My operating principle: reversion to the mean
My dataset: every national poll from June 1 to today
My prediction: Obama wins Nov. 6 by 2.5 points over Romney
Granted, Romney has had a bad month. So what was the mean Obama's lead in polls taken prior to September? 2.3
August was Romney's best month so far. Mean Obama lead in polls taken that month: 2.1
So Romney at his best is 2.1 below Obama; Romney at his worse is 2.9 below Obama. I'm going to go with that as the floor and the ceiling and split the difference.
And things do not get much better for Romney if you look only at likely voters: among likely voters Obama has a mean lead of 2.7 points.
What should Romney hope for? Well, Obama's smallest lead was in August among likely voters, at 0.8. If the Democratic enthusiasm post-convention and post-47 percent speech is not maintained, then Romney is in much better shape.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment