Session 7: Polling Locations

Thursday, July 28 4:25 PM-5:30 PM

Chair: David Kimball, University of Missouri-St. Louis

Discussant: Michael Dickerson, Mecklenburg County, NC Board of Elections

Paper #1: When and Why Does Moving Polling Places Move Access for Some More than Others?

This study examines the immediate and protracted effects of changes to polling place locations on voter turnout, particularly among historically marginalized segments of the American electorate. With the 2013 landmark U.S. Supreme Court decision, Shelby County vs. Holder, opening the door for unfettered electoral reforms, such increased likelihood of abrupt modifications to election administration gives rise to the very real danger that such measures suppress or deny participation in the electoral process. For that reason, I identify as well as interrogate which jurisdictions are most likely to move or close voting locations; this, to get a better idea of when or why they do become so prevalent; and most importantly, whose (i.e., voter demographic) polling places are being changed. Implicit in this analysis is the concern that communities of color are the most directly and adversely impacted.

  • Alejandro (Alex) Flores, MIT Election Data and Science Lab

Paper #2: Vote at any Polling Place: A Case Study of St. Louis County, Missouri

In 2020, St. Louis County, Missouri instituted a new procedure allowing Election Day voters to cast their ballot at any of the county’s 230 polling places. The county also created an App for voters to check voting lines at each polling place. This change was accelerated by the coronavirus pandemic, in which private venues backed out of polling place commitments. In the November 2020 election roughly 30 percent of Election Day voters cast a ballot at a polling place outside their home precinct. We believe St. Louis County is the first local jurisdiction to adopt this type of system, at least among states that do not allow early voting. We rely on interviews with key election staff and data on election administration and voting behavior to document the new voting procedures. We use a systems framework to examine this case of local election administration – changes in one part of the system affect other areas too. The new voting process in St. Louis required new voting technology, but also changes in polling place siting and staffing, poll book procedures, as well as other technological and personnel adjustments. All of these changes had to work seamlessly for the new system to succeed. We also document anticipated and unanticipated changes in voting and administrative outcomes. This case illustrates the interdependent nature of many features of election administration.

  • David Kimball, University of Missouri-St. Louis
  • Anita Manion, University of Missouri-St. Louis
  • Joseph Anthony, Oklahoma State University
  • Adriano Udani, University of Missouri-St. Louis

Paper #3: A Substitution Effect, a Turnout Effect or Both? Changes in Distance to the Early Voting Site and Voter Turnout

Relatively small changes in polling location can have large effects on voter behavior. For example, the farther voters live from a polling place, the less likely they are to cast a vote (Dyck and Gimpel 2005). These effects are particularly strong among people from lower income precincts and who do not have access to cars (Haspel and Knotts 2005). While these studies tell us a great deal about how voters respond to stable, long term costs of voting, we know very little about how changes to precinct location influence voter behavior.

This paper uses a natural experiment resulting from North Carolina’s “Uniform and Expanded Early Voting Act” to investigate how voters respond to changes in early voting sites. Early voting was mandated to be open from 7 AM until 7 PM on weekdays, which increased the number of hours available statewide. But this change also brought about a decline in the total number of early voting locations within 70 of 100 counties. Using multivariate models combining geographic placement and distance changes, voter history and registration data, this paper demonstrates that distance changes in early vote locations reduces the probability that a voter will cast an early ballot, especially among demographic groups with traditionally lower turnout rates.

These results augment existing work by examining the effects of changes, rather than static poll placement, and by examining the effects not simply on the choice of whether to vote, but also how a voter chooses to vote (early, mail, or election day).

  • Michael Bitzer, Catawba College
  • Tyler Duke, The Raleigh News and Observer
  • Christopher Cooper, Western Carolina University