Here's the latest Pew Poll on the Presidential race, which has Obama up 10 over Romney. The first inkling that something is very very skewed with this poll is on the second line, where it shows Romney only +1 among men. No way. No f-in way. Men are overwhelmingly against Obama. Every poll I've seen, every election over the past thirty years, and every gut instinct tells me that men overwhelmingly hate President You-Didn't-Build-That.
So how can this poll say the opposite? Well, page down to the line where it says the party identification. Notably, they don't break it out by percentages... you have to actually do the math. But what you find out is that, in a poll of roughly 2,000 respondents, they are using splits of D/R/I of 43%-24%-32%. A NINETEEN POINT DEMOCRATIC ADVANTAGE!
In other words, if you used a split that looks more like the even split in 2010, or even like the D+6 split of 2008, Romney would likely be winning.
Just another reason not to even look at polls unless you know the splits and whether they are polling likely voters or registered voters (more Dems, but more people who won't show up to vote), or adults (even more Dems, and even more who won't show up to vote). The gold standard should be an even split or a very small advantage to Dems in a poll of at least 1,000 likely voters.
Oh, and it's also worth noting that the liberal media will report this poll uncritically, as here, on Politico. If a guy like me sitting in Wauwatosa, Wisconsin can figure this out, you know that professional political analysts ought to be able to. Which is why more and more smart people are tuning out the professional political pundit class, and realizing that they are simply propagandists for the Left.
No comments:
Post a Comment