You are on page 1of 25

2020 Racial Profiling Data

Analysis for the State of Texas

PRELIMINARY REPORT
April 23, 2021

Institute for Predictive Analytics in Criminal Justice (IPAC)

Dr. Alex del Carmen, Tarleton State University


Dr. Fei Luo, Texas A&M International University
Dr. Wendi Pollock, Texas A&M University-Corpus Christi
Dr. Kimberly Chism, Prairie View A&M University
Dr. Camille Gibson, Prairie View A&M University
Dr. Deborah Sibila, Texas A&M University, Corpus Christi
Dr. Durant Frantzen, Texas A&M University-San Antonio
Dr. Dwight Steward, EmployStats and IPAC Fellow
Dr. Tom Petrowski, Tarleton State University
Dr. Brandi Copes, Justice Administration Department
Institute for Predictive Analytics in Criminal Justice (IPAC) Page 1
Institute for Predictive Analytics in Criminal Justice (IPAC) Page 2
TABLE OF CONTENTS

I. Background 6

II. Researchers 8

III. Research Goals 10

IV. Key Findings 11

V. Preliminary Findings 12

VI. Limitations and Recommendations 21

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 3


TABLE OF FIGURES

Figure 1 12
Gender of individuals stopped by Texas law enforcement officers in 2020

Figure 2 12
Race of individuals stopped by Texas law enforcement officers in 2020

Figure 3 13
Reasons for stops by Texas law enforcement officers in 2020

Figure 4 14
Location of stops by Texas law enforcement officers

Figure 5 15
Racial breakdown of individuals searched by Texas law enforcement officers

Figure 6 15
Reasons for the searches conducted by Texas law enforcement officers

Figure 7 16
Racial breakdown of those with contraband found during searches

Figure 8 16
Racial breakdown of individuals arrested due to contraband

Figure 9 17
Traffic stop outcomes

Figure 10 18
Reasons for the arrests made by Texas law enforcement officers

Figure 11 18
Racial breakdown of physical force involved

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 4


Institute for Predictive Analytics in Criminal Justice (IPAC) Page 5
I. Background
Arrest and Death of Sandra Bland
On July 10, 2015, Sandra Bland, a 28-year-old African American woman, was pulled over in Waller County by
Texas State Trooper Brian Encinia for a minor traffic violation. The stop, captured on dashcam and Bland’s cell
phone, escalated into a verbal, then physical confrontation. Bland was arrested for allegedly assaulting a police
officer. Her cell-phone record of the encounter, released years after the event, revealed that Encinia’s actions
during the stop significantly contributed to the events that led to Bland’s arrest. Following the arrest, Bland was
taken to the Waller County Jail and processed by county jail personnel. She remained incarcerated following her
arrest for several days pending the collection of bail money by family members and friends. She was found dead
in her jail cell July 12 just three days after her arrest.

Aftermath
Multiple agencies, including the Texas Rangers and the FBI, initiated investigations. Bland’s death was ruled a
suicide by the Harris County Institute of Forensic Sciences in Houston. The Texas Rangers, the lead investigative
agency, examined the circumstances surrounding Bland’s arrest and subsequent death to determine if any
criminal violations had occurred. The results of the investigation were turned over to Waller County prosecutors
and ultimately the grand jury. In January 2016 a grand jury indicted Encinia for perjury, a Class A misdemeanor.
Encinia was charged with falsifying parts of his report pertaining to the arrest. In early March 2016 the Texas
Department of Public Safety (TDPS) fired Encinia. In May 2017 the perjury charge against him was dismissed
as part of a plea agreement. Encinia agreed to surrender his license as a peace officer and never to pursue a
career in law enforcement. No other indictments were issued by the grand jury. The Waller County Sheriff
appointed an independent commission to investigate the circumstances surrounding Bland’s death. The
commission’s report, which was not a criminal investigation, was issued April 12, 2016, and called out systemic
failures with the Waller County Jail, including poor training and violation of policies mandating the monitoring
of prisoners. In September 2016 Sandra Bland’s family settled a wrongful death suit against TDPS, Encinia,
Waller County and two jailers for $1.9 million.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 6


Sandra Bland Act
In 2017 Texas legislators introduced a criminal justice reform bill to address issues raised by the Sandra Bland
incident. Senate Bill 1849, renamed the Sandra Bland Act, went into effect Sept. 1, 2017. The legislation requires a
change to corrections and police policy when dealing with individuals having substance abuse or mental health
concerns. The legislation “mandates county jails divert people with mental health and substance abuse issues
toward treatment, makes it easier for defendants to receive a personal bond if they have a mental illness or
intellectual disability, and requires that independent law enforcement agencies investigate jail deaths” (The
Texas Tribune, June 15, 2017).

A major goal of the Sandra Bland Act is to collect accurate data on traffic stops and to strengthen racial profiling
law. Texas law enforcement officers must complete comprehensive racial profiling training and 40 hours of
de-escalation training. The Texas Occupations Code (1701.164) and the Texas Code of Criminal Procedure also
require law enforcement agencies to compile motor vehicle stop reports and submit incident-based data to the
Texas Commission on Law Enforcement (TCOLE) or the governing body of each county or municipality served
by the agency by March 1 of each year (Article 2.134). In 2019, 2,540 law enforcement agencies submitted
reports; the number of agencies increased to 2,699 in 2020. In 2020, 1,195 agencies filed exempt status, and
1,504 filed full status. A total of 6,299,241 stops were reported in 2020, with the per-agency average being
2,334. The number of stops was 0 for agencies that filed exempt status, and TDPS reported the most stops
(1,667,553). The Houston Police Department (217,288) and San Antonio Police Department (138,180)
represent the next two largest number of stops in 2020.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 7


II. Researchers
A research team at Texas A&M University’s Institute for Predictive Analytics in Criminal Justice (IPAC)
analyzed the 2020 Sandra Bland racial profiling data obtained from TCOLE and reports the major findings
from the data.

Ten members of the team participated in this report:

Alex del Carmen, PhD, is a Professor, Associate Dean of the School of Criminology
and Director of IPAC at Tarleton State University. He received his PhD from Florida
State University’s College of Criminology and Criminal Justice. He was named a
Fulbright Specialist by the U.S. State Department in 2018 and has authored over 50
refereed articles and 10 books. His most recent book is Racial Profiling in Policing:
Beyond the Basics with Kendall Hunt. Dr. del Carmen has trained thousands of police
officers, including all of the Texas police chiefs, and worked on two of the largest police
reform cases in the United States, as a federal monitor. He currently is working as a
Special Master for the U.S. Courts in Puerto Rico on the Puerto Rico Police Reform Case.

Fei Luo, PhD, is an Assistant Professor of Criminal Justice at Texas A&M International
University. She earned her PhD from Sam Houston State University in 2016. Her
research interests include policing, race/ethnicity, gun policy, immigration and crime,
intimate partner violence, and the application of quantitative methodology in empirical
research. Her recent publications have appeared in American Journal of Criminal Justice;
Policing: An International Journal of Police Strategies and Management; and
International Journal of Offender Therapy and Comparative Criminology.

Wendi Pollock, PhD, received her BS and MS in criminal justice from Sul Ross State
University and her PhD in criminal justice at Sam Houston State University. She is an
Associate Professor of Criminal Justice at Texas A&M University-Corpus Christi. Her
research interests include the examination of issues involved with police/public interactions,
studies on the short- and long-term impacts of arrest, diversity issues in the American
criminal justice system, and, more generally, studies that involve using advanced statistical
analysis to better understand issues in criminal justice.

Kimberly Chism, PhD, is an Assistant Professor of Justice Studies in the College of


Juvenile Justice and Psychology at Prairie View A&M University. She earned her BS
and MS in criminology from California State University, Fresno and her PhD in criminal
justice from Sam Houston State University. She has conducted research in policing,
corrections, law, juvenile justice and criminological theory.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 8


Camille Gibson, PhD, CRC, is Interim Dean of the College of Juvenile Justice and
Psychology at Prairie View A&M University and Executive Director of the Texas Juvenile
Crime Prevention Center. She is a member of the Texas Racial Equity Collaborative.
Her research interests include youth and police interactions, police misconduct, and the
dynamics of homicide. She is a former elected board member of the Academy of Criminal
Justice Sciences and a Niederhoffer Fellowship recipient.

Deborah Sibila, PhD, is an Assistant Professor of Criminal Justice at Texas A&M


University-Corpus Christi. She earned her PhD in criminal justice from Sam Houston
State University in 2016. Her research interests include drug policy, issues related to
immigrant and gender offending/victimization, intimate relationship violence and
policing. She is a retired federal agent with 26 years of military and federal law
enforcement experience.

Durant Frantzen, PhD, is a Professor of Criminology and Criminal Justice and the
Department Chair of Social Sciences at Texas A&M University-San Antonio. He obtained
his PhD in criminal justice from Sam Houston State University in 2006. His research
focuses on offender rehabilitation and the evaluation of substance abuse treatment
programs. He has worked closely with the Bexar County Specialty Courts and the Bexar
County Jail-Reentry Programs Division in recent years as an evaluator and consultant.

Dwight Steward is an IPAC Fellow and an economist. He works with the IPAC director,
Dr. Alex del Carmen, on the statistical and content analysis of police procedures and
policies related to police use of force and racial profiling. Dr. Steward also works with
Dr. del Carmen and IPAC on developing donor and sponsorship relationships within
Tarleton State University, The Texas A&M University System, police agencies and
community stakeholders.

Tom Petrowski, JD, is a Visiting Assistant Professor in the School of Criminology at


Tarleton State University, an attorney and a law enforcement consultant. He retired from
the FBI after 23 years in the FBI legal program and diverse operational assignments in
the criminal investigative and national security programs. He currently is the Assistant
Special Master to the U.S. District Court in Puerto Rico supporting the U.S. DOJ Consent
Decree against the Puerto Rican Police Department.

Brandi Copes serves as the Racial Disparities and Fairness Administrator for the Harris
County Justice Administration Department. In this role, she works to identify and address
existing disparities in the Harris County criminal justice system across partner agencies.
In addition to working with government stakeholders, her work centers on community
engagement as a guiding principle. Brandi earned a BA in sociology and political science
from Loyola University New Orleans and a master’s degree in public administration from
the Cornell Institute of Public Affairs at Cornell University.
Institute for Predictive Analytics in Criminal Justice (IPAC) Page 9
III. Research Goals

The IPAC research team will examine the following for potential policy implications:

• Gender distribution of stops conducted by • The most common reasons for searches
Texas law enforcement officers conducted by Texas law enforcement
officers
• Racial breakdown of stops conducted by
Texas law enforcement officers • Racial distribution of stops in which
contraband was found as the result of
• Primary reasons for stops conducted by a search
Texas law enforcement officers and the
racial breakdown of stops by reason • Frequency of all outcomes of a stop,
including arrest and the use of force, and
• The most common locations for stops breakdown of those outcomes by race
conducted by Texas law enforcement
officers • Comparison between 2019 and 2020 data
and the potential impact of the pandemic on
• Percentage of stops involving a search and stops conducted by Texas law enforcement
the racial breakdown of stops involving a officers
search

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 10


IV. Key Findings

• Consistent with the crime pattern, males • Whites accounted for the largest percentage
were more likely to be stopped (about 66%) of cases (36.2%) in which contraband was
than females. Compared to 2019, the discovered during the searches.
percentage of females stopped increased
(from 31% to 34%). • Hispanics accounted for the largest percentage
(37.1%) of individuals who were arrested
• Whites were more likely to be stopped, due to contraband.
followed by Hispanics, Blacks, Asian
Americans and Native Americans. • Hispanics and Blacks were more likely to
be arrested than Whites if contraband was
• About 74.6% of the stops were for moving discovered.
traffic violations.
• No widespread use of physical force was
• More Blacks were stopped due to law detected. Only about 1% of the cases
violations, more Hispanics were stopped involved force. Among all valid cases
due to preexisting knowledge and vehicle reported, the largest percentage of cases
traffic violations, and more Whites were with physical force involved Whites (44.4%).
stopped due to moving traffic violations.
• Compared to 2019, significantly fewer stops
• City streets were the most frequent location were made in 2020 (10,679,600 vs
where stops were made (44.7%). 6,299,241). That’s a 41% decrease in stops
made. The pandemic in 2020 is likely a
• Hispanics accounted for the largest major reason for the reduced stops.
percentage of individuals who were
searched (38.8%).

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 11


V. Preliminary Findings
Gender distribution of the stops
• Among all 6,299,241 stops reported in the data, 34.0%
(2,142,176) involved a female and 65.9% a male
(4,152,446). About 1% of the stops have an unknown
or unreported gender.

• The average number of females stopped was about 794


with a minimum of 0 and a maximum of 471,725 by
TDPS. The mean number of males stopped was 1,538.
The minimum number was 0 and the maximum
1,195,828, also reported by TDPS.

• The gender distribution is similar to the 2019 data in


which more than twice the stops involved a male. In
2019, 69% of the stops involved a male and 31% a female. Figure 1. Gender of individuals stopped by
Texas law enforcement officers in 2020

Racial breakdown of the stops


• The stops based on all race categories add up to
7,198,971, which exceeded the reported total stops by
899,730. Analysis reveals that 120 of the agencies
reported incorrect or inconsistent information where the
numbers do not add up to the total reported. Among the
2,579 agencies that reported consistent information,
45.8% of the stops (2,732,077) involved a person identified
as White, 34.6% (2,061,559) involved a person identified
as Hispanic, 16.5% (985,311) involved a person identified
as Black, 2.9% (170,049) involved a person identified as
Asian American, and 0.3% (17,591) of the stops involved a
person identified as Native American.

• Compared to 2019, there was an increase in the number


of White (34% in 2019 vs 45.8% in 2020) and Black
(14% vs 16.5%) individuals stopped, and a slight
decrease in the number of Hispanics stopped (39% vs Figure 2. Race of individuals stopped by
Texas law enforcement officers in 2020
34.6%). Stops of people classified as Native American or
Asian American remained constant (less than a percentage
point of change between 2019 and 2020).
Institute for Predictive Analytics in Criminal Justice (IPAC) Page 12
Reasons for the stops
• Four major reasons were identified: law violation,
preexisting knowledge, moving traffic violation and
vehicle traffic violation. Moving traffic violations
accounted for the largest percentage — about 74.6%
of the stops (4,696,732). Vehicle traffic violations
accounted for 19.6% (1,236,236) of the stops. The
next most common reason for stops was a law
violation — 4.3% (273,094). About 1.3% (82,607)
of the cases involved some form of preexisting
knowledge. The reason was not reported/unknown
for 0.2% (10,572) of the cases.

• Compared to 2019, the biggest change was in stops


for an unknown reason and in stops for moving
traffic violations. In 2019, 11.7% of the stops were
for an unknown or unreported reason. This percentage
dropped to 0.2% in 2020. Stops for a moving Figure 3. Reasons for stops by Texas law
violation amounted to 57% of all stops in 2019; enforcement officers in 2020

this percentage increased to 74.6% in 2020.

Reasons by race
Law violation (273,094)
• The analysis reveals that 106,514 of the stops conducted due to violations of law involved a White
individual, and 1,975,568 involved a Black individual. About 93,905 of the law violations involved an
Hispanic, 4,182 involved an Asian American, and 720 involved a Native American. Noting that the
number of law violation stops involving Blacks exceeded the overall total law violations (273,094),
which indicates a potential error in the report regarding the Black race category.

Preexisting knowledge (82,607)


• Among all stops conducted due to preexisting knowledge, Hispanics accounted for the largest percentage
(39.2%, or 32,385 stops), followed by Whites (37.1%, or 30,659 stops). About 19.5% (16,120) of the stops
involved Blacks, 1.5% (1,258) involved Asian Americans, and 0.2% (158) involved a Native Amercan.
About 2.5% (2,027) of the cases had an unknown race for this reason.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 13


Reasons by race (continued)
Moving traffic violation (4,696,732)
• For stops made due to a moving traffic violation, 42.2% (1,981,518) involved White individuals, 34.3%
(1,612,026) involved Hispanics, 14.7% (688,112) involved Black individuals, 2.9% (134,233) involved
Asian Americans, and 0.3% (12,973) involved Native Americans. About 5.7% of the cases (267,870) had
an unknown race for this category. Markedly, given that Whites are 42% of the Texas population per 2018
census numbers (Texas Demographics Center, 2020), one could conclude that they do not experience
profiling while driving. Hispanics are underrepresented in these stops at 34.3% while they are 40%
of the Texas population. So, too, are Asian Americans underrepresented; they are stopped at 2.9% but
represent 5% of the state’s population. Black individuals, however, are overrepresented in stops while
driving, at 12% of the Texas population, but representing 14.7% of those stopped while driving.

Vehicle traffic violation (1,247,529)


• For stops reported due to vehicle traffic violations, 42.1% (524,782) involved Hispanics, 38.7% (482,643)
involved Whites, 17.0% (212,605) involved Black individuals, 2.0% (24,696) involved Asian Americans,
and 0.2% (2,803) involved Native Americans. The underrepresentation of Whites and slight overrepresentation
of Hispanics and Blacks given their presence in the Texas population raises the question of whether some
races and ethnicities are profiled in terms of cultural or socioeconomic leanings toward certain vehicles. In
other words, these stops are not necessarily “color-blind” (Delgado, 2018). Nevertheless, the research team
noticed that the number reported for all race categories (1,247,529) exceeded the total number of vehicle
traffic violations (1,236,236). This indicates a potential error in one or more of the race categories for vehicle
traffic violations.

Location of the stops


• Among all stops reported, the most frequent
location was a city street, which accounted
for 44.7% (2,813,458) of stops, followed by
U.S. highways at 27.4% (1,727,524), state
highways at 17.8% (1,120,642), county roads
at 5.9% (370,012) and private property at
3.8% (240,588). A small percentage of stops
were conducted at unknown locations (0.4%,
or 27,017).

Figure 4. Location of stops by Texas law enforcement officers

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 14


Stops that involved searches
• Among stops conducted, 5.8% (364,447) involved a
search.

• Of those stops that involved a search, 341,825 reported


the race of the individuals who were searched. This is
a vast improvement in reporting from 2019. In 2019
race was reported for 2.6% of all searches conducted.
In 2020, race was reported for 93.8% of all searches
conducted. Hispanics accounted for the largest
percentage of individuals who were searched (38.8%,
or 132,660), followed by Whites (36.2%, or 123,771),
Blacks (23.9%, or 81,777), Asian Americans ( 0.9%)
and Native Americans ( 0.1%).

Figure 5. Racial breakdown of individuals searches by


Texas law enforcement officers

Reasons for searches


• A total of 364,853 searches reported a reason. About
41.9% (152,950) of the searches were justified by
probable cause, 27.4% (99,909) were searched with
consent, 14.1% (51,499) were incident to arrest,
12.6% (46,024) were due to inventory, and about
4% (14,471) involved contraband. All searches with
a reported reason exceeded the total searches
reported by the agencies. It is possible that some
reasons presented were reported more than once
by some agencies or some cases involved more
than one reason.

Figure 6. Reasons for the searches conducted by


Texas law enforcement officers

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 15


Contraband discovered
• Among all searches conducted, 43.6% (159,039)
resulted in contraband being found. This percentage
is higher than in 2019, when about 34.3% of searches
resulted in contraband.

• In terms of racial distribution of individuals with


contraband, most searches involved Whites at 36.2%
(57,541), followed by Hispanics at 31.8% (50,552),
Blacks at 24.3% (38,610), Asian Americans at 0.8%
(1,340) and Native Americans at about 0.2% (251) of
searches. About 6.8% (10,745) did not report a race.

• Among all cases where contraband was discovered,


34% resulted in arrests (54,141). Of all individuals
arrested, Hispanics were the largest percentage Figure 7. Racial breakdown of those with
(37.1%), followed by Whites (36.0%) and Blacks contraband found during searches

(25.9%). Asian Americans and Native Americans


were 0.8% and 0.2%, respectively. Hispanics
accounted for 31.8% of the individuals, but they
accounted for 37.1% of the arrests.

• Regarding the percentage of arrests made within


each racial category due to contraband, 40.5% of
Hispanics with contraband were arrested, 37.8% of
Blacks with contraband were arrested, and 32.7% of
Whites with contraband were arrested. The percentage
of Asian Americans and Native Americans was 30.2%
and 16.2%, respectively. The results indicate that
Hispanics and Blacks were more likely than Whites
to be arrested due to contraband.

Figure 8. Racial breakdown of individuals


arrested due to contraband

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 16


Contraband hit rates
• Contraband hit rates were first suggested by Nobel Prize-winning economist Gary Becker (1957),
as a method to test bias in search decisions by police. Becker proposed that if police do not
discriminate, then contraband hit rates should be consistent. If, however, a group of people have
a lower hit rate, it would suggest that group is being searched with less evidence, and presumably
greater bias. For the present analysis, contraband hit rates were calculated per 1,000 searches
conducted.

Those rates:
◊ Black hit rate = 472 contraband hits per 1,000 searches (47.2%)
◊ White hit rate = 465 contraband hits per 1,000 searches (46.5%)
◊ Asian American hit rate = 431 contraband hits per 1,000 searches (43.1%)
◊ Hispanic hit rate = 381 contraband hits per 1,000 searches (38.1%)
◊ Native American hit rate = 494 contraband hits per 1,000 searches (49.4%)

• A word of caution regarding these numbers. Information on race was missing for slightly over 6%
of all searches conducted. Race data was again missing for 6.8% of cases in which contraband was
found. As a result, some error is possible in the calculation of hit rates, especially when examining
Asian American or Native American data, as the number of searches and cases where contraband
was found was very low for these groups. Accordingly, for these individuals, hit rates were especially
vulnerable to error from missing data. Further, the literature indicates that hit rate information can be
most telling when examined at the county level to see where geographically disparate profiling needs
to be addressed (The Economist, 2017).

Stop outcomes
• A total of 6,293,885 stop outcomes were reported,
with citations at 41.3% (2,599,645) and written
warnings 40.4% (2,540,958) the most common.
Verbal warnings accounted for 15.5% (978,301)
of the outcomes, and arrests accounted for 1.8%
(114,086). About 5% (30,140) of the stops resulted
in a written warning and arrest. A similar
percentage was found for citation and arrest
(about 5%, or 30,755).

Figure 9. Traffic stop outcomes

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 17


Reason for arrests
• Among all valid reasons reported for the arrests (176,237),
more than half were based on penal code violations (50.5%,
or 89,016), and 24.9% (43,802) were based on traffic law
violations. Outstanding warrants accounted for 23.4% of
the arrest basis, and violations of city ordinance represented
1.2% of the arrest reason.

Figure 10. Reason for the arrests made by


Texas law enforcement officers

Use of physical force


• Only about 1% (6,757) of the cases involved physical force.
Among all valid cases for racial distribution (6,557), 44.4%
(2,911) of the cases with physical force involved Whites,
27.2% (1,782) involved Hispanics, 26.3% (1,722) involved
Blacks, 1.7% (110) involved Asian Americans, and 0.5%
(32) involved Native Americans.

Figure 11. Racial breakdown of physical


force involved

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 18


Complaints
In 2020 a total of 4,047 complaints of racial profiling were reported to Texas law enforcement agencies, and only
0.2% (10) of the cases resulted in disciplinary action. Compared to 2019, the number of complaints significantly
increased from 522 to 4,047. The percentage of cases that resulted in disciplinary action also increased, from 0%
to 0.2%.

Analysis
This report includes data from law enforcement agencies on demographic and outcome related information for
motor vehicle stops in Texas for 2020. The data highlights some important trends in comparison to the state’s
overall demographics. First, Black motorists were more likely to be stopped by law enforcement compared
to White (non-Hispanic) and Hispanic motorists. Specifically, 16.9% of stops involved motorists identified as
Black, but according to the latest Census Bureau statistics (2019) only 12.9% of the Texas population is Black.
Similarly, White (non-Hispanic) drivers were stopped at a rate exceeding their representation in the population.
Census figures reflect that 41.2% of the state is White; however, 45.8% of drivers stopped were White. Conversely,
Hispanic motorists were less likely to be stopped by law enforcement. Census data shows that 39.7% of the
state is Hispanic, yet only 34.6% of stops involved Hispanic motorists. Those identifying as Asian American
and American Indian were also less likely to be stopped relative to their representation in the population. While
Asian Americans make up 5.2% of the population, only 2.3% of stops involved Asian American drivers. The
difference was similar to that involving American Indians (population=1%, stops=.3%).

Another notable finding relates to search hit rates. This measure has been commonly used to assess bias in
police searches. Moreover, if officers make search decisions based on relevant legal criteria and these decisions
are being applied equitably across racial and ethnic groups, the hit rates should be similar. On the other hand, if
one group is searched more frequently but has the same number of hits, the hit rate will be lower compared to
the other groups. Therefore, groups with lower hit rates suggest that bias may have impacted the search decision
and that police are unfairly targeting those groups.

Our study revealed that the hit rate for Hispanic motorists was the lowest across all groups. When viewed together
with the fact that Hispanics were stopped less frequently than Blacks or Whites, these findings suggest a greater
element of implicit bias present in police interactions with Hispanic motorists. Given that probable cause
requirements to search are more stringent than less intrusive interactions such as pat-downs or frisks, it is
reasonable to assume that implicit bias would be less of a factor in decisions to search rather than to frisk.
Hence, hit rates that reflect searches only rather than frisks may be a function of the fact that some searches
are required by law (e.g., active warrant, inventory search or preexisting knowledge justifies the search).
Alternatively, it could also mean that more explicit bias exists in probable cause searches. The racial and ethnic
distribution of contraband found was greater for Whites (36.2%) and Hispanics (31.8%) compared to Blacks
(24.3%). These numbers suggest that while Hispanics are being searched more frequently, there is a lower
likelihood that police find contraband.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 19


Further analysis of traffic stops reinforces results relating to stops and hit rates. The decision to arrest was
examined in cases where officers discovered contraband. Here again, these numbers should be comparable
across groups as the officer has discretion to arrest or to let the motorist go with a warning. The findings
showed that Hispanics (40.5%) were arrested more frequently than the other groups; Blacks (37.8%) were
next, then Whites (32.7%).

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 20


VI. Limitations and Recommendations
Compared to the 2019 data, the quality of the 2020 data significantly improved. More agencies participated in the
report, and fewer cases were missing. However, the research team still detected inaccuracies and inconsistencies in
the data reported. The findings should be interpreted with caution. Compared to other racial groups, more Blacks
were stopped due to law violations, and more Hispanics were stopped due to preexisting knowledge and vehicle
traffic violations. But the communities of these stops are not evident. This makes it possible that because of
racial and ethnic profiles, Texas law enforcement officers are more likely to find what they seek among particular
races and ethnicities in specific communities. While discretion and profiling are inherent in policing, bias in
their application, can mean disparate outcomes. Similarly, that Hispanics account for the largest percentage of
individuals searched and that they have the lowest hit rate from searches is in keeping with the theoretical notion
of the behaviors of instruments of law, in that those perceived to be most marginalized — often immigrants and
foreign language speakers — are vulnerable to experiencing the heavier hand of the law.

For the most accurate understanding of racial profiling in Texas, we recommend the Legislature amend the law
regarding report submission. A standardized reporting process should be established to collect valid and accurate
data. Our preliminary analysis revealed that about 120 of the agencies reported inconsistent data on race. Many
other inconsistencies also were detected. Although more and more agencies are reporting data on racial profiling,
no standard measure audits the accuracy. The conclusions drawn from this report cannot be interpreted with full
confidence until the data is guaranteed to be accurate.

It also would be good practices to require law enforcement agencies to report individual-level data instead of
aggregated departmental data. Individual-level data allows for a more detailed analysis of racial profiling patterns.
It also can ensure accuracy of the total cases reported.

In addition, TCOLE data should be analyzed annually to provide implications for law enforcement agencies and to
improve the data collection process. TCOLE should partner with research centers such as IPAC to analyze the data
and provide useful implications for legislative and law enforcement agencies.

Finally, too many inconsistencies exist in the information on race. Due to the aggregated nature of the data, the
researchers cannot find the reason for these inconsistencies, but possibly individuals, especially those of Hispanic
origin, reported more than one race. For example, an individual may be both Hispanic and White. The U.S. census
has separated the race and ethnicity questions because a person of Hispanic origin may be of any race. We
recommend separating the race and ethnicity questions in order to generate more accurate data. It would also
help to report officers’ race and ethnicity so we can examine how this influences their stop decisions.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 21


Other recommendations
• Include a category for frisks in addition to searches. Ideally the categories would be frisk only, frisk and
search, or search only. Searches are sometimes required by law (as the result of a vehicle inventory, for
example). If implicit stereotypes of race and ethnicity influence officers’ post-stop assessments, they should
have a stronger effect on less intrusive actions such as frisks.

• Include a category for probable cause (non-warrant) searches. Some searches are incident to arrest because
of an active warrant on the individual. Other searches incident to arrest are the result of probable cause
observed by the officer. This would enable more focus on the purely discretionary searches.

• Include the racial composition of the police department’s jurisdiction with the data. This would enable an
assessment of whether stop and other outcomes vary based on the racial composition of the community.

• Include a variable that differentiates the city’s population. This would generate a more place-based analysis.
Research has shown that racial profiling is more prevalent in more homogenous communities, which in Texas
tend to be rural.

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 22


References

Becker, G. (1957).
The Economics of Discrimination. Chicago: University of Chicago Press.
https://press.uchicago.edu/ucp/books/book/chicago/E/bo22415931.html

Delgado, D. (2018).
“My deputies arrest anyone who breaks the law”: Understanding How Color-blind Discourse and
Reasonable Suspicion Facilitate Racist Policing.
Journal of the Sociology of Race and Ethnicity 4(4), 541-554.
https://doi.org/10.1177/2332649218756135

Findings: The results of our nationwide analysis of traffic stops and searches (2021).
Retrieved March 29, 2021, from: https://openpolicing.stanford.edu/findings/

Stanford Open Sentencing Project (2012).


Police Data Suggests Black and Hispanic Drivers Are Searched More Often than Whites.
https://slate.com/technology/2017/06/ statistical-analysis-of-data-from-20-states-suggests-evidence-of-racially-biased-policing.html

Texas Demographics Center (2020).


Texas Demographic Trends & the Upcoming 2020 Census.
https://demographics.texas.gov/Resources/Presentations/ OSD/2020/ 2020_03_04_MetropolitanBreakfastClub.pdf

Institute for Predictive Analytics in Criminal Justice (IPAC) Page 23


Institute for Predictive Analytics in Criminal Justice (IPAC) Page 24
DR. ALEX DEL CARMEN
Director,
Institute for Predictive Analytics in Criminal Justice (IPAC)

254-968-9106
DELCARMEN@tarleton.edu
Institute for Predictive Analytics in Criminal Justice (IPAC) Page 25

You might also like