Electoral geographers are mostly concerned with mapping the responses of voters to different political candidates, while they also work to explain the factors that influence those responses. Yet most studies do not consider how different geographic contexts can affect the political perceptions of voters. In particular, people who live in close proximity to one another may come to embrace similar beliefs and values, while broader social and economic processes may divide these individuals into separate camps. Thus, electoral studies performed at the local level may produce different results than those done at the regional or national level. In exploring how different scales of analysis can give different interpretations of voting patterns, this research gathered data from a series of elections that took place in Virginia. These elections, which occurred between 2003 and 2006, span a variety of federal and state offices, with each presiding over a certain geographic jurisdiction. The study proceeded to map the results of each election in terms of three different types of geographic areas: precincts, counties, and legislative districts. The maps displayed the majority winners within each of these areas, giving a rough indication of the bases of support for each of the different candidates. The study then determined the number of instances where two neighboring areas both favored the same candidate, as well as the number of cases where they voted for opposing candidates. These data helped to shed light on the autocorrelation structure of voting patterns in Virginia, revealing how people in the same general vicinity tend to vote together. Overall, the results of this study demonstrate that smaller geographic units (e.g., precincts) exhibit greater autocorrelation in voting than do larger areas. This observation agrees with the concept of sectionalism, which asserts that location and culture are key influences on voting behavior. However, the data also suggest that class differences are a major source of electoral cleavage, as people from different social and economic backgrounds tend to settle in different areas. The use of multiple scales of analysis thus presents multiple explanations for the voting trend of a given location. / Master of Science