We can reconcile this with the actual data. Two questions that need to be answered:
1. What percentage of their registered sampling has early voted, versus the percentage of actual registered voters that have actually voted?
Answer: According to GMU, it looks like only 3% of Ohio’s 7.8M voters have actually voted. So why are they 20% of the RV sample in PPP?
2. How does the registration of those sampled early voters compare with actual early voter records?
Answer: Assuming all Ds vote for Obama, and assuming Is make the difference, PPP’s sampling of early voters shows that Obama has captured every Democrat, every Independent, and a good portion of Republican.
The specious early voting claims also distort the likely voter screen.
This discussion of flawed methodology of course assumes that they are even polling anyone at all, and not just making up numbers. Not accusing them, but what would keep them - or any pollster - from doing something like that? I seem to recall that Daily Kos had a pollster who got making up results that would always show the Democrat with a ridiculous lead.