Evaluating Redistricting in Pennsylvania using Monte Carlo Simulation
Open Access
- Author:
- Netznik, Nathaniel
- Area of Honors:
- Mathematical Sciences
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- Joseph Brian Adams, Thesis Supervisor
Ronald Walker, Thesis Honors Advisor
Jeremy Blum, Faculty Reader - Keywords:
- Gerrymandering
Redistricting
Monte Carlo Simulation
Simulation
Probability
Elections
Pennsylvania - Abstract:
- For over 200 years, the integrity of elections in the United States has been threatened by the practice of partisan gerrymandering – drawing an electoral district map in such a way that one political party is favored over another. Opponents have spent several decades challenging allegedly gerrymandered district maps in both state and federal courts. These attempts have been largely unsuccessful. In these cases, the courts have commonly declared that they desire a more rigorous standard by which district maps can be evaluated before overruling them due to suspected gerrymandering. Researchers have developed mathematical metrics, statistical tests, and computer simulations for achieving a possible standard. We sought to build upon this research by developing an alternative computer simulation. Expanding upon our previous research, we developed a model for evaluating the fairness of a district map given its real-world outcome. Provided geographic and voter data, our model generates a sequence of randomly drawn district maps by joining counties and precincts into a given number of districts that align reasonably with certain federal and Pennsylvania state-level requirements. The model then determines the number of Democratic and Republican districts in each map. The model compiles these counts into probability distribution for the number of Democratic and Republican districts in a randomly drawn map. This model, provided appropriate data, can be used to evaluate a real-world election result. Our model determined that the 2016 House of Representatives and 2016 State House elections were fair. However, we identified potential limitations that should be addressed in a future model.