During the height of the COVID-19 pandemic in 2020, the United States saw its total number of incarcerated people decrease by 17%, the largest and fastest decline in American history. The decline was due to a variety of factors, including states releasing prisoners early to avoid overcrowding and spreading the disease, as well as an overall slowdown in the criminal justice system. But new research into the pandemic's impact on prison populations has unveiled a troubling pattern: white incarcerated people disproportionately benefited from the decrease, while the percentage of prisoners of color spiked. To help give a better understanding of these racial disparities in prisons, Samuel Scarpino, director of AI and life sciences at Northeastern University, spoke with GBH's All Things Considered host Arun Rath. This transcript has been lightly edited for clarity.

Arun Rath: I sketched out broadly the numbers that we were talking about here, but tell us a bit more detail: How did you get this data in the first place, and what did it show?

Samuel Scarpino: One of the big contributions of our study are the data that we were able to assemble. The data that were released as a part of our work represent the most granular picture of who is incarcerated in the United States that's ever been made public. This required web scraping, filing, multiple Freedom of Information Act requests, and then quite a bit of work lining up data across all 50 U.S. states and the District of Columbia. In order to reveal the racial disparity, it required a huge amount of effort on the part of our team to build this dataset, and part of the reason that we wanted to release it publicly was so that the work can continue in a crowdsourced fashion. Many individuals will be able to work with these data to identify other kinds of patterns.

But in broad strokes, what we saw was the largest reduction in incarcerated population in human history. This was driven in part by early releases, court closures, etc. But during this sharp decline, what we saw was an increase in the racial inequity, such that the percentage of incarcerated Black people went up over the course of this dramatic reduction in the incarcerated population.

Rath: Drilling into those numbers, was there a sense of it being worse in some areas than others, and worse in some prisons than others? What more detail was revealed?

Scarpino: Across the U.S., we estimated that there were 15,000 Black individuals who should have already been released if there was no racial inequity. That's across the entire country during this period of time in the middle of 2020, when we saw this really dramatic drop in the number of people who are incarcerated. Just to say that again: these are 15,000 individuals who would not be incarcerated if they were white. The only reason they were still in prison was because of the color of their skin.

Now, different states saw different amounts of this disparity. Interestingly, in Massachusetts, we're one of a few states where the number of incarcerated individuals has actually been declining fairly steadily over the past few years. But even in Massachusetts, we still saw the percent of the incarcerated population who are not white increase during this period in 2020.

Rath: Those numbers are staggering. When you say 15,000 people that would not still be in prison if they were white — I don't know what to say other than that's kind of staggering. Obviously, we know what makes them different. But beyond that, how do we figure out what were the specific causes that were leading to this horrendous disparity?

Scarpino: Well, I want to stress that that 15,000 are individuals who are Black. There was a similar signal in individuals who are nonwhite and not Black. The number isn't as large as 15,000, but that 15,000 number does not represent all of the individuals who would have been released if they were white.

Rath: Can you tell us a bit more about those racial groups?

Scarpino: Individuals who are Latino/Latina, that group also experienced a disparity during this same time period. We saw disparities for individuals who are in Native American groups. Essentially across all groups that are not white, we saw these disparities. Now, the largest disparity in terms of both percentages and numbers occurred in the individuals who are Black and incarcerated, that group experienced the largest disparity and inequity, but it occurred across all individuals who were who are not white in the U.S. prison system.

Rath: In terms of taking this on, is there an avenue to attack this? Could you take this data to the Department of Justice or a federal judge and say this needs some kind of remedy?

Scarpino: Well, we certainly hope that one of the outcomes of this work is remediation of the racial inequity. One of the things we were able to do was to dig in to the different kinds of mechanisms that were at play. There was certainly an effect of who was released. There was certainly an effect of the closure of court systems and racial inequity associated with cash bail. We know that traffic stops were disproportionately nonwhite, and that increased during the lockdowns. But by far the biggest driver was racial disparity in sentencing.

The way that we think about this is that it wasn't necessarily who was released from prison during COVID, nor really who was put into prison during COVID, it was who was left behind. These are individuals who were given longer sentences essentially because of the color of their skin — and that is something that we have to remediate in this country. There was a federal commission on sentencing that found a 20% increase in average sentence length across all different kinds of reasons that individuals were incarcerated solely due to the color of one's skin. And we, of course, also understand the huge disparities that exist for things like crack versus powder cocaine in sentencing.

What we need to do is remediate this. Our hope and what we will be driving forward from this study is addressing the inequity with respect to sentencing and ideally focusing on individuals who should be released because they have served more time than they should have if their skin color was different.

Rath: Well, it's an exceptionally important story, and we really appreciate you bringing this data to our attention and breaking it down for us. Thank you so much.