Tripp County will need court order to investigate mismatch between hand count and audit
Hand counting experience ranges from hiccups to headaches for three SD counties
Tripp County will be the first South Dakota county in two decades to hand count ballots in an election, according to the Secretary of State’s Office.
Barb Desersa was awake 40 hours straight between Election Day and Wednesday.
Nearly a quarter of that time was spent supervising volunteers hand-counting ballots for Tripp County.
“We left at 5:30 a.m.,” said Desersa, who serves as auditor for Tripp and Todd counties. “We went home, showered and then went back to work.”
Several races had to be recounted by volunteer counting boards – sometimes three or four times that night. When the counting stopped around 5 a.m. Desersa and her team “pushed them to the state” and left. The majority of the volunteers had already worked 12-hour shifts as poll workers.
Tripp County was the first in South Dakota county to hand count its ballots since 2004, according to the Secretary of State’s Office. County commissioners leaned on a little-used state law to overrule Desersa’s authority after hearing from a group of South Dakotans who spent months pushing commissioners across the state to force hand counts.
On election night and into the following morning, Desersa compared the hand counted votes to a tabulator to prove tabulating machines are accurate.
“I couldn’t believe how long it took,” Desersa added. “We had problems balancing, and there were more voters than normal (for a midterm election). It was just such a long ballot; it took everyone a long time to do each ballot and everybody was tired.”
And they’re not done counting.
A Thursday vote canvas revealed a discrepancy in a single precinct between the number of official, completed ballots recorded in poll books and the number of audited ballots. It is unclear at this point how the error occurred.
The county now needs a court order to reopen that precinct’s ballot box to investigate. Deputy Auditor Marcia Haukaas said it wouldn’t have become a week-long process if the county had used its tabulator as an official reporting method, though this issue would’ve popped up no matter what.
“We wouldn’t have been able to officially record that precinct done until we figured out the problem that night — we would have recounted how many ballots were unused, recounted all the used ballots and then we would have ensured everything was legit,” Haukaas said, adding that it wouldn’t have required a court order to check the ballots again that night. “Now we’re going backwards and trying to have people remember things that happened on a 23-hour day with no sleep afterward.”
The intended court order will have no effect on the official tally, but is meant to put both the auditor’s office and county commissioners’ minds at ease.
The precinct in question had 356 names written in the poll book, and hand counters recorded votes for that same number of ballots. But the tabulator ran through 281 votes. The auditor’s office had recorded that only 313 ballots should have been used in the precinct, Haukaas said.
Haukaas hopes it’s human error.
“I don’t want people thinking the machine is doing something wrong, because it was just off in this one precinct but was right on all the other 12 precincts,” Haukaas said. “This is just a mystery right now.”
While Tripp County was the only one to send hand-counted results to the Secretary of State’s Office, Fall River County and Butte County also hand-counted a small number of votes after running ballots through tabulator machines.
The same group that convinced Tripp County Commissioners to order a hand count tried to persuade county commissioners in Butte and Fall River counties to completely hand count votes, The two county auditors and their commissioners, however, opted to partially count votes as an “audit” of the system instead.
“I wanted to try and figure out what the process is and how we can use this data and experience,” said Butte County Auditor Annie Capp. “If the state moves that way (to require hand-counting ballots), we’ll have a better idea of what we need to do.”
It took Butte County volunteers about an hour to prepare and hand count three races on 270 ballots, each from a different precinct. A county-wide hand count would have taken several hours, Capp said.
The hand count for one precinct was inaccurate, but Capp didn’t make the volunteers recount.
“This was an experiment. I didn’t want them to overthink or overwork,” Capp said. “The first time they counted one of the races they had seven more tick marks than they had ballots.”
For Fall River, it took two and a half hours to count three races on 331 ballots from two of the county’s smallest precincts. All hand-counted counties had between three and five counting boards for a precinct.
Fall River Auditor Sue Ganje, who has served as auditor since 2005, has been involved with elections since 1984, when she was an election clerk. She remembers when the state hand counted all its ballots, so she had an easier time running the show.
Instead of recruiting counters from her pool of precinct workers, she recruited “fresh eyes” to count after the polls closed.
The counters were off by one vote for one race in each precinct compared to the tabulators, Ganje said, which is normal based on interpretation by counters.
“Hand counters determine the voter’s intent, whereas the tabulator goes through the scanner and just scans the ovals,” Ganje said. “The board did very well.”
Ganje and her team were up until 2 a.m. running ballots from Fall River and Oglala Lakota County through tabulator machines and partially hand counting.
Both Capp and Ganje were happy with the results of their audits. One of Ganje’s counters had expressed doubts about South Dakota’s election process and tabulator machines.
“I felt good having that person on the board,” Ganje said. “I wanted them to know what the process was like.”
To assuage doubts, Ganje said the state should implement post-election partial audits, similar to what happened in Fall River and Butte counties.
“If that’s what it would take to get the confidence back with voters, maybe that’s what we need to do,” Ganje said.
Capp said both tabulators and hand counting should be used for checks and balances.
“We can’t just trust the machine is doing everything. People program it. It still has a lot of human interaction, and humans can make errors,” Capp said. “I think it’s great to have those checks and balances.”
Desersa believes that the tabulators should be enough. With the test decks, demonstrations auditors run before the election and current checks and balances in place, she said she’d prefer a post-election audit that doesn’t involve hand-counting every ballot in her county – or across the state.
“We’ll have time to forget about it before the next election, but I think it’s pointless unless the whole state orders it. It just doesn’t make sense to me,” Desersa said. “Why go back in time to hand counting when we have machines that are perfectly capable of tabulating the vote?”
Our stories may be republished online or in print under Creative Commons license CC BY-NC-ND 4.0. We ask that you edit only for style or to shorten, provide proper attribution and link to our web site. Please see our republishing guidelines for use of photos and graphics.