OK, this is related - so I'm just going to crosspost this sequence of posts from the Polls etc thread. Plus, that'll line the maps up nicely underneath each other. (I did change/add some wording here and there.)
OK, so both pollster.com and fivethirtyeight.com not only keep track of all the polls that come out for every state, but also compute a trendline or projection on the basis of those polls.
I wanted to draw the current electoral map on the basis of the data on those two sites. Specifically, to draw the map according to different thresholds for what to consider a toss-up state.
If nothing else, it's a good opportunity to practice the Google Charts thing Craven presented
on a more advanced scale. (I've used the web tools he linked in to make simple pie and bar charts, but want to now try out following the instructions in the Google Chart developer's guide
on making maps.
OK. So coloured in on the following map are all the states where both pollster.com and fivethirtyeight.com currently give one of the two parties an advantage of at least 10%. These are, like, rock-solid safe states (note that Iowa, which went for Bush in '04, is already in there for the Dems!):
That's a very red looking map, but in fact it's 160 Electoral Votes for the Dems and 128 for the Republicans. You need 270 to win, of course.
OK, let's tighten it up a notch. States that both sites have one or the other candidate ahead in by at least 7%:
North Dakota, Georgia and Louisiana make it into the Republican camp; Oregon and Maine onto the Democratic side. That's 171 Electoral Votes for the Dems and 155 for the Republicans.
Both still far removed from 270, then. But all the states not colored in now are arguably in play: what's a 7% gap? One candidate needs to convince 3,5% of the voters to switch parties.
Tighten it up another notch: states where both sites have one of the candidates ahead by at least 5%:
That puts Washington and New Jersey in Obama's column, making for 197 Electoral Votes for the Dems and still 155 for the Republicans.
Another notch to states where both sites have one of the candidates ahead by at least 3%:
That puts Michigan, Minnesota and Wisconsin in Obama's column and West-Virginia in McCain's, making for 239 Electoral Votes for the Dems and 160 for the Republicans.
Still 31 short for Obama, and 110 short for McCain. Looking ever better for Obama.
Are you feeling the suspense?
States where both sites have one of the candidates ahead by at least 2%:
That puts Pennsylvania and Colorado in Obama's column and Montana, Missouri and North-Carolina in McCain's, making for 269 Electoral Votes for the Dems and 189 for the Republicans.
This of course would put Obama on the edge of being elected -- except that 269 Electoral Votes makes for a very tortured post-elections process
So basically, what Obama's painless election comes down to is getting at least one of the states left undivided on this map as well.
And there's the rub.
Which of the remaining states are currently most promising?
In Florida and Indiana, both pollster.com and fivethirtyeight.com give McCain the advantage, if with a narrow advantage. In IN, McCain leads by 2,5% according to pollster and 1,7% according to fivethirtyeight. In Florida, he's given the edge by 3,3% and 0,6% respectively.
So that leaves New Hampshire, Ohio, Virginia and Nevada. In all these states, fivethirtyeight.com currently accords the advantage to Obama. But pollster.com disagrees: it currently gives the edge to McCain in every one of them.
On balance, McCain gets Ohio and Nevada, and Obama gets Virginia and New Hampshire:
VA Obama +2,4 McCain +1,0
NH Obama +1,8 McCain +1,2
NV Obama +0,2 McCain +1,5
OH Obama +0,3 McCain +2,9
Virginia actually looks better than NH right now. The latest two polls out of NH both have McCain with a 2-3 point lead there, and one of them is the locally respected UNH poll. The four last polls from Virginia on the other hand have split, two with Obama comfortably ahead (by 6 and 8 points) and two with McCain narrowly in front (by 2 points). The pollsters showing Obama ahead also have a little more entrenched a reputation (ABC/WaPo and Survey USA) than the two with McCain ahead (Rasmussen and ARG). But obviously, it's still pretty much a coin's toss.
And of course, if the state that's Obama's best chance of getting 269+1 is down to a coin's toss, that is starting to make that other thread about the complicated consequences of a 269-269 result in the Electoral College
- [nimh: i.e., this one] uncomfortably topical.