Abstract by Jacob Murri
Gerrymandering in Utah: Markov-Chain Monte Carlo Methods with Gibbs Sampling
Partisan gerrymandering occurs when the boundaries of an electoral constituency are manipulated in order to favor a political party. In Gill v. Whitford, the Supreme Court wrestled with the question of how to measure a district map’s partisan bias and to create a standard for when a map infringes on voters’ rights. We attempt to address this question in the case of Utah by using Markov Chain Monte Carlo methods to construct a large ensemble of alternative district maps that satisfy the legal requirements of contiguous districts with equal population. We present a probabilistic acceptance function which uses Gibbs sampling to ensure that randomly generated district maps are sufficiently compact and have near-equal population, so as to form an appropriate comparison to the original map. Then the original map’s partisan bias may be compared to the ensemble to determine the extent to which partisan gerrymandering has occurred.
Note: our group (Cristina Lange, Annika King, and I, in that order) would like our presentations to be adjacent