Wednesday, November 26, 2008

Hanging Chads 2008

During the 2000 Florida recount, voters were appalled to discover how many punched card ballots were regularly discarded. There was a groundswell movement to replace them with something better. The first reaction was to switch to voting machines. They offer several advantages. There are not overvotes (voting for two candidates) and they can warn about undervotes (not voting for all candidates). Counting is swift and accurate.

But as the 2004 election neared, some Democrats noticed that the voting machine manufacturers were Republicans and the head of one of them promised to "do whatever it takes to re-elect President Bush". They ignored the fact that he said this at a fund-raiser and assumed that he meant that his voting machines could be manipulated to steal the election. After Bush won the 2004 vote we saw several stories where someone found suspicious voting pattern that might indicate fraud. None of these stood up to scrutiny but the controversy still caused a change in voting machine requirements. They would have to print a log at the same time that the vote was recorded. The voter could check this to be sure that his vote had been recorded properly. If there was a question about the machine's accuracy, the log could be compared with the tabulated count.

This should have ended it but Democrats were still convinced that voting machines could not be trusted. They settled on optical mark ballots as being the best way to vote. Ohio Secretary of State Brunner even commissioned a study that proved that, given unlimited access to voting machines, it was possible to alter the results. The study did not evaluate optical mark ballots but Brunner still maintained that it proved that the paper ballots were better.

As envisioned by Brunner, the voter would fill out a ballot by filling in the circle next to the candidate's name. The voter would then feed the ballot into a scanner which would verify that there were not overcounts or undercounts. The ballots would then be collected and transferred to a central location where they would be counted.

So how did that work out? A judicial recount in Florida came up with conflicting results. After hand-counting the ballots the officials found that the optical mark readers were inconsistent. Some ballots would be approved even though the circle wasn't completely filled out. Others were rejected for no discernible reason.

Things are even worse in Minnesota. Look at the challenged ballots here and say that voter intent in always clear with optical mark ballots. Control of the Senate may hing on whether a ballot should be disqualified because the voter wrote "Lizard Man" as a write-in but actually voted for a regular candidate.

Optical mark is just as bad as punched card. While voting machines have their own problems, they are still the most reliable method of counting votes that we have available. That said, the voting machine companies should open their code to independent analysis.

No comments: