On Sundays, staff writer Paul Wood spotlights a high-tech difference maker. This week, a pioneer working to improve the nation's voting systems. WENDY K. TAM CHO is a professor in two departments, political science and statistics, as well as a senior research scientist in the National Center for Supercomputing Applications. She is working on political redistricting in collaboration with Yan Liu. Cho won the Guggenheim Foundation fellowships last year for her research.
With Yan Liu, you have generated 800 million voter-district maps. You make use of Blue Waters. How does supercomputing work with your political-science research?
My work on Blue Waters is driven by a desire to discover creative and important ways for social scientific progress and societal advances to march alongside scientific and technological growth. I think it has become obvious that the power of information and computing can have an extensive and surprising reach into many realms of life. Our capacities to compile, organize, analyze and disseminate information has increased dramatically and facilitated the creation of many tools to connect citizens and automate human tasks. The role of computing and computing tools are not usually connected to the social sciences, but these very same capacities can shed light onto our governance structures, ideally enabling us to improve our democratic society. I have also worked on a Blue Waters project that re-envisions statistical modeling with the capabilities of massively parallel computing power. This model allows us to make more informative statistical inferences and can be applied to a wide variety of social science (as well as physical science) problems. Both of these projects are unique in that they are in areas and disciplines that have not traditionally been connected to high performance computing.
You are a political scientist doing very high-tech work. Do you also have an interest in coding?
I received my undergraduate degrees in political science as well as applied mathematics, where my applied field was computer science. So, I have a long-standing interest in coding and computer science as well as mathematics. In my research, I frequently combine insights from these seemingly varied, but actually quite compatible and complementary fields. All of my research can be aptly described as highly interdisciplinary. My background enables me to view problems through many different lenses, which has been an enormous asset for me and has fueled my unique approach to many problems. The redistricting project is highly interdisciplinary. Yan Liu adds an enormous amount of additional and critical insights and perspective. He brings many talents to the project as an extremely talented computer scientist as well as geographer. His insights into how to combine spatial insights into the computational algorithms have greatly increased the efficiency of our algorithm, which has been additionally bolstered by adapting our algorithm to take advantage of massively parallel computing platforms. Together, with our combined skills and interests, we have been able to weave a highly interdisciplinary project that combines novel research that advances both the social and computational sciences.
Why do we redistrict?
Reapportionment is mandated by the Constitution to occur after the decennial census. Redistricting follows to ensure that the value of each person's vote is roughly equivalent by equalizing the number of voters in each district.
Does gerrymandering work to damage true democracy?
In most states, the majority party of the state legislature assumes the primary role in creating a redistricting plan. It is a partisan process with politicians enjoying wide latitude in how to draw district lines. The result is a redistricting process that is easily manipulated. Future election results are essentially known even before votes are cast, when the district composition is highly skewed toward one party or the other. In these gerrymandered districts where incumbent politicians essentially hand-pick their voters, elections are not competitive. It is not, as we wish to assume in a democracy, the voters choosing their representatives. Indeed, the average re-election rate in the House of over the last four decades (1972-2012) is 93.6 percent.
By contrast, the average congressional approval rating from the 122 Gallup Polls conducted over the last decade is approximately 15 percent.
Does the Supreme Court restrict gerrymandering?
Yes, both partisan and racial gerrymandering are unconstitutional. From a legal vantage point, the court has made it clear that, at some point, partisan gerrymandering offends our sense of justice and the ideals of democracy. Identifying that precise point is another matter. While the Supreme Court allows the use of partisan information in the redistricting process, the predominant motivation for the electoral district cannot be partisan. How one determines whether partisan motivations are the predominant motive versus one of many allowable motives is difficult to ascertain. We do not have a tool or standard to make judgments like this despite the Supreme Court's articulated framework for decision making in this regard. This is where our computational model comes into play. It is simple to determine the role of partisan information in a computational algorithm that creates electoral maps. Partisan information can be used not at all in the construction of maps, or it can be used at varying levels in relation to other factors. If we are able to characterize what non-partisan maps look like, then we can make judgments about whether or how the map in dispute might have used partisan information.
The concept is simple enough, but the execution is not simple because the computational problem is massive. To enumerate the set of partitions for, say, 55 units and six districts, would require an astronomical effort. There are more than 10 to the 39th power possibilities. In an actual redistricting problem, the number of units can be more than 700,000, presenting a formidable computing problem. Using a computer like Blue Waters helps us begin to untangle how to approach these massively large computational problems.
TECH TIDBITS from WENDY TAM CHO
Do you have any wearable electronics?
I don't have any wearable electronics. I like to collect data, but I also like to remain reasonably "disconnected" from electronics just to give myself some of the peace and relaxation that comes with moments that are distracted only by the enjoyment of one's own thoughts and imagination.
Do you prefer a book or Kindle? What are you reading now?
I much prefer traditional books to electronic books. I have yet to read an e-book in its entirety. I just finished reading "A Mathematician's Apology," by the great number theorist, George Hardy.
What, if any, are your favorite social media? Can you use them in your research?
Social media is immensely powerful and ubiquitous these days. I have toyed with using social-media data for my research, but the data are sufficiently restricted, so that it is not easy to integrate it into interesting questions.