As Mitofsky has stated, exit polls have always been weighted to match the actual vote. Now we know the reasons why. His statement, though literally true, is a canard. In fact, the vote count should be re-weighted to match the exit polls.
Mitofsky would have us believe that the vote counts are correct and therefore preliminary exit poll numbers must be re-weighted to match the actual vote. That's why his final, weighted exit polls match the votes - but they are wrong, because they assume that the vote counts are accurate, which is not true.
The re-weighted exit polls are "correct" only if one defines correct as exactly matching bogus vote counts.
Mitofsky never suggested that the vote counts could be anything but perfect. Yet we know that millions of ballots are spoiled in every election, especially in minority precincts. We learned about that from Greg Palast in Florida 2000.
We must assume that if the final exit polls are weighted to match bogus vote counts, then the preliminary exit polls (like the 2004 NEP of 13,047) must be very accurate indeed. In fact, Mitofsky/Edison claim that the preliminary 13,047 WP poll was accurate to within a 1% MOE.
Looking at the last five elections, the preliminary exit polls, which reflect the true intent of the voter, have been "re-weighted" to match bogus vote counts. And it's always the Democrats who lose the majority of the spoiled votes. As many as 4-5 million are estimated to be spoiled, of which probably more than 75% are Democratic votes.http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=341940
Let's assume that of 4 million spoiled ballots, 3 million are Democratic, 1 million Repub. That's 2 million net votes lost to the Dems, or 2% of 100 million votes. In a close election, it makes all the difference.
But we have not even considered BBV, the 21st century equivalent of a punch card undervote/overvote - but without a paper trail to recount. We can therefore conclude that, in the last 5 elections, the preliminary Democratic exit polls exceeded the actual votes by 2-4% - and were very close to the true intent of the voter.
The NEP/WP site displays the pristine National exit poll of 13,047 respondents. The demographic category weights indicate a 51-48% Kerry victory, confirmed by state exit polls which showed Kerry winning most of the battleground states, including Ohio and Florida.
The 2004 preliminary (13,074 respondents) PartyID mix of 38/35/27 matched the final exit poll (39/35/27) in 1996-2000. The fact that this was changed to 37/37/26 in the final (13,660) NEP poll is what we used to call finagling - or cooking the books- to match the Bush 51-48% tally. THAT IS A MAJOR SMOKING GUN.
BUT SMOKING GUN #1 IS STILL THE 21 MILLION NEW VOTERS WHICH KERRY WON BY 59-39%. THERE IS NO WAY IN HELL KERRY COULD HAVE LOST, SINCE HE STARTED WITH A MAJORITY BASE OF DEMOCRATIC VOTERS TO BEGIN WITH.
The argument the naysayers and even so-called Democratic polling bloggers have used, that the exit polls are not random samples and that the votes counts must be 100% correct, is logically and factually false, by definition. On the contrary, we know that the recorded votes do not include those lost to ballot spoilage or machine "glitches", 99% of which favored Bush.
We must therefore conclude that Dukakis, Clinton, Gore and Kerry all did better than the vote counts indicated, because millions of their votes were spoiled ballots - not to mention the unknown number stolen in cyberspace.
And finally, what about those who were disenfranchised by corrupt Secretaries of State, who never had the chance to vote and exit- polled?
So that's it.
The unexplained exit poll discrepancy - explained.