Delivered-To: dfarber+@xxxxxxxxxxxxxxxxxx
Date: Tue, 16 Sep 2003 14:13:54 -0700
From: "Robert M. McClure" <rmm@xxxxxxxxxx>
Subject: Re: [IP] CalTech/MIT Study on Voting Technologies
X-Sender: rmm@xxxxxxxxxx
To: dave@xxxxxxxxxx
At 03:29 PM 9/16/03 -0400, you wrote:
Delivered-To: dfarber+@xxxxxxxxxxxxxxxxxx
Date: Tue, 16 Sep 2003 12:45:50 -0400
From: "Ronald L. Rivest" <rivest@xxxxxxx>
To: dave@xxxxxxxxxx
Just for the record, the CalTech/MIT report
on voting technologies is posted at:
http://www.vote.caltech.edu/Reports/
Cheers,
Ron Rivest
I am familiar with this report, which is well done. It does,
however, raise more questions than it answers. Just why,
for instance should punch card machine miss more votes
than paper ballots, and why should electronic machines
and lever machines miss still more. If one were to take
this report as gospel, then we clearly would not be moving
in the direction of electronic voting machines. More
important, perhaps, is why the undervote varies quite
dramatically from location to location.
My original question is still unanswered. So the next
question is "Why is not someone running a controlled
experiment?" It should not be too hard to set one up,
and there are plenty of machines which are rarely used
with which to experiment.
It seems to me that the focus on missing a few percent
of the vote (by any technology) is swamped by the 50%
of the people who don't vote anyway. Unless we have
some reason to believe that some particular component
of the electorate is being systematically disenfranchised
then 2-5% missed votes is minor noise compared to the
much larger number who don't bother to go to the polls
at all.
I am not trying to ignore the questions of fraud and
security, just trying to get at one issue at a time. This
came up because of the recent 9th Circuit Court
decision essentially claiming that punch card systems
were inferior to electronic voting systems, a proposition
for which I don't find any real support.
Anybody game for defining a controlled experiment to
determine which of the systems currently in use are
defective and in which ways?
Bob