Report finds George Floyd's cause of death was inaccurately listed in DOJ database

NINA MOINI: Under the Federal Death in Custody Reporting Act, anyone who dies in law enforcement custody must be reported to the Department of Justice. A Marshall Project investigation into the data found that many in-custody deaths are reported inaccurately to the federal government or not reported at all.

One of many deaths that was reported inaccurately was the death of George Floyd. Marshall Project reporters Anna Flagg and Ilica Mahajan worked on this story and join me now to share their reporting. Thank you both so much for coming on today.

ILICA MAHAJAN: Thanks for having us.

NINA MOINI: Anna, I'd love to start with you, if I could. What piqued your interest in reporting on this topic?

ANNA FLAGG: Well, I found out towards the end of, I think it was 2023, actually, that the mortality in correctional institutions data set, which was the federal data set published of all deaths that occur in prisons and jails, was going to be ended.

And that was a data set that had been going on for almost 20 years. And that was a data set that was responsive to the Death in Custody Reporting Act.

And at the time, I wasn't aware of the law behind it and just how all of that worked. But just the fact that data set was ending just seemed like a huge red flag to me and a huge loss for people who are trying to analyze deaths in custody and prevent them in future. So that kind of just kicked us off looking at what's been going on with this law and its implementation and the data.

NINA MOINI: And Anna, would you just share really quick for folks what the Federal Death in Custody Reporting Act just specifically is?

ANNA FLAGG: Yeah, the Death in Custody Reporting Act was passed in 2000. And it mandates that the federal government count the number of people who die in state prisons, local jails, and in arrest-related interactions with the police.

And so it's been around for 25 years. But in all that time, it's just never been fully implemented or enforced. And for that reason, we just don't have a workable, functioning system to keep track of those deaths. And so we don't how many people are dying in these custodial settings.

NINA MOINI: Ilica, would you share how-- was it easy to get your hands on this data or what was that process like?

ILICA MAHAJAN: Well, it depends on what you mean by easy or hard. So we got a hold of the data slightly by accident. So late last year, the DOJ, specifically the BJA, the Bureau of Justice Assistance Administration, posted on their website summary tables of death-in-custody data.

And from those summary tables, we discovered that you could actually click into the summary tables to get row level data that I don't believe they meant to publish. So it was entirely collected by accident.

So we were able to download the row level data from their summary tables and then analyze that data to discover all of the things that were missing and inaccurate in the data set.

NINA MOINI: And obviously, here in Minnesota, the murder of George Floyd is something that is still very resonant with a lot of people. Shocking to find out that something that was so high profile would have been reported in the wrong way. Ilica, were you surprised by that? And what was off there?

ILICA MAHAJAN: Absolutely. So there are different ways that DCRA labels the quote, unquote "manner of death." So one of the categories is death attributed to use of force by law enforcement or corrections officer.

So when we pulled up George Floyd's row in the data set, that's what we expected to find. But the data mislabeled his death as a homicide, which is usually reserved for, for example, an incident between two or more incarcerated individuals that result in a death, which is, by their definition, a killing by someone other than a police officer.

NINA MOINI: So Ilica, tell us a little more about what your reporting found is the central issue or the problem when these numbers are not reported correctly, these cases are not reported correctly. Is it an issue of, then, accountability? Or what's the deeper problem?

ILICA MAHAJAN: Yeah. I mean, so one of the reasons that this law was even passed in the first case is that in order to take this problem seriously and address the issues with in-custody deaths, you have to have a good understanding of what is happening in order to make good suggestions or recommendations about how to reduce or prevent in-custody deaths.

And so if you don't have very good data, how are you going to make good recommendations or solve a lot of potential problems in custody settings? So some of the stuff that we found is that there's over 600 missing names that we just found from media reports.

We found entire states that had failed to report almost any deaths in either their prisons or their jails. For example, there's only one jail death reported for the four-year time span in Mississippi.

So if you don't have a lot of this data, how are you going to identify root causes of death? How are you going to tackle that if you don't have a great understanding of what's happening?

NINA MOINI: So Anna, I wonder if there are people actively working on this to make sure that these numbers are more reflective.

ANNA FLAGG: Yeah, I mean, the way that DCRA has been implemented since the beginning has been really I guess I would say uneven. So there are times in the past 25 years when the data collection was better and then times when progress was made and then times when we lost ground.

And right now, we're at a moment where the deaths in custody data is worse than it was 10 years ago. And I think that's because of a lot of things. One is with shifting administrations you have shifting priorities. And the Department of Justice kind of pursues its priorities based on what's going on in the White House to a certain extent.

But this problem does span multiple administrations. So it's not because one administration didn't care about it. The problem is that in all this time, there has never been a sustained effort for a long enough period of time to get this program up and running and workable and accurate.

And there are a lot of steps forward and then steps back. So there was a time when the Bureau of Justice Statistics, the BJS, was running the data collection. And that's when the data was probably at its best. So they were collecting quite high quality data from prisons and had good coverage in jails as well.

They were still struggling with the arrest-related deaths from police departments, but they had proposed a redesigned plan to try to improve that. But then that whole plan just got scrapped.

The whole thing got moved over to a different agency. Ilica mentioned the Bureau of Justice Assistance, which is a grant-making agency that doesn't have the kind of statistical expertise and experience with the DCRA data collection that would allow them to keep the program running.

So, yeah, it's been a lot of stops and starts. And there hasn't been that sustained prioritization to get it done.

NINA MOINI: So with DCRA, again, the Death in Custody Reporting Act, for anyone joining us, what is the responsibility of those individual law enforcement agencies that are failing to accurately report the deaths in custody, Anna? Do you think that the government needs better oversight there? What did you find there?

ANNA FLAGG: Every agency is supposed to report their deaths in custody according to the law. The Department of Justice has a penalty power that it can use that was written into DCRA in its 2014 reauthorization, which is a connection to this grant funding.

It's called a JAG grant, Justice Assistance Grant. And that's some federal law enforcement funding that is handed out to all the states. And the Department of Justice can choose, if they want, to withhold a portion of that funding from a state that has poor DCRA reporting.

But they've had this power for the last 10 years, ever since 2014, and they've never used it against a single state despite these widespread problems that have been known about for a long time. Despite the widespread problems in reporting of the data across so many different states, the Department of Justice has chosen not to use its penalty power.

So this is a law that basically has never been implemented. And it's never really been enforced.

NINA MOINI: So they would have to get more serious about enforcing it, you're saying, probably, for law enforcement to be more diligent about reporting it. Ilica, I wonder if you could zoom into Minnesota before we let you both go and just let us here in Minnesota how the state is doing in terms of reporting this data.

ILICA MAHAJAN: So in the four-year time period between fiscal year 2020 and fiscal year 2023, the data set reported 157 total deaths. Now, I want to remind you that this is what the data set is reporting. This is probably not all of the deaths that have happened in custody, so we should take that number as an "at least."

Out of those, about 38% of them happened under state or local law enforcement custody. 29% or about 30% were in jail. More than 25% were in prison. And then there were 5% of them that are either unknown or happening in what is a community corrections setting.

And so most of these deaths happening in state or local law enforcement settings is sort of surprising because in Oregon, for example, the breakdown was that 65% of the deaths in custody had happened in prison. So that is what I was sort of expecting to find here. But it was surprising to see that most of them reported were happening under state and local law enforcement.

In 19% of the cases, we had either unknown or an unspecified or other manner of death. So almost one out of five cases, we don't have information about what the manner of death was. And so these things were never-- nobody went back and filled them back out to update the data set.

NINA MOINI: So room for improvement all around. Thank you both so much for stopping by Minnesota Now, for pouring into this data for us all and sharing your reporting with us. I appreciate it.

ANNA FLAGG: Thanks for having us.

ILICA MAHAJAN: Thanks for having us.

NINA MOINI: Thank you. That was Anna Flagg, senior data reporter with the Marshall Project, and Ilica Mahajan, a computational journalist with the Marshall Project.