[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
CalTech/MIT: Paper Ballots More Reliable Than DRE
On Feb 11, 2004, at 4:40 PM, Nicholas Bernstein wrote:
The idea that a paper ballot marked and read by a computer is somehow
more reliable than an electronic one (that doesn't have to go through
this conversion) is ridiculous. I'm more than happy to talk about the
difficulties in signal processing and the costs of technology if
anyone's interested.
I believe the CalTech/MIT Study (2001) finds paper ballots to be more
reliable than DREs:
"Our measure of reliability is the fraction of total ballots cast for
which no presidential preference was counted. We call this the
'residual vote.'"
See page 10 of:
<http://www.hss.caltech.edu/ ~voting/CalTech_MIT_Report_Version2.pdf>
According to Linda Franz's summary of the results:
From that report: Residual Votes as a Percent of all Ballots Cast,
1988-2000
(Residual ballots are overvotes and undervotes)
There will be three figures:
President, Governor & Senator, and the average of the two
Paper Ballot 1.8%, 3.3% - 2.55%
Punch Card 2.5%, 4.7% - 3.60%
Optical Scan 1.5%, 3.5% - 2.50%
Lever Machine 1.5%, 7.6% - 4.55%
Electronic 2.3%, 5.9% - 4.10%
DRE)
While they admit that Electronic Voting had the most potential for
improvement at the time (although already after years of field use),
they cite several serious concerns in their explanation for why DREs
may always remain less reliable than paper ballots, including human
factors (some people just don't understand computers, apparently, as
crazy as that sounds to most of us).
It's 18 pages and I've only skimmed it, but it is certainly a powerful
rebuttal to the idea that new computer systems are inherently better
than traditional methods.
Linda Franz also asked a good question of the academics, including Dr.
Jones:
"Does a 'statistically significant' percentage sample of the ballots
need to be greater than the percentage of residual votes?"
To which I'm curious to find out the answer.
Joe