Disenfranchised voters in Chicago and Cook County. I'm shocked...not.
chicagotribune.com
Spoiled vote tied to defect in device
Ballot recorder was misaligned
By Douglas Holt
Tribune staff reporter
Published July 28, 2001
Defects in election equipment used in Chicago and suburban Cook County
may have caused tens of thousands of votes in last fall's presidential election
to be thrown out, according to interviews and documents obtained by the Tribune.
The trouble occurred in a low-tech piece of equipment called a vote recorder, the device into which voters
slide punch-card ballots. Out of roughly 50,000 vote recorders used in Chicago and Cook County suburbs,
half were made with templates that failed to line up perfectly with holes on ballot cards used last fall,
according to a study completed in May by the wholesaler that provided the equipment.
Almost 123,000 ballots in the city and county in
November lacked a recordable vote for president,
about 6 percent of all those cast and twice the
error rate in the 1996 presidential election.
The alignment problem could account for nearly
one-third of that number, or more than 39,000
spoiled presidential votes, according to Michael
Kreloff, policy director for Cook County Clerk David
Orr, who runs elections in the suburbs.
The unusually high undercount tally here helped
earn Illinois the distinction of having the most
problem-plagued election of any state in the nation.
Shortly after the election, city and county officials
blamed the problem on a high turnout of new voters
unfamiliar with the idiosyncrasies of the punch-card
system. But the study is the first hard evidence
that faulty equipment may have played a significant
role in the problems.
In the system used in the city and county in November, voters inserted a card with 456
perforations--enough for all the candidates and questions on the lengthy ballot--into a slot beneath a
molded plastic template. Holes in the template were supposed to match up with the perforations on the
card.
Voters marked their ballot by using a metal stylus to punch through the template and break perforations
on the card.
But in half the machines, the alignment was off by about the width of three human hairs, according to the
study by Election Data Corp. of Valley Center, Calif., the wholesaler that provided vote recorders used by
the city and county. That increased chances that perforations were not cleanly punched out and therefore
not recorded.
Election Data delivered the two-page study to city and county election officials in May but urged them to
keep its contents confidential, according to a company memo. The Tribune obtained the study and the
memo through a Freedom of Information Act request filed with the city and county.
"It was off just enough to open the window for the possibility that someone could punch erroneously," said
John Ahmann, a consultant hired by Election Data to conduct the study. "Nobody caught it. They
should've caught it, but they didn't."
Ahmann said a manufacturing defect in the machines may have led to problems. Another factor, he said,
was the way machines were inspected off the assembly line. Quality inspection workers checked their
accuracy in the factory, where conditions were hotter than in polling places. The plastic templates may
have shrunk slightly when they cooled, Ahmann said.
Pinholes, indentations, chads
After the election, city and county officials scrutinized ballots in a sampling of precincts and concluded
that about 2 percent showed clear signs that a voter had tried to make a punch for president--pinholes,
indentations or partially removed tabs of paper called "chads"--but it couldn't be read by automatic
tabulating equipment. Ahmann said the alignment problems were almost surely the cause.
Projecting that 2 percent error rate out over all the 1.9 million ballots cast in the city and county last fall,
Kreloff estimated that nearly 39,000 of those ballots could have been afflicted by hole-punching problems.
Cook County and Chicago election officials said Ahmann's findings present strong evidence that
equipment bought last year as part of a $25 million upgrade appears to have been defective. But they said
they will know more in a few weeks when a study of the machines by the Illinois Institute of Technology is
expected to be completed.
The problems identified by Ahmann appear to trace to a stamping machine at HK Plastics in Oceanside,
Calif., a subcontractor that produced the vote recorders for Election Data. The machine cranked out two
templates at a time. One of the molds was fine. But the other produced templates with holes that were
out of alignment by a microscopic amount, Ahmann said.
Election Data owner Dick Stephens said he was not fully convinced the defective parts made a difference
in the vote. But he stressed that his company takes full responsibility for fixing the voting devices at no
charge to Chicago or Cook County.
"I've got some bad parts in there. There's nothing I can say about that," he said, adding that he was
"damned unhappy that anything went out that wasn't in spec."
In response to Ahmann's findings, Orr on June 20 sent a stern letter to Election Works of St. Charles, a
dealer operated by Stephens' daughter that sold the city and county the vote recorders. Sandra Hed, who
runs the company, could not be reached for comment.
"We are providing you with formal notice of what now appears to be a `defect in workmanship or material'
in equipment provided to us," Orr wrote."We look to you for solutions, damages and remedies as are
lawfully and explicitly authorized under our contract with you."
Unread rate `inexplicably high'
The letter said that, although there are always some voters who choose not to vote for a particular office,
in last fall's presidential race "our percentage of unread ballots seemed inexplicably high as compared to
past presidential elections."
Although poor neighborhoods, where education levels are relatively low, had higher percentages of
uncounted votes--suggesting a higher rate of voter errors--the letter said that "there was also a fairly
randomized increase in falloff across the board."
In Schaumburg Township, for example, 1.2 percent of ballots cast in 1996 lacked a presidential vote. That
rate jumped to 4.3 percent last fall.
"It's not the whole answer," Kreloff said of the equipment defects. "But it's a piece of the answer that
probably did affect everybody, even in the high income areas." |
