S
snoscytheI posted this in another thread, but new articles today and yesterday made me revisit with some changes, and I thought it could stand alone. Here are some fun sites for those of you concerned by the "polling" showing Obama pulling ("polling"?) away:
First, a site chronicling just how accurate/inaccurate given pollsters were in 2008:
http://blog.chron.com/txpotomac/2008/11/the-list-which-presidential-polls-were-most-accurate/ " onclick="window.open(this.href);return false;
SEVENTEEN of the polls significantly overreported Obama support in 2008....only 3 overreported for McCain, and those three were in the Top-Five most accurate.
Now consider how much closer this race is in 2012 than it was in 2008. To the extent the pollsters are biased, they have more incentive to be biased this year in a close election. Living in Colorado, my wife and I have started trying an experiment. We are both Republican, but we have noted a pretty notable number of "dropped calls" immediately after answering that we consider ourselves Republican. Over the last 10 days or so, we have taken to answering "independent" or "democrat." We've been dropped twice as independents, but neither of us have ever had a dropped call after responding "democrat." In the interest of full disclosure--we terminate the call ourselves when we get to the "who are you going to vote for" if we've misrepresented our affiliation so as not to further skew data. It's fishy as hell the extent to which people are losing connections in these polls, and we get 2-3 calls a night from pollsters since we're registered to vote in CO.
So, taking that, a fellow has put out a site where he "unskews" polls by evaluating who was polled and controlling for discrepancies in representation. If Democrats are oversampled, he adjusts for what the expected electorate on election day will be.
http://unskewedpolls.com/" onclick="window.open(this.href);return false;
http://www.examiner.com/article/mitt-romney-vs-barack-obama-heavily-skewed-swing-state-polls-1?cid=db_articles " onclick="window.open(this.href);return false;
You'll see he has Romney ahead nationwide by almost 8%, and has Romney winning or tied in all the "Swing States," and tied or within 2% in PA.....which no one has called a swing state before, but where Romney announced was back in play today.
The candidates' behavior is further indicative of the true status of the race. Obama is somewhat frenetically trying to raise money and pleading for donations, skipping things like meeting with Netanyahu to hold campaign rallies. They're lobbing every sort of campaign message into swing states they can think of---emptying their clip in September. Romney on the other hand has not really deviated from his campaign plan other than to address new developments in the Middle East, gaffes, and airplane windows. Despite all the noise generated from the reported polls, it is Obama's campaign that is panicking while Romney's campaing holds to the course. While the media reports what Romney is doing as foolish, stubborn, or naive, to me it is indicative that internal polling and focus grouping that are much more statistically sound than most public polls show the Romney campaign that they are winning and that their message is swaying and holding voters.
What's intriguing is yesterday's Politico article attempting to debunk this possibility:
http://www.politico.com/news/stories/0912/81845.html" onclick="window.open(this.href);return false;
But the author just ties himself in knots. He says pollsters are professional and unbiased, but goes out of his way to point out that Rasmussen is a more conservatively biased poll. (Here, eat your cake and have it too!) Not only does that ignore the empirical evidence that 2008's polls were biased in Obama's favor, but also ignores the empirical evidence that Rasmussen was spot on in 2008.
And then there is today's piece, where pollsters admit under oath they spin and obsfucate:
http://www.politico.com/blogs/under-the-radar/2012/10/a-pollster-under-oath-137100.html " onclick="window.open(this.href);return false;
That's where I'm at.
First, a site chronicling just how accurate/inaccurate given pollsters were in 2008:
SEVENTEEN of the polls significantly overreported Obama support in 2008....only 3 overreported for McCain, and those three were in the Top-Five most accurate.
Now consider how much closer this race is in 2012 than it was in 2008. To the extent the pollsters are biased, they have more incentive to be biased this year in a close election. Living in Colorado, my wife and I have started trying an experiment. We are both Republican, but we have noted a pretty notable number of "dropped calls" immediately after answering that we consider ourselves Republican. Over the last 10 days or so, we have taken to answering "independent" or "democrat." We've been dropped twice as independents, but neither of us have ever had a dropped call after responding "democrat." In the interest of full disclosure--we terminate the call ourselves when we get to the "who are you going to vote for" if we've misrepresented our affiliation so as not to further skew data. It's fishy as hell the extent to which people are losing connections in these polls, and we get 2-3 calls a night from pollsters since we're registered to vote in CO.
So, taking that, a fellow has put out a site where he "unskews" polls by evaluating who was polled and controlling for discrepancies in representation. If Democrats are oversampled, he adjusts for what the expected electorate on election day will be.
http://unskewedpolls.com/" onclick="window.open(this.href);return false;
You'll see he has Romney ahead nationwide by almost 8%, and has Romney winning or tied in all the "Swing States," and tied or within 2% in PA.....which no one has called a swing state before, but where Romney announced was back in play today.
The candidates' behavior is further indicative of the true status of the race. Obama is somewhat frenetically trying to raise money and pleading for donations, skipping things like meeting with Netanyahu to hold campaign rallies. They're lobbing every sort of campaign message into swing states they can think of---emptying their clip in September. Romney on the other hand has not really deviated from his campaign plan other than to address new developments in the Middle East, gaffes, and airplane windows. Despite all the noise generated from the reported polls, it is Obama's campaign that is panicking while Romney's campaing holds to the course. While the media reports what Romney is doing as foolish, stubborn, or naive, to me it is indicative that internal polling and focus grouping that are much more statistically sound than most public polls show the Romney campaign that they are winning and that their message is swaying and holding voters.
What's intriguing is yesterday's Politico article attempting to debunk this possibility:
http://www.politico.com/news/stories/0912/81845.html" onclick="window.open(this.href);return false;
But the author just ties himself in knots. He says pollsters are professional and unbiased, but goes out of his way to point out that Rasmussen is a more conservatively biased poll. (Here, eat your cake and have it too!) Not only does that ignore the empirical evidence that 2008's polls were biased in Obama's favor, but also ignores the empirical evidence that Rasmussen was spot on in 2008.
And then there is today's piece, where pollsters admit under oath they spin and obsfucate:
That's where I'm at.