Reply #41: Some comments [View All]

[papau]

"AND IF THIS WAS DONE WITH A VERIFIABLE DATA SOURCE (THE 2000 VOTE), THEN WE HAVE TO ASSUME THAT OTHER ARBITRARY ADJUSTMENTS HAD TO HAVE BEEN DONE TO ALL THE OTHER UNVERIFIABLE DEMOGRAPHIC WEIGHTS, AS WELL."

We are on the same page here, except I do not see it as a given. However,the above was my conclusion also after my smell test of the adjustments. There is a real "say you voted for the winner" effect - so the percentage of folks claiming in 04 that they voted for Bush in 00 can be expected to be higher than the actual percentage of the vote that Bush got in 00. I'd expect the "error" - just not as large as the FINAL exit poll says it was

"1.75 MILLION BUSH 2000 VOTERS DIED" - the 50 million Bush voters, using 1994 GAR Proj. 2002 mortality, would if average age 50 have a .99122 chance of living 4 years, meaning about 500,000 Bush voters in 2000 would have died by 2004. It does not change your logic - but I'd revise the statement to "450,000 to 500,000 Bush voters were expected to die". No need to go beyond the data you have - which I agree implies that the race was won by Kerry - and that the exit poll adjustments are out of the range of reasonableness.