Chicago will be the first place where punch card ballots will be checked
for overvotes and undervotes and return problem ballots to voters for
correction.  The problem was demonstrated in Florida where many older
persons with low vision and a complicated ballot layout registered two
selections in the same race or left the choice for president blank.

Kelly


From the Chicago Tribune

State gets set to test the future of voting

Some fear delays, loss of secrecy
By Douglas Holt
Tribune staff reporter

March 17, 2002

Voters may think they're going to the polls Tuesday simply to pick
candidates, but they'll also be enlisting in a war on the dimpled,
hanging, swinging
and dangling chads that became the scourge of the 2000 presidential
election.

The primary election will mark the first wide-scale use in the area of
technologies designed to prevent a Florida-like fiasco, rejecting ballots
marked
with too many votes or too few and giving voters a second chance to do it
right.

The mechanisms used to accomplish this differ from place to place, with
Chicago and suburban Cook County grafting unprecedented error detection
safeguards
onto its old punch-card voting equipment and DuPage County switching to
optical scan devices that read hand-marked ballots like a college
entrance exam.
Lake and McHenry Counties earlier adopted the optical scan.

The promise of both approaches is a far lower rate of botched or
incomplete ballots and less ambiguity and doubt when outcomes are close.

Whether they will make the elections a breeze or a nightmare is anyone's
guess, and city election officials have warned of the potential of long
lines if
a lot of ballots get flagged for problems.

In Chicago and suburban Cook County, voters will serve as electoral
guinea pigs. Primary voters will fill out punch cards much like they have
for several
decades, but then feed the cards into a Precinct Ballot Counter 2100, a
device that looks like a small cash register. It will keep a running but
secret
tabulation of votes while also auditing individual ballots for problems.

The device was designed to monitor so-called overvotes in which voters
punch two or more holes in a particular race and spoil their selection.
But election
officials in the city and county intend to also use it to flag the more
common phenomenon of undervotes in which voters intentionally or
accidentally fail
to make a choice in every contested race on the ballot.

That has never been tried with punch card equipment in a major election
in the U.S., and the Nebraska firm that makes the equipment warned
Illinois officials
against it as recently as last year.

General concern

"I certainly wouldn't advocate this in a general election where you are
going to have offices on virtually every ballot that are undervoted,"
Herb Deutsch,
an Election Systems & Software official told the state Board of
Elections. "Every ballot would get returned and make the administration
of the election
process unmanageable."

Deutsch now says his reservations have eased in light of extensive
efforts to educate voters and poll workers about the equipment.

"This is the first major election where this is being attempted," he
said. "I guess this will be an interesting experiment."

Screening undervotes could be cumbersome because voters often choose to
skip marking their ballots in contests for certain offices, especially
relatively
obscure ones they know little about. The problem is especially acute in
Illinois because the state is one of the few that elects all its judges
and there
are typically dozens on the ballot. On Tuesday, 104 candidates will be
seeking Republican and Democratic nominations for 24 judicial seats up
for election
in Cook County in November.

When voters feed their ballots into the counter, it will be programmed to
spit back any with potential errors and point the voter to the problem.
Then voters
will have the option of either redoing their ballot or letting it be
counted as is. An overvote in any particular race does not void the rest
of the ballot,
nor does skipping a race.

Election officials here acknowledge that screening undervotes has the
potential to slow down things. Langdon Neal, chairman of the city Board
of Elections,
warned Chicago aldermen in a memo last fall that "the new ballot
screening procedures may cause some congestion in active polling places."

But Neal and other officials said they opted to overlook those concerns
because undervotes were at the heart of glitches that spoiled a record
number of
ballots in the 2000 election.

More than 120,000 ballots in the city and county failed to register a
vote for president, an error rate double that of past elections and
partly caused
by defective voting equipment purchased in 1999. In the city's worst
performing precincts, one out of three ballots did not show a recordable
vote for
president.

Problem precincts

A post-election study of the city's 125 most problem-plagued precincts
found people had tried to cast votes in more than half the cases where no
presidential
choice was recorded by counting equipment. Those ballots suffered from
the same kind of vote-canceling problems that plagued the Florida
presidential tally--hole
punches called chads that remained stuck in some way to the ballot.

If a similar problem crops up Tuesday, it will be flagged by the
second-chance equipment as an undervote.

Critics of the new procedure say it is fraught with the potential to
violate ballot secrecy and provide an avenue for partisan election judges
to lead voters
unfamiliar with what to do next into voting a certain way.

Under the system, election judges are supposed to keep their eyes and
hands off a completed ballot even if it has been kicked back by the error
detection
equipment. But it didn't always work that way during a test run in
municipal elections last April where there were few races and issues on
the ballot.

Voting in Palos Township, Elaine Roupas, chair of the State Board of
Elections, had her ballot rejected due to an alignment problem. Election
judges, she
said, "looked at my ballot to see what was wrong. They were conferring
with each other trying to figure out what the problem was. In the
meantime, there
went my privacy."

She chalked up her experience to growing pains and praised the county's
efforts as likely to make the election more fair and accurate.

Chicago authorities launched a voter education program about the new
system and bought 1,900 television monitors to be placed at precincts and
play a continuous,
two-minute tutorial on how to cast a ballot and screen it for errors.

In the video, former television reporter Bob Wallace demonstrates how to
use a stylus with a firm, up-and-down motion to make a clean hole on a
punch card.
"Remember," he says, "we don't want no chads hanging around Chicago."The
video also emphasizes that voters aren't required to vote in every
office, and
don't have to fix their ballots if there's an overvote. "IT'S YOUR
CHOICE," the video reads in big block letters.

Citing cost and logistical concerns, Cook County Clerk David Orr, who
oversees elections in the suburbs, opted against video instructions but
has mounted
his own education drive.

The error-detection technology to be deployed Tuesday was available for
the 2000 election, but officials were prevented from using it because
state Republican
leaders repeatedly blocked legislation needed to authorize its use. After
the 2000 election, the state Democratic Party won a court order to have
the equipment
turned on.

One key cause behind the high percentage of lost votes in 2000 was traced
to a faulty part in tens of thousands of vote recorders in the city and
county.
Recorders are the devices into which voters slide punch-card ballots.
Roughly half the recorders were found with badly manufactured templates,
plastic
sheets nestled within the machines with precisely drilled holes that
guide ballot punches.

Alignment problems

The more out-of-alignment the holes, the less likely a vote recorder
would work, engineers at the Illinois Institute of Technology found as
part of a study
last summer ordered by city election authorities. Even under pristine lab
conditions, some flawed vote recorders produced 10 percent error rates.

"All the voters that went to that one machine had a 10 percent chance of
doing it wrong even if they used it perfectly," said Michael Hites, IIT's
chief
technology officer.

He said the error rate should have been less than one-half of 1 percent.
"We were very surprised there was an apparent mechanical defect," he
said.

The manufacturer, Election Data Corp. of Valley Center, Calif., agreed to
remanufacture new templates at no cost for all recorders that will be in
use Tuesday.
IIT has conducted tests on the new templates, and Hites said the chance
of error due to mechanical defect has been all but eliminated.

"From everything we've done," he said, "I wouldn't want to trust myself
to punch those cards without running it through a scanner to verify I've
done it
right."


VICUG-L is the Visually Impaired Computer User Group List.
To join or leave the list, send a message to
[log in to unmask]  In the body of the message, simply type
"subscribe vicug-l" or "unsubscribe vicug-l" without the quotations.
 VICUG-L is archived on the World Wide Web at
http://maelstrom.stjohns.edu/archives/vicug-l.html