Margins of Error
Today is the first day of the trial to determine whether the election of Christine Gregoire as Washington State's governor should be overturned. This is something I'll be following closely and with great interest.
For those of you who may not be familiar with the political battle occurring here in Washington State, GOP challenger Dino Rossi filed a lawsuit in January, challenging the third and final (hand) count of the votes in Washington State, which tipped the governor's election to Gregoire after Rossi had won the initial two machine counts, which, to me, is like saying, "Hey, thanks Blue Gene, but we know how unreliable your calculations can be, so we'll take it from here." I mean, wasn't the whole point of machine counting to rescue us from the tedium and inherent unreliability of human hand-counting?
"Every single time a machine counts a vote, it does so in the same programmed manner, by exactly the same inbuilt standards, over and over again. Machines are usually constructed to execute specific functions. Vote-counting machines are designed for the specific single purpose of counting votes. If you doubt a counting machine's accuracy to count, try your pocket calculator -- or an ATM machine."
There's an argument being floated that machine counts are more precise while hand counts are more accurate, but this doesn't take into account the pull of human passions when political results are at stake. While humans don't suffer from the issues that machines do ("hanging chads," anyone?), machines don't suffer from the issues that humans do when performing tedious and repetitive tasks - fatigue, clumsiness, forgetfulness, etc. There's a reason that "Human Error" is a well-known term, and why machines are increasingly used for counting and tabulation tasks, despite their margin of error.
"While automation provides predictable, consistent performance, it lacks judgment, adaptability and logic. While humans provide judgment, adaptability and logic, we are unpredictable, inconsistent and subject to emotions and motivation."
Proponents of hand-counting point to the adaptability and judgement of humans as necessary to the vote-counting process, while neglecting to mention that it's the adaptability and judgement of humans that resulted in thousands of misplaced and lost-then-found ballots that have plagued Washington State's gubernatorial election (not to mention the issue of provisional ballots being fed through voting machines without proper vetting), casting long shadows of doubt and suspicion across the King County Board of Elections. Even Dean Logan, the elections director of King County and a Democrat, admitted to reporters, "Every time you have human judgment and frailty enter into the process, it will change the result . . . when you're talking about close to 900,000 pieces of paper, I think the machine count is going to be more accurate than a manual count."
Yet a manual recount is exactly what Washington voters got, after TWO machine counts that came up with Dino Rossi as the winner -- a hand-count with results so suspect that we're now in the midst of a legal battle that's a black-eye upon the lax nature and slip-shop regulation of Washington State's electoral process.
Throughout human history, almost every major technological advance has been met with resistance: "Ovens, it was once said, would take the charm out of cooking; industrial machinery would cause unemployment; in vitro fertilization would subvert nature . . . at an instinctive level, especially if the technology does something that we do, but does it better, we feel threatened." So critics of the machine-counts and machine voting technology (mostly democrats, natch) are flogging the tired old 1960's Star Trek era meme that humans are somehow better and more reliable at these kinds of things, despite the fact that the demand for more and more technology to replace error-prone human activity flies directly in the face of such sentiments.
I've mentioned the Luddites before, right?
"There's a disconnect between the headlines and the actuality, which is that the machines are working pretty well, but that they're part of a process that involves people and procedures," said Bob Cohen, senior vice president of the Information Technology Association of America (ITAA), and a recent CalTech-MIT Voting Technology Project came to the conclusion that while improved administrative practices and better voter education contributed to more accurate voting results in 2004, it was the significant overhaul of older voting technologies that "made a substantial positive difference in 2004," with the states that engaged in the most significant election technology reform efforts seeing the largest increase in vote tabulation accuracy.
But, somehow, it's the hand-count that matters.
This is a legal challenge that has passions running high on both sides of the political spectrum, as it's a battle for Party control of the Governor's mansion here in Washington State. But this election was a grave disappointment for anyone interested in serious election reform, and I'm hoping that the court rules for a whole new election in 2006 rather than letting this one stand. I'm not at all confident that the court will see fit to make such a ruling, but the electoral process in King County has proved itself corrupt and unreliable, and allowing this last election to stand would be a judicial statement in favor of shoddy electoral procedure.
"No one knows who won this election," said Rossi supporter Brian Thomas, 60, of Seattle. "I hope the judge just nullifies this election so there will be a revote in November." Unfortunately, Chelan County Superior Court Judge John Bridges is only human.



Comments
Back when it looked like Rossi was the winner (based on the machine counts), many democrats (including myself) argued that Rossi's victory was within the margin of error, and that we should re-vote.
Now that the tables have been turned, and Gregoire has a thin margin of victory, many people have changed their position on the subject. It makes me embarrassed of my fellow liberals. A re-vote in a staggeringly close race is the just thing to do, even when "your side" is on top at the moment. Damn hypocrites. But I digress.
Your whole argument of which is more accurate, precise, predictable, passionate, whatever is pointless. For there to be any confidence in elections, the margin of victory has to be greater than the margin of error.
The way to fix this short term? Lets vote again. You know, the whole will-of-the-people thing? I'll vote for Gregoire again, even though I'm unhappy with how she has handled this.
The way to fix this long term? Lower the margin of error. As it will never be zero, make the margin of error a known quantity, and if any election doesn't win by that margin, we do it again. Easy.
Instant runoff voting and other reforms could help, but both major political parties won't ever go for it.
_______________________
Homocon sez:
I agree. The margin of error, taking into consideration both machine and human error, is too close on this one. There have been points made that a recount was completely unnecessary in the first place, as the margin was too close to really know who the winner might actually be. Your idea of establishing regulations as to acceptable, known margins is spot-on..
And while I understand your objections to the "man vs. machine" argument, I can assure you that the argument as a whole, especially regarding voting tabulation, is most definitely not "pointless" -- reverting to a hand-count after two machine counts is, however, just that.
That said, both sides should be happy about a revote, as it has the only real possibility of clearing up the matter. I'll vote for Rossi again, and be delighted to do so.
Posted by: Jim Thatcher | May 24, 2005 7:18 PM