Wednesday, October 29, 2008

Morning Polls

There are some signs for worry this morning (for those who enjoy this Democratic pasttime) as the national trackers do continue to narrow. The good news is Gallup has NOT narrowed today. The Likely Voter Model II (which is the model I trust the most) remains at Obama +7 and has Obama at 51%. The "traditional" model moves from Obama +2 to Obama +3. And their registered voters model moves from Obama +7 to Obama +9. For those seeking to worry, the Rasmussen tracker in particular is worrisome as it moves from Obama +5 to Obama +3 and it is a tracker I have a lot of respect for. The lead dropped by 3 the day before. The good news is that Obama is still at 50%. The Hotline poll moves from Obama +8 to Obama +7. R2000 moves from Obama +7 to Obama +6 but Obama is still at 50% in that tracker as well. One thing to remember about all this is that we have (hoped against but) expected the national horserace numbers to narrow a bit. Followers of Nate Silver's site, for instance, know that the popular vote percentage he tracks (and which is reported over to the right) is a "projected" popular vote number. His model builds in an expected 1-2 points of narrowing before Election Day. We may simply be seeing that take hold now.

On the other hand, the election is really determined at the state level and, even more specifically, in the battlegrounds. And it is hard to find real signs of a narrowing race in the state-level polling. I'm not making this up. Take a look:

Colorado (AP/GfK) - Obama +9
Florida (AP/GfK) - Obama +2
Florida (Quinnipiac) - Obama +2
Nevada (AP/GfK) - Obama +12
New Hampshire (AP/GfK) - Obama +18
North Carolina (AP/GfK) - Obama +2
Ohio (AP/GfK) - Obama +7
Ohio (Marist) - Obama +3
Ohio (Quinnipiac) - Obama +9
Pennsylvania (AP/GfK) - Obama +12
Pennsylvania (F&M) - Obama +13
Pennsylvania (Marist) - Obama +14
Pennsylvania (Quinnipiac) - Obama +12
Virginia (AP/GfK) - Obama +7
Washington (Strategic Vision) - Obama +12
Washington (SurveyUSA) - Obama +17

Let's start by reminding ourselves of the kind of needle that John McCain needs to thread. John Kerry won 252 electoral votes in 2004. Will Obama lose any of those states? The McCain campaign is arguing they think they can pull off upsets in Pennsylvania and New Hampshire. The polling above
suggests otherwise. Dipping into states Bush won in 2004, Obama has strong leads in both Iowa and New Mexico. That puts Obama at 264 electoral votes. That means Obama would then need to win any ONE state out of the group of Colorado, Florida, Georgia, Indiana, Missouri, North Carolina, Ohio, or Virginia. In addition, if Obama were to win in just Nevada, it would throw the election into the House of Representatives.

The polling above combined with polling from the last couple of days shows Obama with big leads in Colorado and Virginia and smaller leads in Nevada, Ohio, and Florida (Obama's lead in Florida appears to have narrowed by a couple of points but he still leads) and probably about tied in Indiana, Missouri, and North Carolina. Again, Obama needs just ONE of these states IF he holds onto Pennsylvania.

The various pollsters above are a mixed bag in my view. I don't think too much of the AP/GfK poll but Quinnipiac is very good. Regardless, we're not seeing a lot of variation. With the exception of Ohio, we don't really see any disagreement at all and even the worst reading in Ohio has Obama ahead. SurveyUSA (the gold standard in state polling to me) had Obama up by 4 in Ohio yesterday.

One negative point to make: Obama is visiting Iowa on Friday. I have no idea why he's doing this. He did cancel an Iowa event a week ago when he went to visit his ailing grandmother so maybe this is a makeup event but I'd say we should be doing makeups after the election. Maybe Obama's internal polls are closer than the publicly-released polls? I don't know but I don't think this is good news either way. On the other hand, any day when you see a candidate's pollster on television all over the place trying to explain why the race is "functionally tied," that's not a good sign for that candidate. Bill McInturff (McCain's pollster) is a good pollster but the logic he's using to explain why the race is "functionally tied" is worthy of a master contortionist (more on that in a post to come momentarily).

Bottom line: Is the race narrowing at the national level? It appears so. Is McCain closing the gap in the critical states he needs? I see almost no evidence of that.

6 days to go: Breathe in, breathe out.

No comments: