Simulating elections under alternative electoral systems
How do election results change when a government changes its electoral system? In my dissertation I develop a computational social science toolkit that produces specific numerical estimates of how a Single-Member District (SMD) country's election results might respond if it switched to a different electoral system. First, I use historical election results and public opinion data to impute preferences in the electorate. Then, I translate game theoretic vote choice models into the computer, and derive new models of voter behaviour, to simulate the electors' vote choices under different rules for translating votes into seats. I apply this toolkit to estimate recent elections in Canada, the United States, and the United Kingdom under proportional and ranked choice systems. The distributions of plausible results under these systems has immediate implications for one of the most important policy conversations in SMD countries.
You can view my 2020 PolMeth poster about this project here.
History of democracies and regime types
I am working on several collaborative projects to create or analyze large crossnational datasets on the history of democracy and political regimes. In a book chapter, my coauthors and I study the effects of institutions on democracy, using a survival model to re-evaluate major hypotheses like Juan Linz's perils of presidentialism. For this analysis, we created an original classification of the existence and selection method of presidents and prime ministers across 17,484 country-years. We also discovered that by dichotomizing the two most prominent numerical classifications of the democratic level of political regimes, The Polity Project polity score and the Varieties of Democracy polyarchy index, it is possible to create a binary dataset that closely agrees with many previous dichotomous classifications of democracy. You can read the latest working paper from that project here. These three projects are joint work with Allen Hicken and Fabricio Vasselai.
Computational social science methods
I have two ongoing projects related to the methodology of computational formal models.
First, I am working on defining a notion of equilibrium in iterative election games. In the period before a large democratic election, thousands of electors, with diverse preferences and very different ways of making decisions, update their ideas about which candidates are competitive and how they should vote. Iterative computer models allow us to study this process without abstracting away the notion of time, and without enforcing the homogeneity that is often required for pencil-and-paper solvability. But this raises a new methodological challenge: how do we know when the model has converged to an election result? Using the example of electors calculating their pivotal probability by consulting a sequence of polls, I demonstrate that certain broad categories of iterative election models produce cyclic behaviours in the electors' strategic intentions, and that those cycles obey simple rules. This simplifies the problem of using formal models to study the process of electors arriving at a vote decision over time.
Here you can view my poster about that project from the 2020 Society for Industrial and Applied Math conference, or download a draft of the paper.
Second, with Walter Mebane and Fabricio Vasselai, I have been studying the effects of strategic voting on the behaviour of the election forensics toolkit. We use Agent-Based Models to check whether or not strategic voting causes false positives in the toolkit's detection of fraud. Here you can read the 2019 PolMeth paper about this project.
I maintain a number of active side projects. At the end of 2019 I noticed that only about 15% of the biographies of political scientists on Wikipedia were about women, while nearly half of them were about Americans, which is a severe bias for one of the most-read websites in the world. So I made one contribution every day of 2020 to the fairness of political science coverage on Wikipedia, focusing on writing new pages about political scientists from groups that are underrepresented on Wikipedia. These efforts only modestly changed the situation, bringing the proportion of political scientist biographies on Wikipedia about women from about 16% up to about 20%, and including the first Wikipedia biography of a political scientist from about half a dozen countries. Wikipedia's content, which is probably the primary way that most people learn about political science, is still very far from appropriately representing the field. I highly encourage anyone interested in helping out to please contact me.
Another active side project of mine draws from my undergraduate education in astrophysics to simulate the very long-term future of low-mass stars using the incredible Modules for Experiments in Stellar Astrophysics. This stellar evolution project is joint work with Fred Adams.
Finally, I've done some work in recreational mathematics, defining a game that translates times of day into simple math problems and then writing a computational solution to the game.