Map of the Month: Who Votes for Mayor?

By Chris Bousquet • October 31, 2017

What percentage of eligible residents would you guess voted in the most recent Dallas mayoral race? 50 percent? Maybe 40?

In fact, the number is 6.1 percent. This problem of poor turnout in mayoral races is not specific to Dallas, but affects many cities across the country. In State College, PA, turnout in the last election was 9 percent, in Palm Beach, FL it was 10 percent, and in New York City it was 14 percent. Across the 30 largest cities in the United States, the average turnout was around 20 percent.

Persistently abysmal voter turnout was the impetus for the “Who Votes for Mayor?” mapping project out of Portland State University, the most recent winner of Harvard’s Map of the Month. “About eight out of ten Americans live in metropolitan areas,” explained Assistant Professor of Urban Studies and Planning Jason Jurjevich, which means mayoral elections affect a huge proportion of Americans. And yet, “very few people are voting in local elections,” he said. The map greatly enhances understanding of this critical policy issue via rigorous data collection and a sleek UX, and for that reason was chosen as Map of the Month.

It was no secret that turnout in mayoral races is low, but Jurjevich wanted to fill in gaps in the existing data. “There’s not a lot of data on the geographic variability in mayoral voting and turnout tends to be selective in mayoral elections,” he explained. He hoped to gain a more accurate picture of voting behavior and better understand the distribution of voting across cities and demographics. “We wanted to answer questions like ‘Are voters representative of a city’s population?’ and ‘Which groups are over or under-represented at the ballot box?’” he explained.

Graphic showing that only 6.1% of eligible voters turned out to vote: average voter age is 62 and average adult age is 48 years old.
Voter turnout in Dallas's most recent mayoral election via Who Votes for Mayor?

In 2015, Jurjevich and Phil Keisling, Director of the Center for Public Service at Portland State University, received funding from the Knight Foundation to launch a pilot in four U.S. cities— Charlotte, NC, Detroit, Portland, OR, and St. Paul, MN—to begin answering these questions. A year later, the project expanded to include the 50 cities—the 30 largest cities in the United States and 20 cities supported by the Knight Foundation.

Jurjevich and Keisling headed a research team that worked with cities, counties, and states in order to collect voter records from the most recent mayoral election. They then geocoded this data by census tract, and mapped turnout on top of Census and American Community Survey demographic data using a MapBox platform. The result was a census-tract level map of voter turnout that allows users to visualize variation in turnout based on geographic and demographic indicators.

The project was not only intended for Jurjevich and Keisling’s analysis, but was also an opportunity to put data in the hands of residents. “It wasn’t just about producing the data for a report, but also making a website to visualize data, and allow people to download and play with data,” Jurjevich explained. With this in mind, the researchers partnered with communications firm Brink Communications in order to design an attractive and accessible platform for users.

Interface showing NYC voting ages and maps

The resulting website allows users to visualize the geographic distribution of voter turnout in each city and compare this to demographic patterns. On one tab, users can see an overview of the city’s most recent election. For example, in New York City, 13.8 percent of eligible voters participated in the most recent election, which the site displays as a visualization out of one hundred residents for emphasis. This voting population is comprised mostly of older residents—30.7 percent of registered voters over the age of 65 voted, while only 7.6 percent of registered voters aged 18 to 34. Moreover, 7.6 percent of voting age residents in New York City live in a voting desert: a census tract where voter turnout among eligible citizens is less than 50 percent of the citywide average.

Graphic showing percentage of residents voting in New York City's most recent elections: 13.8%
Percentage of residents voting in New York City's most recent elections

Moreover, users can compare factors of their choosing on an interactive map. For example, users can see that in Seattle, voter turnout tends to be higher in neighborhoods in the northern part of the city, which happen to be neighborhoods with higher education levels and fewer people of color.

Comparison between distribution of voter turnout and percentage of residents with a bachelor's degree or above in Seattle with more voters having higher education
Comparison between distribution of voter turnout and percentage of residents with a bachelor's degree or above in Seattle neighborhoods.

Comparison between distribution of voter turnout and percentage of people of color in Seattle, with fewer people of color voting
Comparison between distribution of voter turnout and percentage of minorities in Seattle neighborhoods.

The creators of the site hope that cities will use this data to examine voter turnout and voter outreach processes. For one, Jurjevich hoped to call attention to the shockingly low turnout across cities—in 18 of the 50 cities studied, turnout was below 15 percent.

According to Jurjevich, one of the likely causes of this poor turnout is the timing of elections. In a blog post written by Jurjevich and Keisling, the authors explain that in the Progressive era, many cities scheduled off-cycle mayoral elections—elections in odd-numbered years that do not align with elections to fill higher-profile posts such as governor, senator, or U.S. president—as a means of focusing citizen attention on municipal politics. However, this practice had the unintended consequence of reducing turnout: of the 18 cities with turnout below 15 percent, 14 held mayoral elections in non-national election years and 16 held elections on different dates than national elections. “We’re trying to call attention to the timing of elections,” said Jurjevich. “We hope that cities will reevaluate and look at timing mayoral elections with federal, regional, or state elections”

Revising voter registration rules is another potential means of improving turnout. “In every state, except for North Dakota, you have to register to vote in mayoral elections,” Jurjevich explained. Pursuing interventions like automatic enrollment or “motor voter” laws—where residents register to vote when they apply for their driver’s license at the state Department of Motor Vehicles (DMV)—can ensure that more voting age residents register to vote.

And finally, cities need to ensure that those registered to vote actually cast a ballot—in Dallas, for instance, only 7.3 percent of even registered voters voted. Jurjevich identified different means of ballot delivery as a potential solution to this problem. Oregon and Washington both have instituted vote by mail, where the state sends registered voters a ballot and residents send it back via mail.

Portland—which was number one in term of turnout in the study with 59.4 percent of eligible voters—demonstrates the potentially massive influence of these interventions. Portland has made conscious efforts to improve turnout: “We have vote by mail, elections in a presidential year, and a generally engaged constituency,” Jurjevich explained. “We didn’t set up a statistical analysis to determine the effects at play, but three times the average suggests something unique,” he continued.

Voter turnout across U.S. cities with Portland as highest and Dallas as lowest
Voter turnout across U.S. cities

In addition to inspiring interventions to improve voter turnout overall, the map creators hope cities will use this information to perform targeted outreach in areas with anemic voting patterns. “We also wanted to pinpoint voting deserts and oases—places with lower and higher than average turnout,” said Jurjevich. In Portland, Jurjevich explained, there was lower than average turnout in many areas of lower socioeconomic status and higher concentrations of underrepresented groups, emphasizing the need to improve outreach and voter education in these communities.

In the City of Austin, the Austin Eco Network—a news organization focused on educating Austin residents on local issues—has sought to do just that. The Eco Network recently launched a project called Civics 101, through which the organization informs residents on the importance of local government via a series of events with community members. As a part of this project, the organization has used the Who Votes for Mayor? map in order to emphasize the dire state of voter turnout in Austin, where only 13.3 percent of eligible voters cast ballots in the last election.

According to Amy Stansbury, Editor-in-Chief of the Eco Network, the age disparities revealed by the map provided important insights on where to target outreach. “There’s a whole generation that didn’t vote in the last election,” she said. “A lot of millenials don’t follow local news. Some don’t even know who the Mayor is.” Indeed, only 7.8 percent of registered voters aged 18 to 34 voted in Austin’s latest election, compared to 29.7 percent of eligible voters over the age of 65, giving voters 65 and older seven times greater electoral clout than voters aged 18-34. And yet, Austin has the highest percentage of millenials for any U.S. metro area, making young voters a critical part of the electorate.

As a result, the Austin Eco Network has focused its efforts on the younger generation. “We have happy hour events with music and beer where people present on local government,” said Stansbury. “We’ll ask people, ‘What are the issues affecting your community?’ and explain the huge influence local government has in those areas.”

In order to stress the extreme need for more young voters, the Eco Network has also presented visualizations from Who Votes for Mayor? in unique ways. “We’ll use beer cans to present the information—like this number of beers cans represents the number of people under 30 in Austin, and then these cans are the percentage of those people that voted in the last election. We get some of the biggest reactions from this,” Stansbury continued. Seeing the massive age disparities in turnout tends to energize the younger attendees of these events.

Based on insights from the map, the Eco Network has also made an effort to work particularly closely with traditionally underrepresented communities. One of the largest voting deserts is in East Austin, an area where the voting age population is about 75 percent minority. Stansbury said the Eco Network has focused on this eastern part of the city, “reaching out to very different groups to find a common bond.” The goal is to emphasize that in these communities—which are experiencing rapid gentrification and growing inequality—participation in local elections is more important than ever.

Map showing East Austin's lack of voting locations and high percentage of residents of color
East Austin includes both one of the city's largest voting deserts and one of the highest percentages of minority residents.

To some, the picture of mayoral voting painted by Who Votes for Mayor? may seem like an insurmountable challenge. And yet, Stansbury explained that with efforts to galvanize voting, “it doesn’t take much to get the ball rolling.” “People vote because their friends vote,” she said, which means that convincing one person to vote can have reverberating effects throughout a community. Using resources like the Who Votes for Mayor? map to identify areas and groups with low turnout, cities can effectively target outreach efforts and watch natural social forces amplify their effects.

Like many of the other maps we have featured, Who Votes for Mayor? illustrates the power of place-based visualizations to target interventions in the areas in greatest need. However, this particular map is also a reminder of the salience of even non-geographic visualizations—like the jarring infographics that show voter turnout out of 100 people. In addition to providing valuable insight to inform interventions, data visualization can make problems more apparent and inspire a sense of urgency among residents and policymakers.

About the Author

Chris Bousquet

Chris Bousquet is a PhD student in philosophy at Syracuse University. Prior to that, Chris was a Research Assistant/Writer for Data-Smart City Solutions. Chris holds a bachelor’s degree from Hamilton College.