That party breakdown is not hideously bad. Some firms think the breakdown is about 38-38-24.
Is the party ID by registration or self-identification? Self IDs will likely skew away from Dems for a while following the convention.
There are two firms that track Party ID.
Gallup says the GOP actually has a 0.3% advantage over their last 40,000 interviews.
PEW says the Dems have about a 4% lead.
The two firms ask a marginally different question.
PEW just recently made a small adjustment to their sampling methodology to account for changes in response rates in various telephone exchanges (caller ID type stuff) which should (I think) bring the PEW numbers back in line with the Gallup Numbers.
The "true" breakout of party ID is an unknown.
We could debate it till the cows come home, and while there is actually a consensus on how to deal with the issue in polling, it would take far, far too long to explain and type out
If you use the PEW +4 weight then Bush is up 6%
If you go with an even split Bush is up 11%
Take your pick