Commentary: Diebold VP Says Company’s Machines Recorded Tallies Accurately in Test By DAVE BYRD

Tuesday August 30, 2005

A recent guest editorial in your paper inaccurately criticized Secretary of State Bruce McPherson about the latest developments in California’s move to electronic voting machines. The piece misused several figures reported by the Associated Press and Con tra Costa Times about a recent testing of Diebold Election Systems, Inc.’s (DESI) AccuVote-TSX with AccuView Printer Module election voting machines. The author also sarcastically accused a respected public official of poor math skills. One of the misstated facts claimed that during the dry-run test of the Diebold election system, McPherson’s office reported a 10 percent failure rate (the guest writer wondered if the failure rate was actually higher, which it was not). In fact, in that test, 10,720 votes were recorded on 96 voting machines with 100 percent accuracy. Despite 11 paper jams and 21 other problems on the new machine-printer combination, not a single ballot was lost. 

The State of California has understandably asked DESI to fix the printers bef ore the new machines can be used in real, live elections, and we’re making those changes now. Sadly, there has been too much of this type of misinformation in media coverage of electronic voting machines in California. Media coverage of the test earlier t his month sensationalized the results so that any reader would imagine complete chaos and a failure in tabulating results. One newspaper even reported that the paper jams caused long lines, “causing voters to give up and go home,” when the actual test con sisted of a handful of volunteers voting repeatedly on the machines in a warehouse. 

Paper jams on these printers occurred in roughly one out of 1000 cases, which while not perfect, is consistent with similar tests on receipt printing for ATMs and cash re gisters. DESI is now working to make minor adjustments to the printing units to improve their performance, and is also working to reduce or eliminate the screen freezes. 

Do these problems need to be fixed? Absolutely, but let’s not lose sight of the bene fits. A recent CalTech/MIT Voting Technology Project report says that the national average rate for residual (missing) votes in the 2000 elections—conducted overwhelmingly on punch card and paper ballot systems—was 1.9 percent. Applying that ratio to this test of 10,720 ballots would mean 203 lost votes. Again, this test did not lose a single vote. 

No system is perfect, but in comparing touchscreen machines to punch cards, mechanical lever machines, optical scan, or plain old-fashioned paper ballots, a half-dozen studies including the CalTech/MIT report show that your vote is most likely to be counted if you vote on a touchscreen. 

With these kinds of results, the contest between touchscreens and hanging chads should be a landslide. And public officials who are helping California vote more effectively with these machines should be praised, not ridiculed on the pages of your newspaper. 


Dave Byrd is vice president of business operations for Diebold Election Systems, Inc.