Geolocation data show black voters wait longer to vote

Credit: Associated Press

Brett Nelson | Dec 30, 2019

Sections Economics

Collections Politics Race

Free and fair elections are the cornerstone of democracy, but voting in the United States is more convenient for some than others. Long waits discourage all voters and shake their faith in the political process.

Anecdotes and surveys, such as the Cooperative Congressional Election Study, have suggested that black voters endure longer wait times than white voters. A deeper statistical dive—from University of California at Los Angeles’s M. Keith Chen and UCLA postdoctoral research fellow Ryne Rohla, Carnegie Mellon’s Kareem Haggag, and Chicago Booth’s Devin G. Pope—further demonstrates racial disparity in wait times and suggests how to address the issue.

In the 2016 presidential election, just 56 percent of the US’s 245 million eligible voters cast their ballots—one of the lowest turnouts among developed countries for a national election. According to a Pew Research Center survey, 25 out of 32 countries in the Organisation for Economic Co-operation and Development (for which data were available) boasted stronger voter-participation rates in recent elections.

One culprit for the lackluster engagement could be long wait times at the polls. Some research has suggested that long wait times may lead potential voters to leave before voting, and can undermine confidence in the political process. The issue of wait times rose to the national stage when President Barack Obama addressed it during his election-victory speech in 2012. In its 2014 report, the Presidential Commission on Election Administration concluded that “no voter should have to wait more than a half an hour” to cast a ballot. 

To measure wait times during the 2016 election, Chen, Haggag, Pope, and Rohla analyzed geolocation data—captured in the form of pings gathered from voters’ smartphones. The pings corresponded to the location of each phone throughout the day, and were recorded every five minutes or so, any time a mobile application such as a navigation or weather app requested information about a phone’s whereabouts. Using ping data from 150,000 smartphones across 46 states, combined with US Census data from each polling site, the researchers were able to document wait times across the country and break them down by neighborhood racial composition.

The analysis indicates that voters spent an average of 19 minutes at the polls, and 18 percent spent more than half an hour. However, residents of entirely black neighborhoods waited 29 percent longer to vote and were 74 percent more likely to spend more than 30 minutes. The racial disparities persisted independent of differences in population, poverty rates, and other variables, the researchers find.

Pinpointing a precise mechanism for the problem proved more elusive, though Chen, Haggag, Pope, and Rohla were able to dispel a handful of theories. For example, the researchers wondered if black voters, including the many who have inflexible job schedules, were more likely to hit the polls in large bunches, leading to longer wait times—but after examining the data, they find any congestion effects had only a modest impact on the analysis.    

They also examined wait times in areas where Republican-controlled election-official offices might be tempted to inconvenience black voters, who tend to support Democrats. Again, the data didn’t bear this out. “If anything, we find larger disparities in areas that have a lower Republican vote share,” write the researchers.

Chen, Haggag, Pope, and Rohla concede that wrestling with racial disparity at the polls—and ultimately offering suitable prescriptions to fix it—would involve additional investigative research, including a hard look at the amount of resources available to manage specific polling sites.