For the second consecutive legislative session, the 88th Legislature is considering recommendations for re-thinking the role of the Texas Commission on Law Enforcement, or TCOLE. The Texas Sunset Advisory Commission said the state’s approach to regulating law enforcement was “toothless,” “fundamentally broken” and “largely ineffective.”
At Texas 2036, we have been supporting recommendations through our own work and research, which culminated in the Texas Law Enforcement Data Landscape. The report is a deep dive into how TCOLE collects and manages various law enforcement data and how Texas handles so-called “wandering officers.”
Here’s a look at one part of the report that focuses on how law enforcement agencies report motor vehicle stop data to TCOLE and how it maintains and publishes that data. Virtually all agencies in Texas report annually to TCOLE how many stops they conducted. They also report data about the suspects, for example, race and gender; the outcome, including verbal warnings, arrests and citations; and other circumstances, such as if there was a search, if contraband was found and if physical force was used.
The report found significant deficiencies in both data quality and data management, and offered the following high-level findings and recommendations:
Data Quality
1. The data reporting meets basic statutory requirements; however, the available information is unreliable in helping ascertain the nature and extent of racial profiling during motor vehicle stops.
The data collected and reported must contain all data variables necessary to conduct a comprehensive analysis. This includes the data needed to conduct the Gold standard analysis recommended by the CPE/NYU Guidebook and discussed in this report. Finally, an analysis that imparts actionable insights and targeted solutions requires incident-based, micro-level data, and more complex statistical techniques.
Recommendations:
A. The Texas Legislature should consider adopting additional data collection fields that are not expressly required under statute but offer more comprehensive information needed to conduct a robust analysis. This includes the collection and reporting of demographic information on the officer making the stop.
B. TCOLE should collect the incident-based, micro-level data to provide the opportunity for more robust and meaningful statistical analysis to the extent allowable under current law. TCOLE should consider partnering with academic institutions to perform comparative analyses for LEAs and to conduct an annual statewide statistical analysis of the incident-based micro-data. The analyses should utilize bivariate and multivariate techniques that employ summary and descriptive statistics and inferential analyses, including methods such as regression analysis.
C. TCOLE should consider releasing the raw data for analysis by stakeholders in law enforcement and the broader public.
2. There are inaccuracies in the Excel spreadsheet maintained on TCOLE’s website. In addition to collecting a complete data set, data quality also requires accurate data to ensure data integrity.
Benchmark’s preliminary assessment included an initial review of various annual report narratives and TCOLE’s Excel spreadsheet populated with the aggregate data for 2021, reported by LEAs. That assessment reveals that the data reported and maintained in Excel format on the TCOLE website is inaccurate in some instances. For example, there appear to be errors in the stop data reported for use of force, injury to the suspect, injury to the officer, or injury to both. The use of force resulting in physical injury is the only use of force reported to TCOLE. Simple errors erode data quality and undermine the integrity of the data set.
Recommendations:
A. Considering the significance of auditing practices and procedures to the quality of the data and validity of any ensuing analysis, the Texas Legislature should expand the statutory auditing requirement to include the audit of all racial profiling data collected and reported by LEAs to TCOLE, instead of an audit that only verifies data is reported on the race or ethnicity of the driver. This is needed to ensure the data is accurate, complete and includes all data points.
B. TCOLE should continue to hold chief administrators accountable for ensuring their LEAs: (1) submit a report for all applicable motor vehicle stops, and (2) the report is accurate and contains no missing data. LEAs should also conduct regular and systematic audits to correct reporting errors, missing data, etc., which may include an audit committee.
C. TCOLE should embed data validation formulas in the reporting spreadsheet to help eliminate logical inconsistencies and missing data. TCOLE should also include data audit standards in its guidelines for data compilation and reporting, including how to handle incident-based reports with missing or inaccurate data and the extent to which LEAs should include that incident in the aggregated data set.
D. TCOLE should develop and conduct training on the data audit standards set forth in the recommended guidelines. It should also develop and conduct training on when and how to collect and report motor vehicle stop data, including complex scenarios such as reporting requirements when a passenger is arrested because of the stop, etc.
E. TCOLE should utilize established standards to evaluate LEA performance based upon metrics such that agencies which meet and exceed standards are acknowledged and those which fall below standards are trained and remediated.
Data Management Practices
3. TCOLE currently does not have general rule-making authority to set new rules and standards related to the required fields for racial profiling reports.
Recommendation:
The Texas Legislature should consider granting TCOLE explicit rulemaking authority over the required fields for racial profiling reports such that LEAs submit data to withstand academic scrutiny.
4. In 2017, the Sandra Bland Act explicitly required TCOLE to revise its guidelines that provide the standards for compiling and reporting information required under Article 2.134 “better withstand academic scrutiny,” by September 1, 2018.
The report includes the aggregated data set and comparative analysis of that data. TCOLE’s instructions, sample forms, and spreadsheet do not provide the direction necessary to address the data quality and enforcement issues identified in this report.
Recommendation:
TCOLE should develop comprehensive guidelines with sufficient detail to address the data quality issues raised in this report, including standards to ensure the data collected and reported by LEAs is consistent with the statutory requirement. The guidelines should include information on: (1) how TCOLE tracks LEA reporting compliance, (2) the LEA behavior that manifests the intent needed to trigger disciplinary action, and (3) the disciplinary process and potential outcomes.
5. TCOLE currently requires LEAs to submit the detailed written racial profiling policies adopted by LEAs under Article 2.132(b).
Recommendation:
TCOLE should create a sample Racial Profiling/Motor Vehicle Stop Data policy to ensure LEA policies are consistent with the statute and contain all of the requisite elements, including the manner and process by which an individual may file a complaint alleging racial profiling, the specific information officers must collect upon making a motor vehicle stop, standards for reviewing video and audio documentation, and the improvements the LEA can make in its practices and policies based on its review of the stop data collected.
6. Transparency for the public is required by section 5.08(2) of the Sandra Bland Act.
The Act mandates TCOLE to provide, in a downloadable format, all information reported under statute, which specifically requires aggregated, incident-based data compiled during the prior year. In addition to the aggregated data, the statute clarifies that the report must also include a comparative analysis and information regarding complaints alleging a peace officer engaged i racial profiling. TCOLE does provide a downloadable Excel spreadsheet with the aggregated data. It does not provide the comparative analyses submitted by the LEAs, in downloadable or any other format. In addition to the aggregated data, comparative analysis, and information on complaints, TCOLE must also provide a glossary of terms so the public can readily understand the report. The glossary is not available.
Recommendation:
TCOLE should make LEAs’ Motor Vehicle Stop Data and the required glossary should be available to the public. The data should be free and publicly available in tabular form and in several open formats whenever possible. Since most members of the public may not have the skills to analyze the data, TCOLE should provide high level visualizations that summarize the data in a meaningful way.