This is the same pollster who did Portait of America polls back in 2000. Sorry - not buying. He was way off back then. Had us thinking W landslide.
True, but he was much closer in 2004, after he changed his methodology to weigh his results by party affiliation. To be sure, most pollsters pretty much nailed the 2004 election as a 2-3 point Bush win, so I don't know how much stock to put into this result. Likewise, I don't know if that was dumb luck or if he improved his robot phone methodololgy that he uses. I don't remember how good his state-by-state polling was, either.
Again, in the current poll I very strongly suspect here is that the problem is a huge outlier day for Obama on Tuesday. It drops off tomorrow.