At the Jan. 7 Lebanon County Commissioners meeting, the county’s Director of Elections and Voter Registration, Michael Anderson, refuted allegations made by state Representatives Frank Ryan (R-101) and Russ Diamond (R-102) that more presidential votes were counted last November, statewide and in the county, than were actually cast.
At the end of the meeting, Diamond, who attended via Zoom, expressed satisfaction with Anderson’s explanations. Ryan did not attend.
On Dec. 28, 2020, Ryan issued a press release claiming that “the Department of State shows that 6,962,607 total ballots were reported as being cast [statewide, in the Nov. 3, 2020 election], while DoS/SURE system records indicate that only 6,760,230 total voters actually voted.”
“The difference of 202,377 more votes cast than voters voting . . . in the presidential race, adds up to an alarming discrepancy . . .” Ryan claimed.
On Dec. 30, Diamond claimed on Facebook that a similar discrepancy existed in Lebanon County’s presidential election voting.
Addressing the Commissioners who were sitting as the County Board of Elections, Anderson explained that the actual count of ballots done by each county to determine a winner was separate and distinct from the SURE system, which has nothing to do with determining winners and losers.
Determining a winner versus determining who is eligible to vote: apples and oranges
“SURE” stands for Statewide Uniform Registry of Electors, ‘registry’ being the crucial word.
In a telephone interview with LebTown on Jan. 4, Anderson pointed out that SURE is not designed to determine election winners, and never has been. Instead, it is simply a database of registered voters, and a historical record of who has voted in past elections.
Two different processes, two different purposes. This, according to Anderson, is what Ryan and Diamond didn’t understand.
Whether the SURE totals are right or wrong, they have no direct connection to exactly how many ballots have been cast in a given election, or whether those ballots have been counted accurately.
How are ballots actually counted and reported in order to determine a winner?
With the recent advent of universal mail-in voting, the process of accurately counting all votes properly cast now involves two procedures instead of one, but paper ballots and an actual physical count of every ballot cast are at the heart of each.
Votes cast in person at a polling place
According to Anderson, throughout election day the number of voters who have signed into poll books (which results in a “numbered list of voters”) at a polling place is compared to the total number of ballots scanned by the polling place’s scanner to ensure they are equal. If not, the difference is logged and an election official determines the cause.
After the polls close and all the county’s poll books and paper ballots have been collected at the election office, they are again reconciled by election officials to be sure that the county-wide number of actual ballots cast at polling places equals the county-wide total number of signatures in all the poll books.
“That reconciliation is done after every election,” said Anderson, “and the computation is read out loud at a public meeting [of the Board of Elections] that both a Republican and Democratic representative attend.”
Anderson explained that mail-in ballots – “no-excuse,” traditional absentee, and overseas/military – are counted and reconciled by comparing the number of actual ballots returned against the number of actual outer envelopes returned.
Again, Anderson said, “this is done at a public meeting” where observers from each party are present.
When both types of ballots have been counted and reconciled to the Board of Election’s satisfaction, the county’s results are published on its website, announced to the media, and, more importantly, reported to Pennsylvania election officials.
That combined total – in person plus mail-in – would have become part of the “6,962,607 total ballots . . . reported as being cast” statewide cited by Representative Ryan in his Dec. 28 press release.
That, according to Anderson, is the actual count of actual votes that determines an election’s winner. And, that count has no direct connection to how totals are determined and reported on SURE.
The SURE system is just a record of who is registered to vote, and whether they voted in past elections
The importance of knowing if someone who shows up to vote is registered or not is obvious. But it’s also important to keep track of who has voted in past elections, because that determines if they can vote in the next election, or if they should be purged from the rolls for inactivity.
After each election, but as a separate task from the actual counting of votes to certify a winner, Anderson said that he and his staff compile data about who has voted (but not how they voted) in that election.
This is done by manually, one at a time, scanning the bar codes next to voters’ signatures in the poll book they signed at the polling place, and by manually, one at a time, scanning the bar codes on the outside of every envelope that was used to return mail-in ballots. “It takes, usually, about two weeks,” Anderson said.
The result is a countywide record of who voted in the last election.
This is then sent to SURE at the Pennsylvania Department of State, and the SURE database gets updated and is used to determine voter eligibility in the next election.
Anderson explained that there is no need for a county to complete SURE updates at the same time it counts ballots, or even shortly after that, as long as the totals are updated and reported in time for the next election. This explains part of the difference between “votes cast” and “SURE totals.”
As Ryan admitted, his data was incomplete when he made his concerns public, because several counties had yet to report any post- Nov. 3 SURE data to the state.
Anderson added that post-election scanning for the SURE system would be quicker and less prone to human error if paper polls books were replaced by electronic poll books.
Why the difference? Anderson answers Diamond’s questions
Diamond expressed satisfaction with Anderson’s explanations of the counting and reconciliation process, the distinction between counting votes to determine a winner and seeing who voted for registration verification, and why updating SURE system data might be delayed.
“I understand that there is somewhat of a lagging nature of the SURE system,” he conceded. And, “I have always had great faith in the people of Lebanon County to count the votes accurately, and I’m not questioning that.”
Diamond had a specific concern about two county precincts where the difference between votes actually cast and the SURE total was unusually large.
Anderson said provisional ballots played a roll at these precincts. Provisional ballots are issued when a voter shows up to vote, but there is a question about his or her eligibility. The ballot is reviewed by the Board of Elections after election day, and either accepted and counted, or rejected.
Anderson explained that the two precincts cited by Diamond were newly split, and now had two polling places, where before there had been one. For example, the old Swatara Township precinct was now Swatara North and Swatara South. Instead of everyone voting at the township building, half still voted there, but half now voted at a fire company.
Some Swatara voters didn’t know this, and showed up where they voted in years gone by – at the township building – when they should have gone to the fire company. They were then given a provisional ballot and filled it out there, at the “wrong” polling place.
Those provisional ballots were ultimately approved and counted, but credited to the “wrong” polling place’s totals, because that’s where the provisional ballot was issued. As a result, the “wrong” polling place showed more votes cast than it had registered voters, but the total number of votes cast didn’t change.
Anderson added that often a voter will fill out a provisional ballot, but also mistakenly sign the poll book. So until the error is caught, there will be one less ballot than shown in the poll book totals.
He also said that, during the tedious two week process of manually scanning polls books and envelopes for SURE statistics, some ballots don’t scan, and a few manual errors are inadvertently made by overworked election workers.
Anderson assured Diamond that those errors will be caught and fixed in the SURE system in time for the next election.
Diamond ended by expressing his satisfaction. “You’ve answered a lot of my questions, and I really do appreciate it.”
Lebanon County is also participating in a “risk limiting audit”
In addition to having counted and reconciled every ballot cast in the Nov. 3, 2020 election, Anderson said that the county, along with 63 others, is voluntarily participating in a statewide “risk limiting audit.”
A risk limiting audit is a way of confirming that computers tallied an election accurately. It involves manually comparing a statistical sample of paper ballots to the computers’ records for those same ballots.
For Lebanon County, Anderson says 515 ballots cast in the November election have been selected at random to be compared to computer totals for discrepancies. If any are found, further reconciliations and audits can be done.
The Pennsylvania Department of State has hired an outside firm to conduct the audit.
Questions about this story? Suggestions for a future LebTown article? Reach our newsroom using the contact form below and we’ll do our best to get back to you.
Do you support local news?
If you believe that Lebanon County needs independent, high-quality journalism, consider joining LebTown as a member. Your support will go directly towards stories like this and you will be helping ensure that our community has a reliable news source for years to come.
Learn more about membership and join now here.