District 16 recount reveals that one county was plagued with thumb flubs and other voting errors – and it’s not the first time

District 16 recount reveals that one county was plagued with thumb flubs and other voting errors – and it’s not the first time

In the end, it came down to 19 ballots in Santa Clara County that ultimately made the difference in the hotly contested Congressional District 16 race — ballots that were never counted the first time around due to simple human error.

While a change equal to just a fraction of a percent of votes is unlikely to shift the entire results of most elections, a once-in-a-generation perfect tie in this case exposed the gaffes and fumbles by one county in the tabulation process.

What followed the March primary was two months of counting and recounting more than 182,000 votes in the race to replace retiring U.S. Rep. Anna Eshoo — a stunning saga that finally came to a close on Wednesday afternoon, with Assemblymember Evan Low emerging victorious over Santa Clara County Supervisor Joe Simitian by a margin of just five votes.

The veteran politicos had previously tied for second behind former San Jose Mayor Sam Liccardo until a voter-requested recount — which cost the Super PAC funding it about $270,000 — decisively changed the outcome of the election.

In San Mateo County, the machine recount affirmed the original tally of 38,416 votes, according to Jim Irizarry, the county’s assistant chief elections officer. One additional vote counted for Low came from a batch of seven challenged ballots that previously had missing information about postmark dates.

“Elections have to be perfect,” Irizarry said. “That’s the bottom line. There is no room for error.”

But Santa Clara County, which makes up the majority of District 16 and rescanned 143,726 ballots, saw much more variance between the two counts, adding 46 votes overall. In the end, Low picked up 11 votes, while Simitian gained 7. Each of the other candidates also gained somewhere between 1 and 7 votes, except for Julie Lythcott-Haims and Greg Lin Tanaka, whose totals did not change.

So why did Santa Clara County see such a change while San Mateo County essentially duplicated its results? Some of that comes down to a thumb flub.

Santa Clara County admitted that 19 ballots that should have been counted never made it into the original tally. Twelve ballots which had been scanned were inadvertently excluded from the count because the operator accidentally pressed the wrong button, canceling the batch instead of accepting it. The remaining seven were either misfed or jammed the machine and had to be scanned again.

Similar issues while scanning the ballots also caused operators to accidentally double-count three ballots. The county accepted seven challenged ballots, as well.

Assistant Registrar of Voters Matt Moreles said they didn’t have a record of which way those 19 ballots swung, and whether it would have broken the tie in the original count to begin with.

Philip Stark, a statistics professor at UC Berkeley who has developed election auditing processes, said counties must have strong cross-checks in place to avoid election errors. Counties can also voluntarily conduct additional audits, beyond the manual hand review of one percent of the vote that’s required by law following an election.

“The goal at the end of the day should be to determine whether, despite what went wrong, there’s strong evidence that the reported outcome is right,” Stark said of election audits. “That’s the whole notion of evidence-based elections. Stuff will go wrong, despite that, do you still have strong evidence that the outcome is right? If not, then morally you should have a do over.”

Moreles acknowledges that the 19 ballots “slipped through the cracks.”

“I think it’s important to put it in context, we’re talking about a very tiny number of ballots relative to the entire universe of ballots we have to process for every election,” he said.

The assistant registrar said that while they have checks in place, the Congressional District 16 recount has been a learning process, and they are now looking at what additional controls they can put into place to avoid future mistakes.

The Santa Clara County Registrar of Voters has been plagued with a number of issues over the last decade-and-a-half. In 2017, they were the subject of a state audit, requested by none other than Low, that found from 2010 to 2016 the county had identified at least 26 errors in election materials. Meanwhile, election officials in Alameda County told auditors they had less than five errors in a seven-year period while San Bernardino County had one and Orange County election officials reported none.

The errors ranged from mapping issues that caused the ROV to provide some voters with voter information guides and ballots for districts they didn’t live in to omitting key information in election documents. In one case, the ROV in 2014 didn’t publish an argument against a ballot measure in the voter guide that was was sent out to more than 415,000 voters.

In another instance in 2018 that occurred after the audit, a voter was accidentally unregistered when an employee was trying to resolve a duplication in registration records.

The audit forced Santa Clara County to ensure “procedures were strengthened, and our documentation was stronger,” Moreles said.

In recent years, as election integrity has been questioned, there’s been a movement sweeping across the nation to ditch the use of tabulation machines in elections and instead opt for counting ballots by hand.

But Kim Alexander, the president of nonpartisan election watchdog group California Voter Foundation, said that isn’t a solution. If the ballots in the Congressional District 16 were recounted by hand, she thinks “the changes in the results would have been astronomical, because people do make mistakes.”

“I think some people might react and say, ‘we can’t trust technology’…but what came through to me was that we have imperfect human beings running elections who are going to make errors,” she said. “And we have to have voting systems that are resilient and can protect against risk.”