Posted on 11/05/2012 6:08:44 AM PST by George N
At the risk of being an "Eeyore" and called a troll I think a lot of the tooth mashing over the Dem leaning polling samples is off. I hope I'm wrong but I think the outrage is mostly a result of a lack of technical understanding. It's always dangerous when amateur analysts start tearing into poll internals. This year it seems like laymen want to start weighting polls at their own whim. There's a reason pollsters historicLly do not weight by party ID.
That doesn't mean that Dems will outperform or even match 2008 turnout. They won't. But the Dem party ID advantage in many polls is the result of Soft GOP voters being more likely to ID as indies this year. You'll notice that in most polls where Romney leads big amongst indies the party ID gap is the most favorable for Dems. When the indie vote is closer the party ID is tighter. Again, when GOP leaning indies call themselves Republicans they tighten party ID. When they call thenselves indies Romney wins indies bi but suffers in party ID.
I don't know why Republicans are less likely to ID as such this year. Bu it doesn't change he sxientific math. Pollsters know with fair certainty what the sex/age/racial composition of the electorate will be. So when you weight that way and get a scientifically representative sample the top line number will be right. At least with the 95% lvl of confidence.
This isn't to say a massive GOP turnout couldn't fly under theradar or that pollsters don't struggle with urnout models and intensity. This is gonna be a close election but the polls point to a soight Obama edge. Let's not fool ourselves. We need an upset.
Aha! IBTZ. Another score to add to my ever increasing numbers... Yes!
The Gallup organization went out of it's way recently to explain that party affiliation isn't important to them in their seven day average poll, and that they do not use that information to make adjustments in their polling. From what I understand, most pollsters do not follow them. The fact that they have (had) Romney up by 5, before they suspended their polling, is significant.
I see the snark has started early for some.
I think you’re making the apriori assumption that the pollsters are honest, unbiased, and truly interested in sampling the will of the voters.
I reject every bit of that. Polling has become the most cost-effective form of propaganda and unregulated campaign influencing.
Prove me wrong.
On the route that I walk several times a week for exercise, there are many, many Romney yard signs, one Jill Stein, an NO Obama signs. In one driveway, however, there are TWO Priuses and both have Obama bumper stickers - that are from 2008.
ROLL TIDE
Some people on this site are not amatuers when it comes to statistics and polling.
Actually, most pollsters do not weight their results based on Party ID. They weight to meet certain demographics (usually to make sure the demographics roughly match the area being polled), but do not weight to meet a certain Party ID number - they just report the Party ID of their samples, as collected. The problem is that, for whatever reason, those samples this year have massively, and consistently, over-stated the number of Dems (and understated the number of Republicans). Whether that's because conservatives are less likely to talk to media pollsters, more likely to be out of the house (y'know, at work), or for some other reason, the samples seem to be off. It's a sampling problem, not an artificial weighing problem.
Its interesting to note the 2 tones of the campaigns currently.
Romney is upbeat and positive.
Zero is negative and "re"vengeful.
That there tells you that Zero is angered. He's ALREADY lost and knows it. He wants is drones to go out and get revenge for this loss.
That attitude is very telling.
Also what is telling is early voter turn out. Dem numbers are way down. Which means that Dems are no where near as energized as they were last time around.
Last time, IIRC, 95% of blacks voted for Obama. Thats ALMOST unrealistically high for ANY voting group. That means Zero has NOWHERE to go but down in this one area. Between voter turn out for blacks, and the fact that those who do vote, Zero has lost ground in this demographic.
Last time McCain ran a TERRIBLE campaign. Most of those were voting for Sarah. But more significantly, there were numerous "REPUBLICANS" who voted for Zero. Some because they were tired of Bush. Some because they didn't like McCain. Some because they WANTED to vote for the "black" man.
ALL that is out the window.
Todays republicans can best be described as broken glass republicans. We are willing to drag ourselves across broken glass to vote. And it will be against Zero.
Just look at the reports of the campaign stops. Zero has a few thousand show up. Romney has 10 thousand or TENS of thousands show up. Heck, Zero had a stop with Dave Matthews and Bill Clinton. People stuck around for Bill Clinton to speak, but left when Zero spoke. That tells you that Dems ARE NOT fired up about their guy. AT ALL!
Last election voter turnout was unusually high. Dems were energized. That ain't the case now. Zero had high voting percentages in key demographics. Its also said that the evangelical vote stayed home last time, and still an overall high voter turnout.
EVs AND Catholics are energized because of Zero, they won't be staying home this time. And all those key demographics that Zero held last time, those numbers have slipped. He held high numbers, but there is nowhere but down for him to go.
ESPECIALLY amongst Indies.
Heck, just look at approval numbers. 1 month after taking office, Zero's overall approval numbers were at 65%. Today, nowhere near that. More telling, is look at the strongly approve vs strongly disapprove (these are the people who are motivated to vote) Strongly disapprove outnumber Strongly approve 44% to 36%.
Notice how close The strongly disapprove of 44% is to overall approve of 50%.
We are also seeing strong momentum for Romney, which is telling.
Also, Sandy was an OPPORTUNITY for Zero, and he dropped it. His sycophants might try yo say he did well, but the everyday voter doesn't think so.
Even so, I'm much happier, and confident, seeing Mitt drawing massive crowds than Bamster's measly 2,800.
This seems to be the underpinning of your analysis. Do you have some study showing this?
As to party ID, why are pollsters concerned with representative samplings of whites, Asians, Hispanics, and blacks, men and women, etc., but not party ID? If party ID is irrelevant, then wouldn't a sample from the DNC convention be accurate as long as you had ethnic and gender ratios nailed down, even though you would end up with a 100% preference for the Dim?
If a poll was taken only of white voters, the Dims would go ballistic, since everyone knows they tend to vote Republican in disproportionate numbers (though not nearly as disproportionate as blacks vote for Dims) and ignoring that fact could skew the result. That argues for accurate ethnic balancing. Likewise, as to party ID, people who identify as Republicans vote overwhelmingly for the Republican candidate, and Dims vote overwhelmingly for Dims.
Any group can fluctuate from election from election, and the poll that is going to be most accurate tries to pick up on the fluctuations.
Back to your theory that “soft” Republicans are tending to identify as independents this time around: it seems to me that phenomenon would be more likely to occur when a party candidate is very unpopular. Thus, I could see that happening after eight years of George Bush. I do not see a Republican being “soft” after four years of a Marxist community organizer. I think that over-sampling Dims in polling obviously affects the outcome. There is certainly no way to know what the party split (R, D, I) will be in the final vote count, but as has been pointed out on this board and elsewhere numerous times, sampling Dims at a higher relative turnout than 2008 is highly unlikely. Methinks you are falling into the error of being an Eeyore and, possibly, a troll. Chin up!
McCain got big turnouts AFTER Sarah was on the ticket. McCain NEEDED to be seen with Sarah. She was the star of that show.
When polls are calling some states to be a more democrat turnout that 2008, then they are living in a bubble with their head up their arse. In 08, there was Bush fatigue, a revolt from fiscal conservatives, war weariness, a bailout, a loved and hated VP candidate, a media in the tank for Obama, "history", and the worst organization from a candidate that I've seen since I've followed politics.
I also know a little about polling as well from my PolySci days at Michigan State.
IATZ...
In your seminar did they address who pays for these “free” polls, and how they justify the expense to themselves and/or the people who give them the money to spend on them?
zotzotzotzotzotzotzotzotzot...................por favor.
“For your entertainment, Freepers:
MSNBC on the night of the Wisconsin Recall leading with their TOO CLOSE TO CALL crap and quickly realizing their over-sampled Dim polls were wrong. (But then cherry-picking results that were favorable to Zero from the SAME poll.)
ENJOY!!
http://www.youtube.com/watch?v=pchwcD4IPzs
The polls are worthless. They totally depend on who answers them with out slamming the phone down the moment you find out it’s another stupid pollster calling.
Crowd turn out is the best indicator. Obama draws 2000 and Romney draws 30,000.
There’s massive excitement over Romney right now. People like him.
I believer Rush and Dick Morris are right, it’s going to be a landslide for Romney.
And I believe Rush is right when he says that the libs will throw -0- under the bus as soon as he loses. There will be much satisfaction in watching that happen.
Last night on FOX, Brit Hume said crowd sizes aren't a good predictor and that McGovern and Mondale had huge crowds right before they lost.
Byron York also said he attended both a Romney and Obama rally in Ohio this weekend and that he feels there is equal enthusiasm for both candidates now. This coincides with the Gallup poll saying Obama has closed the enthusiasm gap.
I think the Obama voters haven't paid much attention to the last 4 years at all or the campaign up to this point. They have short attention spans. But the Democrat GOTV has managed to gin them up in the last week and that's why the polls have tightened. I really don't think it's about the hurricane. It's also a result of Romney going soft in the last debate and his first and only decisive debate win becoming a distant memory. And it's a result of it being 2 weeks since a debate where Romney gets to appear unfiltered to the public. The media retook control in the last 2 weeks and got to back up Obama's campaign and make Romney into a cold and distant figure again. An easy Romney win has slipped away to a coin toss. This isn't an election anyone can predict at this point because the slightest move in either direction can decide it now.
My sense is the polls are usually right this close before an election. Romney has an uphill battle tomorrow to avoid narrowly losing most of the swing states and the election. This is the first electoral showdown between a now rejuvenated Obama-mania and the Tea Party revolt. It's too close to call but it's Romney's fundamental, long-term deficits as a candidate, some of them deserved, some undeserved, that are responsible for making the race as close as it is.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.