Bush will take OH and hold FL. Rassmussen (which was spot-on in predicting the 2002 elections and which uses the best formulas) shows a steady Bush gain week-over-week.
FL won’t be close. Bush will hit at least the magical 50% number by next week’s polls. Almost every poll with consistent and non-biased sampling has had Bush holding a steady lead in both states.
The NBC and AP polls use suspect methodologies. And any poll that won’t publish its internals isn’t worth considering. Also, there’s a major difference in sampling bias in the regular AP poll versus the CNN/USA Today/Gallup poll. Simply put, the networks are interested in making it appear as if the race is closer than it in fact really is for various reasons. Advertising, viewer interest etc. I won’t address the perception of bias towards any one candidate other than to say that both campaigns feel predjudiced against.
In averaging of all polls over the last 3 months, Bush has maintained a consistent 3-8 point spread over Kerry. Whereas Bush has ranged anywhere from 46 to 51% support, Kerry has never managed greater than 47% support in his best week, and is averaging less than 45% right now, even with the Nader factor.
That, plus the fact that far more Kerry supporters think that Bush will win on November 2nd (20+ percent) and that Bush is doing better than expected on the black vote (8% in '00, and averaging as high as 18%…which will most likely fall to 13%, but is still a significant increase) could serve to depress voter turnout for Kerry if Bush gets to the 50% support mark, as has already happened in a few of the major polls.
Kerry also has a structural disadvantage, “Red” state versus “Blue” state, in the Electoral College. Bush’s “Red” states have gained an additional 5 EC votes since '00 due to population shifts noted in the '00 Census. Kerry absolutely has to have either OH or FL to have a shot. And if he loses MN (which has been trending demographically to the Republicans over the last decade and which is showing as a dead heat, right now) and loses WI, which is showing Bush up by 4 points, Kerry’s dead in the water.
Also, the candidates internal polls (which they base all campaigning decisions on), must be telling them something, else why would Bush spend time in New Jersey and other Kerry “Blue” states, and Kerry not venture outside of safe “Red” states and pull advertising and support from AZ, MO and a few others? The internals are forcing Kerry to stay on the defensive, mainly, and try to keep his flock from straying. Keep an eye on where candidates spend ad money and pour workers into or pull ad dollars and pull workers out of. That’ll give you the best indicator of what each camp thinks of their chances in a particular state.
Lastly, WI went for Reagan in 1980 and 1984, and the exurban migration of populations away from the larger cities show a definite Republican tilt. Same goes for MN. All of this forces Kerry to waste time and money in states that he could once take for granted, leaving less time to work in OH and FL. Plus, he’s had to have Bill Clinton go to formerly-safe PA, and Philadelphia in particular, in an effort aimed at keeping PA in the “Blue” column (some polls show a dead heat or Kerry only up by single digits ranging down to 1-4 points).
FL and OH, boys. It’ll be a long night for Shrub if he can’t hold 'em. But I don’t think that’s going to be a problem. My prediction still stands: Bush 53% of the vote, Kerry 46% and Nader et al 1% 
Bush 275 or more electoral votes, Kerry 265 or less.
