Thursday, July 02, 2020

TCOLE screwed up new racial profiling data required in 2017 Sandra Bland Act

Many kudos to the Houston Chronicle's Eric Dexheimer for identifying a big screw up by the Texas Commission on Law Enforcement regarding racial-profiling data collection by Texas law enforcement agencies.

In 2017, the Texas Legislature expanded the data collected to include after-stop data involving arrests, searches, contraband discovered, and use of force. But when the Texas Commission on Law Enforcement went to implement the statute's requirements via rule making, they excised the "racial" data from the racial profiling reports, making it impossible to use them to assess discriminatory practices.

At first Grits feared we may have screwed up the language in Rep. Garnet Coleman's bill, but at this point all agree the errors arose in the agency rule making process.

By all appearances, this was a good-faith error, and TCOLE has pledged to fix the problem. Normally, Grits might be more skeptical, but TCOLE's chief, Kim Vickers, operates an unusually open and accountable shop, in my experience. Considering they're fixing the problem as soon as it was identified, I'm inclined to afford him the benefit of the doubt. (Since the agency is currently up for "sunset" review at the Legislature, they have lots of incentive to be cooperative.)

Notably, Just Liberty last year analyzed this data as soon as it came out to help construct arguments for limiting arrests for Class C misdemeanors. The 2017 Sandra Bland Act mandated information on Class C arrests and use of force at traffic stops be reported for the first time, and we reported that data by agency. We found Waco PD led the state in Class C misdemeanor arrests at traffic stops, arresting more than 4.5% of driver stopped for these lowest-level offenses. Houston PD used force most frequently, at about one out of every 188 stops.

This was a short turnaround project - the initial data only appeared mid-session in March. Our goal with that analysis was to counter the opposition's argument that arrests like Sandra Bland's are infrequent, so Just Liberty's report focused on total arrests and use of force incidents without attempting to break them out by race. Apparently, nobody else used the dataset until Eric began poking around, so no one discovered the omission.

It's disappointing these data aren't crunched more often, particularly by local reporters and policy makers overseeing local law-enforcement agencies. Grits must admit, I put a lot of stock in transparency and the power of data to confront significant social problems. But data don't matter if no one is using them.

That's even more frustrating because Texas' so-called racial-profiling data are now particularly robust. Once TCOLE fixes its categorization problem, this data set will become one of the most comprehensive of its type, providing policy makers a statistical window into traffic-enforcement activities that constitute a huge proportion of police interactions with the public.

Particularly when it comes to vehicle searches, the ability to tell whether contraband is found will let researchers and police supervisors drill down to discover discriminatory practices. The old data might show black folks were searched more often, but shed no light on whether that was justified.

These sorts of analyses aren't as viscerally compelling as, say, video of a Minnesota cop with his knee on George Floyd's neck. But they help counter arguments that discriminatory policing is confined to "isolated incidents." And they provide policy makers more aggregate information about what, exactly, their officers do in the field than any other source available.


Gadfly said...

Also,assuming TCOLE properly implements the fix, something like this addresses at least in small part the whole issue of implicit bias. If we can make a few more cops think for just a few more seconds, you know?

Unknown said...

Acts of racial profiling reflect an individual officer's actions and decisions. It is an ecological fallacy to take aggregate data (i.e. Sandra Bland data) and make assumptions about individual officers' decision points. Furthermore, disparate data outcomes do not automatically indicate acts of profiling, nor do they automatically alleviate accusations of profiling. The only way to appropriately confirm profiling is to review & investigate each individual complaint. Each police/citizen interaction must be reviewed and accounted for on its individual merits or failures.

Gritsforbreakfast said...

@3:27, who committed this "ecological fallacy"? Point me to a specific example of someone who has misused this data to characterize "an individual officer's actions and decisions." I haven't seen it.

Gunny Thompson said...

From Unfiltered Minds of Independent Thinkers of the 3rd Grade Dropout Section

You can't kill a Dog by only king the fleas on a dog.

My Dear Brothers and Sisters,nothing personal, but why do we insist in continually going down rabbit holes without a GPS?? Texas continues to be a "Tough on Crime"state. As alleged by Unknown, at 7/03/20: the use of date collected under the Sandra Bland Act is misleading and a False Flag'(my words)

Grits, it's not a matter of misusing the data to malign an individual officer. It's what is not included in the official report. Comes to mind is your recent comments regarding a incomplete report filed with the Texas Attorney General by Austin Police Chief, as required by the current regulation. Are we to believe that another law will make any difference in the manner that they are administered? I have no faith in a system that is corrupt to its core. The only solution is dissolving the entire system and create laws that are not written or approved by the Texas Criminal Court of Appeals and/or the Texas Supreme Court.

As I have previously reported, 143 creates a fourth form of government, taking control from local government which are responsible for actions of its personnel.