The Rendell polling was undoubtedly using the exact same samples as the Casey polling above. I don't see your point.
There wasn't just one. I know of at least three polls that showed Rendell with double-digit leads, two of which put him near +20, and there was that SUSA approval poll that gave him over 60% approval. All of the polls were done by different companies.
I understand this and agree with it, but I don't think you're getting my point.
When we have three surveys that show Rendell up double digits and two surveys which show Casey up double digits, we do not have five "different" polls.
The Quinnipiac and Rasmussen samples were undoubtedly both questioned on both the Casey and Rendell races simultaneously, meaning both were subject to any problems with the samples should there be any. Same with the Franklin and Marshall survey.
So, only two polls have shown Casey up by double digits this month and three have shown Rendell up by double digits, two of those being the same that showed Casey up by double digits.