Boulder County’s counting crawls on  (CO)

Laura Snider    Daily Camera    05 November 2008

A full day after polls closed — and long after Barack Obama claimed victory in the presidential race — fewer than half the ballots cast in Boulder County had been counted, earning the county the dubious distinction of being the slowest vote counter in the state.

If the current counting pace continues around-the-clock, county residents won’t see final results until late Saturday.

The devil, it seems, is in the dust. Election officials believe that tiny particles of paper dust are sticking to the lens on the scanner, creating a vertical line running down the ballot. When the line passes through an empty box, the computer may count a false vote.

To ensure accuracy, election workers are visually checking a picture of each ballot on computers, searching for the troublesome dust line. Though officials hope a couple of technological fixes will speed up the counting, the rate of votes counted Wednesday hung around 1,200 ballots an hour.

“Accuracy is more important than speed,” said Boulder County Clerk and Recorder Hillary Hall. “People expect accuracy — this is the job I was elected to do.”

The molasses speed of the counting process is frustrating Hall, but perhaps more confounding is that there doesn’t appear to be any obvious explanation for why the paper-dust problem showed up now.

“I just don’t know,” she said. “If this is a common problem, I’ve frankly never heard about it.”

Looking for the changed variable

The vast majority of ballots in Boulder County are paper, and the votes are counted by a Kodak optical scanner running software provided by Hart InterCivic.

This year, Hall ran elections using the same machines scanning the same ballots provided by the same vendors — Integrated Voting Solutions — in both the Longmont mail-in election and the primaries without incident. Before ballots were sent out for the general election, Hall’s team ran 10,000 test ballots through the system, again without any dust problems.

Because officials think the dust is linked to the paper, some election workers speculate that the paper vendor must have changed something — perhaps the stock, the ink or the printing presses.

“There were no changes in the paper. We used identical machines, the same personnel and even the same facility,” said Frank Kaplan, election services manager for Integrated Voting Solutions.

Kaplan’s company printed more than 4 million ballots for November’s election, and he has not received any other dust complaints. He said it could have something to do with Hart’s system.

But Hart just pointed the finger back in the direction it came.

“The county needs to talk to their printer,” said Peter Lichtenheld, director of marketing for Hart InterCivic. “Hart did not print the ballots ... and the printer did not use Hart secure ballot stock.”

User error?

Even if there are no physical differences in the equipment or the paper, this election differs in size. In the primary, nearly 36,000 ballots were counted, but the total number of ballots for this election is likely closer to 170,000, putting extra strain on the system.

“Think of it like the Xerox copy machine in your office,” said John Gideon, co-director of VotersUnite!, a non-partisan group that keeps track of voting machine errors across the country. “Usually someone comes by once a month, opens it and cleans it out. The static electricity causes the dust on the paper to get in there and gum everything up.”

When running a massive number of ballots through a scanner, the same principle holds, he said, only it’s accelerated. Someone needs to be in there cleaning out the dust as a regular part of the process.

Computer scientists Dan Wallach, from Rice University, and Douglas Jones, from the University of Iowa, agree that it’s pretty routine maintenance to stop during the ballot-counting process and deal with dust. Both scientists work with ACCURATE, a project funded by the National Science Foundation to study electronic voting systems.

“Dust in optical scan systems is a known problem,” Jones said. “When I was observing pre-election testing in Miami, (scanner) technicians had cans of compressed air, and they blew out the scanners every 200 ballots. ... The problem should be anticipated.”

Pickier than most

Some problems with the Hart scanners were anticipated. Last winter, Secretary of State Mike Coffman de-certified all Hart scanning equipment because the scanners “failed to count votes accurately when there are extraneous marks on the ballot.”

Coffman acknowledged that when a stray mark caused an over-vote, meaning more than one box was counted, the system alerted a human to look at the ballot. But when the stray mark caused a race that had not been voted on at all to look like the voter made a choice, the machine would not flag it as a problem. The same is true with the dust lines.

Coffman’s testing board recommended that he institute regulations requiring county officials to review every ballot, looking for the stray marks. Instead, Coffman chose to re-certify the machines, which are used in 47 Colorado counties, without the extra regulations.

That creates the possibility that some counties using the Hart scanners may be having dust problems and not know it.

“We’re thorough,” Hall said. “We went above and beyond. If we were doing (Colorado’s) normal audit process, we might not even find this error.”

Hall discovered the error Saturday when an election worker noticed there was a lot of “noise” in the digital picture of a ballot. On that particular ballot, the noise didn’t cause any mis-votes, but Hall was concerned. She had her workers dig through ballots that had already been scanned until they found a few errors, triggering a massive effort to look at every single ballot.

According to Gideon of VotersUnite!, that kind of attention to detail is far above what he’s noticed in other parts of the country. In his home state of Washington, nearly all the ballots are scanned on a similar system, but there is no audit process or double-checking.

“We just trust the machines,” he said.

Yakima County in Washington did report a similar dust problem with its scanners in 2004, but it was only caught because a close race for governor triggered a mandatory recount.

According to Neal McBurnett, a local computer scientist who has been helping develop better election auditing processes for years, this type of behavior isn’t unusual. Instead, he thinks what Boulder County is doing is unusual.

“Hillary (Hall) is stepping up to look for errors — in this case, little specks of paper or paper dust,” McBurnett said. “Normally, you just wouldn’t notice.”