Monday, November 23, 2015

DPS racial profiling analysis 'misleading,' say experts

At last week's House County Affairs hearing, five academics issued an analysis accusing the Texas Department of Public Safety of "misleading" data analysis in its annual racial profiling reports.

Reported the Austin Statesman, "According to an analysis of DPS annual reports between 2003 and 2014 conducted by a team of racial profiling experts, the DPS consistently searches black drivers at higher rates than white drivers, who in turn are more likely to be released with only warnings than are minority motorists."

Basically, when DPS pulls drivers over, white folks are slightly more likely to get warnings, black folks are much more likely to get searched, and data on Latinos is probably invalid because DPS miscategorizes them.

See the full analysis, which, the Statesman summarized, "recommended ditching any comparison to population numbers and looking at what happens to drivers after they are stopped to calculate search, citation and arrest rates."

For the record, that's the approach taken a decade ago by UT-Austin economist and statistician Dwight Steward and Molly Totman of the Texas Criminal Justice Coalition (see here and here), whose study found disparities among discretionary searches at stops made by most Texas law enforcement agencies. (N.b., your correspondent is listed as an editor on both reports.)

The reason for comparing after-stop data is that, from a mathematical standpoint, there's a clearly identifiable denominator to use as a baseline - the demographics of drivers stopped - whereas statewide population data tells very little about the stop demographics in any one location, particularly in the non-urban areas DPS generally patrols.

The analysis prepared for the House hearing was particularly strongly worded: “DPS has consistently misinterpreted statistics showing the racial and ethnic breakdowns of the number of Texas drivers who have been subject to traffic stops in a manner that is highly misleading.”

Perhaps this DPS kerfuffle will result in improving data collection on traffic stops to make it easier to analyze those after-stop issues where the data can more effectively measure the use of discretion (who gets a warning, who gets asked for consent to search). “DPS Director Steve McCraw said that if lawmakers prefer that his agency changes the way it analyzes traffic stops, 'then we’re on board, especially if there’s an indication that it’s more accurate,'”  though he said he wasn't sure that's the case.

Well, Colonel, you may rest assured. From Grits' perspective, the central critique being made isn't remotely controversial. Once you look at the data sources closely, it'd be obvious to a C student in a high-school algebra class. So DPS shouldn't have any problem changing the data elements gathered to make those reports more probative when the Texas Legislature reconvenes in 2017. Time will tell.

No comments: