Fulton County, home to Atlanta, was the epicenter of a contentious and racially charged governor’s race last year between Democrat Stacey Abrams, who is African American, and Republican Brian Kemp, who is white. After Kemp, the former secretary of state with oversight of voting, won narrowly, Abrams accused him of orchestrating a coordinated campaign of voter suppression to steal the contest; she pointed to broken or inadequate equipment in the Atlanta area and the resulting long lines as one example. Kemp has denied the accusation.
But the study offers a nuanced portrait of the causes of longer lines in Georgia and elsewhere — and presents the data as a prescriptive opportunity for election administrators to improve wait times by understanding when and where they occur.
“There’s a lot of things going on in Georgia,” said Charles Stewart, a political science professor at MIT and one of the study’s authors. “It was a big surge in turnout. You can go down the list before you get to the malevolent explanations.”
Stewart and the study’s other authors, for instance, cited relatively less political clout as one possible reason minority-heavy precincts did not have enough resources to avoid long lines.
“The affluent neighborhood — if they are experiencing long lines, it might very well be that the county commissioner lives in the neighborhood and knows what buttons to push, and the people who live in that neighborhood have experience asking for things and getting them,” Stewart said. “Minorities might think, ‘When we ask, we don’t get things.’ Or maybe, ‘We don’t have time to press our case.’ ”
The researchers also explained that the longer lines correlated with a higher proportion of renters vs. homeowners. A surge of first-time, younger or highly mobile voters in these communities may have waited longer because more voters are showing up at the wrong polling location or taking longer to fill out a ballot, they said.
Matthew Weil, another of the study’s authors and director of the Bipartisan Policy Center’s Elections Project, said the wait times “shot up” in cases where the minority population reached 90 percent or more. After Georgia, topping the list of the longest wait times were precincts in South Carolina, Nevada and Washington, D.C.
Weil said election officials need to tailor their resource needs precinct by precinct.
“This should allow jurisdictions to use this data and say, ‘This polling place over here had four check-in tables and eight places to vote,’ ” Weil said. “ ‘That wasn’t enough. We better make sure we up the resources in time for 2020.’ ”
The study collected data from a sliver of the nation’s estimated 116,000 polling locations. But the authors said it is the largest study on wait times to date, covering a representative swath of the country’s demographic makeup over 11 states and 18 million registered voters.
Researchers found that while wait times sometimes exceeded 30 minutes in 6 percent of the precincts surveyed, most lines were very short, with roughly a third of the hourly observations reporting no line at all. But addressing problem spots, the study reports, “is not about understanding how long lines will be on average. Instead, election officials must account for how long the line will be at its worst.”
Ideally, election administrators from participating jurisdictions will study the data and redistribute resources accordingly, Stewart said. After the 2016 numbers were published, for instance, election officials in Fairfax County, Va., learned that the worst wait times occurred in the morning, Stewart said. So they asked for more money and hired poll workers to work extra hours at the start of Election Day, he said.
Trickier scenarios will be harder to fix, including states or local jurisdictions with regulations requiring voting equipment to be distributed evenly according to the population of registered voters. That can make it difficult to address the reality that other factors besides population affect long lines, Stewart said.
And states that use election voting machines rather than paper ballots can struggle to nimbly redistribute equipment, because the machines are expensive and there is a finite number of them. Paper ballots enable officials to set up additional voting stations more easily, he said. That limitation applied to Georgia, he said.
Richard Barron, the top election official in Fulton County, said he is eager to look at the study and try to distribute resources to improve wait times. However, he said he is constrained by the county’s supply of electronic voting machines and by the expectation that new machines being deployed next year will slow voters down even more as they get used to them.
There are also political forces at work, he said.
“This county, politically, it’s split,” Barron said. “There’s a Democratic south and a Republican north. I hear about it if we don’t have everything even.”
The good news in the study: The average voter found 7.8 people in line upon arriving at a polling location, and the average wait time was 8.9 minutes. Wait times sometimes exceeded 30 minutes in 4.8 percent of precincts and sometimes exceeded one hour in only 1.5 percent of precincts.
Wait times were longest when polls first opened, when 21.2 people were in line, on average.
The Presidential Commission on Election Administration has recommended that voters should not have to wait more than 30 minutes.