Friday, March 12, 2021

For journalists: How to read a Texas racial profiling report, or, a case study demonstrating discrimination through police search patterns in Amarillo

Grits has started to see local reports on law enforcement racial profiling data from Texas traffic stops, but it's evident most local reporters don't understand what they're looking at. So they ask the department what it means, the chief soft pedals it, reporter files the story, bada bing, bada boom. That's how a local Amarillo TV station covered the report. Reporters seem to have gotten into a habit of doing the same simplistic story every year, even as the data has improved and much more can be gleaned. 

So Grits decided to perform a brief case study on what reporters can learn from a single department's racial profiling report. Grits has already examined the data at the macro level; now let's look at the micro, using Amarillo PD's report as an example. 

Black folks are slightly overrepresented among drivers stopped by police compared to the Amarillo population. A bit over 6% of Amarilloans are black while 11.1% of drivers stopped in 2020 were black. But because Amarillo PD patrols highways that cut through town and people who don't live in the jurisdiction drive on their streets, it's impossible to say with statistical certainty if the number of stops is disproportionate. To be clear, I believe it is, but I've been round this block for 20 years and there's no denominator to be had which doesn't pose some significant statistical problem.

This is what the Amarillo chief meant when he said it's impossible to tell if racial profiling occurred without a "better way of establishing the demographics of the driving public." He's right. Sort of.

But he's not revealing the full picture, and the reporter didn't know enough to demand he do so. If you're trying to identify racial disparities, look at search data, not overall stop rates. That's because for searches and arrests we have an incontestable baseline or denominator: the overall pool of people stopped.

In Amarillo, we do see disparity on this metric. Black folks represented 11.1% of people stopped but 18.3% of people searched. Black folks were also subjected to "consent" searches more frequently than their share of stops, making up 18.6% of all such searches. So Amarillo cops are both seeking and performing searches on black drivers more often than white ones.

The chief's concern about demographics of drivers vanishes once you drill down to this level of detail.

The number of people searched at traffic stops in Amarillo dropped by nearly half in 2020 compared to 2019. Interestingly, searches dropped more than stops, which declined but not nearly that much. Is this because of officers social distancing at stops, a change in policy, or some other reason? I can see dredging up a story on that topic on a slow news day. ;)

The majority of times Amarillo cops searched vehicles at traffic stops, they found nothing (56% of the time). That's pretty typical statewide.

Some data is easier to read on the TCOLE spreadsheet than in tables in a pdf file, particularly regarding contraband results from searches, where there are columns on the spreadsheet breaking out when contraband was found but no arrest was made that aren't in Amarillo PD's prose report.

Notably, though black people were more likely to be searched, Amarillo cops were more likely to find contraband when searching white drivers. Contraband was discovered 43% of the time white drivers were searched but just 38% of the time they went through black people's cars.

This is a key racial profiling indicator. If cops are searching black folks half again as often as their proportion among drivers but disproportionately finding contraband more often on white folks, that signals a discriminatory pattern.

This is one of the key, new calculations made possible by the 2017 Sandra Bland Act, but Texas journalists haven't figured out how to use the new resources available to them.

Searches of black people in Amarillo found contraband 35 times, which would not be too far out of line compared to the proportion of times they were searched (even though it's FAR greater than their proportion of Amarillo's population). Even so, 31 of the 35 times contraband supposedly was found in a black person's car, it was so minor they weren't arrested and were sent on their way. Couple that with the majority of arrests that found nothing, and that's a LOT of drivers subjected to pointless fishing expeditions.

By comparison, contraband was found on white folks in 105 searches, and the driver was arrested in 10 of those episodes.

From a journalistic perspective, I'd consider this a big story. Sure, the racial disparities in arrest rates aren't big. But if 90%-ish times when contraband is found, the driver is let go, what's the point of searching to begin with? Either the definition of "contraband" being employed is so broad as to include things that aren't illegal, or Amarillo cops are being awfully lenient when they find evidence of crimes. IRL, Grits suspects it's the former; this data pattern is a bit of an outlier but most agencies release the majority of drivers on whom contraband is found.

Statewide, black folks were slightly more likely to be arrested than white folks when contraband was found (38% compared to 33%), but the percentage for Hispanic folks was even higher (40%). Big gaps in those categories could mean one group is receiving lenient treatment and might be an avenue for local journalists to pursue.

Another important story from these reports: Amarillo PD arrested 115 drivers for a Class C misdemeanor traffic violations in 2020. With 215 total arrests at traffic stops, more than half of arrests at traffic stops were completely avoidable and likely represent pretext stops where the officer is fishing for some other crime. Those would cease if the Texas George Floyd Act became law. Statewide, Class C arrests were only 21% of the total, so this is happening more in Amarillo than elsewhere. What reporter wouldn't want a local news story linked to pending state legislation heard in the next few weeks? Well, it's available to you, if you know what you're looking at.

Notably, black folks made up just 11 of 115 people arrested for traffic violations at Amarillo traffic stops in 2020, so the Class C arrest ban in the Texas George Floyd Act isn't particularly a racial issue in Amarillo. Or black folks in Amarillo simply "consent" to searches before facing arrest, as noted above. Regardless, I'm pretty sure white Amarilloans don't appreciate being arrested for petty BS any more than black folks do. 

The maximum penalty for Class C misdemeanors is only a fine, not jail time. By arresting someone for a Class C and taking them to jail, police are inflicting a punishment more severe than a judge or jury could do.

Amarillo PD reported using force resulting in serious bodily injury at only two 2020 traffic stops. If local reporters don't know those cases are, maybe someone should look them up? Could be a story there.

There are many reasons for justice-beat reporters to examine Texas racial profiling data now that it's been enhanced via Rep. Garnet Coleman's 2017 Sandra Bland Act. In some instances, yes, reporters may find notable racial disparities: In Amarillo, for example, it appears black drivers are disproportionately searched compared to how often they're pulled over, are asked for consent to search more often than other drivers, and have contraband discovered less frequently than do their white counterparts.

But there are also news stories buried in these data beyond racial discrimination: Most immediately in relation to the Texas George Floyd Act, Class C misdemeanor arrests (both traffic and municipal ordinances), and use of force. It also might be informative to compare data from nearby jurisdictions. Often, as demonstrated in this example, differences in search practices from department to department are more significant or telling than the statistical differences among the races.

Not everything important or interesting in this data is about race. Indeed, the report answers basic questions few reporters ever consider: how many stops are made overall? How many of those people are searched? How many are arrested? How often is contraband found? Turns out, even overall traffic-stop numbers change significantly year by year. 

The decline between 2019 and 2020 stop levels is understandable, but what explains the near doubling from 2017-2018? Traffic stops were declining during this period statewide. If I were a local reporter in Amarillo, I'd be looking into that story, which is interesting irrespective of racial breakdowns. 

In a town of fewer than 200,000 people, both the doubling of traffic stops to the 2018 apex and the less-than-might-be-expected decline in 2020 could be significant. Digging around in the whys of it, combined with poking around TCOLE's statewide spreadsheet and a little traditional legwork, could generate many different types of local stories. This information isn't available from any other source.


Debbie said...

From El Paso last year:

As a result of this journalism, the city of El Paso's police racial justice reform plan, devised in the wake of George Floyd protests, calls for an independent study by a third party to review the police department's racial profiling reports and stats.

Gritsforbreakfast said...

Good job, I hadn't seen that. The search/arrest/contraband data are really underutilized, and the 90% consent search rate you found is BONKERS! Need more of this.

Anonymous said...

Is there data for how often officers ask for consent and are told no? I feel that could be another data point to support or counter the theory of racial disparities in Amarillo.

Gritsforbreakfast said...

That's a missing element. We know how many consent searches are performed, but NOT how many are declined. And of course some of those who say no are arrested and their vehicles searched anyway, like Dillon Puente in Keller, TX.