Document Type



This paper evaluates aggregate-level partisan change in presidential and midterm elections at the county level in Georgia, Ohio, and Texas. Specifically, this analysis focuses on how demographic, electoral, cultural, and economic variables affect the percentage of the electorate voting for the Democratic Party candidates for U.S. President and other statewide offices from 1990 through 2016. In addition, this study conducts sub-state regional analyses using U.S. Census Metropolitan Statistical Areas (MSAs) to assess the local nature of partisan change in the U.S. OLS regression and correlation coefficients, as well as difference of means test results indicate that increases in population density over time and the presence of a county in a large U.S. Census MSA of one million people or more increases average Democratic Party vote percentages. Moreover, increases in the African American population in counties is an important positive factor for Democratic Party average vote percentages. On the other hand, increases in median age and median household income decrease Democratic Party vote percentages. Since 1990, there has been a substantial erosion of Democratic Party support across counties outside of MSAs, particularly in midterm elections. Overall, the results illustrate the growing urban/suburban and rural partisan divide in the U.S. at the county level.