Rating Pollsters

Updating this from Super Tuesday, here's how the Potomac Primary and last night look to me:

• Maryland 2/12 result: Obama 60%, Clinton 36.5%, margin 23.5

SurveyUSA, 2/9-2/10: Obama 55%, Clinton 32%, margin 23
Mason-Dixon, 2/7-2/8: Obama 53%, Clinton 35%, margin 18
Rasmussen, 2/6-2/7: Obama 57 %, Clinton 31 %, margin 26
Lowest error: SurveyUSA, .5%

• Virginia 2/12 result: Obama 63.6%, Clinton 35.4%, margin 28.2

SurveyUSA, 2/9-2/10: Obama 60%, Clinton 38%, margin 22
Mason-Dixon, 2/7-2/8: Obama 53%, Clinton 37%, margin 16
InsiderAdvantage, 2/7-2/7: Obama 52%, Clinton 37%, margin 15
Rasmussen, 2/6-2/7: Obama 55%, Clinton 37%, margin 18
Lowest error: SurveyUSA, 6.2%

• Wisconsin 2/19 result: Obama 58.7%, 41.2%, margin 17.5

Research2000, 2/13-2/14: Obama 47%, Clinton 42%, margin 5
Rasmussen, 2/13-2/13: Obama 47%, Clinton 43 %, margin 4
Lowest error: Rasmussen, 12.5%

No polling in Hawaii or D.C.

Some of this may be affected by timing of the polls as voter sentiment was moving toward Obama, and I wish SurveyUSA had polled Wisconsin to add a data point, but it still looks like they're the best. It also looks like undecideds broke for Obama heavily in all three races, much more so than the exit polls indicated. (The other possibility is that everyone got Clinton's numbers basically right but understated Obama's. I find that unlikely.)

Just for information, with 2 weeks left to go SurveyUSA has Clinton up by 5 in Texas and up 9 in Ohio. Undecideds are 5% in both states.

Also, it occurs to me that with the unions breaking for Obama, the odds of Edwards endorsing him have gone up. If I were Edwards and I planned on endorsing someone, I would do it on Monday.