If the best Romney can manage is a tie, then he's not as strong as he's being made out to be. I don't see how the sample matters much, seeing as Obama's approval is underwater.
The sample matters a lot. This poll is too Democratic (ca. D+8) compared with historical turnout.
The current registration in Florida is 41% DEM, 36% GOP and 23% Others.
If the poll would have this sample (D+5), the results would be:
46-43 Romney
But usually Republicans have better turnout in FL than Democrats, like in 2004 and 2006. Only in 2008 it was D+3, in 2010 it was even. So it is very unlikely that Democrats will enjoy a 8% turnout advantage next year, when 2008 was one of the most favorable years for them.
But the approval still says a lot here, if there were too many Democrats skewing the sample, wouldn't that equal better approval? And even with an adjustment, the numbers don't shift that dramatically.
