And speaking of polls…
Ohio, Ohio, Ohio. Yes, it’s a very important state, and stands an excellent chance of being so again this year. The WaPo crows that Obama is ahead in that state by 8 points among likely voters, with a 7+ Democratic sample.
Remember that Obama won that state by about 4% in 2008. Does it seem likely that he’s doubled his lead from that year? Especially when exit polls in 2010 suggested a +1 Republican advantage, and a Gallup poll of 2012 voter IDs in Ohio indicates only +1 Democrats?
You can also find a poll that directly contradicts the WaPo poll, for what it’s worth. I have no idea about the relative reliability of either pollster, but I certainly know the bias of the WaPo in general. On the other hand, the results of a recent Gravis Marketing/Capital Correspondent Ohio poll (with a 10+ Democratic sample, yet!) show a very different picture: a race that’s virtually tied, and has been for the entire month of September. Note also the large number of “undecideds” in that poll: 10% (pretty stable at that level for the whole month). If the poll is accurate, the Ohio election will be decided on how they end up breaking.
And before you excoriate those 10% by asking how on earth anyone can still be undecided at this point, let me just say that I suspect many of them are people who distrust the MSM, are tired of Obama, and yet feel they don’t know Romney all that well. They may even be waiting for the debates to get a better look at him. Remember the debates?
Naturally the “undecided” category includes a lot of different voters. I suspect this year it is hiding a contingent of Bradley Effect/Wilder Effect voters. It also hides people who plan to vote for one of the two candidates but don’t want to say. And it includes people who might not like either candidate and might not even decide to vote.
Although I’ve often toyed with the idea of lying to pollsters, I’ve not done so yet. I just don’t answer the phone. If I answered, though, instead of lying and telling them I supported Obama, I’d probably tell them I was undecided, even though I most definitely am not.
I think Kurt’s right.
One a side note, Romney will be visiting Ohio tomorrow, with Mike Rowe joining him. Romney responded to an open letter Rowe had written him and it looks like they have some common ground. Awesome.
http://tinyurl.com/dxamy38
Is it just me or do we all go through all this blather about polls every four years? Seems like it is always the same and always just as numbing.
Neo, that and the previous polling post were well worth reading. Thank you.
So what do we know? In 2008 Obama won Ohio 51.5% to 46.9% the balance went to third parties. According to the cnn exit polls in Ohio the turnout was D-39%, R-31%, I-30%. Dems voted O – 89% Mc- 10%. Repubs – McCain 92% O – 8%. Obama won independents 52% – 44%.
So they are predicting that Obama is up by 8 in a state he won by 4.6% with a +8 turnout advantage. Their sample was D-35% R-26% I-35%. So compared to 2008 they are predicting a turnout 5% less Republican, 4% less Democrat, and 5% more independent. Breakdown of the vote by party affiliation is not provided, but we do know that Obama is running behind with independents in other polls, why would this be different in this poll?
There was another poll which showed Ohio within 1%, and in PA the Pittsburgh Tribune Review has just published a Susquehanna Polling Poll which shows Romney and Obama within 1% of each other.
Interestingly, they also publish the entire poll results and two facts jumped out at me even with a cursory look. First, it seems that many of the respondents (40+%) came from the highly pro-Democrat Phila area; that would skew the poll towards Obama.
Second, many of the responses overwhelmingly show a preference for Romney (Who would be better at creating Jobs? Romney. Managing the Economy? Romney. Is the country going in the right direction? No.) Anyone who truly believes that this country is going in the wrong direction can’t possibly be considering a vote for Obama (or they are lying). So there is a disjunct between the specific tactical answers (all against the status quo) and the “who would you vote for” responses.
With all of this, it still shows a neck and neck race and it has yet to take any undervote effect into account.
“”Note also the large number of “undecideds” in that poll: 10%””
These 10% are undecided because they know they have 41 days before they have to finally admit to themselves that they were seriously suckered by democrats in 2008. We’ll affectionately refer to them as the “landsliders” come November 7th. And that name will apply probably because most of them simply won’t show up for the election.
I am amazed that the race appears (emphasis on appears) to be as close as it is. I hope this is another 1980 election, and the closeness is just an illusion. The debates are going to be crucial… Romney better get himself ready to put forward the best presentation of his career to explain why, even if you have doubts about him, he is a far better option than the present president. Obama just cannot have another 4 years.
JL,
“I am amazed that the race appears . . . to be as close as it is.”
Dick Morris has consistently offered two reasons fort his whih I find quite readonable. The first is that pollsters ar emostly
Sorry for the interruption.
. . . pollsters are mostly using a 2008 turnout model. (Neo addresses this in the net thread, “Pity the Poor Pollsters”). Remember thsi model skews toward th Dems.
Secondly, Morris also cautions to look at precisely what entity is commissioning the polls. They have influence on the poll results sometimes simply in how the questions are asked or in what order.
In the “Pity” thread, Neo also addresses the response rate. If you notie the table, the reseponse rate has continually decreased and she questions the accuracy of a poll with only a 10% cooperation rate. I add that the geographic distribution of that response is also important, and especially so if one is only fielding a 10% cooperation rate.
This would tend to belie Neo’s own concern that polls in past elections seem to have been accurate in predicting winners if not margins. If response rates were higher and geographic distributions were broader, these earlier polls might simply have been more accurate barometers of the politcal lay of the land.
If on the other hand, response rate has shrunk to 10% and only the most enthusiastic proponents are willing to respond, we may well be looking as a “false positive.” Add to that the fact that the media is mostly in the tank, and I offer that most current polls are dubious at best.
Go to unskewedpolls.com . There polls are reweighted based upon Rasmussen’s party identification polls. The results show Romney/Ryan leading with ~52% of the vote (which, coincidentally agrees with the Univ of Colorado study).
I’m NOT saying this means Romney/Ryan will unequivocally win; it’s just that this is another skewed poll from yet a different point of view and it yields particularly different results. Only the reader can decide which skew s/he believes to be more representative, and until Nov 7th, when we know for sure, it’s all opinion and speculation anyway.
So the message is clear, don’t be disuaded by published polling results, just get to the the ballot box and vote.
A correction. Neo’s table shows a 14% cooperation rate fro 2012, not the 10% I mentioned above (working from memory is always a dangerous thing fro me).