But it's also worth pointing out that the IBD/TIPP poll shows just a two point lead for Barack. Note also that TIPP was the most accurate pollster of the 2004 election season.
How can you account for the disparities in polls? As Todd Eberly explains over at National Review, two important factors include weighting for party identification, and the treatment of "leaners." The results you get out depend on the numbers you put in.
Which poll(s) are most accurate? It's impossible to know until after election day. But like Ronald Reagan, I don't believe in a fate that falls on us no matter what we do. I do believe in a fate that falls on us if we do nothing.
Allowing worrisome poll results to dampen our enthusiasm or our turnout would make predictions of an Obama landslide a self-fulfilling prophecy. It's neither inevitable nor, to put it mildly, desirable.