As of midnight, Real Clear Politics showed Barack Obama with a 2.9-point lead over Mitt Romney in the average of the most recent six presidential election polls. One of those polls is a P-U production of Pew Research Center which shows Obama up by 8 points among 2,343 registered voters. The preposterous weighting of the sample is 37.1% Democrats, 30.6% Republicans, and 32.3% independents.
Any time a poll reveals the Romney v. Obama breakout in each of those three categories, I can run the results through what I'll tentatively christen the NewsBusters/BizzyBlog Poll Decoder, showing what the result would be using party affiliation results found by Rasmussen as of early September and Gallup as of before the Democratic National Convention. Here's what happens when one removes the stench from Pew's poll:
Taking an average, Romney is down by a half-point at Pew:
The Monmouth University poll, after being processed through the Decoder, as seen here, shows Romney trailing by one point.
Unfortunately, I can't run a poll through the Decoder when it doesn't reveal the Romney v. Obama breakout in each of the three partisan categories. That's the case with the AP-GfK and NBC/Wall Street Journal polls currently listed at RCP, which sampled Dems over Republicans by 48-44 and 47-42, respectively.
The overall point? Numerically, after disinfecting Pew and adjusting Monmouth U, Obama is up by only +1.2 points:
It can't be quantified, but the other two polls, if they could be run through the Decoder, would in all likelihood render the result even tighter.
While the press pretends it isn't so, for all practical purposes it's a dead heat.
Cross-posted at BizzyBlog.com.