Update: The Secretary of the State’s Office was offered four hours to comment on this entry before publication, but said they needed more time to gather information. We will post their response promptly.
New York Times: Voting Machines Are Put To The Test – reviews the optical scan performance in the municipal elections, with information from Secretary Byseiwicz <read>
I found a couple of items quite interesting. The following is not a direct quote from Secretary Bysiewicz but from the context seems to be information she discussed with Times. There is no indication of any other source associated with the information:
In a report after that audit, UConn researchers found, after a hand count, that in at least one race, the machine at a polling place in East Hartford counted six more votes for one candidate than the person had actually received. The discrepancy did not change the outcome, but they warned that it could in the future.
No such problems have turned up in 39 recounts of the Nov. 6 election,
This seems contradictory to the recent news describing recounts that resulted in a change in 17 votes in New Caanan and another error in Riverside that overturned the results. I suppose it depends on what is meant by “No such problems”, these may be different in some ways from those in East Hartford.
Also for the first time we learn that UConn has discovered some issues with the cards tested in the pre-election test. It seems that in some cases either a set of blank cards were sent to registrars or sets inconsistent cards, some blank, were sent to registrars.
Mrs. Bysiewicz said that in addition to reviewing the audit results of this past election, the University of Connecticut audited more than 300 memory cards prior to the election and will audit another batch now that the election is completed. The memory card, which she described as the “brains†of the new machines, posed the biggest security concern for UConn researchers, who warned in October 2006 that if someone gained access to the card, the results of an election could be altered.
Mrs. Bysiewicz said the pre-election audit of the cards found that none of the cards had been tampered with, but that a handful were blank, meaning that they had not been programmed with the proper ballot information.
No report yet on how many blank memory cards were discovered in pre-election testing by the registrars. If the same percentage of 6 out of 300 holds, then if each registrar in 695 towns tested at least two cards then about we could expect that about 28 would have been discovered.
