http://www

http://www.columbiapaper.com/index.php/component/content/article/35-letters/1698-by-virginia-martin

Your vote counts: This county knows who won

Written by VIRGINIA MARTIN

Friday, 17 December 2010

WITHOUT CONFIDENCE in our elections, is this really a democracy?

When you hear about “suspect” elections in one country or another, what do you assume? That it's not a democracy. We know we're a democracy when we can trust our elections.

This year, the first in which every single vote was cast on a paper ballot, Columbia County hand-counted each and every one. Before the election, Commissioner Jason Nastke and I ordered a full hand count so that our county's voters could have complete confidence in the vote totals that we'd certify after the election. Who could distrust hand counts that are done by teams of Democrats and Republicans, with one looking over the shoulder of the other every step of the way?

Columbia is the only county that did this, though. Other counties conducted an “audit” of 3% of their voting machines, which is the minimum required by our new state election law. (If a certain number of machine errors are found, the audit must escalate to 8%; then, if more errors are found, to 20%, and potentially to 100%.)

There's a world of difference between 3% and 100%, and we're not just talking about the time involved or the cost of hiring ballot counters. The key difference is that, with our 100% audit, we commissioners and this county's voting public can have unparalleled confidence in our results.

So, why don't all counties hand-count every single ballot? Because it's more work. It's costly and it's time-consuming. (And isn't that what those expensive new voting machines are for, anyway? I'll address that another day.) The problem is that electronic voting machines can give incorrect results without warning. Results could be wrong and no one would ever know -- unless the paper ballots were hand counted to disprove the machines' results.

That's why audits were introduced to the electoral landscape. Audits are intended to cut costs yet provide confidence, and well-designed risk-limiting audits based on sound statistical principles can do just that. Financial institutions, for example, invest substantial resources in expertly designed audits that will confirm or disprove that their equipment, programs, and practices are working correctly. In elections, a good audit can ensure to a very high degree of confidence -- over 95% is a reasonable standard -- that the candidates the voters chose will be the ones sworn into office.

But competent statisticians and audit experts will tell you that New York State's 3% audit doesn't provide anything near that high degree of confidence.

A 3% audit would have us hand-count all the ballots from just two machines (randomly selected) of the 45 we deployed countywide on Election Day. Consider what that might have meant for this year's two close local elections, for Third County Coroner and for Chatham Town Board.

Initial results from the coroner race showed a 21-vote difference between the two candidates. Out of the 45 scanners deployed, just one unaudited machine, reporting 21 votes incorrectly or switching 11 votes from one candidate to the other, could change the outcome without our knowledge. What's frightening is that, in this scenario with a 3% audit, there would be a less-than-5% chance that our random selection would have chosen that scanner to audit. That would give us a more-than-95% chance of certifying the wrong candidate as the winner.

Neither Commissioner Nastke nor I like those odds. Nor should our voters.

As for the Chatham board race, with an initial 93-vote difference, the state's 3% audit law results in a more-than-87% chance of this race not being audited at all!

Absent a 100% audit, election law provides protection, of sorts, to a losing candidate who's skeptical of the results. The candidate can go before a judge to get a court order to have all the ballots hand counted. But there are no guarantees. In Nassau County's 7th Senate District race, with a difference of only 451 votes out of the more than 85,000 ballots cast, the judge rejected the loser's request. Now, on top of the costs, strain and associated difficulties of going to state Supreme Court and retaining attorneys, expert witnesses and the like, the loser is bearing the not-insignificant costs of appealing the judge's decision.

In last year's scanner pilot program conducted in most counties but not here in Columbia, there were serious problems. In one county, the scanners switched votes from one candidate to the other. That gross failure was discovered only because the Election Night results were wildly different than what everyone expected. The investigation that ensued uncovered a failure on the part of testers to recognize anomalous test results. Neither conspiracy nor maliciousness were involved -- the testers were probably just overworked and overtired. The result was that error-riddled scanners were inadvertently deployed to the polls, where they made a mockery of that election board's attempts to conduct a safe and accurate election.

Columbia County's 100% hand count cost a few thousand dollars more than it probably would have cost us to embark on a 3% audit. But we've had no one questioning the results. Throngs of attorneys didn't encamp for months in Hudson at 401 State St. to conduct legal maneuvers. They didn't need to, because we were counting all the ballots anyway. I'm not saying I don't like attorneys, but let's face it -- if we can conduct our elections without interventions by the legal profession, I think we've accomplished something significant.

It's good that the lawyers know our results are accurate. It's crucial that our voters do. Because confidence in the electoral system is fundamental to our sense that it's a democracy we're living in and that our votes really do count.

Virginia Martin is the Democratic commissioner of the Columbia County Board of Elections.

Jason Nastke is the Republican commissioner.