This is truly the year of the amateur. It was mostly the unfamous and unsung who organized the voter registration and get-out-the-vote drives, started and ran the grassroots (if well-funded) 527 committees like MoveOn.org, wrote the blogs and conceived the websites that did what the traditional media seldom did--such as probing deeply into Bush's personal history, including his military service. And it is ordinary people who are now leading the way in scrutinizing newfangled, secretive voting systems, seeking to insure that a handful of corporations don't, accidentally or deliberately, undermine electoral democracy.
Tomorrow the first recount begins--in New Hampshire, of all places, a state George Bush didn't even win. But in those areas where he did well, sometimes the numbers look decidedly odd. In this case, the person who got the ball rolling was one Ida Briggs, a longtime Michigan software designer and database developer who did a statistical analysis of some election results, and found them perplexing enough to trigger concerns in her mind about the efficacy of the electronic vote tabulation system used.
What she found were striking anomalies--mostly in precincts using paper ballots that were then input via the optical scanning machines manufactured by the controversial vendor Diebold, of North Canton, Ohio. In general, according to Briggs, the "Diebold precincts" showed larger and more frequent deviations from expected voting trends than precincts relying strictly on hand counts, and even than those using an optical-scan counting system from another manufacturer. Creating trend patterns by looking at the 2000 and 2004 elections, she found rural, typically conservative precincts that hand-counted ballots as voting more for Kerry than they did for Gore, while larger, urban precincts using Diebold's AccuVote machines often did the opposite. Of the precincts where Kerry did less well than expected, according to Briggs, 73 percent used optical-scan technology and 62 percent used Diebold machines. Fully 92 percent of all out-of-trend votes were optically scanned. New Hampshire has 301 precincts; 126 of them use Diebold's AccuVote technology.
Referring to the recount advocates, a Diebold spokesman told the Associated Press, "I think they're rushing to judgment."
Briggs became interested in the numbers when, shortly after the election, she saw a study published on the web about statistical anomalies in nonswing states. New Hampshire caught her attention because of the sizable--15 percent--differential between early exit polls and results. It was easy to study, because the state made its data available online. And because New Hampshire was a state Kerry won, no one could claim that the goal of a recount there was to change the election results.
In the era of contracted-out services, companies like Diebold are given unusual amounts of liberty to be self-policing. The problems emerge later, if at all. Diebold has faced intense scrutiny and criticism over malfunctions in its touch-screen voting machines, but it steadfastly insists that its optical scanners have proven reliable during years of use. Diebold continues to be one of several corporations with vast power over the levers of democracy.
Respected analysts have found numerous bugs in Diebold's system codes, and complain that the company has failed to release its most recent revisions, preventing an independent verification of improvements. The company drew particular concern after Walden O'Dell, the Ohio-based company's chief executive officer, penned a fundraising letter for Bush in which he declared himself "committed to helping Ohio deliver its electoral votes to the president next year." And Diebold recently settled a civil suit brought by the State of California alleging that the company sold the state and several counties shoddy voting equipment. Diebold agreed to pay $2.6 million to the state. Nevertheless, there is no evidence that Diebold intentionally tampered with its software.
Briggs is the first to admit that mathematics is a tricky game. "New Hampshire might just be odd--this trend could really be happening," she said. "It's unlikely, but it could be." If, indeed, the numbers are wrong, it doesn't mean deliberate tampering. It could be a programming error, which would be consistent with the fact that the unexpected results affected only the first race on the ballot. In any case, that's what recounts are for.
Once Briggs's eyebrows were raised, she said, she tried contacting the Kerry campaign to see if officials there would call for a recount. With no affirmative response, on Friday, November 5, she called Ralph Nader--with less than four hours remaining before New Hampshire's deadline for recount requests. The Nader people didn't know Briggs, and were wary, but Air America Radio host Randi Rhodes managed to mobilize enough listeners that Nader soon had a twelve-inch stack of imploring faxes. With one minute remaining to deadline, he faxed in a request for a recount. (He also agreed to pay a $2,000 filing fee plus actual costs.)
On Thursday, New Hampshire officials will begin a hand recount of paper ballots in five of eleven large urban precincts--in Manchester and Litchfield--where Bush did surprisingly well. The remaining precincts will be counted soon. If the results prove interesting, recounts could be requested elsewhere besides Ohio, where such a request has already been made by Green Party presidential candidate David Cobb and Libertarian Party candidate Michael Badnarik.
Even a hand recount won't satisfy everyone--and shouldn't. The efficacy of the American voting system is dependent on a lot of things going right--and anecdotal evidence suggests many fruitful avenues of inquiry into things that may have gone wrong. Among these: whether ballots were improperly cast (fraud), and whether legitimate voters were prevented or discouraged from voting. To say nothing of whether, in a country where many people vote based on the most effective television commercials, people really understand what they are voting on and the stakes involved.
Kerry beat Bush in New Hampshire by 340,511 to 331,237 votes, a spread of 50 percent to 49 percent, with Nader taking less than 1 percent. A recount, even if it does establish problems, likely won't change the winner in New Hampshire, and even if it does it will certainly not alter the outcome of the presidential race.
However, if it does show significant inaccuracies generated by the AccuVote equipment and software, it could trigger recounts elsewhere--recounts that could, theoretically, reverse the election.
That's highly unlikely, given Bush's hefty wins in key states, and given the prominent role of other voting technologies. But hand recounts of optically scanned ballots will go a long way toward addressing doubts about that technology and about the vendor. And it will perhaps give others the confidence to request recounts when results go against statistical trends, or common wisdom. At minimum, it will be a start on the road to transparency and accountability.
If it turns out that anomalies are just that, so be it. Then we need to spend more time understanding why people voted--really, truly voted--the way they did.