Political pundits seldom pause to ponder polling places. Unless the lines in a given location are so long they discourage voting, the question of where ballots are cast is usually ignored as irrelevant. But wonks — especially those who straddle political science and social psychology — know better. They argue the physical location of the polls not only affects how many people vote; it may also influence last-minute decisions regarding which box to mark or lever to pull.
As the November election approaches, we offer some recent studies that attempt to think outside the ballot box.
Location, location, location: The house-buyer's maxim also applies to polling places. That's the conclusion of a 2005 study published in the Journal of Politics, which found that "small differences in distance from the polls can have a significant impact on voter turnout." Moshe Haspel of Spelman College and H. Gibbs Knotts of Western Carolina University analyzed the 2001 mayoral election in Atlanta. They "geocoded" (now there's a wonky word) each voter's address and calculated the shortest distance between home and their assigned polling place.
Their first finding was hardly a shocker: While distance to the polling place did influence the likelihood of voting, the impact was much greater for households in which no one owned a car. But the researchers were surprised by a seemingly counterintuitive statistic: Moving the location of a polling place actually increased voter turnout. The researchers noted that, since the previous election, the number of precincts in the city had increased from 160 to 168, shortening some distances between voters' homes and the polls. This factor apparently outweighed "any confusion over the location of the polling place," they concluded.
In This Issue
Lobbying doesn't usually work; fat won't kill you; and the Dead Sea doesn't need to die. Check out those stories, our cover story on oxytocin shaking up the field of economics and much more in the September-October 2010 issue of Miller-McCune magazine.
The story is similar in the suburbs. A 2003 study published in the journal Political Geography looked at turnout for the 2000 presidential election in three suburban Maryland counties. J.G. Gimpel of the University of Maryland found that "for each 1-mile increase in proximity to the polling place, turnout jumps by 0.453 percent, or nearly half a point."
But convenience isn't strictly a matter of miles. Since 2003, voters in Larimer County, Colo., have voted not in their precincts, but at one of 32 Vote Centers, according to a 2008 study published in TheJournal of Politics. Located "away from residential population centers and closer to where people travel on election day to work, shop or recreate," the centers service voters from anywhere in the county, providing them with ballots appropriate for their address.
Robert Stein and Greg Vonnahme of Rice University found voter turnout in the county increased significantly after this new system went into effect, and their analysis strongly suggests this is not a coincidence. The Vote Centers "have a positive and substantial effect on individual electoral population," they write. "Moreover, this effect is substantially greater for infrequent rather than frequent voters." If word of this spreads, vote-by-mail could be supplemented by vote-at-the-mall.
Polling places are, in theory, scrupulously neutral places, devoid of visual cues like campaign signs. But according to two recent studies, the building in which a polling place is located can exert subtle but perhaps decisive influence on how votes are cast.
In a 2008 paper published in the Proceedings of the National Academy of Sciences, three researchers from the Stanford Graduate School of Business analyzed the 2000 general election in Arizona, which included an initiative to raise the state sales tax to support education. In the state's slightly more than 2,000 precincts, the researchers found that 40 percent of votes were cast in churches, 26 percent in schools, 10 percent in community centers and 4 percent each in apartment complexes and government centers.
The researchers suspected voters who had to walk by classroom doors or rows of lockers to cast their ballot would be more likely to vote for the school-funding measure. The numbers showed their hunch was right: "People who voted at schools were more likely to support raising taxes to fund education (55.0 percent) than people who voted at other polling locations (53.09 percent)."
A follow-up laboratory experiment confirmed their theory that the voters had been "primed" with the idea of schooling. Participants shown images of a school were more likely to support increased education funding than those who had seen photos of a church. In contrast, those who viewed the house of worship were more likely to support an initiative to limit stem-cell research — a favorite issue of the religious right.
This same dynamic was documented in a study published earlier this year in the journal Political Psychology. Abraham Rutchick of California State University, Northridge, found that during a 2006 election in South Carolina, a proposed constitutional amendment prohibiting gay marriage was supported by 83 percent of voters who cast their ballots in churches, as opposed to 81.5 percent of those who voted elsewhere.
One obvious way to avoid the influence of polling-place surroundings is to mark your ballot while sitting at your kitchen table. According to new research by Barry Burden and three University of Wisconsin-Madison colleagues, voting ahead of Election Day has more than quadrupled since the early 1990s, increasing from 7 percent of votes in 1992 to 30 percent in 2008. (At last count, 21 states allowed some form of early voting).
But does this mean more people are casting ballots? So far, the research suggests not. "We conclude that early voting appeals most to those who are already most likely to vote," write Burden and his colleagues in a paper presented at the 2010 Chicago Area Behavior Workshop. While "early voting might bring out some new voters, on net it reduces turnout by robbing Election Day of its stimulating effects," they add. "This depressant effect is only offset if Election Day registration is also present to provide a vehicle for the last-minute mobilization of marginal voters."
The state of Oregon serves as a case study for this phenomenon. It began experimenting with voting by mail in the early 1980s and has used the process exclusively since 1998 - although, as Reed College's Paul Gronke pointed out in a 2007 paper, "voters may also return their ballots in person on Election Day, thereby rendering many vote-by-mail voters de facto Election Day voters."
"There are good reasons to adopt early voting," he and his colleagues concluded in the journal Political Science & Politics. "Ballot counting is more accurate, it can save administrative costs and headaches and voters express a high level of satisfaction with the system. If a jurisdiction adopts early voting in the hopes of boosting turnout, however, it is likely to be disappointed. We find that early voting reforms have, at best, a modest effect on turnout."
Priscilla Southwell of the University of Oregon, Eugene, came to a similar conclusion in a 2009 issue of the Social Science Journal. She reports that the effect of voting by mail in primary and general elections is "positive but fairly minimal." However, the format apparently increases voter participation "in low-stimulus special elections where the context is a single candidate race, or when a single or a few ballot measures are involved."
All this hand-wringing about hopes to increase voter turnout assumes American voter participation is disturbingly low. But in academic circles, that is a matter of debate. Writing in the Journal of Theoretical Politics in 2006, Lisa Hill of the University of Adelaide in Australia went so far as to propose compulsory voting, which, she argued, would "serve and protect such important democratic values as representativeness, legitimacy and political equality."
In contrast, Michael McDonald of George Mason University contends that "the much-lamented decline in voter participation is an artifact of poor measurement." He argues on the university's United States Elections Project website that when you look at voter turnout as a percentage of those who are eligible to vote rather than as a percentage of the voting-age population (which also includes non-citizens and felons), we're not doing so badly after all.
In the journal The Forum, McDonald reported that 61.6 percent of eligible voters turned out in the 2008 presidential race, "which marks the third consecutive increase in presidential turnout rates since the modern low point of 51.7 percent in 1996. Turnout is no longer declining — if it ever was — and has reverted to the 'high' levels experienced during the 1950s and 1960s." Perhaps in turbulent times, voting takes on an increased urgency, and concerned citizens are willing to literally go the extra mile.