The partisan ID of this sample is 30.3% Dem, 30.2% Rep, and 26.4% Indy (even). The 2008 Tennessee turnout was 33% Rep, 32% Dem, and 35% Ind (R+1). That means Vanderbilt is banking on a NET larger Democratic turnout in 2012 than in 2008. If you're one of those that believes 2012 turnout will not be as favorable to Obama as it was 4 yrs ago, then this poll is over sampling Democrats a bit. But considering McCain carried this state 57-42% four yrs ago, Romney does appear to be under-performing in this Southern/Appalachian state.
They are not banking on it, because it's a poll of adults. And in TN there are a lot more Democrats registered than Republicans. Which makes the sample of adults more or less correct. Likely voter samples might be a different story though.
Regardless of the number of actual registered voters in TN, the poll sample is 30/30/26 (even), while actual 2008 TN turnout was 32/33/35 (R+1). So yes, it appears that for now, Vanderbilt is banking on a larger Dem turnout than in '08. If they weren't, then why wouldn't the partisan ID of the sample reflect it?
While I certainly do not agree with this poll, you really can't say the pollster is saying turnout will be more Democratic, when we are talking about a 1% difference here
