In our work as advocates for using statistics and data science for the greater good, we often encounter situations in which analytics fail to deliver the benefits they are designed to produce. Data for Good is such a powerful practice because it takes us beyond searching for mathematical answers to problems into empowering the development of real solutions.
One of the most concerning areas in which practice can fall short of promise is the law and courts. This month, Stats4Good looks at some of the ways statisticians and data scientists can support equity and fairness in our legal system.
Why Local Action Matters
Issues in our legal systems can be detected—or caused—by statistics and data science. Especially highly local issues that disproportionately affect certain communities. National studies can explore and document mass incarceration, but acting on those findings is subject to state laws and community-level practices. Local action is needed. This is why Data for Good activists must bridge science and advocacy to support positive action. There are many instances in which our unique skill sets are essential for working with others to develop and implement changes in policies and practices.
After years of working with Department of Justice data, I know lack of completeness and consistency are at the core of many problems—so much so that pulling the data together into a format amenable to analysis is often the greatest challenge. This is why our skills are so important for driving change for the better. To help address these weaknesses, the Prison Policy Initiative provides information and training on data sources.
The Human Stakes of Algorithmic Justice
We have all seen how the explosive growth of AI has led to using it in almost every conceivable scenario. The law is certainly no exception. But the problems with AI algorithms in legal applications are especially acute because of the impacts courts can have on peoples’ lives—contexts in which accuracy, reliability, and potential for bias and hallucinations become especially concerning.
One example of this issue is using risk algorithms to predict recidivism, the likelihood of a person committing another offense in the future. These algorithms are used to set bail, make sentencing recommendations, and decide whether a prisoner is paroled. Such practices have come under heavy criticism for the potential bias against poor defendants, minorities, and racial groups.
A landmark 2016 study by investigative journalists at ProPublica showed systematic disparities by race in the use of one such commercial algorithm in Florida. The local nature of these AI applications and their usage creates a great need for statistical experts to scientifically test them for bias.
Another area of concern is the United States’ longitudinal and sustained increase in incarceration rates. Michelle Alexander’s The New Jim Crow: Mass Incarceration in the Age of Colorblindness covers this subject in detail. The problem has only gotten worse over time.
Through much of the 20th century, the US had incarceration rates in the middle 100s per 100,000 in population. Starting around 1970, this rate gradually rose to more than five times the incarceration rates of Western Europe and Canada.
The past few years in the US have seen an especially sharp rise. During this time, the incarceration rate increased 29% above levels from just 10 years ago—despite crime rates falling over the same period by 14% for violent crime and 38% for property crime, according to FBI data.
Justice advocacy through Data for Good is uniquely qualified to address these issues. By supporting equal justice for all, Data for Good activists are empowered to go beyond questions of law and use science to address problems in our justice system—because if we aren’t treated equally under the law, there is no justice.
Getting Involved
Now is a great time for Data for Good activists to find summer volunteer projects. A great way to start is with an organization you already know or are already involved in. Look for ways statistical science can help with their mission.
Readers, our work together has received some important recognition recently, including this author being named a Fellow of the American Statistical Association. Statistics and data science for the benefit of all has played such an important part in reflecting the impact we’re making. Let us be encouraged and empowered to carry on our work together in Data for Good!
David Corliss
With a PhD in statistical astrophysics, David Corliss works as a data scientist in industry. He serves on the ASA Board as a Council of Chapters representative and is the founder of Peace-Work, a data for good nongovernmental organization.
This column is written for those interested in learning about the world of Data for Good, where statistical analysis is dedicated to good causes that benefit our lives, our communities, and our world. If you would like to know more or have ideas for articles, contact David Corliss.
Leave a Reply