Posted: 11/01/12 10:58
by Dave Mindeman
As I have stated before, I believe in Nate Silver and his computer model for election prediction. When the polls are all over the map, it is good to have an aggregate analysis that utilizes all of the data.
I also don't believe in dismissing any polls. I think it is unwise to just cherry pick the ones that tell you what you want to hear and say that the others are just plain wrong. I do look at the data that provides the basis and I sometimes question the criteria, but most polling institutions are not going to put themselves in the position of being completely discredited.
Nate Silver has been increasing the chances of an Obama victory just about every day over the last few weeks. In fact, yesterday's post gave Obama over 300 electoral votes. That has put Silver in the sights of conservative blowback and a number have now dismissed him as just another ideologue with an agenda.
Even I have started to wonder if Silver is a bit too optimistic about an Obama win, but I have also been looking at another analytic site called Election Projection.com. This is operated by a right wing social conservative, but he has a solid track record of election predictions and analysis. He also has his own model which aggregates the polls and factors in economic and historical data.
Over the same period as Silver's projections, his electoral count has also favored Obama. In fact, he currrently gives Obama 290 electoral votes, even though he gives Romney an edge in the popular vote.
His blog posts have in the recent past have postulated that the national edge for Romney will eventually translate into the swing state polls....giving Romney the electoral edge as well.
However, today, his view changed....
According to the polls, Barack Obama is very likely to win a second term next Tuesday. I had hoped, indeed predicted, that the battleground state polls which have stubbornly shown him ahead would correct their peculiar partisan sampling and begin to paint a different picture. Well, time is running out for that to happen. And with the last round of state polls from CBS/NYT/Quinnipiac continuing to show Obama ahead in Florida, Ohio and Virginia, I have to admit that the polls don't look good for Mitt Romney.
It would seem that he and Silver have reached similar conclusions.
Both of them say that the next few days still could change the mathematics. We have a jobs report tomorrow and the Sandy situation could blow up. But it still is interesting that the mathematical models still point to an Obama win.