Tea Parties Have Good Favorables

[Rowan]

They were good in some states, but off in others, just like Rasmussen. Their national poll had Obama winning by 11 points if I'm not mistaken. Rasmussen had Obama winning by 6.

[Landslide Lyndon]

They were good in some states, but off in others, just like Rasmussen. Their national poll had Obama winning by 11 points if I'm not mistaken. Rasmussen had Obama winning by 6.

No, their election day poll at Daily Kos had Obama winning by 5.

[Brittain33]

First, I've pointed out on the other forum that Rasmussen's results skew Republican.

That's your opinion. He was one of the most accurate pollsters in 2008.

They had Democrats leading on the generic ballot by 6 points a couple of days before the election and 2 points immediately afterward with absurdly high undecideds. (Actual numbers were 47%-41%, 43%-41%.) The House vote on Election Day was 53%-44%.

Lying with undecideds is one of the hallmarks of partisan polls designed for campaigns.

[Rowan]

First, I've pointed out on the other forum that Rasmussen's results skew Republican.

That's your opinion. He was one of the most accurate pollsters in 2008.

They had Democrats leading on the generic ballot by 6 points a couple of days before the election and 2 points immediately afterward with absurdly high undecideds. (Actual numbers were 47%-41%, 43%-41%.) The House vote on Election Day was 53%-44%.

Lying with undecideds is one of the hallmarks of partisan polls designed for campaigns.

What does the number after the election mean? It probably means that after voting in a lot of Democrats, they probably said to themselves, maybe we need a more divided government.

6 points isn't that far off from 9 points. There is a MOE on polls you know.

[Brittain33]

6 points isn't that far off from 9 points. There is a MOE on polls you know.

That's not quite how MOE works. They had Democrats at 47% or 43%; the actual results were 53%, way off the MOE. They were much closer with Republicans, pegging them at 41% when they got 44%. By upping undecideds to an unreasonable number at the expense of Democrats, you get results reassuring to Republicans who you hope will buy your poll results in the future.

There were not 12% undecideds before election day or 16% afterward. If there were any undecideds, they were people who were not going to vote and probably couldn't find the ballot box with both hands.

[Rowan]

6 points isn't that far off from 9 points. There is a MOE on polls you know.

That's not quite how MOE works. They had Democrats at 47% or 43%; the actual results were 53%, way off the MOE. They were much closer with Republicans, pegging them at 41% when they got 44%. By upping undecideds to an unreasonable number at the expense of Democrats, you get results reassuring to Republicans who you hope will buy your poll results in the future.

There were not 12% undecideds before election day or 16% afterward. If there were any undecideds, they were people who were not going to vote and probably couldn't find the ballot box with both hands.

Okay, let's look at all the pollsters generic ballot. This is courtesy of realclearpolitics:

NBC News/Wall St. Jrnl   11/01 - 11/02   48   36   Democrats +12
Gallup   10/31 - 11/02   53   41   Democrats +12
Diageo/Hotline   10/31 - 11/02   44   39   Democrats +5
CBS News/NY Times   10/25 - 10/29   48   36   Democrats +12
GWU/Battleground   10/23 - 10/29   45   41   Democrats +4
Rasmussen Reports   10/20 - 10/26   47   40   Democrats +7

Out of all the pollsters, you can really only claim that Gallup did better.

[Dan the Roman]

My other big problem is the entire use of Likely voters. What is a Likely Voter this far from an election? Definitely not someone concerned primarily about health care, the war, or the right to choose. Its the people who are angry about taxes or believe that Obama is an Al Quada plant born in Kenya.

I am sure if you polled Clinton among likely voters in the sumer of 1998 a lot more than the general population would want him resign or be impeached because they cared more. But the GOP did badly in 1998, why? because those people pissed off everyone else leading to a lot of people who wouldn't otherwise have voted in a midterm coming out to vote against them.

While I will be the first person to condemn the "All Adults" polls the networks ran during the campaign, I would trust one of those over this right now.

[Zarn]

My other big problem is the entire use of Likely voters. What is a Likely Voter this far from an election? Definitely not someone concerned primarily about health care, the war, or the right to choose. Its the people who are angry about taxes or believe that Obama is an Al Quada plant born in Kenya.

I am sure if you polled Clinton among likely voters in the sumer of 1998 a lot more than the general population would want him resign or be impeached because they cared more. But the GOP did badly in 1998, why? because those people pissed off everyone else leading to a lot of people who wouldn't otherwise have voted in a midterm coming out to vote against them.

While I will be the first person to condemn the "All Adults" polls the networks ran during the campaign, I would trust one of those over this right now.

Then you only care for pro-Democratic numbers.

Likely voters are people who have previously voted in elections among a few other things. People have voted in elections before are much more likely to vote, again.

[Rowan]

And if you don't vote, who the hell cares what your opinion is? If you don't exercise your right to vote then your opinion means absolutely nothing.

[Dan the Roman]

My other big problem is the entire use of Likely voters. What is a Likely Voter this far from an election? Definitely not someone concerned primarily about health care, the war, or the right to choose. Its the people who are angry about taxes or believe that Obama is an Al Quada plant born in Kenya.

I am sure if you polled Clinton among likely voters in the sumer of 1998 a lot more than the general population would want him resign or be impeached because they cared more. But the GOP did badly in 1998, why? because those people pissed off everyone else leading to a lot of people who wouldn't otherwise have voted in a midterm coming out to vote against them.

While I will be the first person to condemn the "All Adults" polls the networks ran during the campaign, I would trust one of those over this right now.

Then you only care for pro-Democratic numbers.

Likely voters are people who have previously voted in elections among a few other things. People have voted in elections before are much more likely to vote, again.

I was not criticizing likely voter screens generally. I said that I thought Rasmussen, especially for state level numbers, was better than most of the newspaper polls, and easily on par with Research 2000.

What I was doing was saying how ridiculous it is to do a likely voter screen 19 months before an election. How on earth do you do it? Its not so simple as asking people whether you voted in the last Presidential election because turnout will be 60% of that at best. And I am not sure whether people voting in 2002 or 2006 is indicative of much. And the question, "How likely are you to vote in 2010" is guaranteed to give a skewed sample now, since the people who are most advocated are on the extremes.

Basically I like Rasmussen, I think he has an original method. But because of the method he uses he is much better at polling some things than others, and he gets very different numbers from everyone else. If this were October or September of 2010 I would be inclined to trust Rasmussen. But its not and he always shows genuinely weird numbers in off-years because he does all his polls of likely voters.

[opebo]

Alright, I'm calling BS

Even though you have no good reason to do so.

People generally don't like protests.  I find it hard to believe that people would like protests where a President they strongly approve of is called a ''fascist'' on a pretty regular basis. If other polls start showing similar results, I'll gladly eat my words.

Lots more than 51% of the population are racists..