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.
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.
At least you know what’s coming. But your ignorance is showing. You give far to much credit to the accuracy opand intelligence of the pollsters. Take 2010 or 1980. Now to go pop some pop corn and IBTZ.
BS
nope, no troll here
Over sampling is the LEAST of the polls problems.
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
Total waste of bandwidth.
Then why do some polls are sampling D+1, while others are at D+11?
Sounds to me like they simply don’t have a clue.
Turnout is key. If Republicans and conservatives are as angry as we make ourselves out to be, then we have the numbers to flip this election. If that motivation is flagging, then we were never in the game to start with.
We’ll know very soon.
Thanks for warning us about amateur analysts with your amateur analysis.
You’ve made exactly two (2!) posts since you joined FR in 2007. This kind of strains your credibility here.
So I lack understanding and I am dangerous to question a poll of North Carolina that is D+13?
Among the things she told us:
*Public pollsters heavily oversample urban areas, and therefore de facto get more "Dems" even when they aren't identified as Ds.
*Public pollsters make up samples they "assume" are right, but have very little experience anywhere outside of the beltway.
*Public pollsters, due to the larger samples, employ people who are NOT pollsters but simply questioners. The result is, she said, they get quite lazy. They don't probe to get the real sentiment that will affect the vote (and therefore MAKE the vote) but they accept the first answer given. Likewise, this laziness translates into a tendency at the end of the polling time (out of frustration) to take whomever they get and label that person "a Hispanic," or an "over 40" or whatever. She was adamant that not just political internal pollsters but that corporate pollsters HAVE to be accurate, while the big firms can be close on some, hit on one or two, and tout the one or two they hit on and ignore the five they miss on.
Do you drive a Subaru by chance?
Did Obama just have back to back +30,000 rallies in Ohio and Pennsylvania? No.
The media’s spinning all kinds of garbage, Jim Cramer’s seriously suggesting Obama will pick up Georgia, Texas, Montana, South Dakota, Missouri, Indiana, and North Carolina.
That’s blatant propaganda on their part. Who cares what their “polls” have to say?
More than likely a prius or a bicycle.
Hard to accept an analysis by one who can’t be bothered to proofread their own post. Sloppy composition.
‘Teeth mashing.” Indeed.
I am just going on anecdotal information but when you live in a small town you can definitely get a feel for what is going on.
Obama voters are not energized. They have no signs, they have no bumper stickers.
Many former Obama voters are planning not to vote rather then vote for Romney.
Romney voters are eager to vote, proud if they have already voted.
You can tell who probably voted for Obama because they don’t say who they voted for, Romney voters are proud.
On a last note, even college students aren’t all in love with the Revolutionary this time.
Hard to accept an analysis by one who can’t be bothered to proofread their own post. Sloppy composition.
‘Teeth mashing.’ Indeed.
“...teeth mashing...”
I think you mean “teeth gnashing”...
Product of the govt school system.
Pennsylvania and Ohio
posting dem talking point BS is the reason FR is moving slow.
http://www.politico.com/blogs/charlie-mahtesian/2012/08/dem-registration-down-in-key-states-132150.html
No, Gallup and Rasmussen are two pollsters who do extensive polling on party ID.
Most of the other polls do their polling and then weight the results based upon their assumptions as to the party ratio of those who will actually vote.
It is that latter group that is coming up with the big Democrat advantage—because they are putting it there.
Gallup, which does the biggest polling on party ID has reported a major swing to the GOP.
Folks I’m sure we can debate polls forever. The poster gave us some info. The fact is we can win this election in part because many pollsters have Romney leading or tied. But some on here attacking every negative poll are being pollyannas as well.
I saw a Prius in Minneapolis yesterday with a Romney/Ryan bumper sticker. That was a good omen for what’s coming tomorrow, that and Ryan drawing an overflow crowd of 10,000 at the MSP airport yesterday.
Bye-bye Barry!
I'll go with Michael Barone. He knows what he's talking about, and you don't. If the polls were right in an election like this we'd be talking about things like Carter's disastrous second term and President Kerry.
The only poll that doesn't care about party affiliation had Romney up by 5 or six before they suspended polling. We'll know for sure tomorrow. See ya in two years.
my thoughts exactly.
The ONLY way Obama wins is if there is an over the top to to point of becoming obvious voter fraud.
Repubs have to always win by 5%. Figure 2-3% fraud then enough to not force a recount - of which the Communists will win (always) by finding votes in the trunk of a 51’ Ford located in some barn out in the country somewhere.
>> Jim Cramer’s seriously suggesting Obama will pick up Georgia, Texas, Montana, South Dakota, Missouri, Indiana, and North Carolina.
Jim Cramer is one of my favorite prognosticators, because he is *reliably* wrong. You can literally take his predictions to the bank, as long as you invert ‘em first. :-)
Bunk.
So much of our political lives right now are shaped by "gee, what do other people think?" that when Romney wins, it must totally discredit these organizations as being part of the D machine.
- zot this clown please
How many did you poll?
I recall many news articles from just before the 2010 midterm elections, about how the number of registered Democrats was dwindling, some places by as much as 30%, when you factor that with the poll samplings for a 2008 Democrat turn out, and sprinkle in the evangelicals (24 Million) who stayed home in 2008, I have to think another tsunami is in the making.
So off you go now.
I won't call you names for making this kind of ignorant statement.
Whatever you might think of them, Dick Morris and Michael Barone are NOT "laymen". They are professionals, and conducting, analyzing, and critiquing polling is what they do for a living.
So, how about you? What is YOUR professional background in statistics, polling, and political analysis?
Correct me if I'm wrong, but it seems like YOU are the layman in the room.
I'll just call you "roadkill" instead.
PPP blamed “ We underestimated the amount of voter anger out there” for it’s missed call in Wisconsin.
How do they now justify being wrong, after a +8 Dem Sample last night?
Thanks for posting that. I’m a nervous wreck. That gave me some comic relief.
When you look at the poll numbers by STATE, the there is a consistent swing of 10-14% toward Romney from ‘08 results EXCEPT IN OHIO. The. vote is 10% more Republican this year, Romney wins Ohio, Florida, Virginia, North Carolina, Indiana, and Colorado. Iowa, Wisconsin, New Hampshire, and Pennsylvania are within 1.5%. If Romney, stretches the swing to 12%, he wins all three of those. Even Rasmussen shows today that Blacks are disproportionately higher percentage of early voters. And Romney has held serve with the early vote.
Voter ID in polling is not as important as the result shown and how it corresponds to the national trend. If you want to know what Ohio is going to do, look at Florida. Those two states usually finish within 2% of each other, and with a 5% lead in FL, there is no way Romney loses OH.
You are exactly right. We cannot afford to simply dismiss polls we don't like as biased or inaccurate, label anyone who expresses concern about the election as a troll, then sit back and expect everything to be just fine tomorrow. We need to assume this is going to be a close race, get out there and vote, and convince as many of our friends and neighbors as possible to vote. We cannot take anything for granted. Turnout is key.
“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.”
Repubs have not dropped the party to become indy’s like your theory suggests. That may have been the case back in ‘08 but, the GOP has a party ID edge this year, hence all of the legitimate gripes by GOP’ers. gallup shows more people identifying as Repubs than Dems (you have to scroll to the bottome). So polls that produce D+10 samples are wildly innacurate.
http://www.gallup.com/poll/158399/2012-electorate-looks-like-2008.aspx
I read an article last night that almost verbatim made the point you made but it was written back in september before the new party ID demographics had come out. It maybe made sense back then but not with what we know now
I resist posting negative things because no one can handle it. Sadly the crowd sizes don’t tell the whole story. I’ve attended huge rallies with wildly enthusiastic crowds for example for W, and he lost the state.
We like to forget on here that unfortunately there are huge masses of Americans who line up to vote dem no questions asked, or who are enthusiastic for Obama. This is the demographic revolution he is bringing about. This is why Romney isn’t ten points ahead. And we all have relatives who keep defending Obama and it’s maddening because he doesn’t deserve it.
If Obama wins I liken it to the outrage and tragedy of Clinton’s reelection and sadly I know it can happen.
Still, my prediction is a close to comfortable Romney win. Comfortable if places like Michigan, Minn, PA really are in play. If close, watch for the hispanic vote in places like Nevada, Ohio, etc. To screw us.
Thanks for that!
Do you mean “amateur analysts” such as Dick Morris who has been predicting a Romney landslide because of consistent Democrat oversampling skewing the results. Last night on FOX he claimed to have been working “professionally” as a political consultant for several decades.
Here since 2007 and this is only your 2nd post? I’m beginning to smell the ozone coming in on the wind. I suggest you get your flame retardant jeans washed, as this could be a warm day for you (as long as it lasts...)
Oh, and Arcy, here’s a way to get more bandwidth...
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.