If Florida 2000 was the year of pregnant and hanging chads, California 2008 could be known as the year Joe Six-Pack meets a 10-sided die–a la Dungeons & Dragons, the fantasy role-playing game. The idea is that rolling dice is an easy and efficient way to prove to citizens that precincts chosen for post-election audits are truly random.
The D&D recommendation is one of a few scenarios that academics recommended to a new working group formed by Secretary of State Debra Bowen last month to re-examine the state’s 40-year-old post-election audit law.
“California’s 1 percent audit law is 40 years old,” Bowen said. “I want to know how effective it is and whether there are better models for auditing election results.”
Another recommendation, that California expand the 1 percent fixed audit, however, has some elections officials in large counties worried that they may not have the space, time and manpower to complete the canvassing of elections in 28 days, as required by law. This problem is especially acute, they argue, as the state prepares for three major elections next year.
The first is California’s early presidential primary on February 5, which could have national repercussions if the outcome of the vote is unclear.
Bowen’s decision to establish the six-person working group–made up of academics, an advocate, and elections officials–comes on the heels of another task force the Secretary instituted at the end of May to provide a top to bottom review of California’s voting machines. The result of that review, which could see some electronic voting machines decertified, is scheduled for the end of July. Both groups are scheduled to wrap up around the same time.
“These reviews are about Secretary Bowen’s continued effort to ensure [the voting system’s] security, accuracy, reliability and access,” said Nicole Winger, a spokeswoman for the secretary.
Some registers and county elections clerks, though, worry about the direction the working group seems to be headed. And they question the benefits from instituting a costly increase of manual audits.
“I don’t believe the system is broken,” said Deborah Seiler, registrar of voters for San Diego County. Seiler said that for the last election she had 16 teams of three people working six days a week for 10 hours each day to go through the process of certifying election results. The process took three weeks. “If we bumped up [manual recounts] we would have to go 24-seven, and managerial oversight would suffer,” Seiler said. Seiler said that ensuring more attention is paid to testing machines before the vote is a better course of action than expanding the amount of votes counted in a post-election audit.
Conny McCormack, registrar for the County of Los Angeles, agrees. She said that no matter how large you increase the sample size after a vote in a close race, candidates are going to request a recount anyway. “The beauty of the current law is that if you ask for a recount and it reverses the state’s decision, you get your money back. And 99.9 percent of the time, the state is right,” McCormack said. She added that a raise in audits probably would cost at least a million dollars across the state and require leasing new buildings to conduct public audits.
Academics, voting advocates and statisticians argue, though, that 1 percent is too small a sample in close races for voters to be confident their will is being carried out. “The closer the election gets, the harder it is to detect bad things that can happen that could make a difference in an election,” said Joe Hall, a Ph.D. candidate from UC Berkeley who spoke before the working group two weeks ago. “The amount of precincts that could flip the outcome of a race becomes small quickly,” Hall said.
Hall advocates a system for California based on what proposed federal legislation offers. The bill, HR 811, would require a tiered percentage of manual audits based on the closeness of federal races. If the margin is less than 1 percent, 10 percent of precincts should be audited, Hall said.
Kim Alexander, president of the California Voter Foundation, and a member of the working group, said an increased audit sample will help voters have more faith in the system. She points to a statewide survey her organization conducted of eligible but non-registered voters, which found that nearly one in four felt their votes wouldn’t be counted accurately, so why bother.
Hall acknowledged any increase in audit sample size would have a big impact on large counties. Still, small procedures–like using a 10-sided die to choose random precincts to audit, instead of randomly generated computer codes–also would help ensure voters’ faith in the integrity of the process. “If I’m Joe Six-Pack, all I see is someone hitting return on a computer and a bunch of numbers come up.