Voting machines put to test

Miami-Dade and Broward counties publicly tested their touch-screen voting machines in an effort to restore public trust following reports of glitches with the equipment.

BY JOE MOZINGO Miami Herald   13 August 2004

Hoping to restore public confidence in its voting system, Miami-Dade officials conducted extensive testing on 212 electronic voting machines Friday, allowing the public to take part for the first time.

Late Friday, Miami-Dade election workers were still tabulating results but said they had found no problems.

Broward County conducted a smaller test the minimum required by the state and reported perfect results. Space was so limited, observers had to watch from behind glass, which left some dubious.

''You can almost conclude that the test was passed before it was run,'' said Steve Shin, an observer running for Broward supervisor of elections.

By contrast, the Miami-Dade County Elections Department not only opened its doors to the public, it took steps not required by the state to ensure the machines counted votes as they were cast.

With reporters, observers, auditors and election workers clustering around the machines at election headquarters in Doral, the event started out with a bit of confusion. Election managers even considered starting over again after two ballots, instead of one, were mistakenly cast on a machine.

But auditors reconciled the problem, and Elections Supervisor Constance Kaplan decided to continue.

''We expanded the process and knew it was going to be bumpy at the start,'' Kaplan said. ``But this testing is going to show the computers are accurately programmed and do count the votes.''

There were two prongs of the day's testing, scripted and unscripted.

In the first, 25 teams of two cast ballots on each machine, following scripts doled out by the department. One person cast the votes, the other reviewed it. Each script represented one of the 222 distinct ballot styles for the primary.

The results were then collected from the machines and tabulated to make sure they matched up with the scripts.

''We have to do the scripts because we have to make sure every candidate receives at least one vote using every ballot style,'' said Seth Kaplan, spokesman for the department. ``At the same time, we want to introduce other elements of surprise.''

The element of surprise was what went beyond the state's requirement: Members of the public were invited to vote however they wanted on 16 of the machines, with scribes recording their choices by hand and video cameras shooting from above. These again were matched with the results recorded by the machines.

''They picked machines completely at random for the nonscripted testing,'' said Douglas Jones, a computer science professor at the University of Iowa, hired as a consultant for the Elections Department. ``What I'm seeing here is some of the strongest preelection testing we've ever heard of.''

The purpose of the unscripted test is to weed out potential flaws that might not appear in a scripted situation.

Unfortunately, fewer than a dozen members of the public showed up. Neither Broward nor Palm Beach counties conducted an unscripted test. Palm Beach tested 86 of its machines on Wednesday, reporting 100 percent accuracy.

Jones said he hopes Miami-Dade takes another step beyond its neighboring counties by conducting testing on Election Day the so-called parallel testing reformers have been pushing for. County Manager George Burgess said he was considering it.

''We're looking for what kind of options we might have,'' Burgess said.