It's time to toss this poll right where it belongs... IN THE GARBAGE.
If it had Crist up you'd cry "DOMINATING!"
Anyway, this isn't a partisan issue, the poll is junk because it uses decimals, is weighted by party ID, shows an R+5 sample even though party ID was even in 2010, and showed Romney up 6 in 2012 the day before the Election. Is that really not enough for you?
This is false. The Romney+6 poll was conducted by Braun Research, while this poll was conducted by UF's Bob Graham Center.
To summarize Matt Florell's criticisms, it has a fairly small sample size, weighed by media market (rather unheard of for public polling), failed to weigh by gender and race. Additionally, the problem here isn't that it's weighted by party ID but rather that it chose a strange way of doing so. I wouldn't consider this accurate.