I have regularly described the hypothetical drip-drip scenario to people around here for a few years now. First I've heard of the study below.
CB
[lbo-talk] kjkhoo at softhome.net kjkhoo at softhome.net <mailto:lbo-talk% ________________________________
At 3:55 am -0500 3/11/04, snit snat wrote:
>At 01:50 AM 11/3/2004, joanna bujes wrote:
>>I don't actually believe that a big turnout is consistent with a
>>Bush victory. There was a big turnout, particularly of minority
>>votesrs....and a Bush victory.
>>
>>It's not that I overestimate the American electorate; it's just
>>that it doesn't add up.
>
>
>i don't frickin' believe it either. 1/3 of the population was voting
>early and it still looked like the turnout was up at the polls. Then
>I read that, wherever there is an audit trail, the exit polls match
>up, but wherever there isn't one, they don't?
>
>And, on the radio, people were reporting that they'd vote, check the
>screen and see that everything had been changed to the republicans.
>One woman tried three times and then complained. Sure enough, she
>was doing it right and something was fucking up by the time it got
>to the final page. The reporter--and they really stay away from
>conspiracy theory on this station--said that they figure there was a
>subtle mechanism to throw out votes every so often, drip-drip-drip,
>and maybe no one would notice.
For some comfort, the ff. article appeared in the recent CACM Oct 2004 --
ftp://ftp.cs.yale.edu/pub/TR/tr1285.pdf
with abstract:
We examine the effects of a type of electoral fraud easily perpetrated by someone with access to the system software for a direct-recording electronic voting system. The hypothetical attacker modifies the software to arbitrarily change a small number of votes on each voting machine. We determine the effect of this manipulation on the electoral college results of the 2000 United States Presidential election. We conclude that changing only a single vote per electronic voting machine can change the outcome of the election.