QPP 39: Jeff Asher on the transition to NIBRS data collection

Jeff Asher is a data analyst and he talks about the shift in national crime data reporting and collection from the UCR (Uniform Crime Report) Summary Reporting System (S.R.S.) to the National Incident Based Response System (NIBRS). He wrote about it recently for the New York Times.

Audio:

https://www.spreaker.com/episode/43966173

Short version: incrementally better, but the transition is going to be rough, shooting still aren’t a separate category, and just when we most need accurate year-to-year crime data, there’s going to be a break in the system. Sigh.

Here’s the video.

Transcript:

[00:00:04] Hello and welcome to Quality Policing, I am Peter Moskos, and I am back again with Jeff Asher, who is a data analyst and I’m actually not certain of your full title. Maybe you can. What do you do for a living, Jeff?

[00:00:19] I’m a consultant. I do help organizations to understand data and analysis and, you know, really looking for sort of the bottom of the barrel organizations that have no idea what they’re doing and what we bring them to, sort of the amateur level and help organizations to understand largely in the criminal justice sector, what what data is available to them, how they can organize it and what they can be doing.

[00:00:48] Good, good. I know you have too many because you’re just one of the most, I don’t know, respected.

[00:00:56] I don’t use the word because it’s not because you are a professional analyst, but you kind of do a lot of this on your free time and you look at that and collect it and analyze it in ways that it would be nice if the government were paying you to do. But but you’re an excellent source for all things related to crime data. And I’ve asked you to come on today to talk about the change in the reporting system. Can you explain? Well, maybe I can. And you might know more about this than I do. But so traditionally, the UCR, the uniform crime reports, started around 1930, maybe nineteen one is the National Clearinghouse for Crime Data because of the time it was started. You know, it has arson as one of the part one, the major crimes, which isn’t, you know, so common today. But shootings, for instance, aren’t specifically categorized, which is my annoyance. It’s a voluntary reporting system. Somehow there’s some coercion because most departments do it. But there’s an issue of as the data goes up the chain from how the cop responding to the scene categorizes within the department has to record it and then the department has to send it in a different format, usually to the UCR, which is run by the FBI or somehow related to them.

[00:02:11] And and then, you know, nine months after the year in which the data is supposed to be out there, the UCR finally publishes it. So, for instance, we still don’t have official twenty twenty crime data because it won’t be until September when it’s much less useful. So that is the current system. Did I miss something important there?

[00:02:38] I that that’s about it, it’s big and it’s clunky, and it was it was built, ironically, to make it so that agencies could say to journalists that their crime is not out of control. That was that was a major function of a night of UCR originally. And the way that people think of UCR is really is known as SARS and then suddenly blanking on its first date for the summer reporting system. Sorry about that. And so under the summer reporting system, there are seven categories. Arson’s one of them, but it’s so underreported that the FBI doesn’t even estimate the numbers nationally. And then the second system really is called Nyberg’s, which is what my recent piece was about. It was created in nineteen ninety one and it has been almost experimental to this point. It obviously started with zero agencies this last year. Forty eight percent of agencies in twenty nineteen forty eight, which is the year data for right now. Forty eight percent of agencies that are eligible to report reported via neighbors, 90 percent of agencies report overall. So you’re getting and they you know, they connect all the numbers so that the numbers, numbers match up with the SARS numbers and you get national estimates. But basically, we’ve got 90 percent of agencies are reporting via the normal system that we’ve done some thinking. Twenty nine and less than half as of the last time we reported. We’re doing it via this new system. And then they turned off the SARS. They turned off the old system on January 1st of this year. And now all the twenty, twenty one data agencies are only supposed to report Vietnam.

[00:04:32] That seems like we have bad data for twenty twenty one.

[00:04:36] Well, yeah, it’s, it’s the the best number I could get. I asked the FBI about this in February of twenty twenty and we were talking about say a meeting and then it’s like current affairs happen and we sort of put that to the side. And this task about a year later I contacted them again and said no questions and they didn’t reply.

[00:05:01] I did get an answer from the Bureau of Justice Statistics, which says that there was a survey last April which suggested seventy five percent of agencies wanted to report their neighbors or were committed to reporting their neighbors.

[00:05:13] But we don’t know exactly how many actually did report be inaccurate. So if our if our topline is seventy five percent, that means there’s an extra 15 percent of agencies that aren’t going to be reporting to your neighbors, then we have to account for somehow. And that’s basically going to be a guess.

[00:05:30] So this is probably important to point out that the agencies that tend not to report tend to be smaller agencies. So while 15 percent of agencies may not report, it doesn’t mean that 15 percent of the data is missing. So that’s sort of the good news. The bad news is we don’t know about the random element of the missing data and that creates problems.

[00:05:50] And we don’t we don’t actually know where the agencies are going to be. New Orleans, where I’m from, is not going to report the Nyberg’s because their record management system is so old that they can’t match it up. So that’s a huge technical issue and they’re supposedly getting to do our best. But so there’s a report tonight, hours until they get this new arms and our tech projects and law enforcement agencies can can take a little bit of time out of the advantage. A bunch of states like.

[00:06:21] I’m sorry you can go, but if you can get to me, why is the switch being made?

[00:06:27] Yes, so there’s seven there’s also seven states that are not not complying.

[00:06:32] So until those states are not compliant allegedly this year, but who actually knows that California won’t be reporting good neighbors until their Geiger’s comply? So the advantages are that there’s a lot more details in there. Normally, you get very bare minimum of detail, just offenses happen. And so you would get you know, you could count how many murders there were murder. We sometimes got a little more detailed robberies. You got maybe a little bit of locations, but there’s a lot of details in there. And the example I gave is that you can tell in a state how many kidnapings happened at a daycare center instead of just seven categories of crime where you’re going to get crime numbers, you’re getting, I want to say, 32 categories of crime where these more minor crimes that weren’t considered major by UCR, where now we can actually track them.

[00:07:31] The other advantages that you had, SARS had this thing about hierarchy rule, which meant that only the most serious offense was counted. So we’re counting murder every time.

[00:07:39] But if you robbed me and then you murder me, it’s only counted as a murder. If a burglar breaks into your home and you beat the crap out of them or, you know, it might be considered and I guess that’s not a crime. If a burglar breaks into your home and beat the crap out of you, that’s an aggravated assault. The burglary is not counted. And so under SRS, we were getting an incomplete picture. In theory, this will give us a more complete picture.

[00:08:07] You get like if you want to see how many drugs were seized in a state or in a city over the course of a year, you can you can see over time and you can track things over time.

[00:08:18] So it’s an enormous amount of data. It’s a little bit unwieldy. It’s like each year is like six million rows of data, whereas the entirety of UCR is about a million years. And so almost 60 years worth of UCR is is a million rows. So we’re collecting an enormous amount of data. It’s just not really going to be comparable to what we’ve gotten before, which is challenging because it’s sort of a challenging time to be switching off the crime reporting system.

[00:08:49] Even even with a bad system. There’s there’s great value in consistency. So you can compare year to year. And so this year is going to sort of be a new breaking point in the new start, I guess, is sometime. But it would have been nice if it happened, say, and, you know, twenty fourteen when things were calmer. Yeah, yeah.

[00:09:09] It would be nicer. And it was sort of the burning of bridges when you get to to the foreign land type thing where they they’re just not reporting to the old system. And so, so it’d be nice if we had more there.

[00:09:21] When I look at so there’s this is we can get nerdy and technical. That’s what we’re here for. So to see how it is now when I look at a U. Cigarroa, it’s incident based. So it’s one incident and this relates to the hierarchy rule. But if you’re trying to figure out, say, incidents where there are multiple victims on the way currently, as I mean, you know, there’s I’m just explaining it to the listeners.

[00:09:46] Is there different columns that will say number of victims and then there’s, you know, victim one three nine and it takes up a lot of space and they’re usually zeroes for those things for two through nine. But see, in theory, you can tease out that data. But I think most people, quite frankly, who download using our data don’t tease it out because they just don’t understand it. So you don’t get so cases where there are multiple shooting victims are hard to get. And I think there’s a lot of I think a lot of that data is just wrong. For some reason.

[00:10:13] My is my own experience that multiple victims and so on. But the point is, it’s all in one row, one incident and one wrong, no matter what happened during that incident. Is that going to be different now or how how how does that work?

[00:10:26] Well, certainly each incident and I’ll tell you what the offense was, and you can have up to 10 offenses in an incident and then it’ll tell you the details of that offense and you can break it down. You can see the numbers that were cleared by exception. You can see the weapons that were there. I’m just I’m I’ve got right here. I’ve got the I don’t know for sure my screen. I think so. Let’s see. So this is just the state of Colorado and you can see all of the details that you get that you don’t get with with regular. And it gives you the so this is the this isn’t the rosoff. This is the sort of collective stuff from them, but all types of information that like, you know, it depends on what your research question is. What the thing that you want to know is it might be a little difficult to collect all the data, but the idea is that the data is there. And if you do have something.

[00:11:26] That you want to look for if you want to see the relationship between people who want to see the type of weapons that were held, I want to see the type of drugs that were seized. You see the types of injuries that happen. You want to see incidents that were cleared by steps.

[00:11:39] If you want to see the circumstances of victimization, all sorts of different things that you just weren’t getting with under SRS.

[00:11:47] And it’s so clear by exception means that it’s basically covers everything in which the case is no longer being investigated. But the suspect wasn’t arrested. Often it means the witness is dead.

[00:12:02] Sometimes it means the witness. I mean, that’s not the witness, that the criminal is dead. Sometimes it means the person is in prison for another offense and they’re just not going to bother working on the other offense.

[00:12:13] So it’s everything in which the person is not arrested and potentially charged is cleared by exception.

[00:12:20] Yeah, it’s basically any time where you know who the person was that if you’re not arresting them, if the prosecution is denied for non, not just because we didn’t have probable cause for like sort of for good reasons, that it doesn’t seem like that person did it if the person died. If you murder somebody and you go to Jackson, Mississippi, and murder somebody there and arrested Jackson, and I know that you murdered somebody in Louisiana, but you’re in jail, Jackson, and you’re not coming back because they’re going to try to there it’s clear by exception if you’re dead or if the usually if I’m a homeowner, you break in my house, I shoot and kill you. That’s considered a clear exception there because justifiable, you know, all sorts of Dinkins.

[00:13:11] And are we just this is whole change leave you with hope, is it like given that there are inherent problems with switching and the fact that states aren’t complying, which I didn’t know, actually, it seems like a big deal of California is not compliant. Is this is this good?

[00:13:29] Would you have designed a better system or maybe you helped design this system? For all I know, like, are we pleased with the change if and when it works?

[00:13:41] I mean, yes, I mean, I like it in the challenge, the main challenges that we still have shootings and that it doesn’t match up, it’s not a match up of previous years and we don’t have full compliance when we want it.

[00:13:55] We know that, like the major shootings are the one thing I sort of care about my analysis and one. And so that doesn’t change.

[00:14:04] Exactly. And so it’s yeah, it’s it’s got its positives, but it’s got its negatives as well. It’s certainly an advantage over SARS, but it’s like the type of thing that like asking the 10 years the next the next five to 10 years might suck to be someone that looks at crime data and tries to communicate stories. Where are our burrata faster through the roof? Well, no, they were just not counting on the tests. They were happening during other crime. They were happening, you know, they were then the other things that were happening and they were so really they’re just counting more out of this because the same number happened. And they’re just kind of the more I think the the fun stories that someone like the UK crime data is the stories where you say this is happening. This is why it might be happening. This is an important time. You pay attention to the less fun trends are. Yes, this looks like X, but in reality, it’s just because the data sucks. And, you know, like 80 percent of the stories that I have to tell are working on a piece of hate crimes right now. It’s like, well, yes, hate crimes are up, but that’s just the number of agencies that are reporting it up. And it’s you know, you can there’s no trend in hate crimes. That is that is determinable from the available data. It’s just it’s not something we should be trying to evaluate. So I think it’s going to be the issue with neighbors is building and understanding that the system has changed and that there are advantages for people like me to fight the random Joe Schmo on the streets. It’s probably a while before that person is able to understand exactly what’s going on.

[00:15:41] It doesn’t mean anything about this change. Is anything going to increase the turnaround time? Any chance we’ll get, you know, first half of your data before the year is done? Or is that not part of the plan?

[00:15:54] Well, so last year they started doing quarterly data, so they have quarterly data for 20 hours, which they did, as you know, they did quarterly data in the nineteen thirties. It’s amazing that we can’t do the monthly data in the 1930s.

[00:16:08] Oh, so I don’t know how it’s going to be, but it’s all that under stress. So if you’ve got more data to collect, I don’t know if they’re going to do numbers offense’s their divers per quarter. It would be nice. Nighters actually takes longer to compile it out later than SARS. But it’s plausible that if they don’t have to deal with SARS, naggers will be faster to get out. It’s hard to say.

[00:16:38] You know, there could be a funding issue, too, at the federal level, presumably, that they’ll stop cutting money for this type of data collection, at least, which would be good.

[00:16:48] I hope it’s not as slow as the lack of shootings and the lack of timeliness of the two things that the lack of sort of specificity in crime, the three problems that I had with SARS, this sort of specificity issue, it doesn’t inherently stop the other two major issues.

[00:17:05] Yes. Shootings. I mean, in theory, you can get shootings from the Adibi you look for. You look for homicides and aggravated assaults, felony assaults, and then you look at if there’s a weapon in the columns and you know, you can tease it out. The problem is, is when you do that is sometimes sitting simply, don’t you know, there’ll be a crime and the weapon isn’t listed in the initial report or somehow it doesn’t make it up. So that data does not correlate well with when you actually have accurate shooting data from cities. So we know that isn’t good and unfortunately probably won’t get better then. And it’s a lot of it is. Of course, we don’t know we don’t know the departments that record it, you know, do it correctly and those that don’t. And is if you’re a small if you are running a small police department and listening to this, do you know at that on the input level what that what that means for whoever is delegated to compile and turn over this data? Any any advice for small police departments that have to deal with this change?

[00:18:12] No, I mean, I guess automating it as best as possible. You know, there are some agencies that count SARS fighting on this system in the areas that are set up to be automated because there’s no hierarchy or you just report everything that happens.

[00:18:26] So if you have a form that then in there you can do all this on Google forms. I mean, it shouldn’t be it’s more complex, but it shouldn’t be more complicated.

[00:18:38] And so hopefully, if agencies are able to automate that, it makes everything a lot easier. And in theory, automating everything makes it possible to get it out quicker.

[00:18:49] But, you know, so it will work in making a new form that has these categories, which I don’t think which it realistically isn’t going to happen because, you know.

[00:19:02] Well, you got the DOJ is putting it together. You know, they’re set forms to do. It’s just a matter, I think, where there’s a will, there’s. And so I wonder if there’s a welcome DOJ to get this happen. I hope there is.

[00:19:15] Yeah. All right. Anything else? I want to keep this short and sweet.

[00:19:21] It’s a niche topic. It’s very niche. Yeah. I said the story I had early this morning. We like it.

[00:19:28] So had. And so you had a piece last week, earlier this week in The New York Times. That people can feel free to Google Jeff Asher New York Times. You will see a lot of stuff for just your prolific in that way, which is good. Other man is good to see you. Hang in there. And thanks for thanks for being on here. Thanks for having me. This has been quality policing. And I am Peter Moskos and many thanks to Jeff Asher.