The manual on recounts Election procedures aren't what they used to be

Question: What do you think the term "manual recount" means in 2004?

If you flash back to Bush vs. Gore and answer, "An incredibly laborious, contentious, but thorough review of every single ballot," you're wrong.

These days in Florida, manual recounts are supposed to be done only on certain undervotes and overvotes a tiny sliver of the total ballots that are cast.

The change stems from reforms passed by the Legislature in 2001, which eliminated a previous option to "manually recount all ballots" a phrase redolent of the 2000 election fiasco.

The deletion may have seemed like a good idea in 2001. But today, as the vulnerabilities of touchscreen technology crowd into the headlines, the Legislature's decision looks shortsighted.

In a close vote, a hands-on review of every ballot not just those with undervotes and overvotes would provide a welcome opportunity to scrutinize the machines' performance. A full manual recount would be labor-intensive (whether counties use touchscreen or optical-scan voting systems), but in the long run, it could boost confidence in elections and yield smoother ones in the future.

Touchscreen trouble

The demise of a full manual recount is troubling, but the following is worse: The Division of Elections has decided that touchscreens shouldn't undergo any manual recounts not even those limited to un- dervotes and overvotes. (An overvote occurs, for example, when a citizen marks two candidates instead of one. In undervotes, the citizen appears to have made no ion at all in a race.)

The logic behind the division's decision is circuitous, at best, and it goes like this: Since touchscreens don't allow an overvote (the machine won't accept the vote until the error is rectified), there shouldn't be any overvotes to recount. As for undervotes, it's fruitless to manually recount them because the touchscreens' electronic log contains no scribbling, scratching or other signs of voter intent signs that are sometimes visible on paper ballots. (The law allows a garbled ballot to be counted if the voter's choice is discernible.)

With its ill-advised decision, the Division of Elections seems to be excusing touchscreens, rather than forcing them to live up to the letter of the law.

Even when limited to undervotes and overvotes, a touchscreen manual recount could be useful, enabling officials to spot anomalies, polling mix-ups and other glitches which are a real concern, despite officials' expressed confidence in the technology. Furthermore, consistency and uniformity among jurisdictions are critically important to statewide elections; when optical- scan counties perform manual recounts and touchscreen counties don't, there is no such thing as a level playing field.

Reforming the reforms

The manual-recount issues are a sign that the state's 2001 election reforms are already behind the times. While the revised statutes now set lofty standards for fairness, accuracy, consistency and voter access, they fail to address some fresh problems that have cropped up in the transition to touchscreens and new procedures. The problems may be resolved over time, but the state's primary is just six weeks away.

Security and accuracy doubts concerning touchscreens have fanned a movement for a voter-verified paper trail, but there's no pragmatic way to implement one by Election Day. Such a system, and the rules to administer it, will take time to develop.

Until that day, touchscreen credibility can be furthered by well-prepared poll workers, independent monitoring, rigorous machine testing, and a plenitude of systemic checks and balances that help detect and prevent error. Manual-recount requirements should be included in that arsenal.

As for the abandoned option of a total manual recount, legislators should revisit this decision. Despite the logistical and accuracy problems inherent in a massive hand count, it is an important way to publicly examine election results and to restore confidence in them.