This is an analysis of the Gender Demographic, comparing the National Exit Poll (13047 respondents) vs. weighted State Exit polls (73,000 respondents).
As you can see below, the numbers match very closely. The weighted state poll voting split (Male 46.35%/ Female 53.65%) is within 0.50% of the National Exit Poll (46/54). Since the 46/54 is rounded, it's probably
The 53.65/46.35 gender split is within 0.14% of the 53.51/46.49 census survey of 60,000 voters. The survey said 125.7mm voted, but only 122.3mm votes were recorded. The gender demographic has a 0.30% margin of error.
What happened? Where did those 3.4 million votes go? Lost in cyberspace? Spoiled punch cards? In the trash?
According to the National Exit Poll, Kerry won 50.78% of the vote. If we force the state horizontal vote percentage sum to 100% by allocating the small differences to Kerry (see below), he wins the weighted exit
polls with 51.01%.
If the discrepancies are not allocated, he wins with 50.53%.
Finally, if we allocate the discrepancies equally to Bush and Kerry, he ends up with 50.77% - a .01% deviation from his 50.78% in the National Exit Poll. Let's assume this was the case.
THE DIFFERENCE BETWEEN THE NATIONAL AND STATE EXIT POLLS IS ONE-HUNDRETH OF A PERCENT. HOW CLOSE CAN YOU GET? HOW MUCH CONFIRMATION DO YOU NEED? HOW CAN ALL THESE POLLS (13047 + 73000 RESPONDENTS) BE WRONG?
FOR THE UPTEENTH TIME, THEY ARE NOT WRONG. THE FINAL EXIT POLL (13660) RELEASED AT 1:25PM WAS WRONG - BECAUSE IT WAS MATCHED TO A CORRUPTED VOTE.
The 12:22am National exit poll (13047 respondents) had a 1.0% margin of error. Individual state exit polls have MoE's which vary from 2 to 4%, depending on the sample-size. If you want to consider the state
exit polls as one BIG poll of 73,000 respondents, then the combined MoE is 0.37%.
Go ahead, naysayers. Apply your 50% "design effect" and get the MoE up to 0.55%. The problem is that Kerry won the national and state exit polls by 50.8-48.2%, a 2.6% margin, five times the full 51-state
MoE (73,000 respondents) and almost three times the National MoE (13047 respondents).
Assuming a 0.70% MoE, the probability that Kerry would deviate from 50.78% in the polls to 48.28% in the vote (a 2.5% move), is the following:
1-NORMDIST(0.5078, 0.4828, 0.007/1.96,TRUE) = 1.28808E-12
In other words, 1 in 776,348,841,126. One in 776 Billion. No amount of reluctant Bush responders (rBr) or Gore voter "false recall" (fGr ) will get you there.
Of course, the naysayers on that other forum want to stop talking about exit polls. I wonder why. Well, too bad, we won't. The exit polls are gifts that just keep on giving.
By The Numbers