So is time to "unskew" every poll now (even PPP)?
Raising legitimate questions about sample composition and methodology are not unskewing, especially when no one is trying to say what the poll actually shows. I'm sure that Clinton was +2 with the sample they got, they just got a sample that looks almost nothing like the likely 2016 electorate.
Pretty much this. If you want to poll the 2016 electorate and have a properly-represented estimation, then you have to do it with 2016 and not 2012 numbers. Laughing at or antagonizing other members for not taking the toplines at face-value and questioning the methodology and procedure just underscores the limited understanding that poster may have of scientific sampling methods (@jfern).