Yesterday Sarasota County, FL, began their state-run test audits of the paperless ES&S touch-screen systems used in Florida's 13th District House race where 18,000 votes disappeared, resulting in a 369 vote margin between Christine Jennings (D) and Vern Buchanan (R) to fill the seat vacated by former Secretary of State of Katherine Harris.
The first day of testing revealed miscounting errors on all four of the machines used during the test, according to the Sarasota Herald-Tribune and at least two citizen observers who blogged their account of the day's testing.
State officials who convened the panel blamed the miscounting on...wait for it...human error!
Of course they did. The panel itself has already been plagued with controversy after Election Integrity organizations criticized the selection of several members assigned to the commission, including the very partisan Alec Yasinsac, who appeared on the steps of Florida Supreme Court in 2000 wearing a "Bush Won" button, and David Drury, the state official who certified these particular systems in the first place for the state of Florida, declaring them safe and accurate for use in an election.
The first day of testing employed machines that were not actually used in the November 7th election (Friday's testing will include machines actually used that day) and were not performed under normal voting conditions (e.g., the machines were hung from the wall, instead of placed on stands to approximate the height they would be for most voters --- important, since these machines tend to register votes differently for tall voters than they do for other voters.)
Nonetheless, errors in the counting were registered by all of the machines used in testing. The results differed from the votes in "scripted" ballots that state testers had punched into them. Ironically, the errors ended up giving Jennings a higher vote total than she should have had according to the prepared voting scripts.
The Herald-Trib --- who, by the way, has reinstated its FL-13 Special Coverage page since we pointed out its removal a few days ago --- is reporting today that, "All four voting machines that officials used to simulate the Nov. 7 election had miscounts, and three of them had miscounts in the District 13 race."
A diarist over at DailyKos who has been on-site, covering the day-by-day details of the goings-on --- and who has a user name we well approve of --- "BeTheMedia" describes yesterday's results this way:
...
While today's procedure may have tested for a narrow range of possibilities, it failed to test broadly for a wider range of possibilities that could have been revealed by more rigorous and thorough testing. It is valid and reasonable for testing to try to make a system fail. That's what thorough testing is supposed to do.
The tests today did indeed revealed discrepancies. At least three votes recorded as undervotes in the election changed to Jennings votes in today's "audit." That's comes out to an approximate 7% shift in a race decided by less than two-tenths of one percent. It will be quite interesting to see how the State Division of Elections explains this shift.
And, of course, it is interesting to see how the State explained the shift, though their explanation is no longer much of a suprise...