Stop the buck

Management, not system, at heart of election fiasco

Editorial   DailyCamera  Boulder News  November 5, 2004

More than two days after most voters in Boulder County marked their ballots, there was understandable outrage, intense water-cooler conversation and general amazement that, for all intents and purposes, their votes meant nothing.

While the entire rest of Colorado and the nation had reported final tallies long before, as of Thursday afternoon the Boulder County Clerk and Recorder's Office was still counting at a glacial pace. By 1:30 p.m., only 89 out of 227 precincts ? just 39 percent ? had been reported.

The delay has not been inconsequential. Final results in at least two important races, for state Senate District 17 and the University of Colorado regent-at-large seat, still aren't in.

Perhaps we should have seen this coming. Last August, the county's new paper-ballot, optical-scanning system manufactured by Hart Intercivic rendered similarly slow results in a primary election with considerably fewer voters.

At that time, some officials who had in 2003 planned to buy a "direct record electronic" system, i.e. touch-screen type machines, privately put the blame on citizens and the media (including this newspaper) for demanding a system with a paper trail.

Yet 34 of Colorado's 65 counties, and hundreds of other jurisdictions around the country, used paper ballots and optical scanners in this election without serious problems. County officials ? some of whose sole job is to run competent elections ? must shoulder the blame.

Given that many counties using DRE technology on Tuesday reported troubling glitches, we still endorse using a system that uses a recountable paper ballot. And while we agree that accuracy is more important than speed, we certainly never imagined that county voters would have to wait this long to learn final results.

Many things have contributed to this fiasco, but management, not the system, is at the root of the problem.

First, it appears that the county has insufficient hardware and staff to do the job. For example, it takes only minutes to scan a precinct's worth of ballots, but resolution of problem ballots takes much longer. The county apparently was plugging up scanners for resolution when scanned problem ballots might have been exported to a PC, so counting could continue on the limited number of scanners. Also, it appears that the county did not have enough memory cards to record results from scanners.

There also have been issues with the ballots themselves. Literally thousands of undamaged ballots have been tagged as damaged by scanners. And when the scanner says a ballot is damaged, it must be examined by hand, slowing the count.

Strangely, and contrary to past election practices, results that were available (few as they were) were not being continually posted by the clerk's office. The county also reportedly took time to group early-voting ballots with their precinct of origin, rather than creating a kind of "virtual" precinct as has been done in the past.

Many voters found the ballots themselves clumsy, and "coloring" in overlarge boxes with ball-point pens not only potentially compromised privacy, but also contributed to long lines in many precincts.

It seems likely that all of these problems are fixable ? optical-scan systems, including those manufactured by Hart Intercivic, work just fine elsewhere, after all ? and we expect the county will work hard to identify and correct them. If this system ? which the clerk's office ed ? isn't up to the task, we'll simply have to find money for a new one.

But what's clear right now is that the clerk's office wasn't prepared for this election, and those in charge, whether elected or hired, must be held accountable.