[E-voting] experience with VVAT
cansbro at eircom.net
Fri Mar 24 18:44:11 GMT 2006
You've just discounted a huge issue--the supposed "undervotes". Those
weren't undervotes. There were massive irregularities in Nevada,
particularly on the Indian reservations, which is where those
"undervotes" occurred. People waited for hours to vote in that
Presidential election. The "undervotes" almost certainly indicate
malfeasance or malfunctioning--but no way to prove it, unfortunately.
There have been quite a number of articles written about the problems on
the Indian reservations--lots of disenfranchisement, through
worse-quality machines than elsewhere, and highly unusual undervotes or
highly unusual undervotes, neither of which fit in with any previous
pattern or with common sense.
Ditto for the machine malfunction--don't just ignore it as if it doesn't
matter. We go brain-dead when there are problems with a computer--we
just assume, "well, that's life, it's inevitable that these problems
occur" and we never look closer. This is itself a big danger of voting
that uses a black box.
Clearly, the VVAT in Nevada--one of the best-case states--had serious
problems. I haven't touched on the ones that were seriously problemmatic.
That said, you are correct that many if not all of these issues could
theoretically be solved by proper design, proper legislation, and proper
The design doesn't worry me. (I'm assuming here that Nedap/Powervote
have higher integrity than all the US vendors. That may not be a wise
assumption, but I'll give them the benefit of the doubt for the sake of
the argument.) What really worries me is legislation concerning paper
ballot as the official ballot, and probable restrictions to
hand-counting them. And that will lead to sloppy procedures (out of
practice) when they are hand-counted, and probably less interest in
ensuring adequate observation.
We're looking at considering using e-voting machines without the
necessary legislation agreed. Any legislation would almost certainly
favour the electronic ballot (e.g. by limiting chances at counting the
paper), otherwise why bother having the electronic devices in the first
place?! Which is the whole point. Why introduce multiple
vulnerabilities when it's not necessary and there are no real benefits
to voters? Voters don't need to know the results a day or two earlier.
Michael McMahon wrote:
> Catherine Ansbro wrote:
>> You should know better than to believe what the Secretary of State says.
>> Even in Nevada in 2004 there were problems. See here:
> There are six issues identified on that page.
> 1) 10,000 undervotes for the presidential race. This does not look
> like a VVAT issue to me.
> I thought that DRE software was supposed to prevent undervotes, or
> at least to bring it
> to the voters attention. I guess if the voter really doesn't make a
> choice then it has to be
> an undervote.
> 2) 271 votes not counted because machines in test mode. This raises an
> interesting issue.
> People need to learn that they have to check the VVAT print-out,
> and they need to learn
> to recognise what the print-out is supposed to look like. For
> example, if the print-out
> has "Test ballot" or words to that effect, then that clearly is a
> warning. If the VVAT printout
> looks authentic, then it shouldn't matter whether the software is
> working in test mode or not,
> because the print-out is (or should be) the official record.
> 3) The "toilet-roll" style printout. There is a genuine concern here
> relating to ballot secrecy, and
> that needs to be dealt with by improving the technology.
> 4) registration fraud. Not related to VVAT
> 5) Machine malfunction. Issue is similar to 2) Again, voters need to
> satisfy themselves that the
> printout is correct, before they press the "cast-vote" or whatever
> the equivalent action is.
> 6) same as 4)
> I think the problems in the US stem from the unseemly rush to add VVAT
> without fully thinking
> through exactly how it should be done. The legal anomalies relating to
> what happens to the paper ballots
> are a good example of this. It is ludicrous to go to the expense of
> mandating VVAT, but then
> not checking the paper record afterwards.
> The audit has to be a standard part of every election, and has to be
> done openly and transparently.
> Another thing I'd like to see, is a fallback procedure where if a
> printer jams, or other problem occurs,
> then a voter can cancel the procedure (so long as the vote is not
> confirmed yet) and
> can then fill out a paper ballot manually.
> - Michael.
More information about the E-voting