Tuesday, December 05, 2006

Florida again: the mysterious case of the vanishing votes

The widening civil war in Iraq (or, as President Bush prefers to calls it, “the young democracy”) will continue to dominate the news this week, as we await the official release of the Baker-Hamilton recommendations. In the meantime, however, a fascinating – albeit underreported – story is unfolding right here in the old democracy, one of those high-tech nightmares that voting experts have long been warning us about.

In a print column last month, after noting that nine separate reports have found fundamental flaws in the new-fangled electronic touch screen machines, I concluded that “a festering problem could well become a future crisis.” Well, it turns out the crisis is at hand right now, on a blessedly small scale, in the southern Florida congressional district that includes Sarasota. We can at least be grateful that, on Nov. 7, control of the U.S. House of Representatives didn’t hinge on the results of a single seat, because, if it had, Sarasota right now would be ground zero in a national psychodrama.

Just do the math: Back on Nov. 20, Florida state election officials decreed that Republican congressional candidate Vern Buchanan was the winner in the 13th congressional district, topping Democrat Christine Jennings by 369 votes. The problem – the enduring mystery, actually – is that nearly 18,000 touch screen voters in Sarasota County went to the polls, and chose their favorite candidates in all the major races…except in the hotly-contested, high-publicized congressional race. According to the touch screen machines, 18,000 people somehow skipped the Buchanan-Jennings contest.

All told, about 15 percent of the voters in Sarasota decided not to choose between Buchanan and Jennings – according to the machines, anyway. Yet elsewhere in the congressional district, the percentage of people who skipped that race was much lower, anywhere from two to five percent. Nobody has yet explained this stark discrepancy, but it’s clear that Jennings, the Democrat, has a major stake in finding out what happened – because, as the Orlando Sentinel has already reported, those 18,000 voters were predominantly Democratic, strongly backing virtually all the other Democratic candidates, up and down the ballot.

Many voters have come forward in recent days to complain that they tried to vote for Jennings, but discovered that their preference was not recorded when the machine displayed a review screen. Jennings is suing for a new election, and she is suing the touch screen manufacturer, Electronic Systems & Software, alleging “evidence of machine malfunction.” The state’s initial probes have not uncovered any malfunctions, but the authors of those aforementioned nine reports (at Stanford and Princeton and Johns Hopkins, among other esteemed locales) have all warned that these machines are prone to either lose votes or simply fail to register votes.

And the post-election probe is hindered by the fact that these machines lack any kind of backup paper trail. Well, what a surprise. Congress has spent the last three years sitting on a bill that would require paper trails on the new touch screens, and now here we are. In addition, we now have a new federal draft report, issued last Thursday, which concludes that paperless electronic voting machines “cannot be made secure.” In the words of the National Institute of Standards and Technology, a federal agency, the absence of a paper trail “is one of the main reasons behind continued questions about voting system security and diminished public confidence in elections.”

ES&S, the touch screen manufacturer, issued this statement the other day, absolving themselves of blame for the mystery results in Sarasota: "Testing and recounts conducted by Sarasota County and the state have both shown that the touch-screen voting system used in Sarasota County accurately records and tabulates voter selections. Specifically, Sarasota County has conducted a recount - as required by state law. That recount confirmed election day results from the touch screen machines.”

The problem, however, is that this recount was essentially worthless. All it means is that the elections officials recounted the actual votes that the Sarasota machines had already recorded, as opposed to all those they may have failed to record. These are precisely the kinds of “software dependent” machines that NIST, the federal agency, wants to abolish – because they provide no paper trail that could help auditors determine what really happened.

The Republicans and their supporters, meanwhile, have come up with a number of curious arguments that seek to explain away the mystery of the “undervotes”. Buchanan, the certified winner of the election, suggested to the Associated Press that 18,000 Sarasota voters chose to skip the race out of protest, simply because they were “turned off by negative campaigning.” The flaw in that explanation is obvious. Why would predominantly Democratic voters in one county be three to six times more likely to be “turned off by negative campaigning” than the voters in the congressional district’s other three counties?

Charles Stewart III, a voting technology expert and political scientist at the Massachusetts Institute of Technology, has filed a brief on Jennings’ behalf. He summarily rejects the negative-campaign argument: “Evidence that such alternative explanations were causing high undervote rates would have shown up throughout the district, not just in a single county, and not just in one type of voting machine in one county.”

Then we have the Wall Street Journal editorial page, an outpost of the conservative media that has long been known for coming up short on fundamental factual homework. Here’s what the editorialists said last Friday: “if anyone ought to be complaining about undervotes, it’s the GOP. Sarasota is the largest and most Republican country in the district…which makes it more likely that it was Republicans who declined to vote in the congressional race, not Democrats.”

Somehow the editorialists missed the Orlando Sentinel’s report, eight days earlier, which documented that the undervoters were predominantly Democratic. And the Journal neglected to mention that, among all the recorded votes, Democrat Jennings defeated Republican Buchanan in Sarasota County by 52 to 47 percent -- thereby feeding Jennings’ argument that she might have won the election if all votes had been properly counted. Stewart, the MIT expert, agrees; if not for “factors related to machine malfunction,” he contends that Jennings would have won the election, albeit narrowly.

The general public may not be watching this story closely, but the political community certainly is. The topic came up last Thursday at a Washington confab, while I was in attendance. Larry Sabato, the noted University of Virginia pundit, was incensed at the prospect that so many votes had been lost: "It's really outrageous....Imagine how you would feel if that happened in your state or congressional district."

And one of his panelists, former Republican National Committee attorney Michael Toner, while not necessarily endorsing the machine-glitch scenario, said that the United States lags far behind other western democracies "in the professionalization of its elections." Most of the people manning the polling places, he said, are basically "70-year-old volunteers who are making maybe $6 an hour," and are therefore ill-qualified to master and oversee these patently flawed touch screen machines. (Actually, it's worse than that. The feds have yet to put in place a certification procedure that would vet - or question - the reliability of these machines.)

So what happens next in Florida? A state audit of the machines is in the works, Buchanan is setting up his Washington office, while Jennings (who refuses to concede) is demanding that ES&S give up its “source codes,” the proprietary information that would allow independent investigators to probe inside the actual machines. What a mess. Let’s just hope that the latest Florida flap is not just a gruesome dry run for the 2008 presidential election.

And we have been warned enough already; as the Government Accountability Office, the investigative arm of Congress concluded one year ago, “problems with the security and reliability of electronic voting systems (are) potentially affecting the reliability of future elections, and voter confidence in the accuracy of the vote count.”

Let’s hope the Iraqis with the purple fingers don’t hear about this one.