How the GOP Gamed the System in Florida
The point man for the state in compiling those lists was Emmett "Bucky" Mitchell IV, an assistant general counsel to the Florida Division of Elections, who within a week of the November 2000 election was given a senior attorney's job in the state Department of Education. In an interview with The Nation, Mitchell claimed he had exercised restraint in producing the purge lists. "The division always had the policy to err on the side of caution," he insisted. But reports from county supervisors, correspondence between state officials and DBT employees, and testimony before the Civil Rights Commission tell a completely different story.
By March 1999, four months after contracts had been signed, DBT officials already had doubts about the state's ground rules. According to testimony by ChoicePoint/DBT vice president George Bruder, a person could be included on the list if his or her name, date of birth and/or Social Security number closely approximated that of a known felon. In other words, in a state with 16 million people, where many individuals share approximate names and also dates of birth, exact matches were not necessary.
In March of 1999, Thorogood expressed her doubts about those guidelines in an e-mail to Mitchell: "Unfortunately, programming in this fashion may supply you with false positives," she said, referring to names of people who did not belong on the felons list. "This seems to be the approach you would prefer to choose, rather than miss any positive true matches." Mitchell made the state's position clear in his answer to Thorogood on March 23: "Obviously, we want to capture more names that possibly aren't matches and let the supervisors [of elections] make a final determination rather than exclude certain matches altogether," Mitchell wrote. In other words, the lists were designed to include people who were not felons, some of whom eventually fell through the cracks and were unfairly purged.
When supervisors began to complain about errors, Bruder said his company told the Divison of Elections that they were caused by the loose parameters set by the search, but Mitchell ordered no substantial change in the parameters despite recommendations by DBT. "After submitting them they were not acted on by the state," said James Lee, a spokesman for ChoicePoint/DBT. In fact, the next year, as the presidential election approached, the state asked that the parameters be loosened, according to Lee. Instead of 90 percent of the letters in the name of a person on the purge list having to match with those of someone on the voting rolls, the standard was loosened to 80 percent. Although such matches were often eliminated when Social Security numbers or other data were also checked, such information was not always available, and more innocent individuals were included on the felons list.
The state officials were not content to include only former Florida prisoners. They also asked DBT to use its national databases to provide the names of felons from other states who might have moved to Florida and registered. But some of those came from the thirty-six states that have automatic restoration of civil rights, including the right to vote. More than 2,000 such individuals were included on the state's purge lists. Following press and public attention to the situation after the election, the state quietly changed its policy [see Gregory Palast, "Florida's 'Disappeared Voters,'" February 5].
In May 2000 the process went totally awry. Some 8,000 names, mostly those of former Texas prisoners who were included on a DBT list, turned out never to have been convicted of more than a misdemeanor. The new elections director, Clay Roberts, later claimed the error had been caught in time and that none of those individuals lost their rights. But Mitchell admitted that other lists of alleged felons supplied to DBT by the Florida Department of Law Enforcement also contained errors, among them the inclusion of many people convicted only of misdemeanors.
In time, an appeals process was instituted, but in some cases it required ordinary citizens to be fingerprinted in order to prove they weren't the felons they were accused of being. In the end, out of 4,847 people who appealed, 2,430 were judged not to be convicted felons. As Civil Rights Commission attorney Bernard Quarterman put it during testimony in Miami on February 16, "They were guilty until proven innocent."
Elections supervisors in the counties, who had never been consulted about how to assemble the purge lists, battled with the mandate from Tallahassee. "Our experience with the lists is that they are frequently erroneous," Leon County Elections Supervisor Ion Sancho testified before the Civil Rights Commission in Tallahassee. Sancho said he was sent one list with 690 names on it but after detailed checking by his office only thirty-three people were sent letters asking them to prove their eligibility to vote.