I am not good at providing links and such. In a google search it says that there is a business week article that states that of the voters who have voted in Ohio this year, that 45% of the registered democrats had voted. One would expect that virtually all democrats would have voted for Obama. If approx a third of all who will vote have voted early in and almost half the democrats have voted, it would seem that this bodes very well for Romney. ??
Hmm, tricky question - too many unknowns.
But let us assume that D makes up 35% of the voters, then 45% x 35% is about 16% which is 50% of the early voters.
If we then assume that I makes up 30% of the votes, and only a third of them has voted early and furthermore that they vote 55 - 45 for Romney, then they will make up 30% x 45%x 33% of the D vote in the early sample = 14%
D: 50% + 14 = 64% which is pretty close to what RAS got in the sample of the early voters.
Of course looking at the whole picture with a GOP 35% and the I’s splitting 55-45 in favour of R it looks good!