Its partisan split skews much too far to the left.
Is there any reason to believe that Democrats enjoy an 8-point advantage in a state that consistently swings between the parties, mirroring the national vote? I don’t think so.
Quinnipiac may also be missing a massive trend: It has chosen not to poll PA anymore at all, saying that it’s “In the bag for Obama.” Both campaigns clearly disagree, as they’re both spending money and time in Pennsylvania in the closing days to capture it. It’s a must-win for Obama, and a nice-t0-have for Romney.
Update: Public Policy Trolling gets in on the massive unnatural splits act too.
Update: Ed Morrissey has more here. Bottom line, the Q poll gets Obama into the lead by undercounting Republicans compared to 2008 turnout. That is simply not in line with the 2012 reality.
PA “in the bag for O” ???
I can’t even watch TV (not that I would, anyway…): it’s political ads with pieces of programs in between, on both local and national channels. Zillions are being expended by both campaigns and PACs etc.
The table you provided and the link provides the raw sample by party identification. Translating to unweighted percentages:
R: 31%
D: 36%
I: 30%
Other: 3%
Leave it up to the reader to think a -2% from R to +1 D and + 1 Other is “skews too much to the left”.
But the leftist media in my town, Toledo, does not print that key statistic showing how skewed the poll is favoring Democrats. It just prints the big headline-grabbing result showing Obama with a big lead in Ohio, when in reality, it’s tied or Romney slightly ahead. Like all of the Mainstream Media, the media in Toledo favor the Democrats and are trying to shape the vote to favor their candidate of choice. They want him reelected and will do everything they can to make that happen. Quinnipiac doesn’t print that percentage until the end of their report. You have to go to the site and scroll to the end to find it. Hardly anyone does that. The newspapers and TV news shows know that for fact, and they can get away with their sloppy reporting.
Jay Cost writes a very interesting analysis today over at the Weekly Standard. Basically, what we’re seeing is a bimodal distribution of poll results, not an even bell-shaped curve as one would expect. Thus both the right and the left think they are ahead, and both have reason to believe that.
He doesn’t go on to explain what’s causing this very odd distribution, but other poll-watchers have said that it is the weighting that’s occurring, as Bryan points out here. Gallup and Rasmussen have been the most aggressive at updating their weighting, but many of the others are expecting party affiliation and Dem turnout to remain the same as 2008 (or to favor the Dems even more).
In other words, we are not looking at a continuum, where the assumptions made by the pollsters just vary a little along the same scale. As Jay points out, one of these set of assumptions is correct, and the other is not. Little anecdotes from boots on the ground about real-time voter enthusiasm leads me to believe Rasmussen and Gallup have got it right, but we shall see.
“…one of these set of assumptions is correct, and the other is not.”
Actually, both sets of assumptions may be wrong. As just one example, the pollsters assume there is no party-specific bias to non-respondents (e.g., people who hang up on pollsters). Is this correct? To paraphrase Rumsfeld, there are several known unknowns and unknown unknowns here, and I am not convinced that any polling organization really has a handle on what’s going on.
In short, no matter what the polls say, be sure you get your ass out there and VOTE. Or if you’re not in a swing state, connect to the Romney campaign site or the FreedomWorks site and start calling people.
Long accurate track records lead me to believe that Gallup and Rasmussen have it right. They live on their reputation, and their reputation is accuracy – the other players all have other motives.
The 2008 partisan weights were driven by Obamania. In 2012, there is no where near the level of enthusiasm for Obama that there was in 2008. Polls weighted more heavily towards Democrats in 2012 would seem destined to failure. Just judging the the relative lack of enthusiasm.
In 2010, the partisan weights shifted heavily in the Republican direction. My sense is that the 2012 actual votes will be weighted between the 2008 numbers and the 2010 numbers. Probably a bit closer to the 2010. Polls that weight the voters in this way would seem to have a better chance of being accurate.
Hewitt posted transcripts of his interviews with Q & Marist officials about their assumptions concerning party id and scewing. He calls their polls “junk.”
http://www.hughhewitt.com/blog/g/181c0976-0020-4f5b-9e60-b4c1350d2e96
Rush is speculating that some of the polls are not tightening as groundwork to declare that a Romney victory is illegitimate.
But there is another explanation. They may not be tightening in order to prove that massive fraud simply reflects an election that matched the polls.
Rush’s excuse doesn’t make a lot of sense because there are all those other polls that show Romney a little ahead in Ohio and ahead nationally. The second reason given here does make sense, because the polls showing Obama ahead can indeed be used as cover for massive vote fraud even though there are other polls showing how many real voters are likely to show up at the polls.
Remember that if there are indications of fraud, the Obama administration gets an excuse to march in and “investigate” with a compliant media ready to back up its “findings”. So the dispute would not have to play out the way it did in 2000.
The hope would be that this sort of massive fraud takes a while to arrange, so if some other unexpected states like MI, PA or WI go for Romney, the OH electoral votes will have been stolen in vain.
If Romney wins by 330 EC votes, the left will spend the next 4 years whining that we stole the election, the nation has reverted to 1958 style racism, and “the dark night of fascism” they have predicted regularly for the last 80 years has fallen. IOW it doesn’t matter how much Romney wins by, the left will work 24/7 to delegitimize him. It’s the style of politics ever since the left turned on LBJ and dethroned Nixon. I truly believe they want to reset history’s clock to 21 Nov. 1963, the last time before 2008 that they were happy with a president, and now we know that that guy stole the election from Nixon the old-fashioned way.
Did you ever think that these polls not only try to demoralize us but give plausability for a Obama victory if they commit massive electronic voter fraud…
How can we prevent that?
Food for thought
So the sample is skewed -2 republican, +1 democrat…..
Meaning that the ‘winning’ margin of 5% is really a ‘neck and neck’ margin of 2%.
Nice spin.
The key to the 2008 election was not just the high pro-Obama turnout, but also the depressed Republican turnout. For this to repeat itself, something has to depress Romney’s support in the face of strong anti-Obama sentiment in his base. The elephant in the room in this regard is anti-Mormom sentiment amoung evangelicals. There certainly are people in that community who will not vote for a Mormon. The question is how many and where. These folks are unlikely to be influenced by anything Romney can do, while Obama’s GOTV could sweep up low enthusiasm Dems.
This is certainly plausible, and I could see a pollster including adjustments for it, particularly if they have results that suppport it. I can also see them not talking about it much if they are. Not that I have reason to believe any of this is true – but this was an issue in the primaries, and it has quietly disappeared from the coverage since then, so I wonder.
Whatever reservations some evangilicals may have about mormonism are more than offset by horror of Obamas policies, so even if they dont want to vote for romney, they will still vote against obama. The few evangilicals who wont are probably in deeply repub states where they wont matter anyway.
This all boils down to Weighting, specifically the 18-34 age group. Getting 18-34 year olds to answer the phone is becoming harder and harder every year (mostly due to cell phones). Looking at the age breakouts,
18-34 age group – 113 (10.2%) error margin +/- 9.22
35-54 age group – 334 (30.1%) error margin +/- 5.36
55 + age group – 640 (57.7%) error margin +/- 3.87
the weighting of those 18-34 year olds is the big issue. Pulling the rough estimates for the 2012 Ohio population breakouts, 18-34 year olds represent 28.8 of the 18+ population (35-54 = 35.5% and 55 + = 35.7%). So Quinnipac has to weight their 10.2% respondents up to a 28.8%. This helps explain the large margin of error.
18-34 year olds do tend to be more associated with the “Democrat” viewpoint. So using Quinnipac’s percentages, one 18-34 year old is representing roughly around 22,500 other 18-34 year olds in Ohio. One 35-54 year old respondent represents roughly 9,400 other 35-54 years olds and one 55+ year old respondent represents roughly 4,900 other 55+ year olds in Ohio.
Using hypotheticals, if you have 55 respondents in the 18-34 age group that say they are Democrats, 25 Republicans and 33 Independents, then that would translate into ~1,232,000 Democrats, ~562,500 Republicans & ~742,500 Independents in the 18-34 age group. I’m guessing in the 35-54 and 55 + categories, you’d find an increasing number of Republicans compared to Democrats. Then you would come up with the Weighted totals of 29% Republican, 37% Democrat & 34% Independent/Other.
18-34YOs also have, by FAR, the lowest voting percentage.
I live in Ohio. All of the all of the elected State officials are Republicans. The General Assembly and the Senate are Republican. Other than 2006-2010, it has been that way for the past 25 years. There is no way that Ohio is D+ anything. No Way.
It was pointed out elsewhere that if you can lose indies by double digits and still carry the state, by definition it isn’t a swing state.
Just ridiculous.
I’m curious if anyone has done the math on projecting the final vote tallies based on turnout assumptions. For example, subtract the % of votes in a state like Ohio based on the decrease in enthusiasm gap in Ohio, increase the republican turnout by their increase, and calculate the independent switch. I’m struggling –and have been for a long time (long prior to the first debate)– to see how this election is anything other than a landslide. I’ve never seen a president this unpopular with real people (as opposed to the media fed Bush hatred of the special-interest and leftover hippie crowd, which was nowhere near as broad).