Don't you think the sample should be looked at when trying to decide what the poll is really saying? If they are just recording what they get then maybe their polling methodology introduces some sort of bias unintentionally.
If the national poll that had an 8% advantage for Democrats is correct then Romney is going to lose. Considering that 2008 had a 7 point lead I find that possibility to be absurdly low. Obama will get less turnout this time around and Republicans more for a whole host of reasons that I outlined earlier.
The problem is how do you adjust things unless you have a more accurate baseline to work from? Or to put it another way, adjusting for party identification requires you to know what percentage of the population identifies with which party from some source outside of (and more accurate than) your poll. Looking at previous years can provide somewhat of a gut check, but unless it highlights and obvious flaw in their methodology, it's hard to come up with something that can actually be changed. And to be perfectly honest, asking a lot of randomly chosen people what their party identification is intuitively seems like it would produce fairly accurate results unless people are lying.
And while the polling organizations should consider that their polls might be non-representative for some reason, you should also consider that party identification might not mean what you think it means. The "over-representation" theory says that the respective percent of Democrats and Republicans in the polls should closely track with how many votes each candidate is going to get. So if self-identified Democrats outnumber Republicans in a race expected to be close, there's something wrong with the sample. Except that's ignoring that a huge number of people identify as "independent" in the polls while the vast majority of the vote goes to either the Democrat or Republican candidate. There's nothing saying Romney supporters HAVE to identify as Republicans, and I wouldn't be surprised if the rise of the Tea Party leads many people who wouldn't vote for Obama no matter what to identify as independent.