Exit Polls: Question re Bush Reluctant Responder Hypothesis (BRRH)

[Bill Bored]

I have a question about USContVotes' claim that Mitofsky's BRRH is false:

USCV says that because Bush-heavy precincts participated more in the exit polls than other precincts, Bush voters were NOT reluctant exit poll responders. Fair enough, but if they are also saying that the vote totals are questionable or false, how do they know which precincts are truly Bush-heavy and which were not? Are they using the same vote totals they are questioning to make this determination, or are they using exit poll data for these precincts, under the assumption that the vote totals are false, or are they restricting their analysis to precincts in which the polls matched the vote totals, or ???