Melinda Henneberger at Huffington Post runs a news item today concerning an August 2006 letter from the voting machine company ES&S to Florida Elections officials warning about a defect in the iVotronic touch-screen voting machine which succeeded in losing the votes of some 18,000 voters in the razor-thin election between Christine Jennings (D) and Vern Buchanan (R) in Florida’s 13th U.S. House Congressional district. Buchanan was provisionally seated, pending a Congressional challenge and state lawsuit filed by Jennings, after he was declared the “winner” by 369 votes.
The Sarasota, FL, Election Supervisor, Kathy Dent, decided against both having their machines patched to take care of the defect and posting a warning notice for voters as advised by ES&S. “No one in the State of Florida updated,” their machines after receiving the letter, Dent told Henneberger. “That’s because it was too close to the election. It was a state decision that it was too late to make changes.”
Notably, as you’ll see in the full letter as posted below, it was CC’d to David Drury, the Chief of Florida’s Bureau of Voting Systems Certification. Drury, responsible for overseeing state certification of voting systems, in an extraordinary conflict of interest as we’ve noted several times previously, was part of the state team commissioned to audit the iVotronic systems used in the election. The commission was empaneled after the state was forced to relent and launch such an investigation in light of the controversy after their initial denials that there was any problem at all with the extraordinarily high undervote rate in the Sarasota-only section of the race.
The ES&S letter, as Henneberger noted, was only seen recently by Jennings’s attorneys since “it was not provided to them by election officials as it should have been under discovery motions in the case,” according to one of her attorneys.
Instead, the legal team came across the document only recently as posted “on a North Carolina-based website on election reform,” according to Henneberger.
That website, as it turns out, is the NC Coalition for Verified Voting as founded by Election Integrity advocate Joyce McCloy. McCloy had attempted, both before and after the election, to get the attention of elections officials and the Jennings legal team. The letter, not linked by Henneberger, is posted here [PDF]. We’ve also posted the letter in full, as well, at the bottom of this article.
McCloy wrote The BRAD BLOG today expressing her frustration at trying to get anyone in Florida to take notice of the letter warning of “slow response times” to voters’ attempts to select candidates on the ballot.
“I sent that memo (and my concerns) to anyone I could think of,” McCloy explained in her email, “and this year to every election reform list serve that I could, posted it on political message boards. I asked and asked – has this bug been fixed?”
Apparently the answer is no; the bug was neither fixed, nor did Dent bother to warn voters in Sarasota about the problem as had been advised in the letter by ES&S.
As Henneberger reports today…
“The improvement will require an update to the firmware, and state-level certification,” in advance of the general election in November, the letter said. Meanwhile, for early voting, “In order to avoid any potential issues at the polls…it is our recommendation that you train your poll workers and voters to expect this slightly delayed response time…We have included with this mailing a sample voting booth instruction sign for your review and use.”
The poster advised voters that it might be necessary to hold their selection on the touch screen for several seconds in order to register a response.
But Dent said she decided against using the posters “because we already had instructions on voting on the touch screens” displayed in the voting booths, “and we hadn’t had that problem.”
Emails from Dent dated August 24 indicate that she not only decided against using the signs warning voters that there might be a delayed response – but instead used signs giving the opposite message: “Touch Screen Voting, Easy as 1, 2, 3.”
Dent continues to maintain that the reason for the inordinately high undervote rate (nearly 15%) was due to a combination of voters choosing not to vote for either candidate (which ignores the fact that the undervote rate was normal in other county tallies for the same race, and even among Sarasota’s own absentee voters who used paper ballots) and bad ballot design.
McCloy rejects the theory that bad ballot design was the culprit, as Mecklenburg County, NC, which also used a version of the ES&S iVotronic, had normal undervote rates despite a ballot design which may have been even worse than Sarasota’s. As McCloy posted in comments late February on the Princeton University computer science professor Ed Felten’s website:
Here is Sarasota FL 13’s ballot http://www.ncvoter.net/downloads/sarasota_ballot_style.pdf
Now take a look at what appears to be a more confusing ballot style for Mecklenburg County, NC, the NC 08 ballot:
http://www.ncvoter.net/downloads/Mecklenburg_2006_ballot.pdf
(notice the nearly hidden placement of the US congressional race?) Meckelnburg had a 4 % undervote rate for that contest.
Here is a memo from the NC State Board of Elections explaining the differences in NC iVotronics and the FL iVotronics, as well as a ballot comparison:
http://www.ncvoter.net/downloads/Sarasota_NC_Ballot_Comparison_06.pdf
Despite the ES&S letter’s warning that “This delayed response to touch may vary from terminal to terminal and also may not occur every single time a terminal is used,” the state’s official audit — which to the surprise of almost nobody was spun by state officials as offering a clean bill of health for the paperless touch-screen systems — does not seem to have investigated this particular issue.
The August 15, 2006 letter from ES&S to Florida Elections Officials follows in full below…
In the immortal word of Charlie Brown, “Aaaaaauuuuuuggggghhhh!”
…and away goes…
Democracy down the drain……….
It should be blindfold and firing squad time for Dent.
Mistakes were made.
NOW HOW ABOUT A RE-VOTE? Yes, that’s right, a precedence setting RE-VOTE!
Brad,
With all due respect, if this particular problem contributed to the undervote in FL-13, it’s impact was trivial at best and not demonstrably outcome altering as a voter would be just as likely to have a problem getting the machine to register a vote for Buchanan, the Republican, as for Jennings, the Democrat.
Clare Ward-Jenkins’ real experiences as a poll worker are, in my opinion, much more helpful for those of us trying to get to the bottom of what went wrong in Florida 13.
Mike Ridgway
#4 Mike Ridgway
The problem is the electronic machine. On it’s own, it can not be validated. It just can’t, it’s physically impossible. I believe your wrong, since it can not be validated one way or another, you can’t really know what the hell it is doing or executing or outputting. Hell there could be a blown logic chip on it and you wouldn’t know. It could very well have chips in it that say one thing on the part number, but are doped at the manufacture level to do something entirely different.
I am not talking about simple keybounce delay timings. Any idiot that has played with the electronics inside a keybord 8048 or tried to make a keypad knows that crap.
…and in local news, where I live, ES&S representative says, “Vote flipping “” there is no such thing as that,” said Willie “Wes” Wesley Jr., a regional accounts manager with Nebraska-based ES&S. “It’s really a calibration issue.” And gives the machines “the green light”…
Get that? An ES&S representative says THEIR machines are “allright”…the fox watching the henhouse!!!
http://www.citizensvoice.com/si...154&rfi=6
With all due respect the follow quote by #4 Mike Ridgway is not correct.
“With all due respect, if this particular problem contributed to the undervote in FL-13, it’s impact was trivial at best and not demonstrably outcome altering as a voter would be just as likely to have a problem getting the machine to register a vote for Buchanan, the Republican, as for Jennings, the Democrat.”
That quote is only true if Buchanan was as likely to get a vote in FL-13 as Jennings was. That is not the case. These machines almost always fail in Democratic strongholds for a reason. The reason is to prevent mostly Democratic votes from being counted. If random gliches were responsible for machine failures those machine failures would be just as likely to occur in Republican strongholds. The fact that this is far from true in real world voting in the USA, tells any expert in probability that somebody has a vested interest in many of the “gliches”.
Mike Ridgeway’s argument is not quite correct, as Bob Young points out. But I think his conclusion is correct. The smoothing filter was applied to every part of the touchscreen. It would have affected every screen touch in every race on every page of the ballot equally, and yet only the CD13 race had the huge undervote. Moreover, it would have affected every jurisdiction in the U.S. that used the same hardware/firmware combination. Thus, the smoothing filter cannot have been anything more than an insignificant contributory cause of the CD13 undervote problem. The code review team considered the smoothing filter hypothesis and were able to eliminate it.
Regarding the conduct of the Florida election officials: It is true that they knew about the filtering problem in advance, and did nothing about it. But their options were limited. It would have been impossible to fix the software and get it recertified before the election with they time they had, and under the circumstances you don’t want to abbreviate that process. Probably every other state would have made the same decision. The truth is that handling software problems that arise within a few months of an election is a profoundly difficult problem, for which there is no generally good solution. You can make a case that this argues for less, or no, software in the vote capturing process. And it is certainly evidence that the testing and certification process is broken, since this problem would likely have been caught, for example, in volume testing. But it is not an argument that officials made the wrong decision not to order a last minute fix of the software in this case; I think that decision was right.
Election Supervisor Dent’s choice not to put a sign in the voting booths warning voters of the filtering problems and the poor responsiveness that might result is certainly questionable, as is an awful lot of her behavior before and after the general election. But these issues really have nothing to do with the CD13 undervote other than adding irony, and I would hope we would keep them separate.
Running some numbers in Fl -13:
In Fl-13 Vern Buchanan got credit for 58,632 votes or 47.24% of the votes that got “counted” there.
In Fl-13 Christine Jennings got credit for 65,487 votes or 52.76% of the votes that got “counted” there.
Based on other under vote rates over 15,000 of the 18,412 Fl-13 voters, who were credited with an under vote, almost certainly intended to vote.
If the “glich” was distributed evenly by the contestants percentage of the Fl-13 vote the result is far from trivial.
47.24% of an additional 15,000 votes amounts to 7,086 more votes to Buchanan.
52.76% of an additional 15,000 votes amounts to 7,914 more votes to Jennings.
That amounts to 872 more votes for Jennings than Buchanan in an election Jennings reportedly lost by 369 votes.
Jennings would likely have been the winner by more 503 votes over Buchanan if not for this “glich” that created over 15,000 extra under votes in Fl-13.
Trivial??? You make the call!
Well, Diebold does it too
“Testing is fine, but I would try to talk them out of using it live. It was a “Florida Thingâ„¢” only. Curtis Chong of the National Federation of the Blind absolutely hates the feature, and I tend to agree with him. I added it under duress, and feel badly any time I see it being used.”
And
“Tari, what on earth are you using VIBS_Undervote for? Turn it off in the AVTS options.”
another E-Mail exchange from the dark side
Link
Florida media weighs in,..
St Petersburg Times
Sarasota officials ignored warning about voting machines
An alert on glitches came ahead of recent election troubles but no one followed up.
By ANITA KUMAR
Published March 15, 2007
Link