Why, it's easy! Just rig the polling sample, that's all! Then put out bold-font headlines about what a roll the president is on, let all the news services pick it up and run with it, and wait for the actual voters to fall into line behind the narrative.
Seriously. I'm not making this up. Want to know how Obama took a 51% to 44% lead over Mitt Romney in the most recent Washington Post/ABC poll? Hot Air:
In 2010′s midterms, CNN exit polls showed a 35/35/30 split...Today’s [poll] has a D/R/I of 34/23/34, adding seven points to that Democratic advantage and presenting a completely unrepresentative, absurd model for the 2012 turnout.
What happens when you switch from a D+4 to a D+11 in measuring Obama’s standing? Suddenly, his job approval goes from 46% to 50% — actually, a rather weak gain given the sampling distortion in the poll. Not coincidentally, the last time Obama hit 50% in this poll was in February, which also had a D+11 sample, after January’s D+7. Adding seven points to the Democratic advantage impacted Obama’s performance in all areas, although perhaps not as much as the editors had hoped:
Economy — Up six points from 38% to 44%
International affairs — Down two points from January’s poll, 47% from 49%
Right/wrong direction – Up three points from January
Now, with a D+11 and Republicans only comprising 23% of the respondents in this poll, one should be seeing huge leads for Obama in the head-to-head matchups. Instead, Obama lead Romney by only eight among general-population adults, 51/43, barely getting into majority territory...
Poor bastard will never know what hits him come November....
UPDATE 5PM: Apparently, to the Left, pointing out the statistical anomalies in this poll is racist. No, seriously...