The folks at PEW have released their last poll of the year and their headline seems to be good news for President Obama and bad news for me who has predicted his defeat:
Barack Obama has edged ahead of Mitt Romney in the final days of the presidential campaign. In the Pew Research Center’s election weekend survey, Obama holds a 48% to 45% lead over Romney among likely voters.
Here is the line from the Poll
Leftist sites are rejoicing at a 3 pt lead with seemingly no time for Romney to change it, I’m sure Nate Silver is already boosting his averages.
Alas for the left, PEW also released their internals and even worse for them I haven’t forgotten how to do basic math so lets look at this last attempt to pretend something is true that isn’t.
First the splits in Party (click to enlarge) :
Pew interviewed 2709 people, 2611 identified themselves as either Republicans Democrats or Independents by their own count 843 Republicans, 1007 democrats and 761 independents. That works out to the following:
Democrats 1007 / 2709 = 37.1%
Republicans 843 / 2709 = 31.1%
Independents 761 / 2709 = 28.8%
So right off the bat we have a poll with a base advantage for Democrats of 6 pts. Even though both Gallup and Rasmussen have told us that the electorate is majority GOP this time around.
That would be enough to laugh this poll off but the mathematical legerdemain doesn’t end there.
All year we were told that this election was going to be who by he who took the independent voters. If you look at this new Pew poll who is leading among independents? Why lo and behold it is….Mitt Romney by 44% to 41% The GOP goes for Romney 91-7 & dems for Obama 94-5
But that doesn’t matter, apparently at this late date the independent vote that was so important just a few months ago doesn’t matter in election 2012 anymore.
That’s strike two but you need three strikes for an out, is there a third one? Guess what, there is!
If we look at the sample in terms of Gender did you know that the electorate in 2012 women are going to outvote men? As there are more women than men that’s not so odd, until you look at the margin… (click to enlarge)
Pew polled 1538 women vs 1171 men. That works out to the following percentages:
Women 1538 /2709 = 56.8
Men 1171 /2709 = 43.2
No wonder the women’s vote is so important! According to PEW apparently women aren’t going to just outvote men in 2012. They are going to do so in this election by 13 whole points! That is pretty interesting, particularly if you look at this chart from a pew study from just 4 years ago
In 2008 Pew reports a 4.2 point turnout advantage over men. Perish the thought, apparently according to THIS poll the 2012 gap is going to make that 2008 figure look positively miniscule…
…if you buy it that is.
OK that’s three strikes, but lets add a throw to first just in case this gets by the Catcher.
Let’s take a closer look at this sample, how representative of the electorate is it? Well one way to find out is to see how this group voted last time around (click to enlarge):
Well look at that a 13 point advantage of Obama voters vs McCain voters, but again this is registered voters, so lets remove those who didn’t vote so we can get down to those likely guys, what would the numbers be?
Obama voters 46 / 86 voters = 53.4%
Ok that seems to mean these voters President Obama took 52.9% of the vote so that’s a full half a point MORE than the president’s share of the vote in 2008
McCain voters 33 / 86 voters = 38.3%
Hmmmm John McCain drew 45.7% of the vote so this figure is a full 7.4 points BELOW John McCain’s vote total from 2008
That a 15.1 point gap in the popular vote that Barack Obama won by 7.2 points, MORE THAN DOUBLE the actual 2008 number. That’s a throw to first to end the inning and the game.
So this is what the Pew poll would have you believe:
#1 We are going to have a D+6 electorate even though all signs indicate the Democrats are not all that thrilled with president Obama but in 2010 the republicans managed historic wins in the states, and the house and big senate wins too.
#2 Even though independents favor Mitt Romney by 3 points in this poll and Republicans are voting 91-7 for Mitt according to their own number the electorate in 2012 is SO democratic that Obama is still up by 3 points
#3 The female electorate on election day is going to outnumber the male electorate by a full 13 points thus negating any advantage that Mitt Romney might have with men.
#4 A sample where the people interviewed supported Obama in 2008 by a figure double than the actual result should be trusted to show us how the electorate will vote in 2012.
There is a phrase to describe this kind of thinking, it’s known as “Willing suspension of disbelief”.
While that is dandy if you are a Doctor Who fan watching the Doctor reverse the polarity of the neutron flow, it doesn’t do a whole lot of good if you are trying to figure out how an election is going to end up in the real world.
The fact that PEW and the left is flogging a D+6 poll with a +13 point sample of women is funny, that this poll shows Romney leading among independents yet has Obama up 3 is funnier, that you are releasing a poll two days before an election that has a sample whose members supported Obama at double the rate of 2008 isn’t just funny it screams one thing:
This is the type of thing you do when you’ve got nothing else, they’ve got NOTHING and we are going to beat them so bad they are going to wonder why they were so willing to sell their credibility on a candidate as bad as this president.
Update: If you think this is trouble for the left consider the new CNN poll. As Breitbart reports:
Moreover, the polls’s crosstabs indicate that Romney is winning self-described independent voters by a giant 59%-37% margin. A 22-point lead among independents virtually guarantees victory for Romney. Yet Democrats are so heavily over-represented in the CNN poll that Romney’s 22-point lead becomes a mere 49%-49% tie.
Here is the sample:
A total of 1,010 adults were interviewed by telephone nationwide by live interviewers calling both landline and cell phones. All respondents were asked questions concerning basic demographics, and the entire sample was weighted to reflect national Census figures for gender, race, age, education, region of country, telephone usage and whether respondents own or rent their homes. Registered voters were asked questions about their likelihood of voting, past voting behavior, and interest in the campaign; based on answers to those questions, 693 respondents were classified as likely voters. Respondents who reported that they had already cast an absentee ballot or voted early were automatically classified as likely voters. Among those likely voters, 41% described themselves as Democrats, 29% described themselves as Independents, and 30% described themselves as Republicans.
Media Credibility day not withstanding, if you are going to skew a poll with a D+11 sample to get a tie you might as well get a D+12 sample and give Obama a lead. If you are going to make a fool of yourself, why not?
It has Mitt & Obama Tied at 49, but only manages this feat with a +d11 sample