Voting Rights Act Research
The Voting Rights Act of 1965 (or simply the “VRA”) is a landmark law in the United States that centers on ensuring electoral opportunity for minority groups, including those defined by race or by a common language. A key element in bringing a VRA lawsuit is a demonstration of racial polarization in elections — does a minority group tend to vote cohesively for certain “candidates of choice,” but see their choices blocked from election by an also-cohesive majority? Due to the secret ballot, making claims about the racial breakdown of voting patterns requires techniques for statistical inference. Another key element is geography, where plaintiffs must show that the minority group could have made up more than half of a district in a reasonably configured districting plan.
In the Roberts Court, there has been a steady erosion of VRA precedent and a ratcheting upwards of the burden on plaintiffs to bring a case. We have a research direction looking at possible futures for the VRA, and racial fairness in political representation more broadly.
The race-blind future?
Under Chief Justice Roberts and a steadily more conservative Supreme Court, there is a clear aversion to the use of race by any agent of the U.S. government — even when race data is used to remediate racial harms. This court famously suspects no use of race data is limited enough for its race-blind ideals of fairness. A high-level overview of how algorithms entered the redistricting story, and how it relates to other areas of discrimination law, can be found in a piece by Moon Duchin and Doug Spencer in the CS/Law symposium: CSLaw 2022
The political science / law team of Jowei Chen and Nick Stephanopoulos published a highly influential law review article called “The Race-Blind Future of Voting Rights” in February 2021. In it, they floated computational techniques — such as those developed in this Lab! — as a less race-conscious way of administering the Voting Rights Act. In a response piece published together with this article, Moon Duchin and Doug Spencer take a close look at the methods and at the high-level goals of the piece, and find some issues. Our response piece, called “Models, Race, and the Law”, and specifically the two-million-map dataset shown in Figure 2, was cited more than a dozen times in the majority and dissenting opinions in the blockbuster Alabama redistricting case Allen v. Milligan.
Measuring electoral opportunity
An electoral district does not have to be majority-minority in order to offer a reasonable opportunity for the minority to elect a preferred candidate. If the cohesion of the minority group for their preferred candidates is high enough, while the majority members “cross over” in support of those candidates often enough, then the district can still be reliably effective.
A paper by Becker-Duchin-Gold-Hirsch develops a new system for scoring a district’s opportunity for minority groups to elect candidates of choice—with this effectiveness score, BDGH are able to generate large ensembles of maps with many effective districts. The main case study is the Texas congressional delegation, where these methods are shown to find maps that provide much more electoral opportunity than the status quo. Furthermore, while it’s a notoriously hard question to evaluate whether a districting plan would pass muster with a court under the VRA, the authors propose a nuanced way to generate ensembles that are VRA-friendly. This paper has now appeared in the Election Law Journal.
Preprint version / Published version
Statistical techniques and related software
The most common technique for determining candidate support by race in voting rights cases is called ecological inference (or EI), which replaced simple linear regression after Gary King and collaborators developed it as a tool starting in the 1990s. There is a collection of R packages implementing EI, but their details and interdependencies were not totally clear to us when we started working in this space. An MGGG team led by Karin Knudson developed a Python package called PyEI to bring cutting-edge computational techniques from the Python world to all of the modeling techniques we could collect.
Knudson-Schoenbach-Becker published a supporting article in JOSS, the Journal of Open-Source Software. You can read it here.