This doctoral dissertation research project will be conducted in two parts. First, a series of agent-based computer simulations will be run. These simulations will build on previous efforts to replicate biased networks that look like actual social networks and will incorporate empirical information about the extent to which the personal networks of participants in existing studies are systematically biased on the basis of personal characteristics or structural properties of relationships (clustering). Second, predictions based on simulations of voter turnout within these biased networks will be tested against recorded voting behavior in a large, urban county in the Southern U.S. A series of interviews with candidates and political parties will be conducted, and information about campaign activity and voter turnout will be mapped using geographic software in order to test the predictions.
This project advances understanding in both the substantive study of voter turnout and in the methodological study of social simulation. Substantively, the project provides a formal model of voter turnout that generates realistic turnout levels, is compatible with both rational self-interest and extra-rational motivation, and can replicate numerous empirical findings. This model provides novel predictions about the best ways to maintain and further increase voter turnout levels in both local and national elections, as well as how to increase the representativeness of the electorate. Methodologically, this project contributes to the specific problem of modeling social context. The intensive modeling efforts detailed above will allow improvements to and verification of a less computationally intensive, more user-friendly approach to modeling social context to be made available to other researchers as a network construction class in a popular simulation toolkit. As a Doctoral Dissertation Research Improvement award, this award also will provide support to enable a promising student to establish a strong independent research career.
