I was just pointed to the excellent Campaign Desk, analyzing campaign coverage, and this entry in particular, a summary and praise of a very good election prediction article with facts and good stats to back it up. A good conclusion reads:

Unlike so many political writers, Cook doesn't try to force-feed a prediction upon his readers. Rather, he takes them by the hand and guides them, fact by incontrovertible fact, through the probabilities at work in any election, and the known facts at work in this election -- and at the end of the trail, he leaves them to make their own way home, informed by real information, not by bluster or one-sided passion or crystal ball readings.

Mark Morford has brought us another great and timely rant, Who The Hell Is "Undecided"? / And why do so many election polls leave you angry and stupefied and drunk? He talks about how ridiculous it is that Kerry was ahead after the DNC and suddenly Bush is ahead after the RNC, and just barely hints at the manipulation going on, that of course the candidates are neck and neck because that gets people to watch TV and buy newspapers.

He also links to an interesting story about a poll that says most people in other countries, by a 5 to 1 margin, want Bush out. Of course, yeah, I believe that, but I still think that poll is as useless as all the others, and even more useless is the way it was reported. Never did they mention the list of 35 countries that were surveyed. 35 is not a lot, it would have been easy to list them. Never did they question that maybe 35 thousand is not a sufficient sampling of 6 billion. Oh well. Again, I'm not saying I disagree, I of course believe with all my heart and head that the people of the world for the most part hate Bush with a burning passion. It's just disturbing how statistics are spun.