I don't put any more credence in particular polls than others except for, perhaps, the National Elections Studies (NES) poll conducted out of the University of Michigan. The NES poll is purely scientific, however, and it is meant to explain electoral outcomes ex post facto rather than predict them. As long as the sampling methodology and survey design follow accepted scientific standards, most "major" polls should yield reasonably accurate results. I think you'll find that aggregating (i.e. averaging) the results of major polls provides a more accurate picture of electoral goings-on.
In that sense, Gallup has been a bit odd since the primaries (an observation Zogby has noted recently) since it consistently provides poll results that differ rather substantially from the mean. That's not to say that Gallup is inaccurate, but it does suggest that they might employ a sampling methodology that differs a bit from most of the other major polls.
If you average poll results, I think the difference between Bush and Kerry before the debate was about 2 to 3% in favor of Bush. I have no idea how that stands now since only a few polling organizations have released comprehensive polling data since the debate.
I personally see this race as Bush's to win or lose. National security is his strong point, and now it has become a central theme in the campaign. In addition, he sits on an economy that has slowly but consistently grown in the last year. However, public perception of economic improvement often lags by as much as three to six months, so any sudden improvements now will probably not help the president.
Another point to keep in mind is that voters tend to evaluate incumbent presidents retrospectively and challengers prospectively. So essentially, Bush must run on his record on national security, the economy, and more. If he can convince voters that he did an acceptable job on key election issues, he'll probably win. Expect him to put a good spin on his record but not waste much time on talking about what new things he will do. On the other hand, Kerry must both discredit Bush's record and demonstrate that he could do better in the future than Bush has done in the past on key election issues. It is not enough for him to demonstrate that Bush has failed in Iraq or on the economy; he must also convince voters that he would do as well or better.
I honestly can't predict an outcome at this point. Bush should hold most of the cards though, so I would say that his probability of winning is greater at this point than Kerry's.
-- Todd/Leviathn
