Electoral Math 10/28/12

Electoral Math Projection:
Obama 277, Romney 206
No Toss-ups:
Obama 281, Romney 257
Minnesota changes from leaning to barely Obama today after a Mason-Dixon poll was released this morning showing Pres. Obama ahead by only 3. This is down from an 8 point Obama lead in the same poll 6 weeks ago. Overall my projection drops to a 5.5% advantage.
Based on my belief that we will not get enough post-debate polling to fully wash out older polls, I’ve looked at attempting to add the effect of the change in national polls to older data. This allows me to project out the state of the race in states today where polling data is sparse. It’s not a particularly elegant method to project the change in the electorate, but it’s as effective a method as I can quantify.
I haven’t yet updated the map to reflect this projection, however based on this effect, my popular vote projection stands at Obama 49.83%, Romney 48.73%. Even with this accounted for it still diverges fairly significantly from national polling data. What’s the difference? Primarily, it’s an issue of projection. Most national polls are limited (even daily trackers) to a couple thousand likely voters. Those voters are then extrapolated across the country using demographic information. While state polls do the same thing, there is less margin of error at the state level.
I tend to think that this projection is closer to the actual state of the electorate than my state poll only projection. We’ll find out soon enough.