Blaming the messenger: Police must face up to profiling data

Criticizing racial profiling data for disparities in how drivers are treated is a bit like blaming the bathroom scale when you overeat instead of improving your diet and exercising more.

Luflkin police chief Larry Brazil said last week that the Texas Criminal Justice Coalition's analysis of his department's racial profiling data was "flawed," but the study's calculations were based entirely on the department's own reporting. Reported the Lufkin Daily News ("Police dispute profiling statistics," March 5):

Black drivers are 2.7 times more likely than whites when stopped by police to be asked to consent to a search of their vehicles during traffic stops in Lufkin, and Latinos are 1.9 times more likely to be asked than whites, according to a report by the Texas Criminal Justice Coalition, using 2004 data collected and released by Lufkin Police.

But when the reporter took the department's original report and crunched the numbers, she came up with the same figures.

Numbers from Lufkin's 147 consent searches, broken down by race, included: 61 whites, 60 blacks and 26 Hispanics, according to the department's report.

The consent search numbers seem fairly level between black and whites, until you consider the difference in total traffic stops by race versus how many of each race's stops ended in searches. Of the 10,850 total stops, 6,727 were white and 2,476 were black.

Of whites pulled over, .91 percent were consent searched. For blacks, the number grows to 2.4 percent — hence the coalition's proportional numbers.

So where's the flaw? I sure don't see it. That's pretty much 8th grade math. Instead, Chief Brazil's comments appear to misunderstand what data was being analyzed in the report.

Brazil cited what he said were potential problems statewide in using census and jurisdictional information to set minority population numbers. He likened it to using the population moving through Diboll on U.S. Highway 59 as a sample representation of that city's racial makeup — a flawed method, Brazil said.

Problem is, the TCJC study doesn't compare traffic stop data to census or population figures - it compares search data to stop data. While it's true there's no generally accepted baseline to determine who is driving on the roads, by definition people who are searched at traffic stops are a subset of who is stopped. And since Lufkin PD officers gather the data themselves about the race of drivers they pull over, that number can be known. If the data is flawed, it could only be because Lufkin PD reported bad information, not because TCJC compared it improperly. They simply took the numbers LPD gave them and did the math.

(Personally I think disputes about what "baseline" or denominator traffic stops should be compared to are overblown - in previous studies that used multiple baselines, disparities remained about the same no matter what the reference point.)

That said, I agree with Chief Brazil "who called for a uniform reporting format and data benchmarks." The chief complained of "a lack of legislative clarity on gathering data. There is no standard collection method from agency to agency, and no standard number constituting racial disproportion, Brazil stated."

Legislation that would have created better standards for gathering and analyzing racial profiling data passed in the Texas Senate in 2005, but did not clear the House because of opposition from police unions who want departments to stop gathering data altogether.

Lufkin PD was one of the agencies that did a good job on the front end implementing Texas' racial profiling law, and their policy was judged one of the better ones in the state in a 2002 study by ACLU of Texas comparing departmental policies to requirements in the law. But data speaks for itself, and Lufkin officers inarguably are more likely to consent search minority drivers than white ones, regardless of the chief's best intentions.

There's an easy solution, though, if the chief is concerned about the perception: Stop wasting time on consent searches where officers don't have probable cause. Groups like TCJC, NAACP, LULAC, ACLU and the National Rifle Association supported legislation to ban consent searches in Texas last year. New Jersey, Minnesota, Rhode Island and the California Highway Patrol have already done away with consent searches because they're coercive and potentially discriminatory.

At a minimum, if the chief thinks those disparities are too high he should begin requiring his officers to obtain written consent before searching - when the Austin PD began requiring written consent, the number of people consenting to searches declined by 63%, and the racial disparity also diminished.

It's probably true, as the Lufkin paper reported, that the disparities in who is consent searched at traffic stops "did not reflect the department's anti-profiling stance." But that doesn't mean they're not accurate. Now the department needs to own up to the numbers they report. If those figures don't reflect the department's "stance" then the chief needs to change his policies to reflect his community's real values. His comments in Sunday's paper amount to blaming the scale for an expanding waistline.

For consent search statistics from other Texas law enforcement agencies, see the report (pdf) from the Texas Criminal Justice Coalition and these area-specific fact sheets analyzing local numbers. Also, see prior Grits coverage of TCJC's report here and here.