I hate to pour water on the obvious joy that Marc gets in hating John Zogby, but this poll really isn't that far off base. Look at this AP analysis of exit poll results as it appeared in the Guardian:
TIME OF DECISION: Almost six in 10 voters decided in the last week, including about two in 10 who decided in the last three days and almost that many who decided Tuesday. Edwards led by a 2-to-1 margin among those who decided in the last three days. Kerry led by 30 points among those who decided in the last month or earlier. Edwards got major newspaper endorsements in the closing days of the campaign and performed well in a debate Sunday night. More than six in 10 independents decided in the last week, and Edwards got half of them.
The Zogby poll was taken on Friday and, as the exit poll results show, Edwards made great strides by Tuesday. By looking at this poll, the results and the exit polls, we see that Edwards was able, in the waning moments of the contest, to take votes away from Kerry in order to secure his suprising second place finish. Further, Zogby's Dean number is only about five points off the final result, and when you factor in margin of error, that's not bad.
Polls should never be designed nor interpreted as if they will predict the winner and the margin, although sometimes they can. They are a snapshot of the race at the time the poll was taken, with a shelf-life of, at max, a couple of days. I think Zogby's analysis (as if his numbers preclude the need for an actual vote) is wrongheaded. But, I don't see this as an example that Zogby's a fraud.
