Modeling Senate Voting Behavior using Mathematical Modeling.

Using purely quantitative methods to analyze and explain Congress’ actions is rare, yet has the potential to be extremely valuable. We aim to model the voting patterns of the Senate by adapting the Ising model and incorporating economic game theory, with each individual senator being a node with a voting-state and a partisanship value. Our assumptions include symmetry between voting yes and no before the introduction of partisanship and a lack of an external field. The coupling constants of the Ising model reflect the nature of the issue at hand and are individualized for each senator, reflecting their own partisanship as well as the positions of senators who interact with them. The stochastic model was run using Monte Carlo simulations in Python, with several different graph types allowing for multiple ways for senators to be connected to one another. By looking at the properties of the graph such as magnetization and correlations between nodes, we hope to gain insight into how the Senate operates.