Editor’s Notes: (1) Democracy counts on voters voting; please vote on November 6! (2) This post is written by three undergraduates who spent much of their summer working on gerrymandering. I invited them to share their experiences applying their mathematical and statistical know-how, and working in a multi-disciplinary group to tackle this societal problem. Redistricting affects the House of Representatives, members of which stand for election every two years. Redistricting is a decennial process and the next round will occur as soon as 2020 Census numbers are established. States should have their new maps in place for the 2022 midterm elections.
Our names are Ruth Buck, Katie Jolly, and Katya Kelly. We spent this past summer as undergraduate fellows at the Voting Rights Data Institute (VRDI), a program of the Metric Geometry and Gerrymandering Group (MGGG). We are a combination of recent graduates and seniors in the geography and applied math/statistics departments at Macalester College in Saint Paul, MN. The program was a group of 52 undergraduate and graduate students interested in applying math and computer science to redistricting and voting rights. Over the course of the summer, participants worked on projects related to graph theory, software development, operationalizing state rules for districting, and a wide variety of other topics.
Each week we filled out a survey about our interests and goals and then were assigned to work on one of many projects. Each project group focused on one aspect of gerrymandering research and then presented their work in an informal session at the end of the week. Some groups had more of a theoretical focus, while other groups were much more applied. Then, the next week the process began again.
We rotated projects for the first three or four weeks and then chose a few to focus on for the remainder of the summer. One group created an open-source software to allow community members to create districting plans interactively while another explored discrete compactness measures. An overarching goal of VRDI was to build software that uses a Markov chain Monte Carlo algorithm to explore the space of possible districting plans to evaluate gerrymandering. Participants worked on this project throughout the entire summer, writing and rewriting thousands of lines of code. There were dozens of other projects, many of which are freely available and open-source on the MGGG GitHub page.
The main project that the three of us worked on was creating a shapefile of Ohio’s precinct boundaries. The precinct is the smallest geographic unit that election results are reported at. To do any sort of meaningful, high-resolution analysis of election results, one has to have access to precinct boundaries. The problem that arises here is that for most states, there is no central repository where precinct boundaries are stored.
While states like Minnesota and California store their precinct boundaries on the Secretary of State website for public use, most states, including Ohio, leave the creation and storage of precinct boundaries completely up to the counties. More populous counties are able to employ GIS specialists to manage their precinct boundaries and other spatial data, but smaller counties often lack the resources.
Our project involved many telephone calls to each of the 88 Ohio counties, digitizing hundreds of PDF and paper maps, and innovating ways to use the Ohio voterfile to interpolate precinct boundaries in areas where no other information was available. A voterfile is a list of registered voters that often also includes information about voting history. In Ohio this data is publicly available, but that is not the case in every state.
Not only were we responsible for coordinating the efforts of sometimes more than twenty people, but we were also responsible for teaching the vast majority of whom who had never used GIS how to digitize maps. We gained skills in organization, documentation, and writing clear, concise help guides.
The role of math and statistics in policy issues is often underappreciated. The work we did this summer made clear to us how intertwined math and politics really are. For example, take some state’s rules for redistricting, which may require that a redrawn district be “compact.” Compactness is a complicated mathematical concept with many acceptable approaches. The decision to take one approach over another can have a significant impact on the political environment of the new district.
In public policy, it’s important to be able to communicate results clearly and effectively. Statistics are important not only in terms of reporting numbers but also for providing meaningful interpretations of what these numbers imply. It’s easy to get lost in technical terms, but being able to explain the work to the appropriate audience is invaluable in the world of voting rights.
Gerrymandering is detrimental to the sustainability of our most basic civil rights. MGGG is active and seeking innovative ideas to help combat partisan gerrymandering in redistricting. With the 2020 census approaching, there is an abundance of work to be done in a variety of disciplines. We encourage anyone who is inspired by the goals of MGGG and concerned about the future of voting rights to get involved with the group. More information is available on their website.
We would like to acknowledge the work of Moon Duchin, Justin Solomon, Daryl DeFord, and Aidan Kestigian in organizing a spectacular summer program. Thank you also to the the MIT Prof. Amar G. Bose Research Grant and the Jonathan M. Tisch College of Civic Life at Tufts University for funding this opportunity. We are excited to see the Voting Rights Data Institute continue and make a meaningful impact on the future of democracy.