It is time for improving the process, surfacing problems, and strengthening the audits. It is not the time to call for weakening the audits and replacing manual counting with machine recounts that can only confirm erroneous results and provide false confidence.
For those who question the value of manual counting in audits and recounts, here is an example from Greenwich, where despite pre-election testing an error apparently slipped through. <read>
While the new optical scan voting machines won raves from poll station managers on Election Day, a machine error nearly cost a newly elected Representative Town Meeting (RTM) member his seat.
When results became official in the RTM races on Nov. 6, it initially appeared that Randall Smith from District 5 in Riverside had not been elected. With a vote total of 424, he was one of three candidates from the district not to be elected to the 19 seats District 5 has on RTM. However, a recount in the district found that votes that had been cast for him had mistakenly been given to write-in candidates by the machines.
By the time the hand recount was done, Mr. Smith had enough votes to make it and will begin his first term next month. This was the first year the optical scan machines were used in Connecticut and now Mr. Smith’s hoping for the state to make improvements to keep the problem from repeating.
Lets hope that it will not take errors discovered in their own elections to have State Senators, Representatives, and other politicians see the need to strengthen the election system and strengthen the audits. Rather than calls to weaken them <read> <read>
There is more good news and not so good news.
This error was not discovered by the audits nor the mandatory recounts. The error was discovered by the vigilance of a local chair. Perhaps it would be discovered another time by an election official. Perhaps not:
Mr. Smith said he gave a lot of the credit to the work of Mary Ferry, District 5’s chairwoman. Ms. Ferry told the Post that when the polls closed on Election Day, she went to District 5’s polling station at Riverside School to copy down the results. It was there that she first noticed the high number of votes going to write-in candidates, which struck her as unusual since there were no officially registered write-in candidates in the district race.
Here we have a specific problem. We have several questions:
- How did the problem go undetected in pre-election testing? Was the memory card tested by UConn before the election?
- Can UConn add this test to its automatic memory card testing? (We have no bones to pick with the UConn testing. It is quite a credit to Dr. Shvartsman and the Secretary of the State’s Office for implementing the testing so quickly). If so, can cards be re-tested to see if this problem appeared elsewhere in the recent election?
- Where did the process for programming the cards at LHS fall down? Or was it a mistake in information transmitted from Greenwich to LHS?
- What similar problems could have occurred across the state?
- What are the odds similar problems would be detected by our current level of audits? Given weaknesses in the audit expansion criteria would these errors have reached the threshold for triggering further investigations? I’m not sure, but to quote a local furniture executive, ” I doubt it”.
