Eliassi-Rad’s article was published this April in Nature. The following is a reposted article from Northeastern Global News with relevant quotations from two of the article’s other co-authors.
By: Cody Mello-Klein
In the decade leading up to 2020, racial inequity in U.S. incarceration was trending down––until the COVID-19 pandemic hit.
In the early days of the pandemic, COVID-19 was hitting prison populations hard, prompting federal and state leaders to release tens of thousands of prisoners. But new data from a team of researchers helmed by experts from Northeastern University, Yale University and the Santa Fe Institute reveals that the mass release didn’t apply equally to people of color. The data, collected and released for the first time, reveals a suite of longstanding, persistent inequities in the criminal legal system and how the pandemic was just the spark that lit the fuse.
“We were able to calculate that 15,000 individuals should have been out of jails and prisons if you had no racial inequity over this three- or four-month period of time in early 2020,” says Samuel Scarpino, director of AI and life sciences in Northeastern’s Institute for Experiential AI. “What we uncovered was the persistent, long-running effects of inequity in sentencing targeted towards Black individuals and also targeted, but to a lesser extent, toward non-white, non-Black individuals.”
For Scarpino and Brennan Klein, a postdoctoral researcher at the Network Science Institute, the project began as a continuation of their data-based work looking at disparities in COVID-19 outcomes. When they started looking into incarcerated populations, they ran into a problem. It’s incredibly difficult to get access to data on the number of people incarcerated in any given state prison system, and when they did get access to it, the data was messy or out of date.
In trying to find the missing data, Scarpino and Klein stumbled on something much larger.
Over the course of 18 months, Klein filed countless Freedom of Information Act requests and dug through numerous agency websites to find the data they were looking for. What he ultimately received was years worth of monthly data for every state’s incarcerated population broken down by race and ethnicity.
“This is the first time a dataset like this has ever been collected and will ever be released publicly,” Scarpino says.
What immediately jumped off the page was that in the 10 years prior to 2020, the relative proportion of non-white prisoners had been slowly but steadily going down, coinciding with a general decline in the incarcerated population. In 2013, 41.6% of the incarcerated people in the U.S. were Black; at the start of the pandemic, it had dipped to 38.9%.
The second thing the researchers noticed was what happened in 2020.
“You go into the pandemic and over the course of the first part of 2020, we had the largest decrease in individuals incarcerated in the history of humanity happening,” Scarpino says. “What ended up occurring was that the proportion non-white [people] went up, reversing a decade-long trend of decreasing racial inequity.”
The mass release of people in U.S. prisons and jails that occurred in the first year of the pandemic dropped the total incarcerated population by 17%, according to Scarpino and Klein. Meanwhile, the proportion of non-white incarcerated people went up by 0.9%, virtually erasing most of the progress made over the last decade. Although it was only a modest 0.9% increase, Klein says even a minor percentage change can make a difference in the lives of thousands of people.
“Because we incarcerate so many people in this country, percentages that don’t necessarily on their surface seem like a huge swing … once you run the math out on those numbers, you end up with a lot of lives that are really severely affected by this inequity,” Klein says.
Klein says if the decrease in the total incarcerated population among white people that happened during the pandemic was applied to the incarcerated Black population, there would have been 15,000 fewer Black people in prison throughout the U.S.
When it comes to the criminal legal system in the U.S., “everything is disproportionate by race and ethnicity,” Scarpino says. But what surprised the researchers was the “key mechanism” that explained how this had happened: sentencing.
“People stop being admitted to jails and prisons because the courts were closed, so nobody’s going in,” Scarpino says. “A bunch of people get released through decarceration for COVID and also get released just because their terms are up. The people left behind are ones that ended up with sentences that were longer than they ‘should have been,’ and these individuals were disproportionately Black.”
According to Scarpino and Klein, there is an active debate among experts in the criminal legal field about what the biggest drivers of inequity are. With this data, they are able to pinpoint a specific factor that, if addressed, could improve the system as a whole.
In a criminal legal system where proving bias or discrimination on a person-to-person level is difficult, that could make a world of difference, Scarpino says. Moving forward, Scarpino and Klein will be collaborating with Harvard University’s Institute on Policing, Incarceration and Public Safety to create a data and legal clinic that will “start both identifying the most compelling cases to put our methods onto and then put people in front judges and in courtrooms to actually start changing these laws,” Klein says. It’s just the start of what the researchers see as the role data can play in helping to affect social change.
“One of the things that we are hoping to do with this paper is to found an entire new field whereby the legal system in the U.S. can remediate racial injustice through big data,” Scarpino says. “We can identify the fact that the whole system is inequitable with respect to even basic things like how long you are sentenced for the same crime and then remediate that through the legal system.”
Cody Mello-Klein is a Northeastern Global News reporter. Email him at email@example.com. Follow him on Twitter @Proelectioneer.