Voting system switch bogus?
J.D. Prose, Beaver County Times Staff
05/26/2005
 
BEAVER - After reviewing preliminary results from last week's primary election, county officials said Wednesday that there wasn't much difference between the accuracy of paper ballots and the county's recently banned electronic voting system.

 
However, Michael Shamos, the Carnegie Mellon University computer professor who failed the Patriot electronic voting system twice this spring following tests he conducted for the state, said county officials were making invalid comparisons.

Earlier this year, Shamos cited problems the Patriot system had with registering votes, which he said could have caused the Patriot to count ballots inaccurately and have a higher-than-normal undervote rate, most notably in last November's presidential election.

An undervote occurs when the number of votes cast in a particular race is less than the total number of votes cast in an entire election. For example, if 100,000 people cast ballots in an election but only 90,000 vote for president, the undervote rate would be 10 percent.

According to unofficial results of last week's primary, 1,002 Democratic voters out of 26,205 who went to the polls, or 3.8 percent, didn't cast a vote in the primary race for county judge between county Solicitor Deborah Kunselman and Assistant District Attorney Kim Tesla.

In the GOP primary, 583 Republican voters out of 10,564, or about 5.5 percent, didn't vote for either Kunselman or Tesla, an understandable result, though, considering that both are registered Democrats.

Beaver County's undervote in last year's presidential race was 5.25 percent - 4,551 votes - using the Patriot system, far above the 1.49 percent average found in a Grove City College study of 24 Pennsylvania counties.

But, Commissioner Charlie Camp said the latest primary figures demonstrate that the county's $1.2 million Patriot system was just as accurate at counting votes as the paper system used in the primary.

The Patriot system wasn't flawed, he said. Instead, voters are more willing to exercise their rights by not voting for candidates they don't like.

"Only God and the voter know why there's an undervote," Camp said.

Elections director Dorene Mandity said the undervote rates in the primary confirm what she's been saying for months: that the Patriot system was working properly.

The undervote "tells me it's the will of the voter," Mandity said. "You clearly see that for sure now. I don't know where the doubt would be when you have the numbers right in front of you."

Shamos, however, said county officials were jumping to conclusions. He said the comparisons were baseless because last week was a primary race and the undervote figures in the Grove City study were from a general election.

In Beaver County in 1996, the last presidential election in which paper ballots were used, the undervote totaled 820, compared to 4,551 last November.

Also, Shamos said county officials don't really know the intent of voters because the county does not have ballot-scanning machines in every precinct that would allow voters to immediately be notified if they didn't vote in a race.

Short of that happening, Shamos said, there's no way for county officials to truly know what voters were thinking or if they purposely didn't vote in a given race.

Still, Camp and Mandity maintain that the results show there's little difference between electronic voting and paper ballots.

The county used paper ballots until November 1998 when it switched to the Patriot system. The undervote in the 1997 Democratic primary race for county judge was 4.5 percent and 5.5 percent for the Republican primary. Those figures are comparable to last year's Patriot system undervote and last week's primary undervote.

"The numbers are speaking," Camp said, "not Charlie."