Publication Type: Conference Paper/Unpublished Manuscript
Abstract: In an effort to add to existing socioeconomic, demographic, and campaign-effects explanations for voter turnout, an individual-level theory of accessibility and turnout is tested here using data on registered voters in four Ohio counties from the 2004 and 2006 General Elections. Using Geographic Information Systems and a combination of individual and aggregate data in a Hierarchical Linear Model and clustered Logit analysis, the results show that decreased polling place accessibility (the combined effects of distance and impedance) has a statistically significant negative impact on turnout. For each additional minute of travel time, a given voter in the sample is approximately 3% less likely to vote when other variables are held constant. While is accessibility is not as influential to turnout as variables such as age and party registration, its effects are demonstrable and statistically significant. The results also suggest that accessibility is a stronger deterrent in non-Presidential (midterm) elections. Analyses of voter turnout which neglect to incorporate polling place accessibility are omitting an important cost of voting.