National polls, and Poblano's multiple regression analysis

After super tuesday, Clinton seemed to surge slightly in national polls, perhaps reflecting the surge of support she got in some of the larger states. Now that super tuesday is behind us, we see the lead closing again. Gallup has only a 2% gap between Clinton (46%) and Obama (44%): http://www.gallup.com/poll/104281/Gallup-Daily-Tracking-Election-2008.aspx

Rasmussen has this at 5%: http://rasmussenreports.com/public_content/politics/election_20082/daily_presidential_tracking_polling_history

Of course, what really matters are the state polls, and we have no data for important states such as Ohio or Texas, but Poblano came up with an imperfect, but pretty good regression model that shows a 10% Obama victory in Ohio, and a 10% Clinton victory in Texas.

See: http://www.dailykos.com/story/2008/2/9/13227/22519/239/453361