Unskewing the polls?

[Denny Crane]

Interesting review of the 2008 voter turnout by American University.

http://i2.cdn.turner.com/cnn/2008/images/11/06/pdf.gansre08turnout.au.pdf

A downturn in the number and percentage of Republican voters going to the polls seemed to be the
primary explanation for the lower than predicted turnout. The percentage of eligible citizens voting
Republican declined to 28.7 percent down 1.3 percentage points from 2004. Democratic turnout
increased by 2.6 percentage points from 28.7 percent of eligibles to 31.3 percent. It was the seventh
straight increase in the Democratic share of the eligible vote since the party’s share dropped to 22.7
percent of eligibles in 1980.

...

“Many people were fooled (including this student of politics although less so than many others) by
this year’s increase in registration (more than 10 million added to the rolls), citizens’ willingness to
stand for hours even in inclement weather to vote early, the likely rise in youth and African
American voting, and the extensive grassroots organizing network of the Obama campaign into
believing that turnout would be substantially higher than in 2004,” said Curtis Gans, CSAE’s
director. “But we failed to realize that the registration increase was driven by Democratic and
independent registration and that the long lines at the polls were mostly populated by Democrats.”

Gans attributed the GOP downturn to three factors: 1) John McCain’s efforts to unite the differing
factions in the Republican Party by the nomination of Governor Sarah Palin as vice-presidential
nominee was a singular failure. By election time many culturally conservative Republicans still did
not see him as one of their own and stayed home, while moderate Republicans saw the nomination
of Palin reckless and worried about McCain’s steadiness. 2) As events moved towards Election Day,
there was a growing perception of a Democratic landslide, discouraging GOP voters. 3) The 2008
election was a mirror image of the 2004 election.

(My note - 28.7% of eligible voters is not the same thing as 28.7% of those who turned out)

 

[Denny Crane]

And this Fordham.edu study says Rasmussen was the most accurate pollster in 2008 (Obama/McCain popular vote).

http://www.fordham.edu/images/acade...ccuracy in the 2008 presidential election.pdf

Not only does it say that, but it also states:

Four of these polls appear to have overestimated McCain support (indicated with a * below), while most polls (17) overestimated Obama strength. Pre-election projections for two organizations’ final polls—Rasmussen and Pew—were perfectly in agreement with the actual election result (**).

So, barfo, is the science behind the 17 wrong polls somehow good?

Rasmussen is biased. My ass.

 

[barfo]

And this Fordham.edu study says Rasmussen was the most accurate pollster in 2008 (Obama/McCain popular vote).

http://www.fordham.edu/images/acade...ccuracy in the 2008 presidential election.pdf

Not only does it say that, but it also states:

Four of these polls appear to have overestimated McCain support (indicated with a * below), while most polls (17) overestimated Obama strength. Pre-election projections for two organizations’ final polls—Rasmussen and Pew—were perfectly in agreement with the actual election result (**).

So, barfo, is the science behind the 17 wrong polls somehow good?

Rasmussen is biased. My ass.

Yes, Rasmussen is biased. Your ass.

Getting one poll correct doesn't prove a pollster is unbiased, except in regards to that one poll. Having a significant history of biased results over a large number of polls, on the other hand, does prove a pollster has a bias. Rasmussen has a significant history of biased results over a large number of polls, therefore, Rasmussen is biased.

barfo

 

[Denny Crane]

Yes, Rasmussen is biased. Your ass.

Getting one poll correct doesn't prove a pollster is unbiased, except in regards to that one poll. Having a significant history of biased results over a large number of polls, on the other hand, does prove a pollster has a bias. Rasmussen has a significant history of biased results over a large number of polls, therefore, Rasmussen is biased.

barfo

But he doesn't have a history of biased results. They're accurate, but you don't like the results.

The bias is in those 17 polls that overestimated Obama strength. You like those poll results, but they're not accurate.

So what exactly are you arguing about?

 

[barfo]

But he doesn't have a history of biased results. They're accurate, but you don't like the results.

The bias is in those 17 polls that overestimated Obama strength. You like those poll results, but they're not accurate.

So what exactly are you arguing about?

I accept your surrender, since you are out of ammunition.

barfo

 

[Denny Crane]

I think you ran out in post #33

 

[barfo]

I think you ran out in post #33

Reread the link in post #17 for the evidence behind post #33.

barfo

 

[Denny Crane]

Reread the link in post #17 for the evidence behind post #33.

barfo

Right. You were out of ammo in post #17. Thanks for correcting me

I'll paraphrase something I read in one of Nate Silver's blog posts this morning:

"People like to cite Rasmussen's considerably good track record (which really is good), but I don't like the way he does his polls."

 

[barfo]

Right. You were out of ammo in post #17. Thanks for correcting me

Since post #17 wasn't even mine, I'll assume you are just talking out your ass today instead of at least feebly trying to make an argument?

I'll paraphrase something I read in one of Nate Silver's blog posts this morning:

"People like to cite Rasmussen's considerably good track record (which really is good), but I don't like the way he does his polls."

I'll paraphrase something I read in your posts today:

"I've got no argument whatsoever, so I'm just going to pretend I'm too dumb to understand that"

barfo

 

[Denny Crane]

Since post #17 wasn't even mine, I'll assume you are just talking out your ass today instead of at least feebly trying to make an argument?



I'll paraphrase something I read in your posts today:

"I've got no argument whatsoever, so I'm just going to pretend I'm too dumb to understand that"

barfo

What I read in your posts is "I refuse to listen to logic. I've made up my mind and no amount of evidence will change it."

 

[barfo]

What I read in your posts is "I refuse to listen to logic. I've made up my mind and no amount of evidence will change it."

Ok.

barfo

 

[Denny Crane]

Post #34 is a good example. Fact: 17 polls overstated Obama's support and were off badly. Do tell, why should we believe those pollsters now?

 

[GriLtCheeZ]

Get a room guys.

 

[barfo]

Post #34 is a good example. Fact: 17 polls overstated Obama's support and were off badly. Do tell, why should we believe those pollsters now?

Since you obviously didn't read it the first time, I'll just repeat what I said before:

Getting one poll correct doesn't prove a pollster is unbiased, except in regards to that one poll. Having a significant history of biased results over a large number of polls, on the other hand, does prove a pollster has a bias. Rasmussen has a significant history of biased results over a large number of polls, therefore, Rasmussen is biased.

Similarly, getting one poll wrong doesn't prove a poster is biased, except in regards to that one poll.

And the others were not all "off badly" as you claim. Your own link says that the average pollster had Obama winning by 7.5% instead of 6.15%. It's not a huge error.

Should we believe pollsters because they were off by 1.36% (well within the stated margin of error for most polls) on one poll? Well, it's not a reason to disbelieve them. I generally think a fair bit of caution is needed when looking at any poll, no matter how reputable the pollster.

barfo

 

[barfo]

Get a room guys.

I thought this was our room. What are you doing in here?

barfo

 

[Denny Crane]

Since you obviously didn't read it the first time, I'll just repeat what I said before:



Similarly, getting one poll wrong doesn't prove a poster is biased, except in regards to that one poll.

And the others were not all "off badly" as you claim. Your own link says that the average pollster had Obama winning by 7.5% instead of 6.15%. It's not a huge error.

Should we believe pollsters because they were off by 1.36% (well within the stated margin of error for most polls) on one poll? Well, it's not a reason to disbelieve them. I generally think a fair bit of caution is needed when looking at any poll, no matter how reputable the pollster.

barfo

17 pollsters using similar methodology (you called it science) and getting it very wrong says a lot about the methodology.

7.5% vs. 6.15% is a ~25% error, which is very significant. In fact, it's about in line with 25% of republicans staying home in 2008.

See post #30.

 

[barfo]

17 pollsters using similar methodology (you called it science) and getting it very wrong says a lot about the methodology.

Are you actually claiming that all pollsters who overestimated Obama's winning margin used methodology similar to each other, but different from Rasmussen? I think that might be an assertion that you don't have the facts to back up. Feel free to prove me wrong.

7.5% vs. 6.15% is a ~25% error, which is very significant. In fact, it's about in line with 25% of republicans staying home in 2008.

See post #30.

Uh, ok. Your initial post in this ridiculous thread was claiming that the pollsters were making errors of 7 or more percentage points. Glad you now agree that was horseshit.

barfo

 

[Denny Crane]

Are you actually claiming that all pollsters who overestimated Obama's winning margin used methodology similar to each other, but different from Rasmussen? I think that might be an assertion that you don't have the facts to back up. Feel free to prove me wrong.

Post #7. And yes, there is a strong similarity used by the polling firms that are in error - weighting the raw data as if 24% of the electorate is republican, based upon 2008 exit polling data. Something Rasmussen does not do.

Uh, ok. Your initial post in this ridiculous thread was claiming that the pollsters were making errors of 7 or more percentage points. Glad you now agree that was horseshit.

barfo

You have a comprehension problem. If the error is 7 or more percentage points, the polls are that biased.

Where the unskewed polls may have an issue is the enthusiasm of the voters. Obviously, the polls can be Romney +7, but if 25% of the republicans stay home and don't vote, he's not going to win by 7.

 

[maxiep]

If turnout matches 2008 in terms of race, party identification and gender, then President Obama will win. It will be by a lower margin this time, but he will win. However, if the composition of the turnout has changed the ways the polls say they have (polls on party identification, intensity of interest, etc.), then Gov. Romney is in much better shape than most of the polls would lead us to believe.

 

[barfo]

Post #7. And yes, there is a strong similarity used by the polling firms that are in error - weighting the raw data as if 24% of the electorate is republican, based upon 2008 exit polling data. Something Rasmussen does not do.

I don't think you've shown that all other polling firms weight the data solely or primarily based on 2008. You've asserted that, but I don't believe you.

You have a comprehension problem. If the error is 7 or more percentage points, the polls are that biased.

Right, but in the one and only poll you are willing to discuss (the one Rasmussen got right, in 2008) the error/bias was nowhere near 7 percentage points.

Where the unskewed polls may have an issue is the enthusiasm of the voters. Obviously, the polls can be Romney +7

Actually, the polls can't be Romney +7, because that would be a bullshit result. That's why NO pollster, not even your beloved Rasmussen, is reporting Romney +7. It's because it isn't true.

, but if 25% of the republicans stay home and don't vote,

or even if they don't...

he's not going to win by 7.

He's not going to win by 7. Or at all, unless something significant happens in the next few weeks.

barfo

 

[barfo]

If turnout matches 2008 in terms of race, party identification and gender, then President Obama will win. It will be by a lower margin this time, but he will win. However, if the composition of the turnout has changed the ways the polls say they have (polls on party identification, intensity of interest, etc.), then Gov. Romney is in much better shape than most of the polls would lead us to believe.

Yes, because pollsters pay no attention to polls, so they have no idea of changing demographics and interest and party ID. They certainly don't try to factor that information into their poll results.

barfo

 

[Denny Crane]

LOL, barfo.

http://www.hughhewitt.com/transcripts.aspx?id=1c1a7295-7ce1-47e7-8074-4ce24952aceb

 

[barfo]

LOL, barfo.

http://www.hughhewitt.com/transcripts.aspx?id=1c1a7295-7ce1-47e7-8074-4ce24952aceb

How is that LOL barfo? Seems to me he's explicitly saying that he wasn't doing what you accused him of (applying the 2008 turnout). He said he just took the answers people gave as their party ID.
That may or may not make him a good pollster, but it isn't any feather in your argument's cap.

barfo

 

[Denny Crane]

How is that LOL barfo? Seems to me he's explicitly saying that he wasn't doing what you accused him of (applying the 2008 turnout). He said he just took the answers people gave as their party ID.
That may or may not make him a good pollster, but it isn't any feather in your argument's cap.

barfo

It's a state poll.

And there you now have 3 pollsters (the other 2 in the video you don't want to watch) saying Democrats are being oversampled.

I like this pollster's answer:

HH: I mean, when does it become unreliable? You know you’ve just put your foot on the slope, so I’m going to push you down it. When does it become unreliable?
PB: Like the Supreme Court and pornography, you know it when you see it.


And this one:

HH: Do you expect Democrats, this is a different question, do you, Peter Brown, expect Democrats to have a nine point registration advantage when the polls close on November 6th in Florida?
PB: Well, first, you don’t mean registration.
HH: I mean, yeah, turnout.
PB: Do I think…I think it is probably unlikely.

 

[Denny Crane]

Gallup surveyed 177,670 adults and finds...

http://www.gallup.com/poll/156437/heavily-democratic-states-concentrated-east.aspx

August 3, 2012
So far in 2012, 44% of all U.S. adults have identified as or lean Democratic and 40% are Republican.

And here's a good explanation of how the pollsters use (but obscure) the 2008 exit polling data:

http://reason.com/blog/2012/09/18/are-public-opinion-polls-exaggeratin-oba

According to Chris Jackson at Ipsos-Reuters, “most research organizations use a combination of prior voting behavior, interest in the election and self-report likelihood to vote to categorize likely voters. ...Some pollsters also use ‘voter lists’ or commercial lists of people who voted in the last election instead of screening these individuals from the population.”

Rasmussen gives a vague explanation here, “The questions involve voting history, interest in the current campaign, and likely voting intentions. Rasmussen Reports determines its partisan weighting targets through a dynamic weighting system that takes into account the state’s voting history, national trends, and recent polling in a particular state or geographic area.”

ABC News explains, they “develop a range of ‘likely voter’ models, employing elements such as self-reported voter registration, intention to vote, attention to the race, past voting, age, respondents’ knowledge of their polling places, and political party identification.”

As Huffington Post’s Mark Blumenthal reports, “CNN has published no explanation of how they select likely voters.”

"Without seeing the demographic composition of the likely voters in each poll, it appears that several polls are extrapolating 2008 turnout beyond what will actually occur in 2012. Consequently, these polls may overestimate Democratic turnout and thus Obama’s lead in the polls. "

 

[barfo]

It's a state poll.

yes.... and your point is?

And there you now have 3 pollsters (the other 2 in the video you don't want to watch) saying Democrats are being oversampled.

No. You are totally misrepresenting what this guy said. He said he just polled people and asked them what party they belonged to, and they said D. That's completely different than adjusting the sample to fit a previously defined D vs. R breakdown.

I like this pollster's answer:

HH: I mean, when does it become unreliable? You know you’ve just put your foot on the slope, so I’m going to push you down it. When does it become unreliable?
PB: Like the Supreme Court and pornography, you know it when you see it.


And this one:

HH: Do you expect Democrats, this is a different question, do you, Peter Brown, expect Democrats to have a nine point registration advantage when the polls close on November 6th in Florida?
PB: Well, first, you don’t mean registration.
HH: I mean, yeah, turnout.
PB: Do I think…I think it is probably unlikely.

He's obviously not very good in the interview, and maybe he doesn't believe his own poll. But this is totally at odds with what you were objecting to before. This isn't 'skewing' of the results. This is reporting the results without any skew.

barfo

 

[barfo]

Gallup surveyed 177,670 adults and finds...

http://www.gallup.com/poll/156437/heavily-democratic-states-concentrated-east.aspx

Wow. The most heavily democratic states are in the east. That's fucking amazing. No one ever would have guessed that, unless they ever looked at any poll or voting data for a minute or two.

August 3, 2012
So far in 2012, 44% of all U.S. adults have identified as or lean Democratic and 40% are Republican.

And here's a good explanation of how the pollsters use (but obscure) the 2008 exit polling data:

http://reason.com/blog/2012/09/18/are-public-opinion-polls-exaggeratin-oba

According to Chris Jackson at Ipsos-Reuters, “most research organizations use a combination of prior voting behavior, interest in the election and self-report likelihood to vote to categorize likely voters. ...Some pollsters also use ‘voter lists’ or commercial lists of people who voted in the last election instead of screening these individuals from the population.”

Rasmussen gives a vague explanation here, “The questions involve voting history, interest in the current campaign, and likely voting intentions. Rasmussen Reports determines its partisan weighting targets through a dynamic weighting system that takes into account the state’s voting history, national trends, and recent polling in a particular state or geographic area.”

ABC News explains, they “develop a range of ‘likely voter’ models, employing elements such as self-reported voter registration, intention to vote, attention to the race, past voting, age, respondents’ knowledge of their polling places, and political party identification.”

As Huffington Post’s Mark Blumenthal reports, “CNN has published no explanation of how they select likely voters.”

"Without seeing the demographic composition of the likely voters in each poll, it appears that several polls are extrapolating 2008 turnout beyond what will actually occur in 2012. Consequently, these polls may overestimate Democratic turnout and thus Obama’s lead in the polls. "

Did you read that link you provided? It says that maybe pollsters are overstating Obama's lead by 1.5 percentage points. Maybe 1.5. Not 7. 1.5.

barfo

 

[chevyrunssometimes]

Dude, if you want to feel good about your chances, just click the heels of your ruby slippers together and repeat the phrase, there's no place like home...much simpler and you won't look half as silly.

 

[Denny Crane]

You answered your own question from post 56 in post 57.

And the 1.5 difference is in likely voters polls, which they are reporting as pretty much even. That is, the pollsters who do report likely voter poll results.

 

[barfo]

I didn't quite realize that we were just rehashing the republican conspiracy theory of the day, but here's a recap on our debate.

http://nationaljournal.com/2012-pre...s-too-many-democrats-in-your-surveys-20120925

barfo