Skip to main content
Open Access Publications from the University of California

UC San Diego

UC San Diego Electronic Theses and Dissertations bannerUC San Diego

Introducing "Clustering:" Redistricting in Geographic Perspective

  • Author(s): Levitt, Justin Mark
  • Advisor(s): Kousser, Thad
  • et al.

Previous research in redistricting has treated geography and institutions as two separate, disconnected questions. Geographic variation in partisanship and race has mostly been treated as a structural question best suited for long-term, big picture discussion (e.g. Rodden 2009, Bishop 2005). As a result, scholars looking at institutional variation between redistricting systems have normally treating geography as control variable distributed randomly throughout the jurisdiction (Masket 2012, McDonald 2006). I bring these two dimensions into dialogue with each other first by looking at how spatial autocorrelation—clustering—shapes tradeoffs within a single geography and second, how legislatures and commissions make different choices when faced with similar levels of clustering. I test the former by simulating plans that maximize each of competitiveness, compactness, and the number of majority-minority seats in two states using the Better Automated ReDistricting (BARD) tool. I show that the more highly clustered a state is by race, the more majority-minority seats can be drawn. Conversely, states highly clustered by party produce fewer competitive seats. I also show that the tradeoffs between compactness and competitiveness are larger in a state more highly clustered by party while the tradeoff between compactness and number of majority-minority seats is smaller in a state more heavily clustered by race. In addition to the direct impacts of clustering, I show that once we control for the level of clustering in a state, redistricting commissions in 2011 produced more compact and competitive maps than their legislative counterparts. However, commissions are more sensitive to the degree of clustering than legislatures and consistently refuse to trade compactness for either competitiveness or more majority-minority seats under extreme levels of clustering. What this demonstrates is that commissions can produce plans that are more compact, competitive, and contain more majority-minority districts—as long as it doesn’t interfere too much with the general appearance of the district. Ultimately, this project shows the danger of studying a geographic phenomenon without thinking about the underlying geography in a systematic way.

Main Content
Current View