Information Technology and Voting

ChelleChelle writes, "In an interview in ACM Queue, Douglas W. Jones and Peter G. Neumann attempt to answer the question: Does technology help or hinder election integrity?" From the article: "Work in this area is as politically loaded as work on evolution or stem cells. Merely claiming that research into election integrity is needed is seen by many politicians as challenging the legitimacy of their elections... One of the problems in public discussions of voting-system integrity is that the different participants tend to point to different threats. Election-system vendors and election officials generally focus on effective defense against outside attackers, usually characterized as hackers. Meanwhile, many public interest groups have focused on the possibility of election officials corrupting the results."

Re:Can't beat paper votes and scrutineers.

[DarkBlackFox]

Where I voted this morning, we had paper ballots that were fed into an optical scan machine (by Diebold). The ballot was handed to me after I checked in, and each polling station had a hidden desktop with a felt-tip pen. All I had to do was fill in the circles corresponding to whatever candidate I wished to cast my vote for (much like a standardized test). When I was done filling in the circles, I took it over to the big machine, where a poll worker watched me insert it, and made sure the machine processed it correctly.

Seemed simple enough. The ballot itself has no personally identifiable information on it, and if there's a dispute or need for a manual recount, the paper ballots are presumably easy enough to get out of the machine. So these machines do have a paper trail, no potential for miscalibrated LCD screens, and have a ballot that's generally hard to screw up (fill in the circle).

The only way to really screw up a system like that is in the optical recognition software, which I'd hope is tested by poll workers before the polls actually open. And even then, with the paper ballots being retained inside, it's easy enough to do a manual recount.

Before this year we had mechanical voting machines, where you'd have to walk in, close the curtain by pulling a big lever, push all the levers down for the candidates you'd like to vote for, then pull the big lever back again to open the curtain and cast the votes. That was also easy enough, but mechanical machines like that are prone to more common failures, which was the primary reason for going with the electronic machines.

Re:Motives

[peacefinder]

Ars Technica [arstechnica.com] detailed a plausible transmission path for a viral hack on a single machine to spread to the precinct, county, or even state level.

Re:Can't beat paper votes and scrutineers.

[Chandon Seldon]

Optical scan vote counting is potentially a good idea, but it leaves a loophole compared to hand counting the ballots.

Consider the following situation from the 2004 election:

In one of the counties in Ohio, computer counted ballots were used. When a presidential candidate challenged the results and suggested a recount, the county election officials first recounted a random 3% of the county votes as required by Ohio state law. When that 3% showed no discrepancies with the computer vote totals, the recount was stopped.

Now, there's a perfectly reasonable statistical argument for the 3% law - except for a couple issues. First, if there was vote tampering limited to a few select polling stations, it will only be noticed if the 3% selected includes those stations. Second, in the Ohio case, the election officials manually selected the 3% that was recounted - note that it not being actually random completely destroys the statistical argument *and* if the officials were in on it would allow them to make sure that there was no chance of known-questionable stations being recounted.

Conclusion: Computer counting is fine. It saves time and man hours. But... if there is *any question at all* about the validity of the computer results, the procedure needs to specify a complete hand count - and that count needs to actually occur. Anything less provides opportunity for voting fraud.

Re:High Treason

[Watson Ladd]

Unfortunately, treason is a very narrow crime in the US. You have to give aid to an enemy of the United States. So it wouldn't stick.

Doug Jones website:

[sakusha]

I have posted links to Doug Jones' website on numerous occasions here on Slashdot and this seems like another good time to post them. His reports on the history and theory of voting are excellent.

In particular, I recommend his essay on Paper Ballots. [uiowa.edu]
A Brief Illustrated History of Voting [uiowa.edu] is another excellent essay.
There are dozens of technical essays on voting systems on Jones' main Voting and Elections site. [uiowa.edu]