IOWA: Obama leads, but Romney up big with indies

A Gravis Marketing survey shows Barack Obama leading Mitt Romney in Iowa, 50%-46%.

The big caveat is that Romney is leading, 48%-36%, with independents, while both candidates pick up +90% of their respective parties.

That would lead you to correctly assume that more Democrats were sampled than Republicans.

There's nothing wrong with that in, say, California, but this sample assumes turnout that just isn't realistic.

In 2008, Dems had a +1% turnout advantage on Republicans in Iowa. This poll suggests a +6% turnout advantage. There's nothing to suggest anything close to that increase.

If you run with 2008 turnout numbers, Romney would win solidly based on this poll. To wit: In 2008, Obama won indies by 15%, but Romney leads by 12% in this survey.

If Republican and Democratic turnout is even in '12 (like it was in '08), then independents are the tiebreaker, and Romney leads by 12% with the group.

I get that voters are self-identifying, but it doesn't pass the smell test (or really any other) that Dems' turnout advantage would be 5x what it was in 2008.