Nicole Patel

Western Washington University
, Bond Hall 225

Abstract

Measuring Gerrymandering

Gerrymandering refers to the act of drawing electoral district boundaries to give an advantage to a certain political party or group. In this talk, we will examine some of the proposed mathematical methods for measuring gerrymandering. We will begin by discussing the efficiency gap, which compares the number of votes “wasted” (either by losing a district or by winning with excess votes) by each party under a given redistricting plan. Since real life redistricting constraints may make it impossible to achieve a low efficiency
gap, simply computing the efficiency gap for a single plan may not be enough to determine if gerrymandering has occurred. Instead, we can consider the distribution of the efficiency gap over all possible redistricting
plans and see if the proposed plan is extreme. To achieve this, a Metropolis-Hastings algorithm can be used to generate a sample of possible plans. We will discuss the theory behind the Metropolis-Hastings algorithm,  which is a Markov Chain Monte Carlo method. Then we will see this algorithm applied to the problem of redistricting.