Warren Mitofsky and Joe Lenski - the principals at the exit poll service - released a report almost half a year ago investigating the bias in the exit poll.  They uncovered numerous factors that were related to the within-precinct error of the exit poll (the actual result in the precinct versus the exit poll result), such as an interviewer effect with younger interviewer more likely to interview younger voters and a bias related to distance that interviewers had to stand from the polling place.  They called for better interviewer training and better relations with state election officials to bring interviewers closer to the polling place.

It is clear from the report that the unweighted exit poll cannot be taken seriously as an validity check against the actual election results.  If one truly believed the exit poll was an accurate measure of the voters, then an interesting fact would have to be explained: While roughly 50% of the precincts had a within-precinct error outside the margin of error of the exit poll favoring Kerry, 25% had a bias within the margin of error, and 25% had a bias outside the margin of error in favor of Bush.  In other words, if one truly believed that the exit poll was an accurate measure of the electorate, then in roughly 25% of the precincts election administrators stole many votes from Bush.  

Were there election administration errors?  Undoubtedly the answer to that question is yes.  The evidence is strong that the butterfly ballot cost Gore the 2000 Florida election and the presidency.  There is strong evidence that election administration errors continue, and they cut both ways.  For just one example, in checking over 2004 official election results for Mississippi, I and others discovered a reporting error that cost Bush 10,000 votes.  But do the exit polls provide evidence of a 7-8% point discrepancy in the 2000 Florida election?  That would represent hundreds of thousands of votes.  No serious academic or media analysis of the 2000 Florida election results - and there have been many - found evidence of a reporting error on this scale.  It is much more likely that survey methodology errors on the exit poll unrelated to simple sampling error are responsible for the discrepancy than some sort of grand conspiracy.  

Michael P. McDonald
Assistant Professor, George Mason University
Visiting Fellow, Brookings Institution
703-993-4191
mmcdon@gmu.edu
elections.gmu.edu

----- Original Message -----
From: "Steven F. Freeman" <stfreema@sas.upenn.edu>
Date: Friday, August 19, 2005 10:10 am
Subject: Re: Who won the 2000 Presidential Election? Again?

I don’t believe that any publications to date have captured the 
full magnitude
of just how far the official count deviated from the intent of the 
electoratein the Florida 2000 presidential election – recall that 
the exit polls
projected a Gore victory margin of 7-8%.

I closely analyzed that election while researching my current book 
on the
integrity of the 2004 US Presidential Election, and have completed 
a working
paper on it. Write to me if you would like a copy.

Steve Freeman
Scholar and Afiliated Faculty
Center for Organizational Dynamics
Graduate School of Arts and Sciences
University of Pennsylvania

best email: sf@alum.mit.edu


Quoting ethan.fletcher@yale.edu:

Professor Lipkin,

It seems to me that to answer this question we must first define it
more precisely.  If the question is: "For which candidate did more
voters (in the deciding state(s)) leave polling places and submit
absentee ballots believing they had voted?" then there is little 

doubt> the answer is Gore.  If the question is a different one--
did election

officials act appropriately? did courts act in good faith and 

fulfill> their proper institutional roles? etc.--then we've begun 
a much

larger, more complicated discussion.

The Miami Herald did a good piece looking at the issue 

relatively soon

after the election.  It ran, I believe, in December 2000 or January
2001.

Sincerely,

Ethan Fletcher
3L
Yale Law School

Quoting RJLipkin@aol.com:

In today's NY  Times, Paul Krugman insists that "the simple truth"
[is that] 'Al Gore won  the 2000 presidential election. Since
I'm unsure whether List members want to  reexamine this issue now,
I'd be happy to receive off-List answers. My question  is this:
Is Krugman's "simple truth" widely accepted by "election-law  

experts"?> >

Thanks.
Bobby

Robert Justin  Lipkin
Professor of Law
Widener University School of  Law
Delaware