently predicting Romney winning 359 – 179.
Now, one of them has a reputation to uphold, the other blogs from his basement…
Dean Chambers has Methodology!!!:
As others have said, ‘Reality has a liberal bias.’
That pretty much sums up the problem with the GOP… if you wish really hard, u can haz unicorns!
@jan:
So, the major critique of Silver is to be found in the comments? That is rather weak tea.
A fair observation, except that a 70% prediction one way still allows for a 30% chance of a different outcome. That would be the point of giving percentages.
Yes, but that is what ALL polling tells one. There is a reason he calls is a “now cast”–one can extrapolate from there, of course (as he does).
@PJ:
Mr. Chambers never learned stats, apparently, and doesn’t seem to know that pundantry is about as accurate as flipping a coin. Or perhaps much less so if one includes Bill Kristol and Dick Morris in the sample.
When did Silver become a pollster? I know math is hard but differentiating between a pollster and a statistician is pretty simple.
@jan:
Polling house races is much different than polling in Presidential elections. Also, in Nate Silver’s final 2010 model he expected Republicans to gain 54/55 seats, but the model showed them leading in enough races to pick up 59. The final result was 63, so your claim seems to be out of place.
KansasMom,
Silver is not a pollster. He does not conduct his own polls. He’s a statistician who analyzes the poll results and other factors and is attempting to determine where the race actually stands at a given point in time.
Silver’s 2010 House prediction also had a pretty wide confidence interval where the outcome was well within in.
Jan,
You are aware that Nate Silver doesn’t actually conduct polls right? Rasmussen is just one of the many polls that go into his model. If Nate Silver has a problem with Rasmussen, it’s not due to ideology, it’s due to historical accuracy. If you actually look at the fivethirtyeight site, you’ll see that all pollsters are rated on their Republican or Democratic lean, and the model weights them as such.
@Doug Mataconis: Take it up with Jan dude. I’ve known the difference for a couple of decades at least, 7th grade social studies, not to mention vocabulary, in action!
That unskewedpolls website would almost be hilarious unintentional parody, if it wasn’t for the sad fact that so many people actually take it seriously. Millions of Conservatives are primed to wake up on the morning of November 7th to find out that President Obama was re-elected, and to feel absolutely convinced that he “stole it”.
One thing that is really facinating about the Republicans is they skipped modern, going from neo-classicists like Buckley to post modernism. One can see it in the Republicans on the Texas school board trying to get creationism taught side by side with evolutions as two possible theories, with equal value.
I have my own questions about Nate’s national averages. It is quite possible that his weight of polls is missing changes that has occured in the last 2 years. In particular, he differers with RealClearPolitics in the national polling average by 0.5% to 1%. But, the idea that all pollsters except Gallup and Rasmussen are suddenly part of a conspiricy are as clear thinking as Rosie when she said George Bush had demoliton charges planted to go off on 9-11 so he could go to war in Iraq.
BTW – I do find it a bit ironic to note that Mr. Krumm, the man who argued in 2009 that American was on the verge of violent anti-government uprisings (if not all out revolution), is finding issue with Nate Silver’s predictive abilities.
@C. Clavin:
I take then serious, I take them very serious. They control the House, they are trying to take control of the Senate, and they just might win the Presidency. This sounds like a very serious situation to me.
Snark aside (I ran across that article trying to understand Mr. Krumm’s credentials), thanks for posting that article Jan. Unfortunately, I’m not enough of a statistician to have an opinion on the quality of critique of Silver’s meta-analysis model.
As you’ve noted, we’ll know whose predictions were more accurate in just a week and a half.
This sums up the right vs left debate in more area than just polling. The left uses facts and analyses, The right uses fantasy and fear.
Nate Silver, like most liberal and leftist celebrities and favorites, might be of average intelligence but is surely not the genius he’s made out to be. Bhahahahahahahahaha, gasp, gasp, hahahahaha, to quote someone from an old thread.
See also:
http://election.princeton.edu/
http://votamatic.org/forecasting-2012/
https://mmicdata.rand.org/alp/?page=election#election-forecast
@jan:
I find it increasingly hard not to be snarky when I comment on your posts. So – yawn. You really need to get out more, if you think that David Brooks, Ross Douthat, and Maureen Dowd (just to cite a few examples) represent a “left-sided opinion..bias”.
For those who had never before heard of Bob Krumm (I certainly hadn’t), here he is in all of his unbiased glory.
Where did the link go? http://polipundit.com/?p=43228
“This is to be compared to Silver, who uses complex and meticulous statistical models to analyze reams and reams of polling data. And yes, I know that many think that Silver works for a left-wing propaganda machine, but the fact that one cannot see that NYT is not the leftward equivalent of Fox News is part of the problem. And yes, I am sure that Silver will vote for Obama. The problem is, the numbers are the numbers and the methodologies are the methodologies.”
LOL, You have got to be kidding. I know you are a LIberal, but good grief! Nate Silver is a left wing hack, whose complex model puts more weight on a month old PPP or Marist OH model that showing Obama up 10 with a goofy and outrageous partisan split, compared to a more recent Rasmussen sample that shows the race in OH tied because he prefers the more partisan left leaning pollsters such as PPP and Marist compared to Rasmussen. ANY model is susceptible to “garbage in, garbage out” so spare us the propaganda.
Most important, at a time when major unaffiliated pollsters such as Gallup, Pew and Rasmussen are finding a surge in Rep voting participation in 2012 rivaling the great Rep year of 2010, Silver’s model is plastered with bogus swing states polls showing Dem turnout versus Repub 10%+. AIN’T. GOING. TO. HAPPEN. IN. 2012. The best the Dems will do versus Repubs this year is 2%+ or even. Certainly not the Dem 7%+ of 2008 or the Dem 9% – 10% in a bunch of polls he gives a lot of weight to. At bottom, Silver is a human being and rabid Obama sychophant – a sychophant who is not beyond giving extra weight in his model to those factors that tip his weighting to the candidate he openly advocates for.
And to suggest that the “NY Times is not the leftward equivalent of Fox” is beyond laughable. The NY Times is an activist organ devoted to the takedown of anything and everything Repub. I realized you were far left but give us a break: The NY Times is every bit as, or more, partisan as Fox.
@jan:
Just for reference, Mr. Krumm’s prediction on November 3rd, 2008
“Final popular vote tally: Obama 49.2%, McCain 48.8%, Other 2%.
Electoral votes: Obama 244, McCain 273, Pennsylvania’s 21 TBD.”
@Smooth Jazz: So, why does RealClearPolitics shows polls from the last week that average Obama about 2.5%? Is _everyone_ who doesn’t agree with you part of the conspiricy. Out of curiosity, do you accept the theory of evolution as an extremely well verified theory that needs to be taught in school and creationism as fantasy, or do you think biologiests also suffer from liberal bias?
@Scott O:
Wow. Given the rather large gap between Mr. Krumm’s prediction:
It takes a lot of chutzpah to use Silver’s 2010 predictions against him. It’s clear Silver was far closer to the actual results in 2010 than Mr Krumm was in 2008.
What’s particularly interesting to me is one of the reasons that Mr. Krumm gives for getting it so wrong — he let is ideological bias get in the way:
As Jan wrote in another thread, I wonder if this might be a prescient statement about the possible weaknesses of his present argument?
@Scott O:
Wow. Given the rather large gap between Mr. Krumm’s prediction:
Given the size of those gaps it takes a lot of chutzpah to use Mr. Silver’s 2010 predictions against him.
What’s particularly interesting to me (espcially given Jan and Smooth Jazz’s arguements for rejecting the NYT’s as to biased to consider) is one of the reasons that Mr. Krumm gives for getting it so wrong — he let is ideological bias get in the way:
As Jan wrote in another thread, I wonder if this might be a prescient statement about the possible weaknesses of his present argument?
*Sigh*
I had two similiar posts caught by the spam filter. Can someone release one or the other?
Besides the “Silver’s a sissy” idiocy, it should be made clear that Mr. Chambers is the proprietor of “Unskewed” Polls, to which he linked four times in the Examiner item, w/o any mention that it is his site.
Might just indicate a basic dishonesty in everything Mr. Chambers does.
This exactly is why i believe the unskewed polls narrative exists. Romney wins? Look at how he won when even the pollsters are against him! Romney loses? There’s no way he lost! Look at the “real” data.
Half of Republicans think ACORN stole the election for Obama in 2008. The boogeyman is different, but the goal is the same.
“So, why does RealClearPolitics shows polls from the last week that average Obama about 2.5%? Is _everyone_ who doesn’t agree with you part of the conspiricy. Out of curiosity, do you accept the theory of evolution as an extremely well verified theory that needs to be taught in school and creationism as fantasy, or do you think biologiests also suffer from liberal bias?”
Because the RCP average includes bogus polls with Dem 10%+ partisan samples based on surveys suggesting 30% of OH has already voted. If the big unaffiliated trackers Gallup & Rasmussen with thousands of respondents in their samples are BOTH showing Romney up 5 overall AND in the swing states, then Obama is not ahead by 2.5% in any swing state. OH is not all that different from the rest of the country, and has averaged Rep vote of 1% over the National Rep average in Presidential elections over the last 30 or so years.
Obama sychophants in the media such as CNN, ABC, MSNBC & partisan pollsters such as PPP have a vested interested in keeping him in the game by trotting our partisan voting samples of Dems at 10% – even as Obama get obliterated among independents by 10% – 20%. If Obama is losing independents to Romney by 10% – 20% he is not winning OH, not matter how many Dem 10%+ polls CNN trots out. Remember, it is barely 2 years since OH elected Gov Kasich & Sen Portman, and those Rep voters will be back out in force in 2012.
Here is a stat that should make you shutter: According to the various Secy of State County officials in OH, as of 10/26 220,000 fewer Democrats have voted early in Ohio compared with 2008. And 30,000 more Republicans have cast their ballots compared with four years ago. That is a 250,000-vote net Rep increase for a state Obama won by 260,000 votes in 2008. That is not based on a bogus CNN & MSNBC poll with a juiced Dem sample; That is a cold hard stat from OH election officials.
Good luck drinking the Nate Silver 99% probability Obama wins koolaid.
@Smooth Jazz:
It is actually at 74.4% at the moment.
Is it really that difficult to get basic facts straight? Putting “538” into Google take that long,
Is it any wonder it is so difficult to take you seriously?
@Smooth Jazz: Let me ask: what is your test for how we will judge these polls after Nov 6th? Is it a case of who wins? The popular vote totals vis-a-vis the polling? The EC predictions? What?
Good post, Steven.
@Smooth Jazz: And, these biased polls were close in the past. So, not only are the polls biased, but the voting must be biased too. Partisian leaning is an _output_ of a poll not an input. That’s been pointed out numerous times. You can repeat a mantra over and over again. That’s how weights of polls are assigned. And, BTW, when your points are clearly counterfactural, are the facts also biased?
As pointed out in the past, when Obama blew the first debate, the Democrats bemoaned it.
I am a Phd physicist. I’ve been doing Monte Carlo analysis professionally for over 30 years. I know how the models work. Wouldn’t you apply better weights to polls that have done well than polls that have shown historical bias. In 2008, the mean poll was within 1% of the vote. That’s close. And now, the same polls are part of the evil left conspiricy. What date did that happen on. And, if Nate is close with his election eve prediction vs. the vote, say he misses two states, and predicts the winner, does that just prove that the election must have been stolen. To ask the question as a scientist would: is your vierw immune from falsification by data?
@Santiago: Thanks.
smooth:
Are you saying Fox is partisan? Left or right? I honestly don’t know what you think about this. Fox’s most recent OH poll (10/18) shows Obama +3. Does this indicate they are “rabid Obama sychophant[sic][s]?”
Here is a fact that should make you “shutter” [sic]: the last time I saw you make this argument I proved that several of your claims were false (link, link). And you never took responsibility for those false claims.
So show us your citation. I’m sure we will discover, again, that you are inventing your own facts.
At his official site (link), I see that Husted’s last press release on this subject is 10/23, and it references data as of 10/19. So why did you claim you have data “as of 10/26?” Husted told you something he didn’t tell the rest of us?
You did the same thing last time: take old data and present it under a phony date. And your argument was phony in several other ways, and I’m sure that’s also true this time.
@DanM:
Aye, and there’s the rub.
“It is actually at 74.4% at the moment.”
Good Lord Dude. I was being facetious with the 99% remark. Of course, I can GOOG Nate Silver at NY Times to get the exact number. I was using the 99% out of this world number to indicate that at a time when the massive, unaffiliated trackers Gallup & Rasmusssen have Romney at 5% ahead, at a time when Obama’s Job Approval in Gallup is collapsing, Silver still has Obama winning a landslide with a ridiculous %. A %, I might add, based on slivers of Obama supposed leads in a few swing states that could tip either way. What would happen to Silver’s model if Romney get a poll that shows he is up 3 in OH or WI & NH & OH? Would Silver project a 70% Romney victory based on the fact that FLA, NC, CO & VA are off the board?
That said, I think the most ominous and dangerous recent development for Obama is the collapse of his Gallup Job Approval number over the past few days:
Obama Job Approval Approve DisApprove
10/20-22/2012 51 45
10/21-23/2012 53 42
10/22-24/2012 51 44
10/23-25/2012 48 47
10/24-26/2012 46 49
As you can see from the above chart, his JA has plummeted from +9 on 10/24 to -3 today. That is a rapid and ominous drop just a few days. He hasn’t been at 46% JA since early Sep, and all of a sudden it happened over the past few days. I wonder what has caused that: His dissembling on Libya, shrillness on the stump, Binders/Big Bird/Bull Sh…/Romnesia/blah blah blah, all of the above and more? Who knows. I do know that a collapsing job approval number is not what Obama wants during the last days of a campaign when he needs every undecided voter he can get.
Gallup was off 5% in their last poll of the 2008 election. They have regestered voters as exactly even nationwide, but a 5% Republican shift on who goes to the polls. The historical shift has averaged 1%-1.5%. They have shown +/- 10 point swings in their likely voter output. They have problems with that. I don’t think they are biased, but when they historically are off the consensus, they have been off by the same amount from the voters. I think the logical conclusion is they have a problematic likely voter model which they tweak constantly. Rasmussen has had a Republican hosue effect in the last several elections of 1% to 2%. So, it’s reasonable to factor that in. So, when Nate uses five polls, including Rasmussen to gauge Ohio, he would include the house effect, just as he had for polls that vary from the mean consistently for the Democrats. Now, in Ohio, Rasmussen has the race even. The average is 2.6% Obama. Just from the polls, a very possible event is a narrow popular vote win for Romney, but an Electoral college win for Obama. Now, I’d be happy, if I’m wrong, to admit it in 10 days. But, you have never answered my questions. Do you accept data that contradicts your theory or must the universe conform to your a priori knowledge of Truth. If facts differ from what you think, are they wrong, or are you? And, I’d really be curious to see if you accept modern physics and biology or think that scientists are part of the evil conspiricy too.
“@Smooth Jazz: Let me ask: what is your test for how we will judge these polls after Nov 6th? Is it a case of who wins? The popular vote totals vis-a-vis the polling? The EC predictions? What?”
The big test for me will be how all these massive Dem 9%+ – 10%+ samples compare to actual turnout. The pollsters sponsored and paid by the big left wing media organs such as Marist (MSNBC), Quinnipiac (CBS), Langer Associates (ABC), etc and partisan pollsters such as PPP have been the most egregious in posting polls with Dems turnout assumptions showing 10% more than Repubs in swing states where Rep & Dem registration have historically been even – At a time when non partisan, well known pollsters such as Gallup, Pew & Rasmussen are finding that Rep enthusiasm, turnout and participation are at historical highs this year, and may equal Dems in overall turnout %s. A development that favors Romney since he is obliterating Obama among independents even in states like OH.
My sense is the post mortem on this election will be brutal if it turns out the media sponsored pollsters juiced their samples all along to boost Dems as conservatives have been saying for months. The ultimate arbiter is the election and elections have a way of vetting pollsters for the long haul. Remember Research 2000?? The DailyKOS sponsored pollster that went out of business when it was revealed they tampered with the numbers to please their KOS benefactors a few years back. The knives are going to be out after this election and the Marists & Quinnipiacs of the world will be under a microscope when this is all over.
“Do you accept data that contradicts your theory”
YES. That said, the fact that Obama’s Gallup Job Approval number has collapsed 12 points in a few days (from +9 to -3) may be more signifacant that what any head to head poll is saying today. The suddenness of this job approval drop is most telling. I would pay attention to that if I were you. No Nate Silver simulation or theory on modern physics will save Obama if that job approval drop is real and continues.
smooth:
Have you seen the latest Fox poll for OH (link)? It shows Obama +3. For party ID, it shows D +8. So that proves Fox are “rabid Obama sychophant[sic][s],” right?
Are you sure? The most recent CNN poll for OH shows Obama +4. From that poll:
As usual, your ‘facts’ are divorced from reality.
And I’m still waiting for you to tell us where you found your “various Secy of State County officials in OH” numbers.
Maybe these words shouldn’t be surprising coming from this odious (and perhaps envious) person who is so delusional that he predicts that Romney will win the election with 321(!) electoral votes…
Oh look, suddenly Gallup is the Holy Grail of polling for Smooth Jazz…how amazing…
@An Interested Party: SmoothJazz, have you any idea of how to even linear regression to a time series? You don’t cherry pick number. For example, when after the disasterous first debate, Gallup had Obama rising from a 2% lead in popularity/unpopularity to an 11% lead in 2 days (with 1 day in common, that probably wasn’t even statistics, it was adjustment. When you evaluate a 3 day trend in a time series you have to look at whether wild swings occured before. That helps you put it in perspective. What I think you are telling me is that you accept data that confirms your prejudices and reject anything that doesn’t. That’s part of magical thinking. You know, like thinking you can cut taxes 20% and take away 4% of deductions for the top 1% and it balances out. There is a reason Bush Sr. called this type of reasoning “voodoo.” You know, we need his type of Republicans around, but your type has pushed them away, leaving people who think that Andrew Mellon had it right.
@Doug Mataconis:
“Silver is not a pollster. He does not conduct his own polls. He’s a statistician who analyzes the poll results and other factors and is attempting to determine where the race actually stands at a given point in time. “
I realize you were answering Kansas Mom’s parsing post. Yes, technically, the word pollster is reserved for only those ‘whose work is taking public opinion polls.’
However, to clarify my own usage, I have always applied the word ‘pollster,’ in a broader layman’s vernacular, meaning anyone who deals with polling, whether it is directly getting the data (polling), interpreting it through some mathematical model, broadly looking at someone else’s data and analyzing trends, or simply knowing the electoral map so well that one’s appraisal of it becomes a valuable and credible compass in signaling possible election results —much like people refer to an entertainer, an umbrella type of reference to the many subsets that comprise an entertainer (actor, musician, vocalist dancer etc.) Therefore I refer to people like Stuart Rothenberg, Charlie Cook, Larry Sabato, Doug Schoen, Dick Morris, Nate Silver as “pollsters”, along with the likes of Rasmussen and others who actually directly poll people for issue or election preferences. Even Michael Barone is someone I would throw into this pollster bin, as he is one of the most astute people around, IMO, as to insightfully gleaning trends, and objectively yielding predictions based upon said trends. He seems to know the background and inside details for almost every county in the nation — or so it seems.
I guess, though, that crossing every t, dotting all the i’s, and abiding by the most accurate and rigid Webster’s dictionary definition of a word is a must on OTB, unless one wants to get slammed or accused of gross ignorance. Nevertheless, in the 21st century of culturally loose jargon, a President who slips into ebonic cadence whenever it suits him, a VP who salutes Iowa when he is in Ohio (and no one here seems to care), it seems to become more an easy tool, or sophomoric indulgence of sheer hairsplitting, in countering another’s opposing POV. But, again, the pettiness of this campaign seems to have trickled down from the President to his partisan supporters, in this waning part of the election season. So, by that measurement alone, it all makes sense.
@jan:
I don’t think that that is a common distinction. I think most people differentiate between people who take polls (pollsters) and those who analyze them.
Pardon?
Yes, that sounds very serious.
@jan:
He denotes, as an example of this, how poorly Silver’s predictions turned out to be in the outcome of House seats taken by the R’s in 2010.
Silver’s predictions in 2010 were quite good, in fact. The problem is that all 435 House seats are up for grabs every two years, and there just aren’t enough polls done to make a solid prediction in a lot of those races. His Senate and governor race predictions were spot on, and he was a bit off, but not much, in the House.
Look at Silver’s site. You see presidential and Senate analysis. No House. Not enough data.
@jan:
I have always applied the word ‘pollster,’ in a broader layman’s vernacular, meaning anyone who deals with polling,
Were you doing that to deliberately make the reader think you didn’t know what you were talking about? It worked.
Another ‘”pollster,” Election Projection, disagrees with Nate Silver. Prediction: “State polls will shift toward Romney before election day.”
Basically, Obama’s job approval numbers are going south (hopefully, and deservedly so, because of the disgraceful Benghazi cover-up), while Romney’s national numbers are trending upward.
Who would have thought MN would be this close!
By her definition, Jan is a pollster.
@jan: Another fellow into “unskewing” I see, Actually, I remember his blog from the past, but do not recall much about him.
@jan:
I’m glad she calls him “President” here, because it’s pretty obvious what Jan prefers to call Obama.
@Steven L. Taylor:
I don’t think that that is a common distinction. I think most people differentiate between people who take polls (pollsters) and those who analyze them.
I don’t.
Believe it or not people do have their own verbal idiosyncrasies and ways of self-expression, including enclosing all facets dealing with polling under one title or term —–> POLLSTER. Again, sometimes the way some of you nitpick terms, swamps any other meaningful discussions dealing with bigger issues of this election. Maybe that is your underlying intention.
“Who would have thought MN would be this close!
Minnesota Star Tribune(Red Star)
Obama 47%
Romney 44%
FWIW: +5D, and this is within the MOE”
Good GAWG, This is a truly ominous development that even Nate Silver cannot simulate away. Forget about the head to head number. The fact that the Dem incumbent is at 47% in a Blue state not on anyone’s radar this close to the election – at a time when his job approval rating is collapsing nationwide probably due to the Bengazi cover up, the shrillness on the campaign, etc. – is a VERY, VERY dangerous development for Obama. The fact that in nearby IA, the influential Des Moines Register has endorsed Romney may even have a spillover influnce in next door MN.
This guy happens to think MN may flip RED on election day. MN is always a tough state for any REP but at least he presents some cogent arguments why MN may shock the world on 11/6:
http://battlegroundwatch.com/2012/10/19/election-night-surprise-why-minnesota-will-turn-red-on-november-6/
According to Rasmussen, MN is Obama +5. But now I guess Rasmussen doesn’t mean anything, right?
Last time Obama did worse than +4 in any poll in MN: never. But keep hope alive.
Who would have thought Smooth would continue to ignore all the questions he was asked!
@jan: The “pollster” you linked to, the one that disagrees with Nate Silver, is projecting that Obama will win with 290 EVs.
@jan:
Well, yes, people do often have idiosyncratic ways of expressing themselves. Of course, once it is made obvious that such idiosyncrasies are actually errors, it behooves said persons to change the deployment of the language in question, yes?
The stubborn refusal to use appropriate terms when one knows the correct ones is problematic, is it not?
This is not nitpicking, it is accuracy.
I am unclear on why you would wish to be willfully inaccurate.
@jukeboxgrad:
It is quite the shocking development, to be sure.
@Scott O: Yes, but in a blog post he writes:
This appears to be based largely on preference and the reading of the Gallup approval numbers (with some “unskewing” tossed in as well).
@Smooth Jazz:
There are multiple problems with this constant refrain about +X D, but a major one is that you are ignoring, and always ignore, the large independent component in these polls. These are not true independents, but rather are made up primarily of strong leaners one way or another (and, indeed, the trend of late has been for more Rep leaner to identify with “independent” in polls). As such, we really don’t know the real R/D breakdown in these samples the way you think we do.
The actual number of voters who will vote for someone other than the R or D at the presidential level is likely to be in the low single digits. The best indicator in all these polls as to the true R/D breakdown in the sample (as it pertains to voting) is how they answer the Obama v. Romney question, not their partisan self-ID (which is not a demographic characteristic that remains static the way race and gender do). All of the +X D business is really missing the point about what the numbers, especially the Independent numbers mean.
The poll’s methodological note states:
This is not a weighting (the way Rasmuessen does), this is the result of how those polled identified themselves. This is a skew only in the sense that those in the sample self-skewed, so to speak.
@Smooth Jazz:
Remember, Minnesota is the state of nuts like Michele Bachmann
@Steven L. Taylor: FWIW, the noted “pollster” 7-11 shows Obama ahead 59 – 41.
@jan:
I am writing a comment on this blog. Thus, I am a BLOGGER.
@jan:
You sound an awful lot like this guy: ‘When I use a word,’ Humpty Dumpty said, in rather a scornful tone, ‘it means just what I choose it to mean — neither more nor less.’
You would look much better if you just said “Sorry, my mistake. I should have used a different word.”
@Smooth Jazz:
I’m calling bs on you, and not for the first time. Just for starters, the Quinnipiac University Polling Institute is NOT “sponsored and paid for” by CBS. I don’t even want to waste the time refuting your other ignorant, low-information, tinfoil hat crap.
I hear you. That’s why people refer to rent-a-cops at the mall as “police officers”…
@ Jan
Scott Elliot, the man behind “Election Projection” is a contributing writer at PJ Media – you know- PJ Media, where our own Eric Florack is a contributing writer. In other words, he is part of the right wing media, and he writes for people who’s standards are low enough that they let bithead publish.
Sorry hon, everyone outside of the Foxverse knows that the Romney surge has run out of gas. There is a reason why every third word on Romney’s website and FB page is “momentum.” It’s because they are trying to create the illusion that they still have something that has been lost.
>He denotes, as an example of this, how poorly Silver’s predictions turned out to be in the outcome of House seats taken by the R’s in 2010.
You and Krumm both seem to have trouble understanding the concept of a probability estimate. When there’s, say, a two-thirds chance of something happening, it means that we would expect it not to happen one out of every three times. In other words, if Silver’s 70% projections always happen, that would suggest a problem with his model, not a vindication of it.
@jan: By your definition, you are a pollster. Dear God, Wittgenstein is rolling in his grave. 🙂
@jan:
It seems possible that you don’t realize that Rasmussen already has no credibility.
As for the rest of your nonsense, I’d suggest you remove yourself from the right wing bubble. your brain seems to have atrophied.
“Nate Silver is a man of very small stature, a thin and effeminate man with a soft-sounding voice that sounds almost exactly like the “Mr. New Castrati” voice used by Rush Limbaugh on his program.”
Translation: “Feel free to ignore virtually every other argument I’ve made on any subject.”
The “new castratti” label comes along as certain other words that he would like to use are no longer fashionable. It’s a tacit admission of not only being wrong, but the perception of being forced to wallow in it, even though he’s probably never kissed a boy. Probably.
Hard to have any sympathy for THAT.
This NY Post columnist support for Romney probably mirrors why many dems and indies are switching their vote this year.
@jan: It is noteworthy that a great deal of the time you respond not with your own argument (then backed with evidence) but, instead, quote the opinions of others.
@anjin-san:
Sorry hon, everyone outside of the Foxverse knows that the Romney surge has run out of gas.
It might be time to give you a reality check. The real news bubble organization, which reports news on nothing but a left-wing slant, is MSNBC. While Fox does lean right, it interviews people from all sides of the political spectrum allowing for diverse views of a diverse news-scape. If you primarily watch MSNBC, though, I truly can see why you feel the way you do about this race, and are unable to see even the possibility of an Obama loss.
BTW, have you even seen the audience size of the Foxverse versus MSNBCverse? The latter is shrinking as more people are tuning out, while the former is growing because more are tuning in. The reason for that is because Fox gives full news coverage and is simply a better and more honest news medium, despite the bad rap it gets from people like you.
@jan:
I wish that this were the case and that, in fact, what people want is a comprehensive coverage of the news. But people don’t; they want to be entertained. Fox (and MSNBC) are mostly infotainment.
If pure news was what people wanted there would be huge demands for news and the prime time programming on all these cable channels would not be commentary programs.
@Steven L. Taylor:
It is noteworthy that a great deal of the time you respond not with your own argument (then backed with evidence) but, instead, quote the opinions of others.
And, Steven, why is that noteworthy?
If I see a column with ‘noteworthy’ comments, why should I summarize when I can simply post an excerpt giving credit to the writer?
Also, FWIW, out of 8 posts (including this one), only 3 of them had excerpts dealing with poll numbers, an election prediction, and the last one needed no argument or augmentation from me, as it plainly spoke for many of us who have been disappointed in Obama’s presidency (which was duly noted in my opening line).
@Steven L. Taylor:
What news stations do you watch, Steven?
@jan:
Two reasons:
1. You often seem to be outsourcing your argument to others (although, I will grant, sometimes quote are of use, but usually a quote alone is less effective).
2. The preponderance of your links (and I don’t just mean in this thread) are to opinion columnists/bloggers engaging in opinion writing. Opinions can’t be used to bolster opinions. Evidence is needed, not just people who agree.
@jan: I used to consume a heavy diet of both Fox with some MSNBC on the side. At the moment I watch almost no cable news, save when there is some live event going on. When it is an event that might spur commentary I often flip around to different outlets to see what different folks are saying.
I get most of my news from reading I also consume quite a bit of NPR (Morning Edition and All Things Considered, but also a lot of various program (On Point, Diane Rehm, Forum, several others–all of which tend to have guests who are from various sides, but actually interested in discussion rather than in playing cartoon roles on the screen).
I do usually watchMeet the Press and Fox News Sunday each week, and occasionally partake of the roundtable on This Week and sometimes watch Fareed Zakaria’s GPS (which is the newsiet and most wonky of all these shows)
In all seriousness, I think that part of national problem is that most of the so-called “news” that people consume really isn’t news, but rather is entertainment. I used to listen to a substantial amount of talk radio (Limbaugh, etc.) and watch a lot of Fox and MSNBC, but after a while I decided I wasn’t learning much and, further, that almost all it wasn’t news, but mostly predominantly opinion and personality.
@Steven L. Taylor:
Americans today can’t agree on what is news. I’d bet if we polled people (leaving out Rasmussen and Gallup) we’d find out that 75% of the people consider opinion shows to be news. Americans do not believe in a common reality.
Also, I’ve stopped watching the Sunday morning talkies – it’s a waste of cable bandwidth to watch. Stuff like – “and today Bob Schieffer speaks with John Cornyn about Mitt Romney’s chances …” Why watch that? What can possibly be learned? It’s empty calories. So I’ve stopped watching MTP, Bob Schieffer and all those shows. I feel like I get better political coverage by watching Jon Stewart and Bill Maher (at least they entertain me.).
@Steven L. Taylor:
What’s the most frustrating — and this isn’t just directed at Jan — is the sheer number of commenters who continue to make this mistake. What’s equally frustrating is that in the past a number of us (in particular Steven and myself) have contributed rather extensive comments explaining the *difference* between fact and opinion and why one is preferable to the other (here’s one example from an exchange we had months ago — sadly the message did not seem to be conveyed).
@jan’s recent link to Scott Elliott’s site is a prime example of this confusion between facts and opinion. It’s particularly noteworthy because it demonstrates how the current facts (as presented by Elliott himself) undercut both his and your *opinion*.
As noted above the fact’s at his site (i.e. that his OWN current electoral projections based on current polling predicts a electoral outcome for Obama (SE: 290 electoral votes) that is exactly inline with Silver’s current projection (S: 295.5).
As you point out, his argument that Obama will lose is based on the *belief* that Democrats are being oversampled:
Again, as I pointed out up thread, by his own admission your other source Mr. Krumm, in 2008 made a wildly off prediction, again predicated on *belief* rather than the facts at hand:
To a point made earlier, one of the primary difference between a professional and an amateur is the attempt to control for personal bias.
We need to separate the models/algorythms in meta-poll analysis (which typically present a sort of fact) from the analysis of those polls (which more opinion enters into) to the predictions (which often have large amounts of belief built into them).
@ Jan
I’ve tuned into MSNBC for perhaps an hour this entire year – is that enough to brainwash me?
@anjin-san:
Possibly. I can actually say that I have never seen the Ed Show and while I have seen clips of Maddow, I have never watch her show, either. I used to (years and years ago) watch Hardball with some regularity, and would occasionally see Countdown or Scarborough Country.
I actually see more of Fox News than any other channel, as it is on wall-to-wall when my in-laws visit or if I go to my grandparents’.
An interesting bit of information about Gallup was in todays Obama’s approval/disapproval rating. It is a three day running average and went from -3% yesterday to even today. If Gallup didn’t get this by readjusting (I’m pretty sure this is among regestered voters, so their likely voter model tweaks don’t get in here), that would mean a 9% difference in approval measured last Wednesday and last Saturday. That gives a flavor for how noisy this measurement is.
@Steven L. Taylor:
To be fair to Fox, during it’s news blocks, they do have solid journalists/anchors on the channel (Shep Smith in particular). The problem is the channel (and the majority of its programs), like much of MSNBC, has adopted the Limbaugh model of blurring news and infotainment.
That said, I think the following two quotes Jan made in different parts of this thread demonstrate the larger issue:
I’m not arguing that the NYT doesn’t have it’s biases (both in terms of editorial, and even how the news paper is laid out). It’s the idea that Fox is somehow better than the NYT’s or the suggestion that it “selectively plac[es] news items […] not for timely relevance or worthiness, but more for how such news might either positively or adversely effect the [Republican] party.”
One *fact* based example that demonstrates Fox’s News Editorial bent in support of Republicans is the recent *fact* that Richard Murdock’s controversial rape comments were not reported *a single time* on Fox in the 24 hours after they were made.
I challenge anyone to present a good reason for this other than based on a desire not to adversely affect the Republican party.
I don’t even own a TV, and my experience with MSNBC as well as Fox is limited to the occasional clip someone links to. I am a regular watcher of The Daily Show via its website, however. I’m not a subscriber to any newspaper, and most of my news comes from what’s linked to at AOL or Google or Yahoo or the blogs I read.
@al-Ameda:
@mattb:
I think that Shep Smith is legit and The Fox Report is fine (although it has long struck me as flashier and more sensationalistic than I prefer, which is a view I held even back when I regularly watched FNC). Still, even if we treat Special Report and the Fox Report as the news block, that is only 2 hours and a substantial portion of SR is dedicated to commentary. The prime time block is nothing but opinion.
@Steven L. Taylor:
No argument there.
@DanM:
And how! Which points to how important it is too look at polls in aggregate and over time, versus overemphasizing any single data-point (or even just looking at two cycles for an individual poll).
@al-Ameda:
Our current media environment is arguably a far more traditional “American” media environment (in terms of politicization) than the one most of us grew up in. The majority of the twentieth century was really a historical aberration when it comes to news and mass media. That “blip” (which led to a broader “common reality”) was created by the convergence of a number of forces, in particular: rapid economic growth, the rise of hierarchical mass broadcast media, the consolidation of newspapers, and the professionalization of journalism.
@mattb:
Indeed, which is the point that so many seem not to understand.
You raise an interesting point, and one that is actually quite depressing.
@DanM:
I was thinking something similar when I saw those numbers from Gallup. In another thread, a commenter said that the plummet in Obama’s approval numbers at Gallup proved – something or other. I thought that Gallup alone was a pretty slender reed to be using to hold up a point, but at least it was an objective data point, rather than just the opinion of someone who held the same viewpoint they did.
One more interesting factoid about Ohio. An average was done of two types of polls: human calls (including cell phones) and robocalls of land lines. The former averaged Obama by 5% the last week or so, the latter called it even. The demographics of folks who respond to a landline is different than those who own a cell phone. They are older and whiter. So, this appears to show a systematic problem with robocalls…which Rasmussen does, but Gallup does not. I put in the latter for completeness, I’m trying to be objective as possible.