Originally Posted by JustPassingThrough

Breaking my Boomer embargo one more time to interject a little factual insight. I looked up the poll you're referring to and found this article, which says:



What they did is to take a sample of "all adults", used a likely voter screen, and then weighted the results to turn it back into a poll of "all adults". Not even registered voters. It's nuts what these pollsters do, the media picks it up, and they get away with it despite their methodology being totally screwed up.

But if they said they would "definitely vote," they must be registered voters, although reflecting the general population.

The RCP average has CO rated as a tie. Two other polls in the past week have shown Romney leading there by 3 and 4 points. FYI.

One is noted as a Republican poll, the other generally leans toward Republicans. I don't know their methodologies. We'll have to see more polls before we know if Colorado has gone over to leaning-red. The national Gallup tracking poll is holding steady now. For our side that is hope that the Romney debate bump is over. We'll be watching!

Originally Posted by Eric the Green

But if they said they would "definitely vote," they must be registered voters, although reflecting the general population.

One is noted as a Republican poll, the other generally leans toward Republicans. I don't know their methodologies. We'll have to see more polls before we know if Colorado has gone over to leaning-red. The national Gallup tracking poll is holding steady now. For our side that is hope that the Romney debate bump is over. We'll be watching!

The Gallup poll is a 7 day average. Only 3 of those days are after the debate as of now. Obama was up by 6 before the debate, now he's up by 3. That means the last 3 days have been roughly Romney +1.

Originally Posted by JustPassingThrough

Breaking my Boomer embargo one more time to interject a little factual insight. I looked up the poll you're referring to and found this article, which says:

It is an opinion piece and not a news item!

What they did is to take a sample of "all adults", used a likely voter screen, and then weighted the results to turn it back into a poll of "all adults". Not even registered voters. It's nuts what these pollsters do, the media picks it up, and they get away with it despite their methodology being totally screwed up.

The RCP average has CO rated as a tie. Two other polls in the past week have shown Romney leading there by 3 and 4 points. FYI.

Selzer has a reputation as an excellent pollster because of adjusting the results to fit the demographic realities of a state -- urban/suburban/rural; white/Latino/black; Boulder vs. Colorado Springs; under 30 vs. over 55; educational and occupational levels.

It asked people whether they were likely Colorado voters based on their personal responses of whether they were going to vote. That's as good as you can get.

RCP would have to untie its average based on this poll alone. Old polls become less relevant over time and new ones closer to the election are more relevant.

Originally Posted by JustPassingThrough

The Gallup poll is a 7 day average. Only 3 of those days are after the debate as of now. Obama was up by 6 before the debate, now he's up by 3. That means the last 3 days have been roughly Romney +1.

Right, but Obama didn't go down today, even though one post debate day was added and one pre-debate day subtracted.

PPP, a reasonably-good pollster, polled likely voters in Virginia a couple of days after the disastrous debate performance of President Obama. Mitt Romney, who had largely lost the attention that he needed to have any reasonable chance to win, finally drew the sort of attention that he needs. Maybe too late, but many Americans had ignored him.

http://www.publicpolicypolling.com/p...se_VA_1007.pdf


(Fair use -- for analyzed of underlying data of a reputable statewide poll).

Virginia Survey Results

Q1 Do you approve or disapprove of President
Barack Obama’s job performance?
Approve .................................................. ........ 50%
Disapprove........................................ .............. 48%
Not sure .................................................. ........ 2%

This is winning range in a state close to the national average in Presidential voting. It is at a 'winning' level'.

Q2 Do you have a favorable or unfavorable opinion
of Mitt Romney?
Favorable......................................... ............... 52%
Unfavorable .................................................. .. 44%
Not sure .................................................. ........ 4%

Although Mitt Romney has respectable favorability, such is not as strong in significance as an approval rating in the same area for an incumbent. Maybe people like him enough to consider voting for him if the incumbent is a failure, but the next question asks whether the incumbent President has been effective enough without offending enough people to win:

Q3 The candidates for President are Democrat
Barack Obama and Republican Mitt Romney. If
the election was today, who would you vote
for?
Barack Obama............................................. ... 50%
Mitt Romney............................................ ........ 47%
Undecided......................................... .............. 3%

Unless one has a bad sample -- one inordinately favoring President Obama -- President Obama wins. PPP is a very good pollster. (Rasmussen is OK when it isn't getting the desired results for FoX Newspeak Channel). Of course two issues usually dominate an election, and they reflect the results closely this time (perhaps coincidentally):

Q4 Do you trust Barack Obama or Mitt Romney
more on the issue of the economy?
Barack Obama............................................. ... 50%
Mitt Romney............................................ ........ 47%
Not sure .................................................. ........ 3%

Q5 Do you trust Barack Obama or Mitt Romney
more on the issue of foreign policy?
Barack Obama............................................. ... 51%
Mitt Romney............................................ ........ 46%
Not sure .................................................. ........ 3%

The alleged weakness of President Obama has been economic performance. This is a 4T, and economic hardships are the norm during a 4T -- either wartime shortages or the consequences of a recent meltdown of the economy. Republicans have stymied any big projects that might succeed at sopping up unemployment, seeming to prefer that more people compete for a limited number of low-paying retail and restaurant jobs.

Q6 Who do you think won the debate between
Barack Obama and Mitt Romney this week?
Barack Obama............................................. ... 28%
Mitt Romney............................................ ........ 61%
Not sure .................................................. ........ 11%

If winning or losing the debate were all that mattered, then President Obama would be on the way to a defeat similar to that of Hoover in 1932 or Carter in 1980.

Originally Posted by Eric the Green

Right, but Obama didn't go down today, even though one post debate day was added and one pre-debate day subtracted.

Now it's a tie, but RCP is now reporting a 3 day average poll, so this is all post-debate.

University of Colorado study still points to Romney win with 330 electoral votes.

http://www.colorado.edu/news/release...win-university

The study has been right in every election since 1980.

Further, the Battleground poll conducted BEFORE the debates showed a virtual tie with Romney winning the Independents 51-35%. Obama won in 2008 by winning the Independents 52-44%.
The Enthusiasm gap is 86% for Repubs and 73% for Dems.
http://www.colorado.edu/news/release...win-university

Again, this was BEFORE the debate.....

Originally Posted by JustPassingThrough

Nobody is taking Nate Silver seriously anymore. The guy is a laughingstock.

In what universe. Only a right wing gasbag would make that comment. Typical right-winger - ignore the facts
you don't like.

On the left, they sometimes get too realistic (i.e. self-Bashing Obama for debate performance, etc).

Originally Posted by JustPassingThrough

Nobody is taking Nate Silver seriously anymore. The guy is a laughingstock.

Sure. In Free Republic, NewsMax, Breitbart, WorldNewsDaily, National Review, and the Weekly Standard.

A couple of new state polls:

PPP (D) Wisconsin 10/4-10/6

Obama: 49
Romney: 47
Spread: Obama +2

Rasmussen Iowa 10/7

Obama: 49
Romney: 48
Spread: Obama +1

Originally Posted by takascar2

In what universe. Only a right wing gasbag would make that comment. Typical right-winger - ignore the facts
you don't like.

On the left, they sometimes get too realistic (i.e. self-Bashing Obama for debate performance, etc).

Silver is a partisan Democrat, who worked for the Obama campaign in 2008. He now works for the New York Times. He has no background in politics or polling. His background is in sports. He does not disclose the details of the model he uses, and it produces results that are laughably far off in favor of the Democrats. His job is two things: propaganda for the Democrats and hopeful optimism for readers of the New York Times.

I've seen pollsters asked about him on TV recently, and they chuckle about him.

Originally Posted by JustPassingThrough

Silver is a partisan Democrat, who worked for the Obama campaign in 2008. He now works for the New York Times. He has no background in politics or polling. His background is in sports.

To the extent that his background is in sports it's not because he won a greatest fan contest for the old home team.
His true background is in statistics.
And his models have been accurate within the standard deviation in every election that I have followed him, which I began doing in 2006.

Originally Posted by JPT

He does not disclose the details of the model he uses

Actually he is quite clear on his methodology.
It is called calculus.
Each day he runs a computer based simulation of the coming national election with enough variables to produce 25001 plausible outcomes.
You can argue that too many of his variables favor the Democrats although his track record as noted above suggests that he is accurate within acceptable social science standards.

Sorry if I sound too academic but that is the point.
If you dismiss his methodology it's not only I that disagree with you, but just about every statistician and analyst.

Originally Posted by herbal tee

Actually he is quite clear on his methodology.
It is called calculus.
Each day he runs a computer based simulation of the coming national election with enough variables to produce 25001 plausible outcomes.
You can argue that too many of his variables favor the Democrats although his track record as noted above suggests that he is accurate within acceptable social science standards.

Sorry if I sound too academic but that is the point.
If you dismiss his methodology it's not only I that disagree with you, but just about every statistician and analyst.

Actually it isn't calculus. It's decision theory, a completely different branch of mathematics that occasionally uses calculus.

Originally Posted by Copperfield

Actually it isn't calculus. It's decision theory, a completely different branch of mathematics that occasionally uses calculus.

Correct.
I got in a hurry and didn't go into detail, thanks.

Originally Posted by The Rani

Woohoo, a tie!
Ok guys, stop posting all these polls so I can have fun munching popcorn on election night.

we just want to extend the fun for several weeks.

Originally Posted by The Rani

Ok then:

Pew: Romney Leads By 4 In Post-Debate Survey

You posted it in big letters.
Must be a hoax.

(no, I know it's true....)

Originally Posted by Eric the Green

You posted it in big letters.
Must be a hoax.

See, Eric? This is why it's worth keeping ya around.

Originally Posted by The Rani

Ok then:

Pew: Romney Leads By 4 In Post-Debate Survey

the realclearpolitics average for Obama is now down to .5 with 47.9 to Romney 47.4. If thats his average on election day regardless of Romney's, Obama loses...make it stop Playwrite, make it stop.....


http://www.realclearpolitics.com/epo...bama-1171.html

Originally Posted by The Rani

Ok then:

Pew: Romney Leads By 4 In Post-Debate Survey

I don't think the headline is big enough.

That is pretty huge news, though. When you consider how skewed towards Obama Pew has been all year:

Jan: +5
Feb: +8
March: +12
April: +4
May: +7
June: +7
July: +10
Sept: +8

Now it's Romney +4. That is a big swing.

Originally Posted by JustPassingThrough

Silver is a partisan Democrat, who worked for the Obama campaign in 2008.

...And he apparently did his job very well. President Obama apparently can understand mathematical models. Calculations are tedious.

He now works for the New York Times.

Which has deep pockets due to its huge revenues from subscriptions and advertising. It's not as if he is some high school senior with a math SAT of 680 as a 'special contributor' to the Tombstone (AZ) Epitaph, a newspaper with a very limited circulation. An entity with deep pockets can reward people well -- but it is also very demanding of those who enjoy its largesse.

A hint: a gossip columnist is not the right person to do this job.

He has no background in politics or polling. His background is in sports.

He is neither a current or former elected official or pollster. His background in sports is in analyzing baseball on a statistical basis as a jock could never do. The statistical techniques applicable to baseball (a pitching staff that gets lots of strikeouts and gives up few walks is going to keep runs down -- so although the Detroit Tigers and New Yankees should play evenly if they meet this post-season, either will be difficult to beat in the World Series).

He does not disclose the details of the model he uses, and it produces results that are laughably far off in favor of the Democrats.

He doesn't show the model that he uses because it is one of mathematical complexity due to the procedures that it uses. Few people have the talent and temperament to use mathematics full of abstract subscripts.

But I'll try to make it simple. Figure that President Obama has a near-lock on all states that either Gore or Kerry won in either 2000 or 2004 and add Nevada, and President Obama has locked up 263 electoral votes. Figure that the only states that count are now Colorado, Florida, Missouri, North Carolina, Ohio, and Virginia. Any one of them puts President Obama over 270. They are different enough that no single pitch easily wins them all, and there is no way of reaching them all simultaneously. Romney must win every one of those states if he is to have a chance to win.

The chance of Obama winning (call that "W(OB)") is 1 less than Romney's chance of winning them all. Let R(XX) represent the chance of Romney winning any critical state with XX being replaced by the postal abbreviation for the state. Then the chance of President Obama winning is

W(OB) = 1- R(CO) x R(FL) x R(MO) x R(NC) x R(OH) x R(VA).

To give some idea of how it works -- even if Romney has roughly an 89% chance in every one of those states, he has only about a 50% chance of winning. Multiply enough numbers smaller than but close to 1 and you can get a very small number. The effect with independent events is multiplicative, as with coin tosses and dice tosses.

So let's see what happens if the values of R(XX) are alike

R(XX)...........................W(OB)

0.95 .......................... 0.265
0.90 .......................... 0.469
0.85 .......................... 0.623
0.80 .......................... 0.738
0.75 .......................... 0.822
0.70 .......................... 0.892
0.65 .......................... 0.925
0.60 .......................... 0.954
0.55 .......................... 0.974
0.50 .......................... 0.984

His job is two things: propaganda for the Democrats and hopeful optimism for readers of the New York Times.

The math does not lie. If it were the other way around, and an Obama victory depended upon him winning all of Michigan, New Hampshire, Oregon, Pennsylvania, Wisconsin, and Washington the math would be the same. The duty of a political campaign is to turn a campaign from a long-shot into a good chance, a good chance to even, even to likelihood, or likelihood into certainty.

I've seen pollsters asked about him on TV recently, and they chuckle about him.

Because he is not a pollster!

Originally Posted by JustPassingThrough

I don't think the headline is big enough.

That is pretty huge news, though. When you consider how skewed towards Obama Pew has been all year:

Jan: +5
Feb: +8
March: +12
April: +4
May: +7
June: +7
July: +10
Sept: +8

Now it's Romney +4. That is a big swing.

Such a swing characterizes a breaking scandal or some diplomatic or military debacle. Misprints happen.

Two polls now have Romney within 3 points of Obama in Michigan. Another has him within 2 points in Pennsylvania.

Originally Posted by JustPassingThrough

Two polls now have Romney within 3 points of Obama in Michigan. Another has him within 2 points in Pennsylvania.

They were junk through September and they are junk now. These guys make Rasmussen look liberal-friendly.

Well geez whizz, why would Romney want this?

(Because he has more things to say in private to his rich supporters that wouldn't go over too well if they were to leak out to the public.)

It's all gonna come down to the 16th (the last debate, about foreign policy/natural security, is hopeless for Obama).

If Obama chokes again, Romney wins by double digits.