Two more crack-pot polls released today.

IndepCentristMA said:

This is exactly why I have said all along RCP is not as accurate as it could be, because it gives too much validity to outliers and smaller independent agencies...

They need to weight surveys by the big guns that have consistently rated in the top 5 in every election, like Rasmussen, Pew, etc. higher... and compare the sample size as a measure as well...

Click to expand...

According to Fordham University, Rasmussen and Pew did not score in the top 5 in 2008. Out of 20 pollsters rated for accuracy, here's how they ranked:

Democracy Corps (D) 51/44 +7
Fox News 50/43 +7
CNN/ORC 53/46 +7
Ipsos/McClatchy 53/46 +7
American Research Group 53/45 +8
IBD/TIPP 52/44 +8
Harris Interactive 52/44 +8
YouGov/Polimetrix 51/45 +6
Pew 52/46 +6
Rasmussen 52/46 +6
NBC/WSJ 51/43 +8
GWU (Lake Tarrence) 49/44 +5
ABC/WaPo 53/44 +9
Diageo/Hotline 50/45 +5
Daily Kos/Research 2000 51/46 +5
Marist 52/43 +9
CBS 51/42 +9
Gallup 55/44 +11
Reuters/C-SPAN/Zogby 54/43 +11
CBS/NYT 52/41 +11

http://www.fordham.edu/images/acade...campaign_/2008 poll accuracy panagopoulos.pdf

Investor's Business Daily also ranked eight pollsters higher than Rasmussen and Pew: IBD/TIPP Takes Top Honors Again - Investors.com

This Republican attack on the polls is doubly moronic. First, as explained several times, it is bad polling practice to try to correct for party identification. And second, the party identification in the Q polls is actually quite close to the actual Ohio results from the 2008 election. So even if they DID correct for party identification, it would make little or no difference in the result.

AdamT said:

This Republican attack on the polls is doubly moronic. First, as explained several times, it is bad polling practice to try to correct for party identification. And second, the party identification in the Q polls is actually quite close to the actual Ohio results from the 2008 election. So even if they DID correct for party identification, it would make little or no difference in the result.

Click to expand...

Youre making the case for the Republicans... because the polls they're railing against did just that... they took a sample, and then corrected their result to skew them for a maximum turnout for the Democrats, and a dull turnout for Republicans... Like you said.... Correcting for party identification is bad practice... which is why the Republicans are saying forget the party identification, go with average turnout for all the election years in the past decade, and you come out with a huge lead for Romney...

IndepCentristMA said:

Youre making the case for the Republicans... because the polls they're railing against did just that... they took a sample, and then corrected their result to skew them for a maximum turnout for the Democrats, and a dull turnout for Republicans... Like you said.... Correcting for party identification is bad practice... which is why the Republicans are saying forget the party identification, go with average turnout for all the election years in the past decade, and you come out with a huge lead for Romney...

Click to expand...

No, that's exactly what they DIDN'T do. But again, even if they had, it wouldn't have mattered in the result.

AdamT said:

No, that's exactly what they DIDN'T do. But again, even if they had, it wouldn't have mattered in the result.

Click to expand...

Again... your bias interferes with reality...

The general polls that aren't adjusted for turnout show a slight Romney lead of like 1% (which already contain a built-in bias which is pro Democrat)...

When those current polls are being adjusted for 2008's numbers Obama has a 2% lead...

When adjusted for the average turnout for the period of 2004, 2006, 2008, and 2010 it shows a large Romney lead of about 5%...

With the fact that enthusiasm is up for Republicans, and down for Democrats, I don't see a 2008 sized turnout... I also don't necessarily see a 2010 sized turnout for Republicans, since most of the 2008 Obama followers were probably even unaware of the mid-term elections...

My guess is we will see a final result which is somewhere between the 1% and the 5%, like a 2-3% point national lead for Romney, of like 51%-48%...

That's the national poll...

Inside the states, the swing state polling, Rasmussen has dominated the accuracy of swing state polling results for the last 2 presidential election cycles... I know you find him to be biased, but you only say that based off 2010 over estimation of tea party influence in some senate races... In state polling I will defer to the polls which have had it as close to correct as it has been, with Rasmussen... (BTW it currently indicates a Romney win... but they are still polling)

IndepCentristMA said:

Again... your bias interferes with reality...

The general polls that aren't adjusted for turnout show a slight Romney lead of like 1% (which already contain a built-in bias which is pro Democrat)...

When those current polls are being adjusted for 2008's numbers Obama has a 2% lead...

When adjusted for the average turnout for the period of 2004, 2006, 2008, and 2010 it shows a large Romney lead of about 5%...

With the fact that enthusiasm is up for Republicans, and down for Democrats, I don't see a 2008 sized turnout... I also don't necessarily see a 2010 sized turnout for Republicans, since most of the 2008 Obama followers were probably even unaware of the mid-term elections...

My guess is we will see a final result which is somewhere between the 1% and the 5%, like a 2-3% point national lead for Romney, of like 51%-48%...

That's the national poll...

Inside the states, the swing state polling, Rasmussen has dominated the accuracy of swing state polling results for the last 2 presidential election cycles... I know you find him to be biased, but you only say that based off 2010 over estimation of tea party influence in some senate races... In state polling I will defer to the polls which have had it as close to correct as it has been, with Rasmussen... (BTW it currently indicates a Romney win... but they are still polling)

Click to expand...

Again, you accusing anyone else of bias is laughable. And it couldn't be more ironic, as you do it in the context of weaving conspircy theories to try to explain away the fact that your man is losing this campaign.

And you are simply wrong on the facts. AFAIK the only major polling organizatin that does adjust for party ID is Rasmussen. Why would you prefer some pollsters GUESS as to who will turn out to the actual responses you are getting from people who answer your polls? If you have 30 polls in Ohio and they consistently show that Democrats are likely to show up in about the sam proportion as they did in '08, why would you ignore that evidence and assume that you know better? It makes no sense.

Numeros said:

According to Fordham University, Rasmussen and Pew did not score in the top 5 in 2008. Out of 20 pollsters rated for accuracy, here's how they ranked:

http://www.fordham.edu/images/acade...campaign_/2008 poll accuracy panagopoulos.pdf

Investor's Business Daily also ranked eight pollsters higher than Rasmussen and Pew: IBD/TIPP Takes Top Honors Again - Investors.com

Click to expand...

That's 1 site and 1 way of ranking them... by how accurate they got the final margin... that's not very accurate to me...

If you'll look... the outcome was 52.9-45.7 for Obama... Rasmussen had it as 52-46... in their poll... not too shabby... What those people in your ratings did is said the results was 53-46, with a spread of 7, and that Rasmussen had it as a gap of 6, they were off by 1...

The one they have ranked #1 there... had it as 44-51... so while they had the margin correct, they were off by a few points on the actual turnout... -1.9 short of Obama's and -1.7 off of McCain's... Wheras Rasmussen was off by -.9 on Obama's and +.3 on McCain's... You tell me which is more accurate...

I also know IBD/TIPP has been tooting their own horn quite a bid, but no one else really is (just an FYI, not a good source to quote an agency's own report about their results being first)... Especially considering, they've been separate as IBD and TIPP on other occasions, and brough down the result of other agencies which faired better without the association of one of them... Gallup, Zogby, and CNN have faired better when they weren't involved with either of those...

However, I'm sure if/when you do a little more research into the rankings of all the agencies and all of the elections, you'll find Rasmussen and Pew's accuracy to the be the best over that course of time...

If it helps, check with the NYT rating of them, and RCP's rating of them...

AdamT said:

Again, you accusing anyone else of bias is laughable. And it couldn't be more ironic, as you do it in the context of weaving conspircy theories to try to explain away the fact that your man is losing this campaign.

And you are simply wrong on the facts. AFAIK the only major polling organizatin that does adjust for party ID is Rasmussen. Why would you prefer some pollsters GUESS as to who will turn out to the actual responses you are getting from people who answer your polls? If you have 30 polls in Ohio and they consistently show that Democrats are likely to show up in about the sam proportion as they did in '08, why would you ignore that evidence and assume that you know better? It makes no sense.

Click to expand...

ZZZZZZ.... same tired rhetoric that we are hearing out of Obama...

In fact, Romney is winning this campaign...

and, now... I know you are AdamT... the DP poster (and likely paid representative for the Democratic Party)...

But...

This was the exact analysis by Carl Rove (who helped the village idiot win 2 elections) and Dick Morris (who helped a Governor of Arkansas become president)...

Now, I know you're bound to cry about it, and accuse them of bias, but neither have a dog in this race... their reputation as an analyst is what is on the line, so the more accurate they are, the better for them...

Carl Rove, I watched throughout the Republican Primary process, and Romney was not his guy to start with... and nothing like his boy GWB... but even during the primary process, Carl Rove was maticulous with the numbers from all the different counties and precincts, and pays attention to this as if it is a science. Dick Morris for his part has been known to do the same thing...

BOTH, Carl Rove and Dick Morris, have broken down the numbers by county and precinct, and both have suggested Romney could win this in a landslide, if the numbers are anything below what they were in 2008 for Obama, which is expected...

Again, I'm sure you value your own opinion more than theirs... I don't... and I doubt many others do...

Grim17 said:

Crack-pot poll #1

What happens when you conduct a state poll in Ohio,and sample 9% more democrats than republicans in the survey, and sample 11% fewer Independents than turned out for the election in 2008?

What you get is a PPP poll showing Obama with a 5 point lead over Romney in Ohio... Or as I like to call it... A friking joke.

********************************************************************

Crack-pot poll #2

This one is a survey of likely voters from National Journal and is listed in the RCP average today. It shows Obama up nationally by 5%, 50 to 45 over Romney... What's wrong with this poll, other than it's small sample size and 4.4% margin of error, is written in the story about it here, where it says:

In its likely-voter model, the Congressional Connection Poll projected that the 2012 electorate will be virtually unchanged from 2008, with Democrats holding an 8 percentage-point advantage among voters (compared with 7 points last time)
​

You got that? They think even a higher percentage of democrats over republicans will vote in this election, compared to 2008... I guess they just ignored the survey Gallup did of more than 9000 voters, which not only didn't show that this election would likely have a 8 point Dem advantage, but projects that Republicans will have a higher turnout this election than the Democrats will.


Just when I thought the polling organizations had started to move toward reality, we get this crap.

Click to expand...

The good thing is that in less than a week we can test your hypothesis. What happens if the elections look like these results? Will you start doubting your partisan sources of information?

IndepCentristMA said:

ZZZZZZ.... same tired rhetoric that we are hearing out of Obama...

In fact, Romney is winning this campaign...

and, now... I know you are AdamT... the DP poster (and likely paid representative for the Democratic Party)...

But...

This was the exact analysis by Carl Rove (who helped the village idiot win 2 elections) and Dick Morris (who helped a Governor of Arkansas become president)...

Now, I know you're bound to cry about it, and accuse them of bias, but neither have a dog in this race... their reputation as an analyst is what is on the line, so the more accurate they are, the better for them...

Carl Rove, I watched throughout the Republican Primary process, and Romney was not his guy to start with... and nothing like his boy GWB... but even during the primary process, Carl Rove was maticulous with the numbers from all the different counties and precincts, and pays attention to this as if it is a science. Dick Morris for his part has been known to do the same thing...

BOTH, Carl Rove and Dick Morris, have broken down the numbers by county and precinct, and both have suggested Romney could win this in a landslide, if the numbers are anything below what they were in 2008 for Obama, which is expected...

Again, I'm sure you value your own opinion more than theirs... I don't... and I doubt many others do...

Click to expand...

Wow, are you for real? You are really trying to suggest that Carl Rove and Dick Morris are unbiased sources?! :lamo

Why do I have the feeling that your account here will be deleted on Nov. 7th?

AdamT said:

Wow, are you for real? You are really trying to suggest that Carl Rove and Dick Morris are unbiased sources?! :lamo

Why do I have the feeling that your account here will be deleted on Nov. 7th?

Click to expand...

I never said they weren't "biased", everyone has a bias...

What I said is that their profession, which they draw their current income from isn't on getting Romney elected, it's on accurately predicting the result of the election... Otherwise they're less likely to get a call in the future regarding these things...

I also said they know the numbers inside and out, and better than most people, having gone through the game numerous times, getting proven results...

So, if I could, I'd put money on the fact that I'll be celebrating a Romney victory Tuesday night... the same way I celebrated the Reagan victory in 80... and Obama supporters like yourself will be on suicide watch... but Obama himself, will be breathing a large sigh of relief, knowing he doesn't have to play this game anymore...

OpportunityCost said:

Yeah absent Gallup, Marist, PPP, Quinnipiac and you want to call that reliable? You took the left leaning pollsters put them all in one place then hung your hat on that for accuracy?
Cmon you can do better than that.

Click to expand...

Reliability can only be measured after the election. Meanwhile, the poll at issue appears to have used valid criteria to determine likely voters, which is what it was trying to measure. It wasn't measuring the voter preferences of registered voters or parties.

Focus.

Kandahar said:

Polling is indeed showing evidence that the number of people who self-identify as Republicans has cratered. Here's a chart showing recent party identification polls. The percentage of voters who call themselves Republicans is only in the mid-20s (compared to the low to mid-30s for Democrats). The fact that the Dem-Rep gap appears to be larger than in 2008 isn't because the polls are systematically excluding conservatives, it's because conservatives are less likely to call themselves Republicans.

Click to expand...

This is the issue - how to measure LIKELY voters (not registered voters). And the OP and its conservative knownothingism and conspiracy theories simply ignores it. What else can conservative do?

IndepCentristMA said:

That's 1 site and 1 way of ranking them... by how accurate they got the final margin... that's not very accurate to me...

Click to expand...

It was good enough for Scott Rasmussen. He prominently featured a citation and link to Fordham's Initial Report on his website. That's the report that listed Rasmussen & Pew as most accurate, based on an incorrect estimate of a 6.15 margin. And he milked that invalid report for all it was worth, keeping the citation and link on his site for more than a year -- long after the official election numbers had been released. Sometime around or after December 2009, when Fordham finally completed their second report based on the correct 7.2 margin, Rasmussen removed the link and citation.

What I also found interesting was Scott Rasmussen's recent appearance on Fox News with Megyn Kelly. She trotted out the old, invalid report and stated that Fordham rated Rasmussen #1 in the 2008 election. Instead of correcting her, Rasmussen just nodded his head and smiled.

As far as Rasmussen and Pew's accuracy in predicting the popular vote in Presidential elections, they did well in 2004. Not quite as well in 2008. I also checked out Rasmussen's 2008 state polling. I wasn't particularly impressed.

Numeros said:

It was good enough for Scott Rasmussen. He prominently featured a citation and link to Fordham's Initial Report on his website. That's the report that listed Rasmussen & Pew as most accurate, based on an incorrect estimate of a 6.15 margin. And he milked that invalid report for all it was worth, keeping the citation and link on his site for more than a year -- long after the official election numbers had been released. Sometime around or after December 2009, when Fordham finally completed their second report based on the correct 7.2 margin, Rasmussen removed the link and citation.

What I also found interesting was Scott Rasmussen's recent appearance on Fox News with Megyn Kelly. She trotted out the old, invalid report and stated that Fordham rated Rasmussen #1 in the 2008 election. Instead of correcting her, Rasmussen just nodded his head and smiled.

As far as Rasmussen and Pew's accuracy in predicting the popular vote in Presidential elections, they did well in 2004. Not quite as well in 2008. I also checked out Rasmussen's 2008 state polling. I wasn't particularly impressed.

Click to expand...

Rasmussen has one of the highest pro-Republican house effects of any pollster. Interesting read here: Simon Jackman: House Effects, Back By Popular Demand

Numeros said:

It was good enough for Scott Rasmussen. He prominently featured a citation and link to Fordham's Initial Report on his website. That's the report that listed Rasmussen & Pew as most accurate, based on an incorrect estimate of a 6.15 margin. And he milked that invalid report for all it was worth, keeping the citation and link on his site for more than a year -- long after the official election numbers had been released. Sometime around or after December 2009, when Fordham finally completed their second report based on the correct 7.2 margin, Rasmussen removed the link and citation.

What I also found interesting was Scott Rasmussen's recent appearance on Fox News with Megyn Kelly. She trotted out the old, invalid report and stated that Fordham rated Rasmussen #1 in the 2008 election. Instead of correcting her, Rasmussen just nodded his head and smiled.

As far as Rasmussen and Pew's accuracy in predicting the popular vote in Presidential elections, they did well in 2004. Not quite as well in 2008. I also checked out Rasmussen's 2008 state polling. I wasn't particularly impressed.

Click to expand...

2008
The List: Which presidential polls were most accurate? | Texas on the Potomac | a Chron.com blog
Low favorables: Democrats rip Rasmussen - Alex Isenstadt - POLITICO.com
FiveThirtyEight: Politics Done Right: Pollster Ratings, v3.1.1
(even Nate Silver, who hates Rasmussen had him ranked up at the top, though he tries to run away from it now)

2004
RealClearPolitics
Who nailed the election results? - Slate Magazine


Rasmussen has been the most accurate over time... so... poor poor Nate Silver... he's let a competitive vendetta get in the way of relevant facts and detailed analysis...

head of joaquin said:

Reliability can only be measured after the election. Meanwhile, the poll at issue appears to have used valid criteria to determine likely voters, which is what it was trying to measure. It wasn't measuring the voter preferences of registered voters or parties.

Focus.

Click to expand...

Yes because the CBS NYT plus 11 Dem samples are so damn reliable. Your posts are like a comedy of errors without the comedy.

IndepCentristMA said:

2008
The List: Which presidential polls were most accurate? | Texas on the Potomac | a Chron.com blog
Low favorables: Democrats rip Rasmussen - Alex Isenstadt - POLITICO.com
FiveThirtyEight: Politics Done Right: Pollster Ratings, v3.1.1
(even Nate Silver, who hates Rasmussen had him ranked up at the top, though he tries to run away from it now)

2004
RealClearPolitics
Who nailed the election results? - Slate Magazine


Rasmussen has been the most accurate over time... so... poor poor Nate Silver... he's let a competitive vendetta get in the way of relevant facts and detailed analysis...

Click to expand...

Interesting links, however, you can throw the first one straight into the trash. It's in reference to the outdated, invalid Fordham Initial Report which I mentioned earlier. It was based on preliminary data from the hours immediately following the election. http://www.fordham.edu/images/acade...ccuracy in the 2008 presidential election.pdf

There was good reason it was titled "Initial Report"; it can take weeks to tabulate all the votes. An ESTIMATE of only 6.15 was used for Obama's winning margin. which looked very good for Pew and Rasmussen with their 6-point, 52/46 projections. But Obama did not win by 6.15; he won by 7.2, which changed everything. See page 5 of the official FEC report for 2008 to verify the popular vote numbers. http://www.fec.gov/pubrec/fe2008/federalelections2008.pdf

Again, the final Fordham report, based on the official numbers, is available at http://www.fordham.edu/images/acade...campaign_/2008 poll accuracy panagopoulos.pdf . Fordham and Pew were not the most accurate. CNN and Ipsos obviously came closer with their 7-point margin, 53/46.

It's really very simple.
7 is closer to 7.2 than 6 is.
53 is closer to 52.9 than 52 is.


Your claim that Rasmussen is most accurate over time doesn't seem to be borne out by the data I've seen so far. Averaged over the last 2 Presidential elections, Pew was more accurate than Rasmussen. According to your link to RCP for 2004, Pew "nailed it" while Rasmussen came in 6th. For 2008, they tied for runner-up spots.

As for Rasmussen's state polling, what I've looked at so far (RCP for November, 2008) has not been particularly accurate. And the report from Nate Silver is dated May, 2008, so I'm not sure what races it pertains to. But if you have more in-depth reports, I'd like to have a look.