Detecting gerrymandering with mathematical rigor

Abstract

In recent years political parties have more and more expertly crafted political districtings to favor one side or another, while at the same time, entirely new techniques to detect and measure these efforts are being developed.

I will discuss a rigorous method which uses Markov chains---random walks---to statistically assess gerrymandering of political districts without requiring heuristic validation of the structures of the Markov chains which arise in the redistricting context. In particular, we will see two examples where this methodology was applied in successful lawsuits which overturned district maps in Pennsylvania and North Carolina.