Realville: NBC joins the “Poll Truther” Brigade…

…with this statement concerning their just released polls:

In Ohio, where there has been a renewed focus by the Romney campaign after the former Massachusetts governor’s strong debate performance, Obama leads 51 percent to 45 percent. That’s a 2-point uptick for Romney.

But the Ohio poll also included an 11-point advantage for self-described Democrats — 40 percent to 29 percent for Republicans. Last week’s poll had a narrower 5-point advantage for Democrats. . (In 2008, the party identification split was 39 percent Democrat and 31 percent Republican, according to exit polls.)

Amazing how the samples matter once Romney’s numbers improved in Ohio, I wonder where they buy their Cheetos?

Meanwhile in VA the previous D+4 sample becomes an R+1 samples and guess what? Romney now has a +1 pt lead! Change the sample and you change the results. How about that!

“But DaTechGuy” you say, “That party identification is psychological, people call themselves more Republican because the GOP is doing better and Romney is doing better.”

Well if that’s the case then we should see change in Florida. While there seems to be universal agreement that the Debate Moved numbers for Romney Florida remains unchanged, how can that be when other Florida Polls have moved…

…what a coincidence, the D vs R sample hasn’t moved and PRESTO the Obama Romney gap doesn’t move, even when the entire universe concludes Obama is in free fall.

Funny how that works isn’t it?

Mind you Florida & VA are the same states the Suffolk decided not to poll anymore because Romney had it cold.

Why do I have a sneaking suspicion that after the next debate (or perhaps after tonight’s VP debate) the party samples will move and the numbers will improve for the Obama campaign?

If people don’t realize they are being played it’s only because they don’t want to know it.

Update: take a look at the party splits vs the poll results from CBS


It’s worth noting that there is no measure of if people “lean” amazing how the results seem to match the samples within a point or two.