Execellent article on polling

[mamooth]

It's normal in elections for each side to declare that polls they don't like are inaccurate, but this year is something special. One side, the conservatives, has hopped on the conspiracy train to crazytown. They're making up crazy stories about past polls being wrong, about the awful Rasmussen polls being right, about Democrats being oversampled, all to support a nutty conspiracy theory about deliberately rigged polls. It's not a very good conspiracy theory, given how it shows ignorance of how polls work.

Reputable pollsters -- that is, more or less everyone but Rasmussen -- don't aim samples at groups or massage the data with bizarre turnout models. The good pollsters simply take a large sample along lines to match census demographics. Party affiliation doesn't enter into the sample choice into in any way. There is no list of Democratic or Republican people or districts that pollsters decide to call. That's just barking lunacy.

Pollsters call their random samples, then they _ask_ about party affiliation. If the polls reported more Democrats, that simply means more random respondants self-identified as Democrats. Sorry conservatives, but it really is that simple. The news for conservatives is worse than they're willing to believe, in that there's been a shift to more people self-identifying as Democrats.

Alas, that's just reality, so it won't matter to the true believers here. They've got their conspiracy theory giving them false hope, so they're clinging to it. And like any good conspiracy, it's impossible to disprove it. After the election outcome matches the polls, the conspiracists will declare it "proves" how the dirty liberal poll-rigging plot was successful in shifting public opinion.

 

[hazlnut]

Apparently, 'excellent' now means confirming my preconceived beliefs without facts or reason.

Santa Claus is excellent to children...

 

[California Girl]

[
The University of Colorado has already called this race for Mitt Romney--based on historic data--and they have been right since 1980.
[

I hate to burst your bubble, but that Colorado study got nothing right.

That study was only recently concocted, working backwards, making the criteria fit the results in hindsight.

Sorry.

Not according to the actual researchers... and they have the data to back it up. They've been running these since 1980. You lied.

 

[NYcarbineer]

[
The University of Colorado has already called this race for Mitt Romney--based on historic data--and they have been right since 1980.
[

I hate to burst your bubble, but that Colorado study got nothing right.

That study was only recently concocted, working backwards, making the criteria fit the results in hindsight.

Sorry.

Not according to the actual researchers... and they have the data to back it up. They've been running these since 1980. You lied.

No. They. Haven't.

 

[NYcarbineer]

You realize your link doesnt prove what you write, right?

I asserted that RCP had Walker up the entire campaign and the link to RCP that I provided backs that claim up.

Does it hurt to be this stupid?

And I've shown your point is irrelevant. The media and polls clearly showed a close race where it didnt exist.

A simple question you won't answer:

How do you know where a race is a month before the election? How do you know what the election results would be if the election were held that day,

since no election is held that day?

Since you're claiming you know, explain to us how you know.

 

[Dr.House]

Do democratics really think turnout will be D+10 on election day?

Srsly?

 

[The Rabbi]

I asserted that RCP had Walker up the entire campaign and the link to RCP that I provided backs that claim up.

Does it hurt to be this stupid?

And I've shown your point is irrelevant. The media and polls clearly showed a close race where it didnt exist.

A simple question you won't answer:

How do you know where a race is a month before the election? How do you know what the election results would be if the election were held that day,

since no election is held that day?

Since you're claiming you know, explain to us how you know.

I'm not claiming anything of the sort, numbskull. I'm claiming the media published polling that showed a close race right up until the last week or so before the Election, where Walker cleaned the guy's clock. And since exit polling showed most voters in WI had made their minds up already months before the election I can only conclude the media tried to push a story line on the public that wasn't true. Just as they are doing now. Remember, WIsconsin was a dry run for Team Obama.

 

[NYcarbineer]

And I've shown your point is irrelevant. The media and polls clearly showed a close race where it didnt exist.

A simple question you won't answer:

How do you know where a race is a month before the election? How do you know what the election results would be if the election were held that day,

since no election is held that day?

Since you're claiming you know, explain to us how you know.

I'm not claiming anything of the sort, numbskull.

Yes you did. You said this:

The media and polls clearly showed a close race where it didnt exist.


The polls showed a 6 point race just before the actual vote. Walker won by 6.

If the polls showed it closer a month before that, how do you know it wasn't?

That's what you're claiming, so how do you know it.

 

[NYcarbineer]

Do democratics really think turnout will be D+10 on election day?

Srsly?

The Foxnews poll is +4 D's, with Obama ahead by 5.

And with Obama winning Independents. So even if you jiggered the +4 down to zero, Obama still wins that poll.

 

[The Rabbi]

A simple question you won't answer:

How do you know where a race is a month before the election? How do you know what the election results would be if the election were held that day,

since no election is held that day?

Since you're claiming you know, explain to us how you know.

I'm not claiming anything of the sort, numbskull.

Yes you did. You said this:

The media and polls clearly showed a close race where it didnt exist.


The polls showed a 6 point race just before the actual vote. Walker won by 6.

If the polls showed it closer a month before that, how do you know it wasn't?

That's what you're claiming, so how do you know it.

Jerk off. I posted numerous articles from before the election all showing the election was neck and neck.
Pay attention or don't post.

 

[Article 15]

Do democratics really think turnout will be D+10 on election day?

Srsly?

Anytime you want to put your sig where your mouth is, House Gimp, you let me know. Until then you are nothing but a coward trying to fool himself into thinking romney has a chance to win this.

Romney is done for. This election was over a couple of weeks ago.

 

[elvis]

What does D+10 mean?

 

[The Rabbi]

Do democratics really think turnout will be D+10 on election day?

Srsly?

Anytime you want to put your sig where your mouth is, House Gimp, you let me know. Until then you are nothing but a coward trying to fool himself into thinking romney has a chance to win this.

Romney is done for. This election was over a couple of weeks ago.

Darn and I forgot to vote.
Oh wait. The election is in November. We're nowhere near there. Obama will flair in the debates and other factors will enter into it.
Of course you've already agreed to leave the site when Obama loses so gloating won't help me much.

 

[Dick Tuck]

Why does the right wing always need a new conspiracy to justify the fact that they're failing?

 

[Article 15]

Do democratics really think turnout will be D+10 on election day?

Srsly?

Anytime you want to put your sig where your mouth is, House Gimp, you let me know. Until then you are nothing but a coward trying to fool himself into thinking romney has a chance to win this.

Romney is done for. This election was over a couple of weeks ago.

Darn and I forgot to vote.
Oh wait. The election is in November. We're nowhere near there. Obama will flair in the debates and other factors will enter into it.
Of course you've already agreed to leave the site when Obama loses so gloating won't help me much.

As usual, you are wrong. I haven't entered into any loser leaves town bets. They are for pussies who are too scared to face the music after a loss.

Anytime you want to agree to a sig bet, I'm your guy. I realize that you are way too much of a pussy to accept such a wager but the offer is there nonetheless.

 

[elvis]

I'd do it but I want Romney and points. Lol.

 

[Article 15]

Why does the right wing always need a new conspiracy to justify the fact that they're failing?

That would involve accepting personal responsibility for something, you know they would never dream of such a thing.

 

[del]

I'd do it but I want Romney and points. Lol.

hold out for twenty

 

[Article 15]

I'd do it but I want Romney and points. Lol.

Lol. Sorry, my friend, no dice. I'm only interested in betting with the chest thumpers who have convinced themselves that Romney is going to win.

 

[M.D. Rawlings]

It's normal in elections for each side to declare that polls they don't like are inaccurate, but this year is something special. One side, the conservatives, has hopped on the conspiracy train to crazytown. They're making up crazy stories about past polls being wrong, about the awful Rasmussen polls being right, about Democrats being oversampled, all to support a nutty conspiracy theory about deliberately rigged polls. It's not a very good conspiracy theory, given how it shows ignorance of how polls work.

Reputable pollsters -- that is, more or less everyone but Rasmussen -- don't aim samples at groups or massage the data with bizarre turnout models. The good pollsters simply take a large sample along lines to match census demographics. Party affiliation doesn't enter into the sample choice into in any way. There is no list of Democratic or Republican people or districts that pollsters decide to call. That's just barking lunacy.

Pollsters call their random samples, then they ask about party affiliation. If the polls reported more Democrats, that simply means more random respondants self-identified as Democrats. Sorry conservatives, but it really is that simple. The news for conservatives is worse than they're willing to believe, in that there's been a shift to more people self-identifying as Democrats.

Alas, that's just reality, so it won't matter to the true believers here. They've got their conspiracy theory giving them false hope, so they're clinging to it. And like any good conspiracy, it's impossible to disprove it. After the election outcome matches the polls, the conspiracists will declare it "proves" how the dirty liberal poll-rigging plot was successful in shifting public opinion.

The awful Rasmussen polls? Now who's being obtuse? The Rasmussen polls have been, pound-for-pound, the most reliable in recent years as measured against actual outcomes. Post-game outcomes. Objective fact! Empirical fact!

Here's another fact: all pollsters weigh and adjust their raw data for what they believe to be demographic anomalies against past experience. There is an element of guesswork, for lack of a better term, involved in the polling process, hopefully within the poll's margin of error.

They weigh for demographics that are thought to be underrepresented in the raw data based on the experience of the recent past, demographics that are thought to still be in play, but could not be contacted.

In this election cycle, those demographics just so happen to be mostly comprised of those who supported Obama in 2008!

Hence, your statements that "[p]arty affiliation doesn't enter into the sample choice", "[p]ollsters call their random samples, then they ask about party affiliation", "if the polls reported more Democrats, that simply means more random respondants self-identified as Democrats" are moot. True but irrelevant with regard to the matter of weighting the raw data against certain demographic concerns, a process that occurs after the raw data is collected. In this instance, the polls, including Rasmussen, by the way, are weighting for a Democratic demographic that was not part of the raw data, not part of the sample, not part of anything but an educated guess.

Sorry, mamooth, but it really is not as simple as you would have it.

LOL!

Hush, child.

Nevertheless, I believe, based on past experience and certain averages of the various polls that Obama is ahead, both nationally and in certain key, swing states that Romney must take in order to win the election. I also believe that the race is closer than many of the polls indicate.

The debates will decide this election.