I awoke this morning to a front page article in our little right wing daily paper here highlighting an AP poll that found that John McCain had narrowed the gap, and that Obama’s lead was only one point, 44-43%. This, said the article, is a result of the third debate and the success of the “Joe the Plumber” talk.
Right away I went to RealClearPolitics, where I follow the polls, to see if the others had noticed the same closing. But no change – Obama by anywhere from 6-12 points. So I judged the AP poll to be an outlier. We’ll see.
However, here is some of the details behind the AP poll (thanks to Americablog.com): They weighted their respondents so that 45% of them were born again/evangelicals. However, in 2004, the evangelicals made up only 23% of the voters, scary enough in itself, but that’s another story. By adding an additional 22% of them to the sample for their poll, AP heavily skewed it in favor of McCain.
Incidentally, there are eight national polls at work, and seven of them show Obama with a large lead. Our Chronicle chooses to run a front page story on the one that doesn’t. Right wingy dingy anyone?
Piece of Mindful 
Plus, a different and more annoying problem than what the Chronicle did (assuming they had other options to go with besides that AP article, which may not be the case) is what the AP is engaging in. News organizations do their own polls and then pretend that poll is the only one that exists. In their world there’s no set of polls to average, so any movement is news. It generates headlines and sells papers, but doesn’t inform us at all.
LikeLike
Anytime I see a poll that promotes its “likely voter” results, I just think of the Tester-Burns race, and smile …
LikeLike
Polls have always had sampling bias, statistical error, and poor question phrasing to confuse the results, but is political bias now also entering the field, do you think? From what you report here, it seems so…
LikeLike
Hard to judge motives, but to introduce such a large number of evangelicals, disproportionate to the population they are surveying, speaks either of gross incompetence or an agenda.
LikeLike
I think I’ve stumbled the real reason for such inequalities in poll results. Via NRO:
“David Johnson, the CEO of Strategic Vision, shared the messages with National Review Online.
One of the messages stated:
My goodness, your polls stinks. There are 3 polls that have Obama by double digits and only yours has Obama down. WOW!. How come your poll is the only one giving Palin high favor ratings? I think you nee dto be careful tonight when you get in your car and might want to check underneath your car. SCRAP YOUR IDIOTIC POLLS OR ELSE!
Another stated:
A poll that gave Sarah Palin and Barack Obama the same favorability rating is wrong off the bat. Be careful going outside tonight because you might not see tomorrow.
A third message stated:
Why would your presidential election poll results be so drastically different from every other reputable poll taken over the same time period? Are they that dumb or are you guys that smart? Smart guys wind up dead.”
At least evangelicals resist putting Simtex under your car seat.
LikeLike