Does access to voting locations affect the choice to vote?

One means of increasing voting participation among populations that have been deliberately disenfranchised in the past is to make it easier for them to reach polling locations. The potential abuse of power posed by determining voting locations is underscored by the well-established effect of distance to polling locations on voter turnout. In this work, we address two open questions related to the effect of access to polling locations on voting participation. First, we extend previous studies by evaluating five voting location accessibility metrics in terms of their relationship to voter turnout. Second, in light of concerns about disenfranchisement of non -white voters, we evaluate whether the effects of voting accessibility on participation varies by race. We evaluate voting in the City of Atlanta, Georgia, during the 2020 general election, which occurred during the COVID-19 pandemic. We find that the likelihood of voting differs by as much as 2% for people located one minute closer to a polling location by car, although this relationship is only evident using one of the five accessibility measures. This estimate may be muted due to the prevalence of absentee voting during the pandemic. We also find that the effects of accessibility vary depending on a voters' race, raising questions about the mechanisms behind the relationship between accessibility and voter participation. Overall, our findings point to the importance of equitable provision of accessible in-person polling locations and transportation to the polls to avoid posing barriers to voting, even in regions that have unrestricted access to mail -in voting.