D.C. elections officials offered an entirely new explanation Tuesday for the major vote-counting delays that plagued the city’s April 1 Democratic primary: The issue was not five mishandled electronic voting machines, but a broad computer network failure.
The network failure was a mystery to elections officials as it unfolded, said Clifford D. Tatum, executive director of the Board of Elections. But its effect was abundantly clear to all involved on election night, when vote-counting — including ballots the city had accumulated during weeks of early voting — did not begin until almost 10.
Deborah Nichols, chairwoman of the elections board, said that at least $2 million in new electronic voting machines and server upgrades — and perhaps another $2 million in computers and other office improvements — would be needed to ensure timely reporting of results in future citywide elections.
“That’s different than what was reported previously,” said D.C. Council member Kenyan R. McDuffie (D-Ward 5), who as leader of the council’s oversight of the board called a hearing to scrutinize problems with the election. “At what point did you determine that there was a problem?”
“I know it sounds preposterous,” Tatum said, explaining that on election night, polling officials never really did determine the problem. “I was standing over the machine, saying . . . ‘What’s taking so long with these results?’ We were waiting and waiting, but it never dawned on us that we were having a major issue.”
The problem tallying results from the city’s touch-screen voting machines — which were first used in District elections in 2010 — was the biggest in a series of cascading election-night failures, Tatum acknowledged during the hearing.
Poll captains’ calls for help went unanswered at elections headquarters because of an overburdened phone system; a coding error repeatedly published inaccurate tallies on the board’s Web site; and even such simple mistakes as distributing pens — and not pencils — to all 143 polling precincts sent staffers scrambling the day before polls opened to keep from botching the paper ballots.
In all, the failures produced an hours-long delay in the tallying process that made it impossible to call D.C. Council member Muriel Bowser’s victory over Mayor Vincent C. Gray until nearly four hours after polls had closed.
The night added to a nearly unbroken string of troubles for the Board of Elections.
In the 2008 presidential primary, officials ran out of paper ballots, and in a city primary later that year, officials found inexplicable discrepancies in vote tallies that muddied the outcomes of two council races.
In the 2010 mayoral primary — the first year for early voting and same-day voter registration — a six-hour delay in reporting triggered accusations of mismanagement.
This year’s problems came even as fewer voters turned out for a D.C. mayoral election in any year since 1998.
Dorothy Brizill, a longtime civic activist and poll watcher, called the board’s troubles this year the worst she’d seen. She also announced that she had formed a new Citizens Committee for Election Reform and asked presidential appointees to chair the effort to keep voting in the nation’s capital from growing into a larger embarrassment.
Nichols bristled at the criticism, calling the April 1 primary a success in terms of accuracy.
“The BOE’s performance during the April 1 election and early voting, although not perfect, was in my assessment successful,” she said. “All the statutory requirements and obligations were met.”
“Timeliness,” Nichols added, “is only one factor.”
She put the blame for the delays squarely on the city’s touch-screen machines, which she said were purchased used and have 12-year-old technology.
To meet the D.C. Council’s demands, she said, the board more than doubled the number of touch-screen machines in use April 1 to more than 300. Keeping those machines functional, however, proved to be complicated.
Officials had misjudged how often the backup tape in each machine should be replaced, and the process flustered poll workers, Tatum said.
On election night, Tatum said, many D.C. poll workers had “admittedly never touched laptops before.”
As results continued to trickle in after 11 p.m., officials told reporters that the problem centered on five electronic voting machines from which results were not properly extracted. When vote-counters noticed the problem on the machine printouts delivered to election headquarters, they stopped releasing electronic results and fanned out across the city to check the machines and retrieve accurate counts.
On Tuesday, Tatum said the results from the five machines did not delay the election night tally.
“The impact was the first hour out of the gate” tallying the early voting results. “The first hour set the whole tone — that pushed everything back,” Tatum said.