Mike specifically mentioned that the “Big Lie” figures were just based on the off-site manipulation of the vote totals from the packet logs. It did not include any manipulation programmed into the machines. The off-site manipulation was necessary because the Trump voters “broke” the algorithm built into the machines. That is why there had to be an injection of Biden votes after the counting pause.
Dr. Framk’s understanding of correlation is the same as any other statistician’s understanding: You run the correlation algorithm and see what you get.
He claimed that correlation of >0.95 was unusal and suspicious. It isn't. Nobody can run it on his data because he won't release it.
However I did build a spreadsheet on Ohio counties using the 1984 Presidential election vs. the 1980 Census. and got a correlation of 0.9993874795
This predates any possibility of internet hookup and most electronic voting machinery.