D.C. voters had to wait four hours after polls closed Tuesday before receiving meaningful election returns after problems with a handful of voting machines led to an unusually lengthy and chaotic tabulation process.
Here’s a timeline of events leading well past midnight, when Board of Elections officials finally explained what went wrong.
8 p.m. Polls close.
Waiting for first ballots to arrive. The election results URL will go live as soon as they do. #ElectionsDC— DCBOEE (@DCBOEE) April 2, 2014
About 9:40 p.m. No results have been announced. Board of Elections spokeswoman Tamara Robinson, who has been shut away from the media room for most of the night, says the first results will be released in about 15 minutes. She says that the results are being counted more slowly than she would have predicted but that nothing has gone wrong.
We appreciate your patience. Precincts are trickling in, and as soon as the votes are counted, we'll post an updated URL.— DCBOEE (@DCBOEE) April 2, 2014
About 9:55 p.m. At the same time, the BOE posts a tabulation on its Web site of about 9,000 ballots and hands out a packet in the media room tallying 10,540 ballots. Clifford Tatum, the board’s executive director, says that the paper packet is more up to date and that a technical problem has held back the Web site from displaying all the results. He does not answer, however, which votes are missing from the online tally and why there are 4,000 fewer early ballots than the board had announced earlier. News outlets, including The Washington Post, report that those 10,540 votes come only from certain precincts.
About 10:40 p.m. Tatum enters the press room, flustered, and says repeatedly, “This is why we don’t rush things.” He seems to be referring to the discrepancy between the numbers on the BOE Web site and the paper packet distributed to reporters by his staff. He hands out a new packet, the first to include actual Election Day votes, but he has printed only one or two copies of it, so most people don’t get one. The packet claims to include results for 55 of 143 precincts, but the vote counts — 11,474 for Bowser and 8,786 for Gray — seems low in a city where more than 130,000 voters cast ballots four years ago.
Tatum answers only a few questions and leaves without explaining that the figures reflect only the paper ballots cast in those precincts and not the electronic ballots, which had not been tallied because of problems with electronic voting machines that he would later explain.
The Bowser campaign members in the room declare victory. Campaign chairman Bill Lightfoot and supporter Terry Lynch clasp hands and pat each other on the back, saying “It’s over.”Asked if he’s sure that Bowser won, Lightfoot says, “I’m going to go out on a limb and say yes.”
From 10 p.m. to 11:20 p.m. For nearly an hour and a half, with the exception of Tatum’s one visit to the press room, all BOE personnel remained behind closed doors. Reporters and campaign teams waiting for results grow increasingly frustrated. Virtually all questions posed to BOE staff, from the people at the front desk to those bringing votes up the elevator, were referred to Robinson. Knocks on closed doors were met with staff members who referred all questions to Robinson. BOE staff consistently stated that Robinson was in a different room and that no one else would speak to the press. These staff members repeatedly said that there was no problem or that they did not know if there was a problem.
About 11:20 p.m. Robinson hands out a packet that claims to include 116 precincts but only counts 45,180 Democratic votes — again, a seemingly low figure for that number of the city’s total precinct count. For the first time, Robinson states that there has been a problem shutting down electronic voting machines in “five or six” precincts, a necessary process before vote counts can be extracted from the machines. The results she is handing out, therefore, include only early votes cast before Election Day and Election Day paper ballots.
About 11:55 p.m. Robinson hands out a packet including votes from 127 precincts. For the first time, these include Election Day results from touch-screen voting machines, she says.
About 12:40 a.m. Robinson names the five precincts that caused problems. Even in that conversation, she initially says with certainty that six machines caused problems. When it becomes clear that she has numbers only for five precincts and can’t seem to answer the question of whether two machines were shut down incorrectly at one of the precincts, she adjusts and says there were definitely five problematic machines. She says the complete count will be finished in 10 minutes.
Board of Elections workers actually drove out to the five problematic precincts to pick up the machines that weren't shut down properly.— Julie Zauzmer (@JulieZauzmer) April 2, 2014
About 12:55 a.m. The next packet of returns arrives and includes 142 out of 143 precincts.
About 1:10 a.m. With the counting complete at last, Tatum emerges and talks to reporters from The Washington Post and WAMU to explain the evening. In the middle of the conversation, he sits down. He says he’s very tired. He says he had hoped to finish by midnight, but he doesn’t think there were any major disappointments in the way the night went.
Tatum explains the process of correctly counting votes from electronic voting machines. The machines first must be shut down. Shutting down an electronic voting machine takes about 10 to 15 minutes, and shutting down two machines is more complicated than simply doing the same process twice. Workers must transfer a memory card from one machine to the next and then print a paper audit showing both machines’ votes totals. Those audits are then delivered to the board’s central office for counting.
“Some of our workers have admittedly never touched laptops before,” @DCBOEE ED Clifford Tatum said, abt teaching poll workers to use tech.— Julie Zauzmer (@JulieZauzmer) April 2, 2014
BOE election workers responsible for counting votes at the board’s headquarters realized something was wrong because machine serial numbers were missing from five of those paper audits — a sign that the vote totals did not include votes from those machines. Several workers drove back to the precincts in question to shut the machines down again and extract the correct totals.
In the sit-down, Tatum blames the problems on the difficulty of training poll workers (“some of our workers have admittedly never touched laptops before”) and says the board will consider asking the D.C. Council for money to buy newer machines that are easier to shut down next year.
Tatum also explains why the very first vote total of the evening did not include all early votes cast before Election Day. He says that the program that converts the board’s ASCII results from ASCII to a CSV file for the Web site broke before it was able to count the paper ballots for early votes, and that an IT guy fixed it during the evening. He says that the 4,000 missing votes were absentee ballots and provisional ballots, which may be counted as late as April 11.
Note: This timeline relies partially on papers distributed by BOE officials that were incorrectly stamped with the wrong time. As a result, time references are approximate. It appeared that some of the timestamps on the vote reports were in the wrong time zone.