TAMPA — Hillsborough County's elections chief is demanding a correction to a study of the 2012 presidential election that found none of his polling places closed on time.
Supervisor of Elections Craig Latimer said the national racial justice group Advancement Project used faulty data to conclude that Hillsborough precincts closed an average of 72 minutes late.
"That simply is not true," Latimer said.
The report prepared by the group was submitted last month to the bi-partisan Presidential Commission on Election Administration, which was looking into reports of long waits to cast ballots during the last election.
The group relied on data recorded when voting results from each precinct were transmitted to the supervisor's central tabulation center. A more valid ending time, while still problematic, would have been when his poll workers shut down machines and spit out reports from them, Latimer said.
And even that doesn't necessarily reflect when the last person voted, he said.
Still, by that measurement, Latimer said, 65 percent of his 347 precincts closed within 30 minutes and nearly 85 percent closed within 45 minutes.
A representative of Advancement Project said it will be happy to look at the records to which Latimer is referring and will amend its report if necessary.
But she says her group stands by the point of the study, which found that precincts with concentrations of minorities and young people tended to close later than those with more white people.
The study's authors, two university professors, used closing times as a "proxy" for how long people had to wait in lines after polls shut down at 7 p.m. The challenge is that elections supervisors aren't required to track wait times and report closing times differently from county to county.
"We're very, very happy to receive the other data," said Katherine Culliton-González, senior attorney and director of the voter protection program for Advancement Project. "We do stand by the conclusion that some people in some precincts waited longer than others. The disparity is going to be the same."
She said Latimer's office never responded to a public records request, so the researchers relied on information gathered by the Orlando Sentinel newspaper for a project it did on the 2012 election.
Latimer said he was not aware of a records request from the rights group researchers or the newspaper.
But he said there's no way the group can draw conclusions about disparities in wait times for voters of different races or ethnicities. The data the group used reflected the time after which precinct workers accepted the final vote, printed out reports from vote counting machines, transported the results by vehicle to regional collection centers and then transmitted them to his central vote-counting office in Brandon.
"I don't know how you can make a determination of which precincts had lines when their information is inaccurate," Latimer said.
Times researcher John Martin contributed to this report. Bill Varian can be reached at email@example.com or (813) 226-3387.