Here are my thoughts on this election:
*My educated guess: this election won't be resolved tomorrow night.
*This election is going to come down to four states, really: Ohio, Florida, Wisconsin, and Iowa. Kerry wins if he gets two of the four (as long as he gets one of Ohio or Florida). Bush needs both Ohio and Florida or a combination of three out of the four to win.
*The polls from the other "swing" states--NH, MN, MO, CO, NV, NM, AR, HI, NJ, MI, and PA--seem to be firmly entrenched on one side or the other. I would be fairly surprised if any of them swing to the other side.
*Yesterday's "The Note" suggests that the early returns on the Iowa absentee ballots are heavily in favor of Kerry, whereas the likely turnout at the polls should be split about 50/50.
Here is the link to the article:
"Twenty-seven percent of Iowa adults surveyed said they had already voted. Kerry leads Bush, 52 percent to 41 percent, among that group of early-bird voters. Among the 73 percent who said they definitely would vote on Tuesday, Kerry and Bush are tied."
*The U.S. Appellate court ruling today was a big win for Kerry. Essentially, it prevents lawyers (or anyone else) from preventing and challenging others from voting at the polls.
Not that anybody cares, but my own $0.02 is that this move by the Ohio Republican party smells of modern-day, Jim Crow-racism, akin to poll taxes.
*Social scientists say that it is rare for incumbents to get more than they get in tracking polls in the week leading up the election (i.e., the polls seem to be an "upper bound" for the incumbents for lots of reasons, such as the public's familiarity with the person and knowing what he/she stands for). If that is true, then Bush should be worried. His "upper bound" in Ohio is ~48%, it is 47% in Florida, Wisconsin, Iowa, et al.
*I'm really skeptical of CNN/Gallup polls or any of the partisan polls. One lesson in how they *lie* with statistics: the sample that Gallup uses has consistenly been 5% to the right of actual 2000 exit polls. In other words, they are making a sampling error by not recalibrating the data to match historical trends, then repackaging it as concrete evidence. This is the same trick that partisan pollsters from both sides use to boost people's impressions of their candidate's chances.