Brit,
Thanks for the comments. It occurs to me that
over time the same printer might use different paper stock, or in this case
Sequoia might have used paper stock rated for a newer scanner that possibly
didn't work as well on their older systems. Probably a number of other
possible permutations that could lead to error.
In any case I think the P1583 standards should
address these issues. I have forwarded your comments to the candidate and Al
and I are scheduled to meet with her next Friday. So far the Denver election
commission has not been willing to give us actual ballots though we have been
allowed to look at them in their office.
Chuck
P.S. Guess I need to be added to the TG2 reflector as well in order to
cover this topic.
At 4:55 PM -0500 12/7/03, Brit Williams wrote:
Charles - You are right. The FEC
Standards do not specify the paper for an optical scan system. This
due to the fact that the vendors use varioius types of readers so paper that
might be fine on one system may not work on another. This is handled
by the vendors specifying the type(s) of paper that must be used on their
system. Then the ITAs qualify the system using the paper the vendor
specifies. These specifications are then contained in the vendors
documentation and in the ITA reports. Jurisdictions using the system
obvioulsy should use the paper specified. Problems happen when, for
whatever reason, they do not.
Most of the vendors qualify printers and
provides a list of qualified printers to the using jurisdictions.
These vendors only guarantee an election if the juridsiction uses a
qualified printer.
The situation you described is entirely
possible. If the paper that was used is too translucent for the
system, then the marks on the other side could interfer with the system
correctly reading marks on the ballot.
If they have some left over blank ballots
you can easily determine whether or not this was happening. The
question then is what to do about it. I do not know how this would be
handled in Colorado, but in Georgia the evidence would be turned
over to the State Election Board and they would decide the appropriate
remedy. If this indeed happened, I would not be surprised if you wind
up reholding the school board race.
Best regards.
Brit
----- Original Message -----
From: Dr. Charles E.
Corry
To: Stephen
Berger ; hdeutsch@essvote.com ; Brit Williams ; stds-1583-tg2@ieee.org
Cc: vernonw74@earthlink.net ; Doug Fletcher ; Al Kolwicz
Sent: Saturday, December 06, 2003 3:07 PM
Subject: Paper stock and possible voting error with
optical scanners
Stephen, Herb, Brit, et al.,
During the November elections Al
Kolwicz and I were contacted by a candidate in a Denver school board race.
She was concerned about the excessive number of undervotes in her race,
~30% compared to 5-7% in the other races.
While a larger undervote in a school
board race isn't unusual in school board races, an investigation turned up
some factors that affect the P1583 standards.
This was a mail election using paper
ballots and the ballots were read using a Sequoia optical scanner system
using a DOS operating system (DOS OS is the only Sequoia scanner system
certified by Colorado at present). The Denver election officials were
quite cooperative and allowed us to inspect the actual ballots used, which
are generated and printed by Sequoia in California.
The problems with respect to the
standards come in because the paper stock used to print the ballots on was
rather more translucent than I expected for an optical scanner. Also, the
ballot was printed on two sides. So the first issue is that I don't recall
seeing any standards that specify how opaque the paper stock must be for
paper ballots.
Secondly, the Sequoia scanner read
the ballot as a mark between a greater than and less than symbol, i.e.,
> < is the blank the voter sees, and they indicate their
vote by a line between the symbols >-< for
their candidate of choice. While we were not able to test the
problem for the usual reasons, a potential error arises with translucent
paper stock because on the reverse side of the ballot there was a solid
line aligned with the candidate we were helping. So looking at a light
with a blank ballot in hand one saw ------>-<--------
We are unable to determine
whether this caused the candidate to receive additional votes, or whether
undervotes resulted if the voter voted for some other candidate. An
undervote might result if the voter marked some other candidate,
particularly if the voter drew a very light line between the symbols for
their candidate. The scanner might then read both the mark on the front of
the ballot and the one on the back of the ballot. As overvotes are
automatically rejected, the election tabulation might show many fewer
votes in this race than were actually cast.
In either case the voting
equipment standards need to include a test of optical scanners for such
malfunctions and specify ballot opacity. As best I can find the draft
standards don't consider these issues at present.
Chuck Corry
--
Charles E. Corry, Ph.D.,
F.G.S.A.
President, Equal Justice Foundation
http://www.ejfi.org/
455 Bear Creek Road
Colorado Springs, Colorado
80906-5820
Telephone: (719) 520-1089
Domestic violence against men:
http://www.dvmen.org
Personal home page:
http://corry.ws
Curriculum vitae:
http://www.marquiswhoswho.net/charleselmocorry/Default.aspx
The
Equal Justice Foundation is a member-supported, non-profit 501(c)(3)
corporation.
Contributions are tax deductible and can be made on the
Web at http://www.ejfi.org/Join.htm
The good men
may do separately is small compared with what they may do
collectively.
Benjamin
Franklin