Cambridge city staff members knew on election night that there was a problem with the results for the School Committee, but did not inform the four-member Board of Election Commissioners who voted to release preliminary unofficial results to the public.
That led to Eugenia Schraa Huh being announced as one of the six School Committee members, ahead of David Weinstein by 115 votes. Friday’s unofficial results, accounting for the thousands of ballots not counted on election day, showed Weinstein elected over Schraa Huh by 95 votes. The error in preliminary results was not a machine miscounting of physical ballots, but rather the result of digital test data not being cleared out of the system.
At some point – it’s unclear when – Election Commission staff realized the problem was data from the “test deck” used to verify the machine tabulators are properly functioning.
Test deck data was mistakenly included in the reported results, according to Jeremy Warnick, a spokesperson for the city of Cambridge. “The possibility of a discrepancy was identified Tuesday night, but ultimately could not be verified until Wednesday,” Warnick said.
One clear indicator of an issue with the ballots on election night was that the results included some write-in votes. The machines that count ballots on election day don’t read write-ins, but the results included 33 write-in votes. A digital file of ballots from the test deck was mistakenly combined with the digital file of real ballots to produce the results that were tabulated and reported on election night.
Warnick said the issue was confirmed in an audit of results on Wednesday.
Problems not conveyed
The four commissioners – two Democratic and two Republican, nominated by their respective city committees and appointed by the city manager to oversee elections – were not told by commission staff about a potential problem on election night.
“The Election Commission felt it was important to release what was available at the time rather than delaying the first round of preliminary unofficial results another day (or longer),” Warnick said. “It was not clear if the test ballots were going to have any impact on the candidates announced in the second set of unofficial results until the write-in and auxiliary ballots were factored in.”
But, write-in and other auxiliary ballots almost never change the preliminary results. For the school committee, many of the 3,000 auxiliary ballots would likely have been blank, because in a typical election about one in ten of Cambridge’s voters don’t cast votes for the school committee.
The city’s voting machine vendor, LHS Associates (who resells Dominion Voting Systems machines and supplies the high-speed tabulator), did not respond to inquiries.
Discrepancies in disclosure
The first public announcement of the problem was Friday evening, November 7, when the results were updated to reflect 5,324 ballots not counted on Tuesday.
Tanya Ford, executive director of the Election Commission, told the commissioners that “some ballots from the test deck that were included in the results. It didn’t change city council, but it changed school committee.”
There also appears to have been discrepancies in how the error was communicated – Schraa Huh was notified about the test deck problems before the Board of Election Commissioners were. But Weinstein said he was not told until after the update.
School Committee member Elizabeth Hudson said she also was aware of the issue by Friday morning. A very pregnant Hudson (due in one week) was at the Senior Center on Friday to hear the updated results, as were Trish Marti, Patty Nolan’s campaign manager, and Robert Winters, a candidate and author of the Cambridge Civic Journal website.
Schraa Huh declined to say how she had learned of the problems in advance. “I’m not going to talk about it at all, I’m sorry,” she said. “I just don’t see that this is an important story.”
Schraa Huh said she didn’t know if she would ask for a recount, but did say she had collected 193 supporters in favor of one, well over the 50 signatures required to make the request. She has until 5 o’clock Monday Nov. 17 to request one. A recount would require all the ballots to be hand-counted.
Hudson said Tuesday that the city’s explanation of the error was not sufficient. She said a number of city councillors and school committee [candidates] have asked for a more complete explanation.
About the test deck
Massachusetts law requires election tabulator machines to be tested at least four days before the election. And the test must include at least 50 ballots marked to “adequately test the computer program” and show the machines will reject overvotes – two votes for the same candidate or two candidates ranked with the same number.
Cambridge municipal elections have 33 precincts, each with two precinct tabulator machines, meaning 66 machines must be tested. Election commission staff posted a notice indicating the test deck would be run beginning on Monday, October 20 at 8:30 a.m., and would continue “from day to day” until completed. (That notice does not appear online on the city calendar.)
There is also a central high-speed tabulator, which was used to process mail-in ballots on Friday, Oct. 31 and also used to process auxiliary ballots in the days following the election, Wednesday Nov. 5 through Friday Nov. 7.
Warnick said the “high speed scanner test data was mistakenly” included in the unofficial results. Election records provided to Cambridge Day by the Election Commission included two types of test deck data: a record of the individual paper ballots as counted by machine and a “results report” for each precinct, the high-speed tabulator’s equivalent of the machine tape printed by the precinct tabulators.
The individual paper test ballots across 33 precincts were 244 per precinct, 125 for city council and 119 for school committee, or 3,927 School Committee ballots. That’s opposed to the 1,370 School Committee ballots that the city reported were incorrectly included on Friday.
The test ballot data is the same in each precinct, 12 number one votes for Richard Harding, 6 number one votes for Eugenia Schraa Huh, and 3 number one votes for David Weinstein, giving Schraa Huh 99 extra number one votes.
Asked about potential unequal treatment of candidates in the test deck, Lesley Waxman, the assistant director of the Election Commission, said, “It wouldn’t be a very good test deck [otherwise].”
The results report also had discrepancies. The tests included 98 ballots per precinct across the city council, school committee and ballot question contests, 42 of which were for school committee. That adds up to 1,386 School Committee ballots, different from the 1,370 ballots the city said were included erroneously.
Warnick said Monday night he would work with the Election Commission to see what contributed to the discrepancies between the three sets of numbers. The city was closed Tuesday for Veterans Day.
This story has been corrected to reflect the correct number of individual test ballots for city council and school committee, and the total number of School Committee ballots. Cambridge Day regrets the error.
A stronger
Please consider making a financial contribution to maintain, expand and improve Cambridge Day.
We are now a 501(c)3 nonprofit and all donations are tax deductible.
Please consider a recurring contribution.
> “I just don’t see that this is an important story.”
I’m sorry, you don’t see that the city knowing about incorrect results, and *telling you* about incorrect results before the commissioners of the Election Commission (or the public in any way), is an important story?
Shoulda stuck with “no comment, sorry.” :p