|
Activity Number:
|
112
- Statistics and Legislative Redistricting
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 9, 2021 : 1:30 PM to 3:20 PM
|
|
Sponsor:
|
Social Statistics Section
|
|
Abstract #316779
|
|
|
Title:
|
Monte Carlo Methods for Revealing Gerrymandering
|
|
Author(s):
|
Gregory J Herschlag* and Jonathan C. Mattingly
|
|
Companies:
|
Department of Mathematics and Duke University
|
|
Keywords:
|
Gerrymandering;
Redistricting;
MCMC;
Fairness;
High-Dimensional
|
|
Abstract:
|
A powerful method for detecting gerrymandering centers around generating large collections of alternative non-partisan redistricting plans. Novel Monte Carlo algorithms are continually being developed to generate these collections. In this talk, I will discuss several of these methods, as well as discuss how the selection of method and constraints on the space may affect the overall conclusions on whether or not a given plan has been gerrymandered. Examples will include non-reversible Markov chains and hierarchical/multi-scale methods. Time permitting, I will also discuss several investigations into the high-dimensional structure of the space of redistricting plans.
|
Authors who are presenting talks have a * after their name.
