UnSkewed Political Projections
Now Even the "Unskewed" Polling Has Obama out in Front
|
Posted Friday, Sept. 28, 2012, at 2:02 PM ET
Mitt Romney arrives on the stage during a campaign rally at the Valley Forge Military Academy and College in Wayne, Pa., Friday
Photograph by Jewel Samad/AFP/Getty Images.
As we explained in yesterday's Slatest PMnewsletter, some conservatives are becoming increasingly vocal regarding their doubts about recent polls that show President Obama pulling out in front of Mitt Romney. In short, they contend that pretty much everyone but the conservative pollsters over at Rasmussen are relying on turnout models that unfairly favor Democrats. If the models were tweaked appropriately, they say, Mitt Romney would have a substantial lead coming down the home stretch.
Or at least that's what their "unskewed" polls had been showing up until yesterday when Fox News released its latest survey, which showed Obama up by 5 points, 48 percent to Romney's 43.
Dean Chambers, who runs unskewedpolls.com, quicky worked his usual magic on the Fox data but this time his "unskewing" wasn't enough to move the numbers in the GOP's favor. The result: Obama up by 2 points. While that's obviously less than the gap the Fox News poll showed, it's nonetheless noteworthy because it's a major departure from the past dozen or so of Chambers' "unskewed" polls that claimed to have Romney well on his way to a historic victory this fall.
"When the data from the Fox News poll is unskewed by weighting their reported percentages between Romney and Obama to the partisan affiliations showed by Rasmussen's extensive data results on that issue, the overall picture of the race is different. With Republicans weighted 34.8 percent, Democrats at 35.2 percent and Independents at 30.0 percent, the results calculate to Obama leading but by a smaller 46 percent to 44 percent margin over Romney"
It's also worth pointing out again that the major pollsters who keep tabs on the race for a living obviously don't see eye to eye with Chambers. Frank Newport, the editor-in-chief over at Gallup, penned a forceful response to the polling critics yesterday, explaining that their main beef with the numbers—what the doubters say is an oversampling of Democrats—actually is just further proof that the president is out in front. More info on that here.
Hey liberals, you're right: The polls are telling us something
As I exposed the polls for being skewed in favor of the Democrats and showing a larger lead than President Obama
has had, if he had one at all, the liberals are screaming that the
polls really do tell us something, and they were quick to defend the
polls when they showed large Obama leads. But looking at the polls now, I
see liberals have a point, the polls really are telling us something
now. Even though they are still skewed, especially the three that came
out yesterday, they are telling us something quite significant. Skewed
as they are, the recent polls are showing Mitt Romney gaining traction.
President Obama enjoyed the largest lead of the three polls in the heavily skewed Quinnipiac poll, which showed him leading 49 percent to 45 percent over Romney. Unskewed, that poll's data showed only a 49 percent to 48 percent lead, within the margin of error, for Obama. Obama leads 49 percent to 46 percent, within the margin of error of 3.4 percent, the new NBC News/Wall Street Journal poll released yesterday. That poll, unskewed, would show a Romney lead of one percent, 49 percent to 48 percent.
The candidates were tied in a new very skewed National Journal poll released yesterday. That poll showed them both at 47 percent, based on surveying an undisclosed number of likely voters in a survey that had a very large 4.5 percent margin of error. I have to wonder why they didn't disclose the margin of error, even Real Clear Politics doesn't list a margin of error this National Journal poll, I can only conclude it must have featured an embarrassingly small sample of likely voters. Unskewed, Mitt Romney takes a 49 percent to 46 percent lead in this poll's data.
I think there are two trends going on here. Some of the pollsters are quietly toning down the degree to which they are skewing their samples to gradually get more accurate results as we get closer to the election, so they will report the most accurate results they are capable of when they release their last few polls before the election. I called and predicted this would happen. The Fox News poll and the CNN/ORC poll I covered, still skewed but less skewed in each instance, are cases in point.
The other trends, as the polls remain skewed somewhat, is that Mitt Romney is actually gaining traction with voters and this is showing up even in these skewed polls. Not to long ago, the Real Clear Politics average of polls had Obama up by more than four percent, I think it was nearly five percent, and that has come down to 3.4 percent in favor of Obama today. As the polls become less skewed closer to the election, so too will the RCP average. In fact, I predict the RCP average will reflect a Romney lead before election day. When both the RCP average and the UnSkewedPoll.com average show Romney leading before the election, which is going to happen, we'll know two things. The polls are fairly accurate, because the pollsters care more about their credibility as the election is near, and that Mitt Romney is going to win the election. I stand behind both predictions.
Those three polls released yesterday show Obama leading by an average of just 2.3 percent, within the margin of error of most polls and essentially tied statistically. That average is 48.3 percent for Obama and 46.0 percent for Romney, leaving 5.7 undecided. If most of those undecided voters vote for Romney, which is likely because the very reason they are likely to be undecided is having decided they aren't voting for Obama but haven't been sold on voting for Romney yet, Romney would win a majority of the popular vote even as indicated by this data.
But remember that data is built on polls that all three over sampled Democrats. If the surveys are re-weighted or as I call it, unskewed, they reveal on average that Romney leads by one percent, or 48.7 percent to 47.7 percent for the president.
So yes, liberals, the polls are telling us something. Romney is gaining strength and voters focus more on the choice at hand, and undecided voters are finding the reasons they will vote for Romney and increasingly so telling the pollsters they're voting for Romney. Liberals can believe, if it comforts them, that the polls in August and September of this year were not skewed as I exposed them to be, but either way they will not be able to deny that Romney takes the lead in the polls in a few weeks before winning on the election day. Liberals, keep denying the polls are skewed, but get ready to join us in saying President Romney. It's going to happen.
President Obama enjoyed the largest lead of the three polls in the heavily skewed Quinnipiac poll, which showed him leading 49 percent to 45 percent over Romney. Unskewed, that poll's data showed only a 49 percent to 48 percent lead, within the margin of error, for Obama. Obama leads 49 percent to 46 percent, within the margin of error of 3.4 percent, the new NBC News/Wall Street Journal poll released yesterday. That poll, unskewed, would show a Romney lead of one percent, 49 percent to 48 percent.
The candidates were tied in a new very skewed National Journal poll released yesterday. That poll showed them both at 47 percent, based on surveying an undisclosed number of likely voters in a survey that had a very large 4.5 percent margin of error. I have to wonder why they didn't disclose the margin of error, even Real Clear Politics doesn't list a margin of error this National Journal poll, I can only conclude it must have featured an embarrassingly small sample of likely voters. Unskewed, Mitt Romney takes a 49 percent to 46 percent lead in this poll's data.
I think there are two trends going on here. Some of the pollsters are quietly toning down the degree to which they are skewing their samples to gradually get more accurate results as we get closer to the election, so they will report the most accurate results they are capable of when they release their last few polls before the election. I called and predicted this would happen. The Fox News poll and the CNN/ORC poll I covered, still skewed but less skewed in each instance, are cases in point.
The other trends, as the polls remain skewed somewhat, is that Mitt Romney is actually gaining traction with voters and this is showing up even in these skewed polls. Not to long ago, the Real Clear Politics average of polls had Obama up by more than four percent, I think it was nearly five percent, and that has come down to 3.4 percent in favor of Obama today. As the polls become less skewed closer to the election, so too will the RCP average. In fact, I predict the RCP average will reflect a Romney lead before election day. When both the RCP average and the UnSkewedPoll.com average show Romney leading before the election, which is going to happen, we'll know two things. The polls are fairly accurate, because the pollsters care more about their credibility as the election is near, and that Mitt Romney is going to win the election. I stand behind both predictions.
Those three polls released yesterday show Obama leading by an average of just 2.3 percent, within the margin of error of most polls and essentially tied statistically. That average is 48.3 percent for Obama and 46.0 percent for Romney, leaving 5.7 undecided. If most of those undecided voters vote for Romney, which is likely because the very reason they are likely to be undecided is having decided they aren't voting for Obama but haven't been sold on voting for Romney yet, Romney would win a majority of the popular vote even as indicated by this data.
But remember that data is built on polls that all three over sampled Democrats. If the surveys are re-weighted or as I call it, unskewed, they reveal on average that Romney leads by one percent, or 48.7 percent to 47.7 percent for the president.
So yes, liberals, the polls are telling us something. Romney is gaining strength and voters focus more on the choice at hand, and undecided voters are finding the reasons they will vote for Romney and increasingly so telling the pollsters they're voting for Romney. Liberals can believe, if it comforts them, that the polls in August and September of this year were not skewed as I exposed them to be, but either way they will not be able to deny that Romney takes the lead in the polls in a few weeks before winning on the election day. Liberals, keep denying the polls are skewed, but get ready to join us in saying President Romney. It's going to happen.
No comments:
Post a Comment