Voting Machine Mess-up Du Jour (Displayed 09/26/04)

Hillsborough County, Florida. September, 2004. Sequoia.
245 "test votes" were determined to be real votes.

One early-voting touch-screen machine was left in test mode during the early voting period. It reported zero votes, since all the votes cast were recorded as test votes. Seventeen days after the primary, the error was caught and the 245 "test" votes were added into the results.*

They brought the machine's results cartridge in for its results to be counted.

And the machine reported zero votes.

I talked to [County Registrar] Buddy Johnson on Monday. He was frank about what happened. The machine was left in test mode; his procedures had not contemplated such a possibility.

"It would not leap out at you," he said, "that this was in test mode." The printout the machine produces indicated a test, Johnson said. But the poll workers had not been trained to look for it.

From another article:**

Early voting began Aug. 16, and on the three machines at Westgate, a total of 1,050 votes were recorded even though about 1,300 people had signed in during the two weeks.

The match between sign-ins and ballots cast is rarely 100 percent. Some people bail out without voting, and others don't complete the ballots. Those are called undervotes.

"It didn't make a lot of sense that there was that many undervotes," Johnson said of the library situation. So officials began looking for other explanations.

They found the answer Friday. On the machine mistakenly left in test mode, the votes were collected in a way that wouldn't be read later when the machine's data cartridge was fed into a tabulating computer.

* Reminder on Election Day - this is not a test. St Petersburg Times. September 21, 2004. By Howard Troxler, Times Columnist.

** 245 Votes In Primary Originally Uncounted. Tampa Tribune. September 18, 2004. By Ted Byrd.

See: Sequoia in the News

The problem with touch screens as vote counters
is that they can be easily manipulated.
~ Mike Devereaux, VP of Sales, ES&S