Count prob­lems with vot­ing ma­chines de­tailed in re­port

The Herald (Rock Hill) - - Local & State - BY BRIS­TOW MARCHANT [email protected]­

In the last elec­tion, some votes in South Carolina got counted twice. Oth­ers were cred­ited to the wrong can­di­date.

Also, one ob­server thinks, the state’s 14-yearold vot­ing ma­chines are start­ing to show their age, pro­duc­ing other er­rors.

Those are some of the con­clu­sions in a re­port re­leased last week by the League of Women Vot­ers of South Carolina.

On Jan. 22, the league will host a pub­lic fo­rum at the Rich­land County Pub-


Lynn Teague, vice pres­i­dent of the League of Women Vot­ers

lic Li­brary on ways to im­prove the state’s elec­tion sys­tem. The group is back­ing ef­forts in the S.C. Leg­is­la­ture to re­quire a paper bal­lot sys­tem.

“Over the years, they’ve made up­grades, and it’s still flawed,” Lynn Teague, vice pres­i­dent of the league, said of the state’s ex­ist­ing vot­ing sys­tem. “They’re still count­ing votes wrong ... and all this with­out some­one de­lib­er­ately try­ing to mess with the sys­tem.”

Both the re­port’s author and the League of Women Vot­ers praise the state Elec­tion Com­mis­sion for mak­ing all of its 2018 elec­tion data pub­licly avail­able.

“South Carolina has a lot more trans­parency,” said Teague. “That’s why we know what hap­pened here.”

The Elec­tion Com­mis­sion doesn’t dis­pute the er­rors re­ported. But, it says, most of the prob­lems cited are the re­sult of hu­man er­ror, not prob- lems with the elec­tion sys­tem.

“Af­ter a deeper in­ves­ti­ga­tion, tak­ing into ac­count the full process, al­most all of them are due to some miss­ing knowl­edge on the part of the user,” said Elec­tion Com­mis­sion spokesman Chris Whit­mire.

Even seem­ing er­rors in how the ma­chines oper- ated “are ei­ther an er­ror in set­ting up the ma­chines or in how the re­sults were pro­cessed,” Whit­mire said.

“When it’s op­er­ated cor­rectly, the sys­tem works,” he said.

But the Elec­tion Com­mis­sion shares the goals of the League of Women

Vot­ers in im­prov­ing the state’s elec­tion sys­tem. The com­mis­sion al­ready has asked for pro­pos­als from ven­dors for new vot­ing sys­tems this spring, pro­pos­als that must in­clude a paper trail of the bal­lots cast.

“This soft­ware was prob­a­bly de­vel­oped around the late ‘90s,” Whit­mire said of the state’s ex­ist­ing vot­ing ma­chines. “It is old. The sys­tem’s old. The hard­ware’s old. We’re look­ing for­ward to hav­ing a new sys­tem.”

State of­fi­cials up­graded the op­er­at­ing sys­tem on the state’s touch­screen vot­ing ma­chines be­fore the 2018 elec­tion. Still, a re­port by Univer­sity of South Carolina pro­fes­sor Dun­can Buell cites er­rors in the vote-count­ing process and in how the ag­ing ma­chines op­er­ated.

Dur­ing June’s statewide pri­mary, for ex­am­ple, Buell says records ap­pear to show bal­lots be­ing counted twice in one Marlboro County precinct, likely as a re­sult of the process of trans­fer­ring the bal­lots from vot­ing ma­chines into the county’s cen­tral com­puter.

“There were four ma­chines that had 148 votes, and then a fifth ma­chine that had five votes,” Buell said.

Be­cause of a mal­func­tion in the fifth ma­chine, its five votes had to be uploaded sep­a­rately. In the process, poll work­ers mis­tak­enly uploaded the rest of the re­sults again, too, Buell said.

That added up to 301 votes, even though only 153 were cast.

“I can see Per­son A say­ing, ‘Up­load the votes from this mem­ory card,’ and walk­ing away, and, in­stead of only up­load­ing that one, Per­son B uploaded all five.”

In Bam­berg County, one precinct in Novem­ber was miss­ing one of two County Coun­cil races, a bad enough prob­lem. But, then, the cen­tral com­puter shifted 420 votes from one race to the other. Races lower down the bal­lot also saw votes shifted be­cause of the er­ror, Buell said.

The USC pro­fes­sor recorded a sim­i­lar er­ror in Beau­fort County in 2010. Then, due to a sim­i­lar vote shift, the last item on vot­ers’ screens — a bal­lot amend­ment — re­ceived no votes ei­ther for or against, even though 725 vot­ers cast bal­lots in the precinct.

Buell points out the shift in Bam­berg didn’t change the out­come of the elec­tion — “both (coun­cil can­di­dates) were un­op­posed.” Also, vot­ers may not have no­ticed if their coun­cil district was miss­ing from the bal­lot.

“If they left off the gov­er­nor’s race, some­body would have no­ticed,” Buell said.

A big­ger con­cern may be an in­creas­ing mem­ory er­ror seen in the re­sults.

A check to see if the same vote ap­pears in a vot­ing ma­chine’s re­dun­dant mem­ory — and, thus, is con­firmed — pro­duced er­rors on just a hand­ful of ter­mi­nals from 2010 to 2016. But that num­ber jumped to 57 ter­mi­nals dur­ing the pri­mary and 71 on Elec­tion Day.

Buell spec­u­lates those er­rors are the re­sult of the age of the vot­ing ma­chines, which the state bought in 2004.

A sim­i­lar er­ror may be re­spon­si­ble for vot­ing ma­chines’ record­ing votes as be­ing cast on days other than Elec­tion Day, which was Nov. 6 last year. Those wrong dates can be chalked up to er­rors — even those votes that were recorded af­ter Elec­tion Day, the USC pro­fes­sor says.

“If the date shows up as Nov. 7, but the card was uploaded Nov. 6, that’s a bug,” he said. “I’d be more wor­ried if the vote was added Nov. 7, and the card was uploaded Nov. 8.”


In Co­lum­bia, for the first time, elec­tronic vot­ing ma­chines will be used for vot­ers to cast their bal­lots.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.