Why The NY Times Poll Is Wrong

[petegz28]

Whether deliberately or not, the New York Times/CBS/Quinnipiac Poll is wrong! It shows Obama carrying Ohio, Florida, and Virginia. In reality, Romney is leading in all three states and will carry them all.

Here’s the deal. The Times is weighting the raw survey data to reflect the ratio of Democrats to Republicans who voted in 2008. True, if we get the same massive turnout among minorities and young people that propelled Obama to victory in 2008, he will win this election and carry these states. But we won’t. All the polling shows that the electorate is now much more Republican and that GOP voters are much more motivated to turn out than their Democratic counterparts.

If we weight the Times results for the average turnout of the past four elections: 2004, 2006, 2008, and 2010, we find Romney winning all three states. Republican pollster (the best of them all) John McLaughlin and I used exit polls from the past four elections to figure out how many Democrats and Republicans actually voted and then we averaged them together. Here are the real numbers:

FLORIDA

NY Times results: Obama +1

Dem/Rep ratio in Times poll: Dems +7

Average ratio Dem/Rep past four elections: Reps +1

Times overstates Dem vote by 8 points

Correct poll result: Romney +7

OHIO

NY Times results: Obama +5

Dem/Rep ratio in Times poll: Dems +8

Average ratio Dem/Rep past four elections: Dems +2

Times overstates Dem vote by 6 points

Correct poll result: Romney +1

VIRGINIA

NY Times results: Obama +2

Dem/Rep ratio in Times poll: Dems +8

Average ratio Dem/Rep past four elections: Reps +1

Times overstates Dem vote by 9 points

Correct poll result: Romney +7

And even these results don’t tell the full story. The Gallup Poll finds that the 2012 election will actually have more Republicans and fewer Democrats voting than any of the past four elections. In 2008, the electorate had 12 points more Democrats and Republicans. In 2004, the electorate was evenly divided. Gallup estimates that in 2012, there will be 3 points more Republicans voting than Democrats.

So the numbers above are not reflective of Romney’s real strength. He will do even better. And add to them the fact that the undecided vote goes against the incumbent, and you are looking at a real landslide.

http://www.dickmorris.com/why-the-ny...poll-is-wrong/

[petegz28]

I think Morris is correct in so much that we are not going to see the turnout for Obama this year as we did in 08. I think that's pretty much a conceded notion among all.

[RINGLEADER]

Who knows at this point. Outside of a sexy scandal or bizarre audio tape I don't think anything moves the line much. Maybe employment numbers, but I doubt it...

[petegz28]

Originally Posted by RINGLEADER

Who knows at this point. Outside of a sexy scandal or bizarre audio tape I don't think anything moves the line much. Maybe employment numbers, but I doubt it...

I don't expect the numbers to be great one way or the others and I do expect last month's numbers to be revised upward

[alnorth]

Originally Posted by petegz28

Here’s the deal. The Times is weighting the raw survey data to reflect the ratio of Democrats to Republicans who voted in 2008.

This is incorrect. No reputable pollster re-weights by party to some expected party distribution anymore. You weight polls by race, age, gender, and sometimes geography if a state has a huge population center.

Dick Morris is an idiot, and is, incredibly, attempting to be even more wrong this year than he was in 2008.

[alnorth]

Dick Morris, September 2008:



Dick Morris, October 2008:



In just one single month, this clown went from "New Jersey is only leaning to Obama, almost every blue state is in play" to "Texas is just leaning and McCain will lose his home state"

Even when he called an Obama blowout, his map looked bizarre. This guy doesn't know what the hell he's doing and is just making stuff up to excite viewers.

[patteeu]

Originally Posted by alnorth

This is incorrect. No reputable pollster re-weights by party to some expected party distribution anymore. You weight polls by race, age, gender, and sometimes geography if a state has a huge population center.

Dick Morris is an idiot, and is, incredibly, attempting to be even more wrong this year than he was in 2008.

That's not true. You're describing how some of the reputable pollsters do things these days, but not all.

[whoman69]

Originally Posted by alnorth

This is incorrect. No reputable pollster re-weights by party to some expected party distribution anymore. You weight polls by race, age, gender, and sometimes geography if a state has a huge population center.

Dick Morris is an idiot, and is, incredibly, attempting to be even more wrong this year than he was in 2008.

People in polls often mention party by whom they are voting for at that moment.

[RINGLEADER]

Originally Posted by alnorth

This is incorrect. No reputable pollster re-weights by party to some expected party distribution anymore. You weight polls by race, age, gender, and sometimes geography if a state has a huge population center.

Dick Morris is an idiot, and is, incredibly, attempting to be even more wrong this year than he was in 2008.

We'll soon know who's right...

[Direckshun]

Dick.

Morris.

...

Christ.

[cdcox]

Romney might win Florida and VA, but he won't take either one of them by anywhere near 7 freaking percent. Which shows you that his Ohio math is just wacked.

[RINGLEADER]

Originally Posted by cdcox

Romney might win Florida and VA, but he won't take either one of them by anywhere near 7 freaking percent. Which shows you that his Ohio math is just wacked.

I agree with this.