FOX National: Obama leads by 7

[AmericanNation]

For a reference point, '04 turnout was 37/37/26%

This is not 2004.

It's worth remembering that 2004 was the best presidential election of the last 24 years for the Republicans. So while Democrats may be a little oversampled here, it would be pretty silly to re-weight the sample so that Ds and Rs were at parity.

If I had to say, I'd say:
1) Republicans are more motivated than 2004
2) Romney is better with independents/moderates than Bush
3) Romney is a better candidate outside of the Bush states than Bush was.   

[AmericanNation]

For a reference point, '04 turnout was 37/37/26%

This is not 2004.

It's worth remembering that 2004 was the best presidential election of the last 24 years for the Republicans. So while Democrats may be a little oversampled here, it would be pretty silly to re-weight the sample so that Ds and Rs were at parity.

If I had to say, I'd say:
1) Republicans are more motivated than 2004
2) Romney is better with independents/moderates than Bush
3) Romney is a better candidate outside of the Bush states than Bush was.   

#3 is still false. Romney is no better in the blue states (perhaps except Massachusetts and Michigan) than Bush was.

You mean it's false now?  Yea, the challenger of an incumbent typically has to do some work over more than a month long time frame. 

Romney is probably better than Bush (given that a campaign goes on more than a month) in Pennsylvania, Ohio, Wisconsin, Michigan, and New Hampshire.       

[AmericanNation]

back to the topic,

This poll is of Registered Voters... So, of the sample that will actually vote the margin is probably 1 to 3 points. 

[AmericanNation]

back to the topic,

This poll is of Registered Voters... So, of the sample that will actually vote the margin is probably 1 to 3 points. 

It doesn't work like that.

Sometimes it does work something like that.  How do you think it works?