So, I was looking at the crosstabs on CNN's snap poll of "registered Americans" after the debate claiming Romney won the debate 69-25. They start on page 8:
Notice anything funny? According to the breakout, all the people surveyed are white, 50+, and from the South. Are they being serious with this? I know the media loves a horserace, and I'll admit Obama was less passionate and shouty than Mitt, but it's pretty hard not to ascribe a motive to their selection process.
Appreciate the rec list bump. Been a long time estranged from the site and figured I would do my due diligence. Romney, to me, came off a bit condescending, angry, and frequently interrupting of the POTUS. I got curious as to what the reaction would be among minorities and women, so figured I'd do some digging. Imagine my surpise to find that no one had bothered to ask them.
From TFA: "A CNN poll of debate-watchers found Mr. Romney very clearly ahead, with 67 percent of registered voters saying he won the debate, against just 25 percent for President Obama.
A CBS News poll of undecided voters who watched the debate found 46 percent siding with Mr. Romney, 22 percent for Mr. Obama and 32 percent saying it was a tie.
Google, which is experimenting with online surveys, found 38.9 percent of respondents saying they thought Mr. Obama performed better in a poll it conducted during the debate, against 35.5 percent for Mr. Romney and 25.6 percent who said it was a draw. But a second poll they conducted after the debate found 47.8 percent of respondents giving Mr. Romney the advantage, against 25.4 percent for Mr. Obama...
The 67 percent of voters in the CNN poll who said that Mr. Romney won Wednesday night’s debate was higher than in any of the network’s surveys of the 2008 debates.
By comparison, 51 percent of debate-watchers in a CNN poll after the first presidential debate in 2008 said that Mr. Obama had the stronger performance. And 58 percent said so after the third and final presidential debate in 2008."
A bit odd how strongly the CNN poll diverges from CBS, Google, and even it's own polling history, no?
12:06 AM PT: To avoid confusion:
Apparently, as noted below in the comments, the N/A on the crosstabs is listed to denote that the sampling range places the margin of error below statistical significance. So, there were presumably under 50's, minorities, and non-southerners sampled. It's still odd, to me at least, that only southern, 50+ whites were sampled enough to get a significant MOE on the results, when I'm sure there was probably a sizeable nationwide debate audience. Maybe that has something to do with the willingness of the original respondents to be bothered with a second phonecall survey. In any event, takeaway for me is that the loss seems pretty evident, but it's certainly fairly mild as it did not even really move this particular block much in terms of their voting decisions. I wouldn't go so far as to cry scandal or shennanigans, but it strikes me as a big oversight. CBS's poll seems much more on point, with about a third of the undecideds breaking mostly towards Romney, which actually lines up pretty accurately with Nate's earlier prediction today of a modest increase in the polls toward the Challenger.