14,000 Overlooked Ballots Discovered on Henrico Flash Drive — Yes, It Was an Innocent Error

by James A. Bacon

The conservative blogosphere is ablaze with anecdotes and accusations of Democratic Party voter fraud. I’ll leave it to others to determine if there is any substance to the charges in in locales ranging from Philadelphia to Nevada. But there is one case very close to home — in my home jurisdiction of Henrico County — that bears comment.

Last night, Republic Nick Freitas held a slender lead — 1,350 votes — over incumbent Democrat Abigail Spanberger in the 7th congressional district race. Then, suddenly, the election scoreboard showed that she had leaped ahead with a 5,134-vote lead. The explanation, according to the Virginia Public Access Project: Election officials discovered 14,616 votes that had been overlooked on election night.

“Officials overlooked the ballots, which were saved on a memory stick mislabeled as ‘provisional ballots,'” VPAP tweeted. This batch of ballots broke 64.0% for Spanberger and 34.8% for Freitas.

The conservative blog Just The News highlighted the sudden shift this morning in a headline: “Virginia Democrat Spanberger surges ahead after ‘overlooked’ ballots found on flash drive: Report.”

I understand the alarm. Given the numbers provided by VPAP, we can calculate that the discovery of mislabeled ballots added about 9,354 to Spanberger’s tally and 5,086 to Freitas’. In other words, the mislabeled ballots netted 4,268 votes for Spanberger. Without those votes, her lead would be less than 1,000 — certainly enough to warrant a recount in an election in which 450,000 votes were cast. But this does not appear to be a case of crooked officials “discovering” missing ballots that never existed.

Here is the explanation that Henrico Registrar Mark Coakley gave the Henrico Citizen:

The ballots in question all involved in-person absentee votes, he said. When a voter casts a ballot in person, he or she scans the ballot into a machine just as voters do on Election Day, and memory sticks in the machine take images of each ballot.

Each machine has two sticks – a primary one and a backup, he said. Each 4 GB stick can hold about 9,000 ballots, while an 8 GB stick can hold about 17,000.

As more sticks began to fill up, Coakley suggested that (instead of buying new ones) officials scan more in-person absentee ballots using a machine that typically scans only provisional ballots.

On election night and well into the early morning hours of Wednesday, Coakley’s team was uploading the data from each memory stick used during the absentee process to produce an overall absentee report to send to the Virginia Department of Elections.

But, he said, the report’s parameters by default were set to collect only the data from absentee machine sticks – and not from the provisional ballot machine. As a result, the ballots that had been saved on the provisional machine memory stick (just shy of 15,000) were not part of that report, Coakley said.

In their haste to send the report to the Virginia Department of Elections shortly after 2:30 a.m. Wednesday, Coakley said, Henrico officials didn’t compare the data to an earlier report they’d produced Tuesday night that reflected a total of all in-person and absentee ballots they had received.

Later in the day Wednesday, officials from the Virginia Public Access Project noticed what appeared to be a 21,000-vote discrepancy between the number of absentee votes the county had reported to the VDOE and the total number it had identified Tuesday night. VPAP Tweeted about the discrepancy, prompting Coakley and his team to review their data immediately.

“When VPAP sent their Tweet, that threw up all sorts of red flags,” Coakley said, “because I knew we didn’t have 21,000 [completed].”

But he and others in the office then took the recording tapes from each machine that was used to scan ballots, compared the total number of ballots reflected on them and compared that with the data they had reported to VDOE. That’s when they realized that the memory stick from the provisional machine had been ignored during their earlier report.

“We would have caught it yesterday afternoon,” Coakley said Thursday, since such a reconciliation of data is a standard part of the reporting process.

Coakley’s seems entirely plausible, and it is noteworthy that the Freitas campaign has neither made an issue of the mislabeled ballots nor even found the incident worth issuing a statement on. I’ll keep my eyes peeled for updates, but at this point this seems to be a non-story. I hope the same can be said of the other allegations of election fraud around the country.