As state legislatures prepare for the once-a-decade electoral district overhaul, a research team has developed a better method of calculating to help identify inappropriate gerrymandering designed to favor specific candidates or political parties.
In an article from Harvard Data Science Review, the researchers describe the improved mathematical methodology of an open source tool called GerryChain (https://github.com/mggg/GerryChain). The tool can help observers detect gerrymandering in a constituency plan by creating a pool, or set, of alternative cards that also meet legal voting criteria. This set of maps can show whether the proposed shot is an extreme outlier – a shot that is highly unusual compared to the standard of unbiased generated shots, and therefore likely to be drawn with partisan goals in mind.
An earlier version of GerryChain was used to analyze maps proposed to remedy Virginia House of Delegates districts that a federal court ruled in 2018 as unconstitutional racial gerrymanders. The updated tool will likely play a role in the next redistribution using the new census data.
“We wanted to create an open-source software tool and make it available to people interested in reform, especially in states where baselines are skewed,” said Daryl DeFord, assistant professor of mathematics at the University of London. Washington State and co-lead author of the study. paper. “It can be an effective way for people to get involved in this process, especially in this year’s redistribution cycle, where there will be plenty of opportunities to report less than optimal behavior.”
The GerryChain tool, first created by a team led by DeFord as part of the Voting Rights Data Institute 2018, has already been downloaded 20,000 times. The new article, written by Deford with Moon Duchin of Tufts University and Justin Solomon of the Massachusetts Institute of Technology, focuses on how the mathematical and computational models implemented in GerryChain can be used to put plans into context. proposed constituencies by creating large samples. other plans valid for comparison. These alternative plans are often used when a voting plan is challenged in court as being unfair, as well as to analyze the potential impacts of a redistribution reform.
For example, the plan adopted in 2010 by the House of Delegates in Virginia had 12 electoral districts with a black voting age population of 55% or more. By comparing this plan to a set of alternative plans that all meet the legal criteria, lawyers have shown the map to be an extreme outlier of what is possible. In other words, he was probably intentionally made to “pack” some districts with a population of black voters to “break” other districts, thus breaking the influence of those voters.
One of the biggest challenges in creating voting cards is the sheer number of possibilities, DeFord said. Many states like Virginia have hundreds of thousands of census blocks. They also have many rules and purposes for structuring constituencies: like keeping them geographically contiguous and compact with units like counties and cities intact. Many states also want to protect “communities of interest,” a term often undefined, but federal voting rights law is explicitly meant to protect minority voters because historically gerrymanders have sought to weaken the effect of their vote. . In addition, several states require that electoral maps be drawn with an attempt at political neutrality.
Even with all of these rules, voting cards can still be drawn in multiple ways.
“There are more doable plans in many states than there are molecules in the universe,” Deford said. “That’s why you want this kind of math tool.”
Since the advent of computers, models have provided the ability to create an array of maps. Prior to the current version of Gerrychain, many models used a data method called a “flip walk” to create alternatives, which involves changing only one assignment at a time, such as an electoral district or census block. Each change has a ripple effect on the other districts, resulting in a different map.
The tool developed by DeFord and his colleagues uses a method called spanning tree recombination or “ReCom” for short. To create an alternate voting card, the method is to take two districts, merge them before separating them again in a different way. This creates a bigger change with multiple voting blocks changing at once.
The calculator can create many alternative voting plans in a matter of hours or days, and it is available free of charge for use by voting reform groups or anyone with knowledge of Python, the data software that runs it. underlies.
The authors stress, however, that computers alone should not create the voting plan that will ultimately be adopted for use. Rather, the ensemble method provides a tool for analyzing baselines and evaluating potential alternatives.
“It’s not some sort of magic black box where you push the button, and you get a collection of perfect shots,” Deford said. “It really does require serious engagement with social scientists and lawyers. Because the rules are written and implemented by people, this is a fundamentally human process.”