Pulaski.County.jpg

[Image courtesy of bluffton.edu]

The latest example of “there is no small stuff” in elections comes to us from Pulaski County (Little Rock) Arkansas – where a small but crucial error in preparing ballots for an upcoming millage election ended up costing the county thousands of dollars when they had to be reprinted. The Arkansas Democrat-Gazette has more:

One number put in the wrong place resulted in a decision Saturday to reprint more than 53,600 ballots before the March 11 Pulaski County special millage election.

The Pulaski County Election Commission — which now holds meetings during each poll-worker training session — voted unanimously Saturday to reprint the ballots after realizing the misprinted forms could not be counted by the voting machines at the precincts.

Election officials were not concerned Saturday that the error would delay the election. But it will cost the commission about $12,800 — at 23.9 cents per card, Election Commission Director Bryan Poe said. The extra cost will come from the organization’s $400,000 printing budget, reducing the amount of any unused portion that reverts to the county’s general fund at the end of each year.

The error here, however, wasn’t an obvious one that would have been caught right away – in fact, it didn’t even involve the content of the ballot itself. It stemmed from the so-called “code channel” on the ballot that allows it to be read by optical scanning machines:

The mistake occurred after a commission worker incorrectly typed a code that produces black boxes along the left side of the ballots. One box was omitted at the top of the “code channel,” which usually has two boxes where the rest of the channel has only one.

“It was one number put in the wrong place,” Commissioner Phil Wyrick said. “It was a human error.”

Unfortunately, the mistake wasn’t caught because the ballots weren’t tested on both models of scanners in use in the County – and specifically, the older ones that are placed in precincts:

Although the county uses two different types of machines — the M-100 and M-650 — to tabulate the votes, the final ballot draft was tested on only the M-650.

The county’s more than 100 polling sites all are equipped with the M-100 machines, allowing votes to be counted on site. The M-650 is located at the commission’s office in downtown Little Rock and is used to tabulate the combined results.

When the new ballots were tested Thursday on the M-100, the machine failed to read the forms.

The Commission did discuss whether or not to just count the ballots centrally, but decided that having the poll scanners function as “glorified voting boxes” would likely be unacceptable to poll watchers on Election Day.

I think there is a “lesson learned” and a larger challenge in play here. The lesson learned, of course, is to test a completed ballot on every model of counting equipment in a jurisdiction – especially older models, which are usually not as flexible or forgiving as newer models.

The larger challenge is to understand that even as election officials embrace the concept of usability in order to make ballots and other materials more accessible to voters, they also need to respect the increasing complexity of vote-counting technology. That complexity means that margins for error in making ballots accessible for machines are smaller than ever. Consequently, setting up systems to verify that people and machines will both have good experiences with a ballot are vitally important.