Rocking the Voters

San Diego — It's no secret that this fall's election season didn't go smoothly for San Diego Registrar of Voters Mikel Haas. First came news that some 5000 absentee ballots, ordered from controversial e-voting vendor Diebold Election Systems, failed to arrive, forcing workers to send out photocopies of the ballot to irate voters. Once returned, votes cast on each ballot had to be transcribed by hand onto forms to be run through the Registrar's optical scanners, further delaying an already late tally. Then on election day complaints piled up about things like Diebold's voting machines breaking down or being plugged into electrical outlets without juice. But that may be only the tip of the iceberg.

Based on recent e-mails obtained from the County under the California Public Records Act, Haas's lieutenants were at loggerheads with Diebold staffers over such issues as lack of poll worker training, which critics say is at the core of many of the Registrar's problems. On September 11, for example, Registrar of Voters chief deputy Len Schultz e-mailed Diebold's Michael Rockenstein to complain that John Decker, another Diebold employee, would not be in San Diego at the time promised. "John Decker just talked to one of my staff," Schultz wrote. "He said that according to his schedule, he is not going to be in San Diego until October 16. You have assured me that he will in fact be here beginning on September 18. I am very concerned that John's schedule does not reflect your assurances and that he will not be here next week. From what I understand, he is making travel arrangements to be back east next week.

"John is a key member of the team that will be significantly re-scripting our training," Schultz continued. "All of our planning has been based on having him available. His absence would have a significant impact. Please make sure that John and his manager are all on board with him being here next week."

Contacted by phone after last week's election, Schultz said that he and Rockenstein subsequently "negotiated back and forth" about the issue and finally agreed that Decker would not come to town until October 10. In spite of his earlier concerns expressed in the e-mail to Rockenstein, Schultz insisted that Decker's late arrival did not result in any training deficiencies.

But that wasn't the first time Registrar staffers expressed concern that election workers might not be up to the job. In an e-mail dated July 21, Jean Olsson, in charge of training "troubleshooters" -- workers who are supposed to track down and resolve problems at the polls before they interfere with voting -- voiced a series of worries to Schultz. "For the troubleshooter training facility," Olsson wrote, "my main concern is that 50 to a classroom is not a feasible plan with regard to troubleshooters receiving the quality training they need to do their job. We found that for the June election that as the class size hit 20-22, having one instructor demonstrating and 2 helping begins to break down."

Olsson also voiced doubts about her ability to find enough qualified workers in time to train them all. "With a big push to recruit 200 people -- we need to prescreen carefully. The 10% turnover in June was very difficult to manage. (I used 30 people from a back-up list in the last 2 weeks before the election after the initial group of 273 applicants were processed and scheduled for class).

"For example, all the last minute rush [of] recruits referred by Edie Greise and Michael Nguyen were no shows or dropped the day of their class. Unless we significantly increase our staffing, 10% of 500 turnover will be impossible to manage in the last 2 weeks before the election, especially from a remote training site. Every person replaced requires a change in a significant number of binder materials, and coordination of training, supply pick-up and redistribution, etc. We really need some script for everyone in the ROV to use to recruit and screen so that we attract the appropriate people."