Racial Profiling: Limited Data Available on Motorist Stops (Letter
Report, 03/13/2000, GAO/GGD-00-41).
Pursuant to a congressional request, GAO provided information on the
racial profiling of motorists, focusing on the: (1) findings and
methodologies of analyses that have been conducted on racial profiling
of motorists; and (2) federal, state, and local data available, or
expected to be available, on motorist stops.
GAO noted that: (1) GAO found no comprehensive, nationwide source of
information that could be used to determine whether race has been a key
factor in motorist stops; (2) the available research is limited to five
quantitative analyses that contain methodological limitations; (3) they
have not provided conclusive empirical data from a social science
standpoint to determine the extent to which racial profiling may occur;
(4) however, the cumulative results of the analyses indicate that in
relation to the populations to which they were compared, African
American motorists in particular, and minority motorists in general,
were proportionately more likely than whites to be stopped on the
roadways studied; (5) data on the relative proportion of minorities
stopped on a roadway, however, is only part of the information needed
from a social science perspective to assess the degree to which racial
profiling may occur; (6) a key limitation of the available analyses is
that they did not fully examine whether different groups may have been
at different levels of risk for being stopped because they differed in
their rates or severity of committing traffic violations; (7) although
GAO has no reason to expect that this occurred, such data would help
determine whether minority motorists are stopped at the same level that
they commit traffic law violations that are likely to prompt stops; (8)
several analyses compared the racial composition of stopped motorists
against that of a different population, but the validity of these
comparison groups was questionable; (9) federal, state, and local
agencies are in various stages of gathering data on motorist stops, and
these efforts should augment the empirical data available from racial
profiling studies; (10) the federal government, which has a limited role
in making motorist stops, is undertaking several efforts to collect
data; (11) in accordance with a presidential directive, three federal
departments are preparing to collect data on the race, ethnicity, and
gender of individuals whom they stop or search; (12) state and local
agencies are in the best position to provide law enforcement data on
motorist stops because most motorist stops are made by state and local
law enforcement officers; (13) a number of state legislatures are
considering bills to require state or local police to collect race and
other data on motorist stops; (14) several local jurisdictions are also
making efforts to collect motorist stop data; and (15) whether the
efforts that are underway will produce the type and quality of
information needed to answer the questions about racial profiling
remains to be seen.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: GGD-00-41
TITLE: Racial Profiling: Limited Data Available on Motorist Stops
DATE: 03/13/2000
SUBJECT: Data collection
Comparative analysis
Blacks
Minorities
Transportation statistics
Racial discrimination
Traffic violations
Statistical methods
Law enforcement information systems
Federal/state relations
IDENTIFIER: DEA Operation Pipeline
DOJ Police Public Contact Survey
DOJ Law Enforcement Management and Administrative
Statistics Survey
DOJ Survey of State Police Agencies
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO report. Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
** A printed copy of this report may be obtained from the GAO **
** Document Distribution Center. For further details, please **
** send an e-mail message to: **
** **
** **
** **
** with the message 'info' in the body. **
******************************************************************
United States General Accounting Office
GAO
Draft Report to the Honorable
James E. Clyburn, Chairman
Congressional Black Caucus
March 2000
GAO/GGD-00-41
RACIAL PROFILING
Limited Data Available on Motorist Stops
Ordering Copies of GAO Reports
The first copy of each GAO report and testimony
is free. Additional copies are $2 each. Orders
should be sent to the following address,
accompanied by a check or money order made out
to the Superintendent of Documents, when
necessary. VISA and MasterCard credit cards are
accepted, also. Orders for 100 or more copies to
be mailed to a single address are discounted 25
percent.
Order by mail:
U.S. General Accounting Office
P.O. Box 37050
Washington, DC 20013
or visit:
Room 1100
700 4th St. NW (corner of 4th and G Sts. NW)
U.S. General Accounting Office
Washington, DC
Orders may also be placed by calling (202) 512-
6000 or by using fax number (202) 512-6061, or
TDD (202) 512-2537.
Each day, GAO issues a list of newly available
reports and testimony. To receive facsimile
copies of the daily list or any list from the
past 30 days, please call (202) 512-6000 using a
touch-tone phone. A recorded menu will provide
information on how to obtain these lists.
Viewing GAO Reports on the Internet
For information on how to access GAO reports on
the INTERNET, send e-mail message with "info" in
the body to:
[email protected]
or visit GAO's World Wide Web Home Page at:
http://www.gao.gov
Reporting Fraud, Waste, and Abuse in Federal
Programs
To contact GAO FraudNET use:
Web site:
http://www.gao.gov/fraudnet/fraudnet.htm
E-Mail: [email protected]
Telephone: 1-800-424-5454 (automated answering
system)
United States
General Accounting Office
Washington, D.C. 20548-0001
Official Business
Penalty for Private Use $300
Address Correction Requested
Bulk Rate
Postage & Fees Paid
GAO
Permit No. G100
(182081)
Contents
Page 221 GAO/GGD-00-41 Racial Profiling
Letter 1
Appendix I 24
Studies of Racial
Characteristics of
Drivers Stopped by
Police
A Summary of Analysis Design, Results, 24
and Limitations
Appendix II 34
Methodological Issues In
Studying Racial
Profiling of Motorists
Appendix III 37
Bureau of Justice
Statistics Police Public
Contact Survey
Appendix IV 44
Federal Law Enforcement
Data Collection
Presidential Directive 44
Appendix V 48
State Legislation and
Proposed State
Legislation to Collect
Traffic Stop Data:
Elements to be Collected
Appendix VI 50
Four Localities Data
Collection Plans
San Diego Police Department 50
San Jose Police Department 50
Alameda Police Department 52
Piedmont Police Department 52
Appendix VII 54
Comments From U.S.
Department of Justice
Appendix VIII 58
GAO Contacts and Staff
Acknowledgments
Tables Table 1: Status of Traffic Stop Bills 15
Introduced in State Legislatures
Table 2: Traffic Stop Data Elements 16
Collected by Four California Police
Departments
Abbreviations
ACLU American Civil Liberties Union
BJS Bureau of Justice Statistics
DEA Drug Enforcement Administration
III Interstate Identification Index
INS Immigration and Naturalization Service
LEMAS Law Enforcement Management and
Administrative Statistics
MSP Maryland State Police
NJSP New Jersey State Police
B-283949
Page 21 GAO/GGD-00-41 Racial Profiling
B-283949
March 13, 2000
The Honorable James E. Clyburn
Chairman, Congressional Black Caucus
Dear Mr. Chairman:
Racial profiling of motorists by law
enforcement-that is, using race as a key factor in
deciding whether to make a traffic stop-is an
issue that has received increased attention in
recent years. Numerous allegations of racial
profiling of motorists have been made and several
lawsuits have been won.
As agreed with your office, this report
provides information on (1) the findings and
methodologies of analyses that have been conducted
on racial profiling of motorists; and (2) federal,
state, and local data currently available, or
expected to be available soon, on motorist stops.
Results in Brief
We found no comprehensive, nationwide source
of information that could be used to determine
whether race has been a key factor in motorist
stops. The available research is currently limited
to five quantitative analyses that contain
methodological limitations; they have not provided
conclusive empirical data from a social science
standpoint to determine the extent to which racial
profiling may occur. However, the cumulative
results of the analyses indicate that in relation
to the populations to which they were compared,
African American motorists in particular, and
minority motorists in general, were
proportionately more likely than whites to be
stopped on the roadways studied. Data on the
relative proportion of minorities stopped on a
roadway, however, is only part of the information
needed from a social science perspective to assess
the degree to which racial profiling may occur.
A key limitation of the available analyses is
that they did not fully examine whether different
groups may have been at different levels of risk
for being stopped because they differed in their
rates and/or severity of committing traffic
violations.1 Although we have no reason to expect
that this occurred, such data would help determine
whether minority motorists are stopped at the same
level that they commit traffic law violations that
are likely to prompt stops. The best studies that
we identified sought to determine the racial
composition of motorists at risk of being stopped
and collected data on the population of travelers
and traffic violators on specific roadways.
However, even these well-designed studies made no
distinction between the seriousness of different
traffic violations, and it is not clear that all
violations are equally likely to prompt a stop.
There appears to be little comparative research on
traffic violations committed by different racial
groups, including possible differences in the type
or seriousness of traffic violations. In addition,
none of the studies provided information on which
traffic violations, if any, were more likely to
prompt a stop. More information is needed to
determine the extent to which race, as opposed to
other factors, may be the reason for the traffic
stop.
Several analyses compared the racial
composition of stopped motorists against that of a
different population, but the validity of these
comparison groups was questionable. In addition,
missing data may have skewed the results of some
analyses. Finally, because only a few locations
have been studied, and these locations were not
selected to be generally representative of
motorist roadways, the results cannot be
generalized to roadways and locations other than
those reviewed. These limitations notwithstanding,
we believe that in order to account for the
disproportion in the reported levels at which
minorities and whites are stopped on the roadways,
(1) police officers would have to be substantially
more likely to record the race of a driver during
motorist stops if the driver was a minority than
if the driver was white, and (2) the rate and/or
severity of traffic violations committed by
minorities would have to be substantially greater
than those committed by whites. We have no reason
to expect that either of these circumstances is
the case.
Federal, state, and local agencies are in
various stages of gathering data on motorist
stops, and these efforts should augment the
empirical data available from racial profiling
studies. The federal government, which has a
limited role in making motorist stops, is
undertaking several efforts to collect data. For
example, the Justice Department's Bureau of
Justice Statistics (BJS) is conducting a national
household survey that should provide aggregate
data on the characteristics of stopped motorists
and the nature of the traffic stops. BJS is also
conducting other surveys that should identify the
motorist stop information maintained by state and
local law enforcement agencies. In accordance with
a presidential directive, three federal
departments are preparing to collect data on the
race, ethnicity, and gender of individuals whom
they stop or search. State and local agencies are
in the best position to provide law enforcement
data on motorist stops because most motorist stops
are made by state and local law enforcement
officers. A number of state legislatures are
considering bills to require state and/or local
police to collect race and other data on motorist
stops, and Connecticut and North Carolina have
passed such legislation. Several local
jurisdictions are also making efforts to collect
motorist stop data.
Given the paucity of available data for assessing
whether and to what extent racial profiling may
exist, current efforts to collect information on
who is stopped and why are steps in the right
direction. Getting more and better data involves a
variety of methodological considerations and
information needs. Whether the efforts that are
currently under way will produce the type and
quality of information needed to answer questions
about racial profiling remains to be seen.
Background
The Fifth and Fourteenth Amendments prohibit
law enforcement officers from engaging in
discriminatory behavior on the basis of
individuals' race, ethnicity, or national origin.
The Fifth Amendment protects against
discrimination by federal law enforcement
officers, and the Equal Protection Clause of the
Fourteenth Amendment protects against
discrimination by state and local law enforcement
officers. Two federal statutes also prohibit
discrimination by law enforcement agencies that
receive federal financial assistance. Title VI of
the Civil Rights Act of 19642 prohibits
discrimination on the basis of race, color, or
national origin by all recipients of federal
financial assistance. The Omnibus Crime Control
and Safe Streets Act of 19683 prohibits
discrimination on the basis of race, color,
national origin, sex, or religion by law
enforcement agencies that receive federal funds
pursuant to that statute. In addition, a 1994
statute grants the Attorney General the authority
to seek injunctive relief when a state or local
law enforcement agency engages in a pattern or
practice of conduct that violates the Constitution
or federal law, regardless of whether the agency
is a recipient of financial assistance.4
The Fourth Amendment guarantees the rights of
people to be secure from unreasonable searches and
seizures. The temporary detention of individuals
during the stop of an automobile by police
constitutes a seizure of persons within the
meaning of the Fourth Amendment. The Supreme Court
recently held that regardless of an officer's
actual motivation, a stop of an automobile is
reasonable and permitted by the Fourth Amendment
when the officer has probable cause to believe
that a traffic violation occurred.5 The Court
noted, however, that the Constitution prohibits
selective enforcement of the law based on
considerations such as race, but the
constitutional basis for objecting to
intentionally discriminatory application of laws
is the equal protection provisions of the
Constitution, not the Fourth Amendment.
Some have expressed concern that the
escalation of this country's war on drugs has
placed minorities at increased risk of
discriminatory treatment by law enforcement. The
allegation is that law enforcement officers stop
minority motorists for minor traffic violations
when, in reality, the stop is a pretext to search
for drugs or other contraband in the vehicle.
In 1986, the Drug Enforcement Administration
(DEA) established Operation Pipeline, a highway
drug interdiction program that trains federal,
state, and local law enforcement personnel on
indicators that officers should look for that
would suggest possible drug trafficking activity
among motorists. In a 1999 report,6 the American
Civil Liberties Union (ACLU) stated that Operation
Pipeline fostered the use of a racially biased
drug courier profile, in part by using training
materials that implicitly encouraged the targeting
of minority motorists. DEA's position is that it
did not and does not teach or advocate using race
as a factor in traffic stops. Further, according
to DEA officials, a 1997 review of Operation
Pipeline by the Justice Department's Civil Rights
Division, which is responsible for the enforcement
of statutory provisions against discrimination,
concluded that Operation Pipeline did not instruct
trainees to use race as a factor in traffic stops.
Representatives of organizations representing
law enforcement officers have stated that racial
profiling is unacceptable. The National
Association of Police Organizations, representing
more than 220,000 officers nationwide, has
expressed opposition to pulling over an
automobile, searching personal property, or
detaining an individual solely on the basis of the
individual's race, ethnicity, gender, or age. The
International Association of Chiefs of Police, one
of the largest organizations representing police
executives, stated that stopping and searching an
individual simply because of race, gender, or
economic level is unlawful and unconstitutional
and should not be tolerated in any police
organization. Neither group supports federally
mandated collection of data on motorist stops.
Lawsuits alleging racial profiling have been
filed in a number of states, including Oklahoma,
New Jersey, Maryland, Illinois, Florida,
Pennsylvania, and Colorado. For example, in
Colorado, a class action suit filed on behalf of
400 individuals asked the court to halt racially
based stops by a Sheriff's Department highway drug
interdiction unit. Traffic infractions were cited
as the reason for stopping the motorists, but
tickets were not issued. The court ruled that
investigatory stops based solely on motorists'
match with specified drug courier indicators
violated the Fourth Amendment's prohibitions
against unreasonable seizures.7 A settlement was
reached that awarded damages to the plaintiffs and
disbanded the drug unit. In another case, a class
action lawsuit filed by ACLU against the Maryland
State Police resulted in a settlement that
included a requirement that the state maintain
computer records of motorist searches. These
records are intended to enable the state to
monitor for any patterns of discrimination.8 In
yet another case, a Superior Court in New Jersey
ruled that the New Jersey State Police engaged in
discriminatory enforcement of the traffic laws.9
The Justice Department's Civil Rights
Division has recently completed investigations in
New Jersey and Montgomery County, MD, which
included reviewing complaints of discriminatory
treatment of motorists. In the New Jersey case,
Justice filed suit in U.S. District Court alleging
that a pattern or practice of discriminatory law
enforcement had occurred. The parties filed a
joint application for entry of a consent decree,
which the judge approved in December 1999. Under
the consent decree, state troopers in New Jersey
will be required to collect data on motorist stops
and searches, including the race, ethnicity, and
gender of motor vehicle drivers. In the Maryland
case, the Justice Department and Montgomery County
signed a Memorandum of Understanding in January
2000 that resolved the issues raised in Justice's
investigation. The agreement included the
requirement that the Montgomery County Police
Department document all traffic stops, including
information on the race, ethnicity, and gender of
drivers.
Lack of empirical information on the existence and
prevalence of racial profiling has led to calls
for local law enforcement to collect data on which
motorists are stopped, and why. To support local
data collection efforts, the Bureau of Justice
Assistance plans to release a Resource Guide in
spring of 2000. The guide is expected to focus on
how data can be collected to monitor for bias in
traffic stops, with specific "lessons learned" and
implementation guidance from communities that have
begun the data collection process.
Scope and Methodology
Our objectives were to provide information on
(1) analyses that have been conducted on racial
profiling of motorists by law enforcement; and (2)
federal, state, and local data currently
available, or expected to be available soon, on
motorist stops.
To obtain information on analyses that have
been conducted on racial profiling of motorists,
we did a search of on-line databases and reviewed
all of the quantitative analyses that we
identified that attempted to address whether law
enforcement officers stop motorists on the basis
of race. We also contacted the authors of the
analyses and obtained references to any other
analysis or research sources they considered to be
pertinent. Our criterion for selecting analyses to
be included in this report was that they provide
quantitative information on motorist stops,
although these analyses might have also measured
searches, arrests, and/or other activities. We
used social science research principles to assess
the methodological adequacy of the available
analyses and to discuss factors that should be
considered in collecting stronger empirical data.
Our review is not intended to constitute a
statement regarding the legal standard for proving
discrimination in this context.
To obtain information on the federal
government's efforts to collect data on racial
profiling of motorists, we reviewed published and
electronic literature and discussed data sources
with officials at the Justice Department's Bureau
of Justice Statistics (BJS), officials in the
office of the Attorney General, academic experts,
the American Civil Liberties Union (ACLU), and
several police associations.
To obtain information on states' efforts to
collect data on racial profiling of motorists, we
conducted Internet searches and reviewed the
literature. We also held discussions with academic
experts, state officials, ACLU officials, and
representatives of the National Conference of
State Legislatures.
To obtain information on local efforts to
collect data on racial profiling of motorists, we
reviewed the literature and held discussions with
academic experts, interest groups, local police
officials, and knowledgeable federal officials. On
the basis of these discussions, we judgmentally
selected several communities that had voluntarily
decided to require their police departments to
collect motorist stop data. In September 1999, we
visited four police departments in California-in
San Diego, San Jose, Alameda, and Piedmont. We
selected these police departments because they
appeared to be furthest along in their plans for
collecting data, could provide examples of
different data collection methods, and varied
greatly in size.
We performed this work from August through
February 2000 in accordance with generally
accepted government auditing standards.
Few Studies of Racial Profiling
We found no comprehensive, nationwide source
of information on motorist stops to support an
analysis of whether race has been a key factor in
law enforcement agencies' traffic stop practices.
We identified five quantitative analyses on racial
profiling that included data on motorist stops.
The quantity and quality of information that these
analyses provided varied, and the findings are
inconclusive for determining whether racial
profiling occurred. Although inconclusive, the
cumulative results of the analyses indicate that
in relation to the populations to which they were
compared, African Americans in particular, and
minorities in general, may have been more likely
to be stopped on the roadways studied.
A key limitation of the available analyses is
that they did not fully examine whether the rates
and/or severity of traffic violations committed by
different groups may have put them at different
levels of risk for being stopped. Such data would
help determine whether minority motorists are
stopped at the same level that they commit traffic
law violations that are likely to prompt stops.
Most analyses either compared the proportion of
minorities among stopped motorists to their
proportion in a different population (e.g., the
U.S. population, the driving age population of a
state) or did not use a benchmark comparison at
all. There appears to be little comparative
research on traffic violations committed by
different racial groups, including possible
differences in the type or seriousness of traffic
violations. Therefore, there are no firm data
indicating either that the types and seriousness
of driving violations committed by whites and
minorities are comparable, nor that they are not.10
Although we have no reason to expect that such
differences exist, collecting research data on
this issue-though difficult to do-could help
eliminate this as a possible explanation for
racial disparities in the stopping of motorists.
The studies with the best research design
collected data on the population of travelers on
sections of interstate highways and on the portion
of those travelers who violated at least one
traffic law. The studies compared the racial
composition of these groups against that of
motorists who were stopped. However, the studies
made no distinction between the seriousness of
different traffic violations. Although violating
any traffic law makes a driver eligible to be
stopped, it is not clear that all violations are
equally likely to prompt a stop.
None of the available research provided
information on which traffic violations, if any,
were more likely to prompt a stop. We recognize
that it is difficult to determine which traffic
violations specifically prompt a law enforcement
officer to stop one motorist rather than another.
Different jurisdictions and officers may use
different criteria, and candid information on the
criteria may be difficult to obtain. Pursuing such
information would be worthwhile, however, as would
analyses that considered the seriousness of the
traffic violation. Below, we summarize the
reported results and our judgment of the key
limitations of each analysis. More detail on each
analysis is provided in appendix I.
� An analysis by Lamberth of motorists
traveling along a segment of the New Jersey
Turnpike11 found the following: (1) 14 percent of
the cars traveling the roadway had an African
American driver or other occupant; (2) 15 percent
of cars exceeding the speed limit by at least 6
miles per hour had an African American driver or
other occupant; (3) of stops where race was noted
by police, 44 percent of the individuals in one
section of the roadway and 35 percent of the
individuals in this section and a larger section
combined were African American.12 Lamberth also
reported that 98 percent of all drivers violated
the speed limit by at least 6 miles per hour. This
study is notable in that it attempted to determine
the percentage and characteristics of drivers who
put themselves at risk for being stopped. However,
we are uncertain whether traveling over the speed
limit by at least 6 miles per hour on a major
highway is the violation for which most police
stops occurred.
� In a similar analysis of motorists traveling
along a segment of Interstate 95 in northeastern
Maryland,13 Lamberth found the following: (1) 17
percent of the cars had an African American
driver; (2) 18 percent of cars exceeding the speed
limit by at least 1 mile per hour or violating
another traffic law14 had an African American
driver; (3) 29 percent of the motorists stopped by
the Maryland State Police were African American.
This study also found that 92 percent of all
motorists were violating the speeding law, 2
percent were violating another traffic law, and 7
percent were not violating any traffic law.15
However, we are uncertain whether Lamberth's
criteria for traffic violations were the basis for
which most police stops were made.
� Another analysis examined motorist stops in
Florida. Using data that were first presented in
1992 in two Florida newspaper articles, Harris16
reported that more than 70 percent of almost 1,100
motorists stopped over a 3-year period in the late
1980s along a segment of Interstate 95 in Volusia
County, FL, were African American or Hispanic. In
comparison, African Americans made up 12 percent
of Florida's driving age population and 15 percent
of Florida drivers convicted of traffic offenses
in 1991. Harris also reported that African
Americans and Hispanics made up 12 percent and 9
percent, respectively, of the U.S. population.
The findings reported by Harris were based on
videotapes of almost 1,100 motorist stops made by
Volusia County Sheriff deputies. However,
videotapes of stops were not made for much of the
3-year period, and sometimes deputies taped over
previous stops. Because no information was
provided on other motorist stops made by the
deputies over the 3-year period, we do not know
whether the videotaped stops were representative
of all stops made during that period. In addition,
no information was provided on drivers who put
themselves at risk for being stopped.
� The Philadelphia ACLU reported that motorists
stopped by Philadelphia police in selected
districts during 2 weeks in 1997 were more likely
to be minority group members than would be
expected from their representation in census data.17
Limitations of this analysis included the use of
census data as a basis for comparison and an
absence of information on drivers who put
themselves at risk for being stopped. In addition,
there were substantial amounts of missing data.
The race of the driver was not recorded for about
half of the approximately 1,500 police stops made
during the 2 weeks.
� The New Jersey Attorney General's Office
reported that African Americans and Hispanics,
respectively, represented 27 percent and 7 percent
of the motorists stopped by New Jersey State
Police on the New Jersey Turnpike.18 Interpreting
these results is difficult because no benchmark
was provided for comparison purposes.
Because of the limited number of analyses and
their methodological limitations, we believe the
available data do not enable firm conclusions to
be made from a social science perspective about
racial profiling. For example, we question the
validity of comparing the racial composition of a
group of stopped motorists on a given roadway in a
given location with the racial composition of a
population that may be vastly different. It would
be more valid to compare the racial
characteristics of stopped motorists with those of
the traveling population who violated similar
traffic laws but were not stopped. This is what
Lamberth did, although we are not certain that the
traffic violations committed by the motorists
observed in his studies were the same as those
that prompted police stops. Nonetheless,
Lamberth's analyses went furthest by attempting to
determine the racial composition of motorists at
risk of being stopped by police as a function of
traveling on the same roadways and violating
traffic laws. We believe that the state of
knowledge about racial profiling would be greater
if Lamberth's well-designed research were
augmented with additional studies looking at the
racial characteristics of persons who commit the
types of violations that may result in stops.
Other significant limitations of the
available analyses were that the results of some
analyses may have been skewed by missing data and
may not have been representative of roadways and
locations other than those reviewed. These
limitations notwithstanding, we believe that in
order to account for the disproportion in the
reported levels at which minorities and whites are
stopped on the roadways, (1) police officers would
have to be substantially more likely to record the
race of a driver during motorist stops if the
driver was a minority than if the driver was
white, and (2) the rate and/or severity of traffic
violations committed by minorities would have to
be substantially greater than those committed by
whites. We have no reason to expect that either of
these circumstances is the case.
Appendix II contains a discussion of some of the
methodological considerations and information
needs involved in getting stronger original data
from empirical research on the racial profiling of
motorists. These include the need for high-quality
data from multiple sources, such as from law
enforcement records, surveys of motorists and
police, and empirical research studies. By high
quality, we mean data that are complete, accurate,
and consistent and that provide specific
information on the characteristics of the stop and
the individuals involved in the stop in comparison
to those who are not stopped. The accumulation of
these data would form a better foundation for
assessing whether, and to what extent, racial
profiling exists on the roadways.
Federal Efforts to Collect Data on Motorist Stops
Although the federal government has a limited
role in making motorist stops, several federal
activities currently planned or under way
represent the first efforts to collect national
level information. The Police Public Contact
Survey conducted by BJS will include information
on the characteristics of individuals reporting
they were subject to traffic stops and other
information about the stop. BJS is also conducting
surveys of state and local law enforcement
agencies to determine what motorist stop data they
maintain. In addition, to help determine whether
federal law enforcement agencies engage in racial
profiling, three federal departments are under a
presidential directive to collect information on
the race, ethnicity, and gender of individuals
whom they stop or search.19
Population Survey of Motorist Contacts With Law
Enforcement
A national household survey now under way asks
respondents to discuss their contacts with police
during motorist stops. As part of BJS' 1999 Police
Public Contact Survey, BJS is conducting
interviews with 90,000 people aged 16 or older to
ask them up to 36 questions pertaining to the most
recent occasion (if any) during the prior 12
months that their motor vehicles were stopped by
police officers. For example, the interview
questions ask for information on the race of the
motorist and police officer, the reason for the
stop, whether a search was conducted, and whether
the officer asked what the person was doing in
that area. (See app. III for the survey questions
to be asked.) BJS completed the survey in December
1999, and expects the results to be available in
September 2000.
Surveys of Motorist Stop Data Collected by Law
Enforcement Agencies
BJS is conducting two surveys in an effort to
determine whether law enforcement agencies collect
stop data that can be used to address the question
of racial profiling. One survey targets state
police agencies; the other survey targets both
state and local law enforcement agencies.
In April 1999, BJS administered a survey of all
state police agencies in the nation. The Survey of
State Police Agencies asked, in general, whether
the agency required its officers to report
demographic information on the driver or other
occupants of every vehicle stopped for a routine
traffic violation. If the agency reported that it
did collect such information, then more detailed
questions were to be answered, such as whether
individual records were kept detailing the
driver's race and immigration status and whether a
search was conducted. BJS issued the results of
the state police survey in February 2000. BJS
found that 3 of the nation's 49 state law
enforcement agencies whose primary duties included
highway patrol reported that they required
officers to collect racial/ethnic data for all
traffic stops. Of the three states, Nebraska and
New Mexico reported storing the racial/ethnic data
electronically, and New Jersey reported that it
did not store the data electronically.
BJS administers the Law Enforcement Management and
Administrative Statistics (LEMAS) survey to a
sample of state and local law enforcement agencies
every 3 to 4 years. The survey collects
information on the budget, salaries, and
administrative practices of the agencies. The 1999
survey included a single question asking if the
agencies collected data on traffic stops. The
survey was sent to a sample of about 3,000
police/sheriff departments and was to include all
agencies with 100 or more employees. The 1999
survey results are expected to be available during
the summer of 2000. According to a BJS official,
the 2000 LEMAS survey will contain more questions
about what records are kept on motorist stops and
whether they contain information on race.
Data Collection on Motorist Contacts With Federal
Law Enforcement
Pursuant to a presidential directive, three
federal departments are to collect data on
contacts between their law enforcement officers
and the public. The directive did not instruct the
departments to focus solely on motorist stops, but
data on motorist stops are to be included.
In June 1999, the President issued a
memorandum on fairness in law enforcement that
addressed the issue of racial profiling. The
memorandum directed the Departments of Justice,
the Interior, and the Treasury to design and
implement a system for collecting and reporting
statistics on the race, ethnicity, and gender of
individuals who are stopped or searched by law
enforcement. The three departments were tasked
with developing data collection plans within 120
days and implementing field tests within 60 days
of finalizing the plans. After 1 year of field
testing, the departments are to report on
complaints received that allege bias in law
enforcement activities, the process for
investigating and resolving complaints, and their
outcome. The memorandum also required a report to
the President within 120 days of the directive
concerning each department's training programs,
policies, and practices regarding the use of race,
ethnicity, and gender in law enforcement
activities, as well as recommendations for
improvement.
The departments submitted data collection
plans and proposed locations for the field tests
to the White House in October 1999. (See app. IV
for the list of data elements to be collected and
all federal data collection test sites.) Federal
law enforcement offices and proposed locations
likely to be involved in motorist stops included
the following:
� INS inspectors at the land border crossing at
Del Rio, TX;
� INS border patrol agents from San Diego, CA;
Yuma, AZ; and El Paso, TX;
� National Park Service officers at eight
national parks; and
� National Park Service officers on three
federally maintained memorial highways.
According to Department of Justice plans,
officials will also pursue a variety of techniques
at some sites to try to determine if the
characteristics of those stopped differed from
populations encountered at the field site in
general.
Several States Proposed Traffic Stop Data
Collection Legislation, but Few Bills Passed
Most traffic stops are made by state and local law
enforcement officers. Consequently, state and
local agencies are in the best position to collect
law enforcement data on the characteristics of
stopped motorists. Several states have introduced
legislation that would require their state and/or
local police departments to collect data on
motorists' traffic stops. However, few bills have
passed.
As of October 15, 1999, at least 15 states had
taken some action to address concerns about racial
profiling of motorists. Two of the 15 states-North
Carolina and Connecticut-enacted legislation
requiring the collection and compilation of data
on motorist traffic stops. Similar legislation
requiring the collection of specific stop data was
introduced in 11 states. The legislation was
pending in 7 of those 11 states and was either not
carried over to the next legislative session or
vetoed in 4. The two remaining states, New Jersey
and Virginia, issued resolutions. New Jersey's
resolution calls for the investigation of racial
profiling, and Virginia's resolutions call for
data on traffic stops to be compiled and analyzed.
See table 1 for a list of the states that had
proposed or enacted traffic stop bills or
resolutions and their status as of October 15,
1999.
All 13 states with data collection legislation
proposed to collect data on driver's race or
ethnicity, the alleged traffic violation that
resulted in a motorist stop, and whether an arrest
was made. Most of these states also proposed to
collect data on age, on whether a search was
conducted, and on whether an oral warning or
citation was issued. The number of data elements
that each state proposed to collect ranged from 6
to 16. For a list of data elements that each of
the 13 states proposed to collect, see appendix V.
North Carolina passed legislation in April 1999
that called for the collection of statistics on a
variety of law enforcement actions. Part of the
legislation detailed what information on routine
traffic stops by state law enforcement officers
should be collected, maintained, and analyzed. All
of the state's approximately 40 state law
enforcement agencies are to collect the data,
although about 90 to 95 percent of all traffic
stops are made by the North Carolina State Highway
Patrol.
Connecticut's legislation passed in June 1999 and
requires collection of certain traffic stop data
on stops made by state as well as local police
departments. In addition, Connecticut's
legislation bans the practice of racial profiling
and calls for the collection of data on complaints
that were generated as a result of law enforcement
officer actions at traffic stops. North Carolina
and Connecticut were both in the process of
developing specifications for data collection.
They planned to begin data collection January 1,
2000.
Table 1: Status of Traffic Stop Bills Introduced
in State Legislatures
State Bill number Date Bill status
introduced
Arkansas HB 1261 January 1999 Referred to
committee;
session
adjourned, no
carryover
California SB 78 December 1998 Vetoed by
Governor 9/99
Connecticu Sub. SB March 1999 Bill became
t 1282 law 6/99 -
Public Act No.
99-198
Florida HB 177 September Pending -
1999 referred to
committee
10/99
Illinois HB 1503 February 1999 Pending -
referred to
committee 3/99
Maryland SB 430 February 1999 Passed House;
session
adjourned; no
carryover,
Massachuse SB 1854 June 1999 Pending -
tts referred to
committee 6/99
New Jersey Concurrent March 1999 Pending -
Resolution referred to
No. 162 Committee 3/99
North SB 76 February 1999 Bill became
Carolina law 4/99 -
Session Law
1999-26
Ohio HB 363 May 99 Pending -
referred to
Committee 5/99
Oklahoma SB 590 February 1999 Referred to
Committee; no
carryover
Pennsylvan HB 873 March 1999 Pending -
ia referred to
Committee 3/99
Rhode SB 131 January 1999 Pending -
Island referred out
of Committee,
5/99
South SB 778 April 1999 Pending -
Carolina Referred to
Committee;
session
adjourned;
bill carried
over
Virginia Joint Both January Both referred
Resolutions 1999 to committee;
736 and 687 session
adjourned; no
carryover
Sources: Professor David Harris, University of
Toledo College of Law; National Conference of
State Legislators; Internet search of state
legislatures; WESTLAW database.
Local Initiatives to Collect Motorist Stop Data
We visited four California police departments-San
Diego, San Jose, Alameda, and Piedmont-to learn
about local efforts to collect traffic stop data.
These departments had either begun or planned to
begin to voluntarily collect traffic stop data.
Some officials told us that their departments were
interested in collecting traffic stop data because
they wanted to address community concerns about
racial profiling. San Jose began collecting data
in June 1999, Alameda and Piedmont began
collecting data in October 1999, and San Diego
began collecting data January 2000.
The departments generally planned to collect
similar data; however, their data collection
methods and plans for analyzing the data differed.
All four police departments planned to collect
data on five data elements: race or ethnicity,
age, and gender of the driver; the reason for the
traffic stop; and whether the stop resulted in a
warning or citation or an arrest. In addition,
Alameda, Piedmont, and San Diego planned to
collect data on searches conducted during traffic
stops. San Diego planned to collect six additional
pieces of information. Table 2 summarizes the data
that the four police departments will collect.
Table 2: Traffic Stop Data Elements Collected by
Four California Police Departments
Data element San Diego San Jose Alameda Piedmont
Driver's X X X X
race or
ethnicity
Age X X X X
Gender X X X X
Reason for X X X X
stop
Location of X
stop
Search X X X
conducted
Legal basis X
of search
Obtain X
consent
search form
Result of X X X X
stop/stop
disposition
(e.g. oral
warning or
citation
Issued,
arrest made)
Property X
seized
Contraband X
found
Officer on X
special
assignment
Total number 11 5 7 6
of data
elements to
be collected
Source: GAO summary of police department
information.
In San Jose, officers use their police radios to
report traffic stop information to the dispatcher,
who then enters the data into a computer system.
Officers can also use mobile computers located in
their patrol cars to report traffic stop
information, and this can be transmitted directly
to the computer system. In San Diego, officers
initially are collecting vehicle stop data using
manually completed forms, and plan later to use a
wireless system to transmit information to the
department's database. The Alameda police
department also planned to use its computer-
assisted dispatch system to collect data, but only
on stops where citations are not issued, such as
stops resulting in warnings or arrests. For stops
in which the motorist receives a citation, traffic
stop data are to be abstracted from patrol
officers' ticket books and from motor officers'
hand-held computer printouts and input into a
citations database. Police officials in Piedmont,
a police department consisting of 21 officers,
decided that manually recording traffic stop
information on paper forms would work best for its
small department.
Three of the four departments indicated that they
expect to analyze their traffic stop data. A
preliminary report, issued in December 1999 and
providing analysis results on data collected
between July and September 1999 in San Jose,
indicated some racial disparity in traffic stops.20
According to the San Jose Police Department, the
differences were due to socioeconomic factors
rather than ethnicity. The report noted that more
police were assigned to areas of San Jose that
generated more police calls, and those
neighborhoods tended to have more minorities.
Because more police were available in these areas
to make traffic stops, more stops were made there
than in districts with a lower police presence.
Within each police district, the stops reportedly
reflected the demographics of the district. In the
report, the San Jose Police Chief emphasized that
more data were needed, along with the cooperation
of the community to analyze what the data mean.
Alameda officials told us they had no current
plans to analyze their data, but the data will be
available should there be a public request. None
of the four departments planned to independently
validate the accuracy of the data provided by the
police officers. They said they rely on the
integrity of the officers and supervisory
oversight to ensure that the data are correct.
Officials from two of the departments reported
that the amount of data to be collected was
limited so as not to be burdensome for officers.
However, a lack of information may limit the types
of analyses possible. For example, the data
collection efforts do not require data on the
specific violation for which a motorist was
stopped, so questions about whether minorities
were stopped more often for less serious
violations cannot be answered. None of four
localities planned to collect this information.
Officials noted, however, that trade-offs needed
to be considered: police officers would be more
likely to record motorist data if the data
collection requirements imposed on them were not
overly detailed or burdensome. For a more detailed
discussion on each of the four police departments'
traffic stop data collection plans, see appendix
VI.
Conclusions
The five quantitative examinations of racial
profiling that we identified did not produce
conclusive findings concerning whether and to what
extent racial profiling exists. Although
methodologically limited, their cumulative results
indicate that in relation to the populations to
which they were compared, African Americans in
particular, and minorities in general, may have
been more likely than whites to be stopped on the
roadways studied. Because of methodological
weaknesses in the existing analyses, we cannot
determine whether the rate at which African
Americans or other minorities are stopped is
disproportionate to the rate at which they commit
violations that put them at risk of being stopped.
Although definitive studies may not be possible,
we believe that more and better research data on
the racial characteristics of persons who commit
the types of violations that may result in stops
could be collected.
To date, little empirical information exists at
the federal, state, or local levels to provide a
clear picture of the existence and/or prevalence
of racial profiling. Data collection efforts that
are currently planned or under way should provide
more data in the next few years to help shed light
on the issue. These efforts are steps in the right
direction. However, it remains to be seen whether
these efforts will produce the type and quality of
information needed for answering questions about
racial profiling.
Agency Comments and Our Evaluation
We requested comments on a draft of this report
from the Justice Department. Based on a January 18
meeting with a Deputy Associate Attorney General
and other Justice officials, and technical
comments provided by Justice, we made changes to
the text as appropriate. In addition, Justice's
Acting Assistant Attorney General for Civil Rights
provided us with written comments, which are
printed in full in appendix VII. Justice agreed
with us that there is a paucity of available data
for assessing whether and to what extent racial
profiling of motorists may exist. Justice also
agreed that current data collection efforts by law
enforcement agencies, as well as additional
research studies, could generate information that
may help answer questions about racial profiling.
Justice felt, however, that our report set too
high a standard for proving that law enforcement
officers discriminate against minority motorists.
We believe that Justice's letter
mischaracterized the conclusion of our report.
Justice states that it disagrees with the "draft
report's conclusion that the only `conclusive
empirical data indicating' the presence of racial
profiling would be data that proved the use of
race to a scientific certainty." Our conclusion,
however, was that the "available research is
currently limited to five quantitative analyses
that contain methodological limitations; they have
not provided conclusive empirical data from a
social science standpoint to determine the extent
to which racial profiling may occur" (page 1). We
also noted that to account for the disproportion
in the reported levels at which minorities and
whites are stopped on roadways, (1) police
officers would have to be substantially more
likely to record the race of a driver during
motorist stops if the driver was a minority than
if the driver was white, and (2) the rate and/or
severity of traffic violations committed by
minorities would have to be substantially greater
than those committed by whites. We do not believe
that our approach to reviewing the research
studies was so rigorous that we required
"scientific certainty" in the data to draw
conclusions about the occurrence of racial
profiling. And we make clear in the report that
our review was not intended to comment on the
legal standard for proving discrimination in this
context (see our Scope and Methodology section).
With respect to Justice's suggestion that we
required research studies to provide scientific
certainty of racial profiling, we would note that
the concept of scientific certainty is generally
not applicable to social science research. This is
because social science research data are generally
imperfect because they are collected in the "real
world" rather than under controlled laboratory
conditions. A fundamental, universally accepted,
social science research principle that we did
incorporate into our assessment of study results
was whether the studies ruled out plausible
alternative explanations for findings. We found
that the available research on the racial
profiling of motorists did not sufficiently rule
out factors other than race-that is, other factors
that may place motorists at risk of being stopped-
that may have accounted for differences in stops.
We observed that the two studies by Professor
Lamberth were well-designed and went further than
others in attempting to determine whether race was
related to traffic violations that increased the
risk of being stopped. But Lamberth established a
criterion in each study that cast the net so wide
that virtually the entire population of motorists
was eligible to be stopped (i.e., traveling at
least 1 and 6 miles above the speed limit,
respectively, on two major interstate highways),
and his studies provided little information about
why motorists actually were stopped. Although law
enforcement officers can use their discretion in
deciding whom to stop, more information is needed
on the actual reasons why they stop motorists
before a firm conclusion can be made that the
reason was race. As we indicate in the report,
current data collection efforts by local, state,
and federal law enforcement agencies may provide
information on the reasons for stops that may help
answer this question.
With respect to what kind of data would be needed
to "prove" the use of race in motorist stops, this
issue was outside the scope of our work. We
recognize that the evidentiary standards that a
court may apply in ruling on an allegation of race-
based selective enforcement of the law may be
different from the social science principles that
we used to review these studies. It was not our
intention to express or imply anything about legal
standards to prove discrimination.
Justice also criticized our work for failing "to
recognize or comment on the extensive scholarly
debate on the subjects of the degree of
statistical certainty, and the extent to which
potential variables must be examined in order to
demonstrate discrimination from a social science
perspective." We did not comment on the matter of
statistical certainty because it was not the basis
for our determination that the available research
on racial profiling is inconclusive. The problems
that we identified with the research studies dealt
primarily with the design of the studies; that is,
using inappropriate or questionable benchmarks to
isolate race from other factors. More and better
data are needed on what traffic violations trigger
stops and whether race is related to them.
Justice agrees that it is important to use an
appropriate benchmark against which to compare the
racial composition of stopped motorists. Justice
disagrees, however, about the importance of
examining whether certain driving behaviors or
characteristics of vehicles may affect the
likelihood of being stopped. In this context,
Justice suggests that we make the unwarranted
assumption in our report that severe traffic
violations account for such a large proportion of
traffic stops that they have a significant effect
on the data. We did not intend, nor do we believe,
that the report makes any assumptions about the
reasons for which motorists are stopped. We simply
believe that if the objective is to determine
whether minority motorists are disproportionately
more likely to be stopped than whites, then it is
important to know what portion of the driving
population on that roadway or in that jurisdiction
commits the traffic offenses for which motorists
are actually stopped-as opposed to being eligible
to be stopped. This is the type of benchmark
information that would isolate, to the extent
possible, race from other variables that could
influence traffic stops.
As arranged with your office, unless you
publicly announce the contents of this letter
earlier, we plan no further distribution until 15
days after the date of this report. At that time,
we will send a copy to other appropriate
congressional parties, the Honorable Janet Reno,
the Attorney General, and to others upon request.
If you or your staff have any questions concerning
this report, please contact me or Evi L. Rezmovic,
Assistant Director, on 202-512-8777. Other key
contributors to this report are listed in appendix
VIII.
Sincerely yours,
Laurie E. Ekstrand
Director, Administration of Justice Issues
General Government Division
_______________________________
1 As used in this report, traffic violations that
can legitimately put motorists at risk of being
stopped include actions by drivers and
characteristics of motor vehicles that constitute
traffic/vehicle code infractions. These could
include, for example, speeding, tailgating,
failing to signal a lane change, driving an
unregistered vehicle, driving with license plates
not clearly visible, failing to dim the vehicle's
high beams when there is oncoming traffic, and
equipment violations.
2 42 U.S.C. 2000d.
3 42 U.S.C. 3789d(c).
4 42 U.S.C. 14141.
5 Whren v. U.S., 116 S. Ct. 1769 (1996).
6 "Driving While Black: Racial Profiling On Our
Nation's Highways." American Civil Liberties
Union, June, 1999.
7 Whitfield v. Board of County Commissioners of
Eagle County, 837 F. Supp. 338 (D. Colo.1993).
8 Both cases are described in David A. Harris,
"Driving While Black" and All Other Traffic
Offenses: The Supreme Court and Pretextual Traffic
Stops." Journal of Criminal Law and Criminology,
Vol. 87, No. 2 (1997), pp. 544-582.
9 New Jersey v. Soto, 734 A.2d 350 (N.J. Super.
Ct. Law Div. (1996)). The court therefore granted
motions to suppress evidence of criminal activity
by motorists that was obtained in these stops.
10 In 1997, the National Highway Traffic Safety
Administration (NHTSA) conducted a large-scale
nationally representative telephone survey of
drivers 16 and older to learn about the public's
experiences and beliefs concerning speeding, and
unsafe driving. Among other questions in a lengthy
interview, respondents were asked whether they had
committed a series of specific unsafe actions
while driving. Demographic data, including race
and ethnicity, were obtained on each respondent.
Although answers to the unsafe or aggressive
driving behavior questions were analyzed by some
demographic characteristics, no analyses by race
or ethnicity of driver were conducted. National
Survey of Speeding and Other Unsafe Driving
Actions, U.S. Department of Transportation,
National Highway Traffic Safety Administration,
September 15, 1998.
11 Lamberth, J.L. (1994, unpublished). Revised
Statistical Analysis of the Incidence of Police
Stops and Arrests of Black Drivers/Travelers On
the New Jersey Turnpike Between Exits Or
Interchanges 1 and 3 From the Years 1988 Through
1991.
12 The race of the driver was not available in two-
thirds of the cases.
13 Report of John Lamberth, Ph.D. from ACLU Freedom
Network, http://www.aclu.org/court/Lamberth.html
14 Lamberth told us that his study noted four other
types of traffic violations in addition to
speeding. The other violations consisted of no
signal for a lane change, unsafe lane change,
weaving, and tailgating.
15 Does not sum to 100 percent due to rounding.
16 David A. Harris, "Driving While Black and All
Other Traffic Offenses: The Supreme Court and
Pretextual Traffic Stops." The Journal of Criminal
Law and Criminology, Vol. 87, No. 2 (1997), pp.
544-582.
17 Plaintiffs' Fourth Monitoring Report: Pedestrian
and Car Stop Audit, Philadelphia Office of the
American Civil Liberties Union, July 1998.
18 Interim Report of the State Police Review Team
Regarding Allegations of Racial Profiling, New
Jersey Attorney General's Office, April 20, 1999.
19 Bills to provide for the collection of data on
traffic stops were introduced in the House and
Senate on April 15, 1999. These bills, H.R. 1443
and S. 821, called for the Justice Department to
study racial profiling by acquiring data on
motorist stops from law enforcement agencies.
Neither bill had passed as of March 1, 2000.
20 Vehicle Stop Demographic Study, San Jose,
California Police Department, December 17, 1999.
For each group, the percent of San Jose residents
and the percent of motorist stops reported were as
follows: Hispanics were 31 percent of residents
and 43 percent of stops; African Americans were
4.5 percent of residents and 7 percent of stops;
whites were 43 percent of residents and 29 percent
of stops; and Asian Americans were 21 percent of
residents and 16 percent of stops.
Appendix I
Studies of Racial Characteristics of Drivers
Stopped by Police
Page 29 GAO/GGD-00-41 Racial Profiling
A Summary of Analysis Design, Results, and
Limitations
As part of our work, we reviewed all available
quantitative analyses that we could identify
pertaining to the use of race as a factor in
motorist stops. This appendix provides a summary
of the design, results, and limitations for each
of the five analyses.
Source
Lamberth, J.L (1994, unpublished). Revised
Statistical Analysis of the Incidence of Police
Stops and Arrests of Black Drivers/ Travelers on
the New Jersey Turnpike Between Exits or
Interchanges 1 and 3 From 1988 Through 1991.
Study Design/Results
This analysis, done as part of a research study
for a court case, provided a comparison of the
races of vehicle occupants who were involved in
traffic stops and arrests, drivers who violated
traffic laws, and motorists in general who
traveled along a segment of the southern end of
the New Jersey Turnpike. The study involved three
types of data collection: (1) direct observation
of motorists from fixed observation points along
the side of the road; (2) a moving survey in which
an observer drove on the roadway and noted the
races of drivers and whether they were speeding;
and (3) obtaining law enforcement records from the
New Jersey State Police (NJSP).
In the first data collection effort, observers
were stationed beside the road. Using binoculars,
they noted the number of cars that passed the
observation point, the race of the driver and/or
any other occupant, and the vehicle's state of
registration. One observer was assigned to each
lane of traffic, and a data recorder was present
to record their observations. Observations were
made in 18 randomly selected 3-hour blocks of time
at 4 locations between 8 a.m. and 8 p.m. over a 2-
week period in June 1993. The author noted that
"most if not all" of the 26 pending cases in
Gloucester County Superior court arose between
these hours. Observers were reported to have been
between 14 and 45 feet from the roadway.
According to the observations, 42,706 cars were
counted as traveling on the turnpike, and the
race(s) of the occupants were recorded for nearly
100 percent. An African American driver and/or
other occupant were in 14 percent of the cars.
Seventy-six percent of the cars were registered
out of state.
In the second data collection effort, a moving
survey was conducted to identify the racial
distribution of all drivers on the road who
violated the speed limit. In this phase, one
observer drove at a constant 60 miles per hour (5
miles per hour above the speed limit at the time),
and he recorded onto a tape recorder the race of
each driver who passed him and whom he passed. The
observer noted all cars that passed him as
violators and all cars that he passed as
nonviolators.
In the moving survey, 1,768 cars were counted.
More than 98 percent were speeding and classified
as "violators." Fifteen percent of the cars
observed speeding had an African American driver
or other occupant.
A third data collection effort involved gathering
data from NJSP. The data included the race of
drivers who were stopped or arrested on randomly
selected days between April 1988 and May 1991
along the section of the Turnpike covered by the
traffic surveys and an additional section of the
roadway. These data included 1,128 arrest reports
from turnpike stops; 2,974 stops from patrol
activity logs from 35 randomly selected days; and
police radio logs from 25 of the selected days.
(The 1988 radio logs had been destroyed.) Of the
2,974 stops, 870 were from the section covered by
the traffic surveys. Data were not provided on the
number of arrests from this section.
Of 1,128 NJSP reports, the race of the
driver/occupants was noted in 1,059 of them.
According to these 1,059 reports, 73 percent of
those arrested were African American. The patrol
logs and radio logs noted 2,974 events as "stops."
Of the 2,974 stops, all but 78 noted the state of
the registration of the car. Twenty-three percent
of the stops were of New Jersey cars.
Lamberth noted that race was "rarely if ever"
noted on the patrol activity logs and that in the
radio logs, race appears about one-third of the
time for the records that had not been destroyed.
(Out of 2,974 stops, race was not noted in 2,041,
or 69 percent of the stops. Of the 870 stops that
were in the sections covered by the traffic
surveys, race was not recorded in 649, or 75
percent of them.) According to the available race
data on all stops, 35 percent of drivers stopped
were African American; 29 percent of all race-
identified stops involved out-of-state African
Americans; and 6 percent of the same stops
involved in-state African Americans. Of the 221
race-identified stops from the section covered by
the traffic surveys, 44 percent of the drivers
were African American.
In a separate analysis, Lamberth examined the race
of individuals who were ticketed by three
different units of the Moorestown, New Jersey
State Police barracks.1 He compared the proportion
of tickets issued to African Americans by the (1)
Radar Unit, which used a remote van and left no
discretion in the hands of patrol officers; (2)
Tactical Patrol Unit, which concentrated on
traffic problems at specific locations on the
roadway and exercised more discretion on whom to
stop than the Radar Unit; and (3) Patrol Unit,
which was responsible for general law enforcement
and exercised the most discretion among the three
units. Lamberth found that African Americans
received 18 percent of the tickets issued by the
Radar Unit, about 24 percent of the tickets issued
by the Tactical Patrol Unit, and about 34 percent
of the tickets issued by the Patrol Unit. These
results suggested that increasing levels of
trooper discretion translated into increasing
percentages of African American stops.
Limitations
Although the data suggest that African Americans
may have been disproportionately represented among
motorists stopped and arrested, because of several
limitations in the study's methodology, this study
does not provide clear evidence of racial
profiling of African American drivers.
First, the percentage of drivers violating traffic
laws was measured by determining the percentage of
drivers who were driving at least 6 miles per hour
over the posted speed limit. The study did not
attempt to distinguish motorists who were driving
6 miles per hour over the speed limit from those
who were speeding more excessively. On the basis
of the criterion used to indicate speeding
violation, the report concluded that 98 percent of
the cars were violating at least one traffic law.
We are uncertain whether this is an adequate
indication of the type or seriousness of traffic
violations that put motorists at risk for being
stopped by police. We also do not know the reasons
for which motorists were stopped.
Second, the traffic surveys and the data on police
stops and arrests were not from comparable time
periods. The police data were from about 2 to 5
years prior to when the traffic surveys were
conducted-the traffic surveys were done in June
1993, and the police data were from randomly
selected days from April 1988 to May 1991.
Third, the observed differences in the percentage
of African Americans ticketed by Radar, Tactical
Patrol, and general Patrol units may or may not
have been due to discriminatory practices on the
part of law enforcement officers. For the Tactical
and general Patrol units, we do not know the
reasons why tickets were issued, nor do we know if
different groups may have been at different levels
of risk for being stopped because they differed in
their rates and/or severity of committing traffic
violations.
Fourth, among stopped vehicles, the occupants'
race was not recorded for three-fourths of cases
along the portion of the highway where the traffic
surveys were conducted; race was not recorded for
two-thirds of cases along a larger portion of the
highway. Therefore, the race of most motorists
stopped is unknown. Statisticians performed
calculations to determine the implications of the
missing data for drawing conclusions about racial
disparities in stops.2 The calculations revealed
that if the probability of having race recorded if
one was African American and stopped was up to
three times greater than if one was white and
stopped, then African Americans were stopped at
higher rates than whites. Because we do not know
what factors affected officers' decisions to
record race, the true extent to which officers
tended to record race for African Americans versus
whites is unknown.
Source
Report of John Lamberth, Ph.D. From ACLU Freedom
Network, http:-//www.aclu.org/court/Lamberth.html
Study Design/Results
This analysis, done as part of a research study
for a court case, provided a comparison between
the racial distribution of motorists stopped by
the Maryland State Police (MSP) on I-95 in
northeastern Maryland, motorists whose cars were
searched by MSP, all motorists on the roadway, and
motorists on the roadway who violated traffic
laws. The study involved two types of data
collection: (1) a moving survey in which a team of
researchers drove on the roadway and noted the
race of drivers and whether they were speeding,
and (2) obtaining law enforcement records from the
Maryland State Police.
In the first data collection effort, a moving
survey was conducted to determine the races of
highway motorists and the races of highway
motorists who violated traffic laws. A team of
observers drove separately at the posted speed
limit (either 55 or 65 miles per hour) and
recorded the race of each driver who passed him or
her and whom he or she passed. The observer noted
all cars who passed him or her as violators and
all cars that he or she passed as nonviolators
(unless they were observed violating some other
traffic law.) Twenty-one observation sessions were
conducted on randomly selected days between 8 a.m.
and 8 p.m. during the period June to July 1996.
In the moving survey, over 5,700 cars were
counted. The author reported that driver's race
was identified for 97 percent of cars. Seventeen
percent of cars had African American drivers, and
76 percent had white drivers. Ninety-three percent
of cars were observed violating traffic laws.
Eighteen percent of the violators were African
American, and 75 percent were white.
In the second data collection effort, data on
motorists traveling a segment of I-95 were
obtained from MSP. These data included information
on (1) motorist stops made between May and
September 1997 in Baltimore, Cecil, and Harford
counties; (2) searches conducted between January
1995 and September 1997; (3) searches by MSP on
roadways outside this corridor; and (4) drug
arrests resulting from these searches.
The MSP data indicated that along the I-95 segment
studied, 11,823 stops were made by MSP between May
and September 1997. Of the 11,823 vehicles
stopped, it was reported that 29 percent had an
African American driver, 2 percent had a Hispanic
driver, 64 percent had a white driver, and 5
percent had a driver of another race/ethnicity.
With respect to searches, 956 motorists were
searched between January 1995 and September 1997.
It was reported that 71 percent were African
American, 6 percent were Hispanic, 21 percent were
white, and 2 percent had a driver of another
race/ethnicity. The proportion of searched cars in
which contraband was found was the same for whites
and African Americans and the same for I-95 as
compared to the rest of Maryland.
In comparison, there were 1,549 motorist searches
outside the I-95 segment. Of these searches, 32
percent were African American, 4 percent were
another minority, and 64 percent were white.
Limitations
Although the data suggest that African Americans
may have been disproportionately represented among
motorists stopped and/or searched, because of
several limitations in the study's methodology,
this study does not provide clear evidence of
racial profiling of African American drivers.
First, we are uncertain whether the study
adequately measured the type or seriousness of
traffic violations that put motorists at risk for
being stopped by police. For example, motorists
who greatly exceed the speed limit, commit certain
types of violations, or commit several violations
simultaneously may be more likely to be stopped
than others. The measure used to determine whether
a car was speeding was whether it was traveling at
any speed over the posted limit. As with the New
Jersey study by the same researcher, this study
did not attempt to distinguish between motorists
who drove 1 mile over the speed limit and those
who sped more excessively. Furthermore, this study
recorded whether traffic violations other than
speeding were committed but treated them as equal
in seriousness and equally likely to prompt a
stop. This may or may not have been a valid
assumption. In addition, we do not know the
reasons for which motorists were stopped.
Second, the data on police stops and police
searches were not from comparable time periods.
The data for stops were from May through September
of 1997, and the data on searches were from
January 1995 through September 1997. Lamberth
noted in a correspondence to us that the stop data
were not provided in time for his initial report.
These problems do not necessarily indicate a
systematic bias, however.
Source
Harris, David A.; Driving While Black and All
Other Traffic Offenses: The Supreme Court and
Pretextual Traffic Stops. Published in The Journal
of Criminal Law and Criminology 87 (2): 1997.
Study Design/Results
The analysis provides quantitative data from
Florida and Maryland. The Florida data first
appeared in two Florida newspaper articles in
1992. The Maryland data were obtained by the
author from lawyers involved in a Maryland
lawsuit.
The journal article compares the racial
characteristics of drivers involved in videotaped
stops on a segment of I-95 in Volusia County, FL,
over 3 years in the late 1980s (obtained from the
County Sheriff's Department by the Orlando
Sentinel) with population and observational data.
It was reported that videotapes of stops were not
made for much of the 3-year period and sometimes
deputies taped over previous stops. More than 70
percent of the persons stopped among nearly 1,100
videotaped stops on I-95 were African American or
Hispanic. African Americans, however, made up 12
percent of the driving age population in Florida,
15 percent of the traffic offenders in Florida in
1991, and 12 percent of the U.S. population.
(Hispanics were 9 percent of the U.S. population.)
Moreover, according to the Orlando Sentinel's
observations of 1,120 vehicles on I-95, about 5
percent of the drivers were dark-skinned.
The article also noted that of the nearly 1,100
stops, 243 were made for swerving, 128 for
exceeding the speed limit by more than 10 mph, 71
for burned-out tag lights, 46 for improper license
tags, 45 for failure to signal, and a smattering
of other offenses. Roughly half of the cars
stopped were searched, 80 percent of the cars
searched belonged to African American or Hispanic
drivers, and African American and Hispanic drivers
were detained for twice as long as whites. Only 9
of the 1,100 drivers stopped received tickets.
In Maryland, the only data provided in the article
are the percentages of African Americans and
Hispanics among 732 motorists stopped and searched
by 12 Maryland State Police officers with drug-
sniffing dogs between January 1995 and June 1996.
The article stated that 75 percent of the persons
searched were African American; and 5 percent were
Hispanic. Of the 12 officers involved, 2 stopped
only African Americans. Over 95 percent of the
drivers stopped by one officer were African
American and 80 percent of the drivers stopped by
six officers were African American.
Limitations
Because of several methodological limitations,
this analysis does not provide clear evidence of
racial profiling of African American or Hispanic
drivers.
For the Florida data, the validity of the
comparisons made is questionable. For example, the
data from the videotaped stops combined African
Americans and Hispanics, but the comparison data
for the driving age population of Florida included
African Americans only. More importantly, no
information was provided on the percentage of
African Americans and Hispanics among traffic
offenders. It is also not clear how accurately
information on "dark-skinned" drivers was
captured. In addition, there was an unknown amount
of missing data because videotapes of stops were
not made for much of the period. Therefore, we do
not know whether the videotaped stops were
representative of all stops.
For the Maryland data, no comparative data are
provided on the percentage of African Americans
and Hispanics among motorists generally, among
stopped motorists, or among motorists who violated
traffic laws. The data for drivers in Maryland
included only motorists who were stopped and
consented to being searched.
Source
Plaintiffs' Fourth Monitoring Report: Pedestrian
and Car Stop Audit, Philadelphia Office of the
American Civil Liberties Union, July 1998.
Study Design/Results
This was an analysis of the racial characteristics
of motorists and pedestrians stopped by the
Philadelphia Police Department in selected
districts and persons stopped by the department's
Narcotics Strike Force.
All police incident reports recording interactions
between police and civilians that involved stops
and investigations of pedestrians or automobiles
in the 8th, 9th, 18th, and 25th Police Districts
for the week of October 6, 1997, were obtained.
Hardcopy and computerized records were reviewed
and coded according to whether tickets or arrests
resulted from the stops and, if not, whether the
record indicated any legal explanation for the
stop. Previously unreported data were also
provided on pedestrian and automobile stops in the
9th, 14th, and 18th Police Districts for the week
of March 7, 1997. All reports filed by the
Narcotics Strike Force for incidents in the 4th,
12th, 17th, 25th, and 35th Police Districts that
involved a pedestrian or a vehicle stop during
August 1997 were obtained. Records were coded in
the same way as described above. Demographic data
for all Philadelphia residents from a 1995 census
were provided as a benchmark for the city as a
whole, and demographic data by census tract from
the 1990 U.S. census were provided as benchmarks
for the district-specific analyses. (The report
mentions that Philadelphia Police Districts
approximately encompass specific census tracts.)
For the week of March 7, there were police records
of 516 motorist stops in the 3 districts. Overall,
the race of the driver was recorded for only 51
percent of these stops, with race being recorded
for between 40 and 58 percent of the stops in the
three districts. For the week of October 6, there
were police records of 1,083 motorist stops in the
4 districts. Overall, race of the driver was
recorded for only 48 percent of these stops, with
race being recorded for between 44 and 46 percent
of the stops in three of the districts. (No
separate data were provided for the 25th District,
and no explanation was given for this omission.)
In both weeks in each district, for stops with
race of driver recorded, the driver was more
likely to be a member of a minority group than
would be expected on the basis of racial
characteristics of the district as indicated by
1990 census tract data. Additionally, for stops
with race recorded, the report indicated that
minorities were more likely than whites to be
involved in stops that were judged as not having a
legally sufficient explanation than in stops
judged to have a legally sufficient explanation
for the March data, but not for the October data.
There were records of 214 stops by the Narcotics
Strike Force in August 1997. (Task Force data were
not presented separately for motorists and
pedestrian stops.) However, the race of the
individual stopped was recorded for only 68
percent of the stops. For stops with race
recorded, the report indicated that minorities
were more likely to be involved in stops judged
not to have a legally sufficient explanation-43
percent African American, 39 percent Hispanic, and
18 percent white-than in stops judged to have a
legally sufficient explanation-33 percent African
American, 47 percent Hispanic, and 20 percent
white.
Limitations
Because of several methodological limitations,
this analysis does not provide clear evidence of
discriminatory targeting of minority drivers.
First, data on the racial characteristics of most
motorists covered in the study were not available.
The absence of these data is a severe limitation
because the race of most drivers stopped is
unknown.
Second, 1990 census tract data were used as
benchmarks for the racial characteristics of the
residents of the selected police districts.
However, as the study notes, these census tract
data were several years old at the time the study
was conducted, and it is unknown how well these
1990 census data portrayed the 1997 population of
these parts of Philadelphia. More importantly, no
information was provided on the race of drivers
who put themselves at risk for being stopped.
Source
Interim Report of the State Police Review Team
Regarding Allegations of Racial Profiling, New
Jersey Attorney General's Office, April, 20, 1999.
Study Design/Results
The report provides the racial characteristics of
drivers stopped, searched, and arrested by the New
Jersey State Police (NJSP) along the New Jersey
Turnpike. Data were obtained from NJSP on the
numbers of stops and searches made by troopers
assigned to the Moorestown and Cranbury police
barracks-two of three barracks assigned to the
turnpike. Motorist stop data were from April 1997
through November 1998 (except February 1998). Data
on motorist searches resulting from stops were
from the same two barracks. Only data on searches
for which motorists gave their consent for the
search were available. Motorist search data were
from selected months in 1994, all months in 1996
except February, and every month from April 1997
to February 1999. Data were obtained on motorist
arrests made by troopers assigned to the Cranbury,
Moorestown, and Newark barracks. Data on these
arrests were from January 1996 through December
1998.
Over 87,000 motorists were stopped by NJSP. Twenty-
seven percent of motorists stopped were African
American, 7 percent were Hispanic, 7 percent were
another minority, and 59 percent were white.
Little difference was reported between the two
NJSP barracks in the racial characteristics of
motorists stopped. Only 627, or less than 1
percent, of these stops involved a search, but the
racial characteristics of the motorists searched
were not reported separately.
Racial characteristics were available for 1,193
motorists who gave consent for searches. Fifty-
three percent of motorists searched were African
American, 24 percent were Hispanic, 1 percent were
another minority, and 21 percent were white.
Little difference was reported between the two
NJSP barracks in the racial characteristics of
motorists searched.
Approximately 2,900 motorists were identified in
the state's Computerized Criminal History Database
as being arrested3 by troopers assigned to all
three barracks. Sixty-two percent of motorists
arrested were African American, 6 percent were of
another minority, and 32 percent were white.
Little difference between the three NJSP barracks
in the racial characteristics of motorists
arrested was reported.
Limitations
Because of several methodological limitations,
this analysis does not provide clear evidence of
racial profiling of minority drivers.
First, direct comparisons between the racial
characteristics of drivers stopped, drivers
searched, and drivers arrested are problematic
because comparable data for stops, searches, and
arrests were not reported. Although there is some
overlap, data for stops, searches, and arrests
were reported for different time periods.
Second, search data were provided for consent
searches only. Data on instances when motorists
denied troopers' search requests were not
available. Without data on denied search requests,
it is not possible to know the racial
characteristics of all motorists from which
nonwarrant and nonprobable cause searches were
requested.
Overall, as the report acknowledges, it is
difficult to interpret the significance of the
study's results because of the absence of any
benchmark data, such as data from a survey to
determine the racial or ethnic characteristics of
turnpike motorists or the racial characteristics
of motorists who put themselves at risk for being
stopped.
_______________________________
1 The analysis was not included in Lamberth's
unpublished report but was cited in the judge's
decision in the related court case (New Jersey v.
Soto, 734 A.2d 350 (N.J. Super. Ct. Law Div.
(1996)).
2 These calculations were performed by two
statisticians, and the Justice Department provided
us a report of their findings.
3 Arrests generally include arrests for more
serious offenses, including all drug-related
arrests, but exclude arrests for drunk driving.
Appendix II
Methodological Issues In Studying Racial Profiling
of Motorists
Page 36 GAO/GGD-00-41 Racial Profiling
Determining whether and to what extent racial
profiling may occur on the nation's roadways is a
complicated task that would require collecting
more and better data than are currently available.
Additional studies using comparison groups that
are similar to the stopped motorist group in terms
of their risk of being stopped for a traffic
violation would contribute to our understanding of
this issue. Federal, state, and local data
collection efforts currently under way should
augment the available information provided that
the data are complete, accurate, consistent, and
specific. To the extent that such data are
gathered by a number of jurisdictions, a more
complete picture of which motorists are stopped
and why may emerge. Surveys of motorists and
police officers and reviews of police protocols
and training guides can also contribute to the
state of knowledge about racial profiling. In our
judgment, such a multifaceted examination of the
issues is the means for developing a full and
meaningful answer to questions about racial
profiling.
We have noted that some of the existing
analyses may have made comparisons that were not
valid. These analyses generally compared the
racial characteristics of motorists who were
stopped with the racial characteristics of a
larger population. The larger population may have
been a state's driving age population or the U.S.
population as a whole, among others. The
limitation of such analyses is that they do not
address whether different groups may have been at
different levels of risk for being stopped because
they differed in their rates and/or severity of
committing traffic violations. Although discretion
may play a part in an officer's decision to pull
over a driver, the justification for initiating a
stop is a violation or infraction committed by
drivers. The available research on racial
profiling, however, has given very little
attention to potential differences across groups
in the relative risk of being stopped.
Lamberth's studies1 have been important steps
in the direction of estimating the relative risks
of being stopped, but they did not provide
conclusive results. In both studies, Lamberth
found that more than 9 out of 10 motorists
violated a traffic law and were thus legally
eligible for being stopped by the police. However,
it is not clear that the driving violations that
made motorists legally eligible for being stopped
were the same violations that would prompt actual
stops by law enforcement officers. For example,
one of Lamberth's studies considered only
speeding, although this type of infraction is not
the only reason that motorists are stopped. The
extent to which motorists exceed the speed limit
and/or the number of violations they commit
simultaneously may also affect their likelihood of
being stopped. Lamberth's other study considered
speeding plus other traffic law violations.
However, this study also did not differentiate
between the type or seriousness of different
violations. For example, motorists who greatly
exceeded the speed limit, committed certain types
of violations, or committed several violations
simultaneously may have been more likely to be
stopped than others. None of the analyses that we
identified examined whether there may be racial
disparities in motorist stops that are related to
the type or seriousness of the traffic violation
committed. We recognize that it is difficult to
determine which traffic violations specifically
prompt a law enforcement officer to stop one
motorist rather than another. Different
jurisdictions and officers may use different
criteria, and candid information on the criteria
may be difficult to obtain. Nonetheless, to
understand the extent to which motorist stops may
have a discriminatory basis, data are needed on
traffic violations-including the type and
seriousness of those violations-that produce stops
and the relative rates at which different groups
of drivers in a particular jurisdiction commit
those violations. Although we have no reason to
expect that there are racial differences in
committing traffic violations, such data would
enable the most appropriate comparisons to be made
in order to answer a key question; that is, how do
the racial characteristics of motorists who are
stopped for a particular traffic violation compare
with the racial characteristics of all drivers who
commit the same violation but are not stopped?
Both observational studies and driver surveys may
be useful in developing such comparative
information.
Federal, state, and local efforts to collect
data on motorist stops should increase the amount
of information on law enforcement practices on the
roadways. However, the usefulness of such data for
addressing research questions about racial
profiling will depend on the extent to which the
data are complete, accurate, consistent, and
sufficiently specific to provide meaningful
information. Although we recognize that no
empirical data are likely to be perfect, it would
be difficult to draw conclusions about racial
profiling if (1) stop data were selectively
recorded, (2) race or other stop information is
inaccurately recorded, (3) different jurisdictions
capture different information, and/or (4) the
information recorded is too broad to understand
what happened. For example, recording "vehicle
code violation" as the reason for the stop-when
such a code can represent anything from failing to
signal a lane change within a designated distance
to a serious speeding offense-could make it
difficult to discern whether and how the traffic
violations for which motorists are stopped differ
between racial groups.
In addition, confidence in the quality of
data would be enhanced if provisions were made to
validate the accuracy and completeness of data
that are collected. Also, it would be constructive
to have a mechanism in place for agencies to
communicate and coordinate with one another to
ensure that they are collecting comparable
information, and at a sufficient level of
specificity, to be useful for answering questions
about racial profiling in a meaningful way.
It could also be instructive to examine
whether there was a correlation between the race
of the law enforcement officer and that of the
stopped motorist. In addition, information is
needed on the extent to which officers exercise
discretion in the process of stopping, citing, and
searching drivers. Toward this end, a review of
established police protocols and training guides
could be useful. In addition, a survey of officers
could provide information on what observations and
judgments they factor into their decisions to make
stops. Although survey data of this sort would be
subject to response biases, including the
possibility that respondents would offer socially
acceptable responses, well-designed surveys of
police officers could be a useful supplement to
official data. Further, in addition to querying
drivers about the frequency with which they were
stopped, cited, and searched, driver surveys could
also ask about how many miles the drivers
typically drove and how often they committed
infractions that were likely to prompt stops. Data
from police records and surveys could then be
compared with them.
_______________________________
1 See appendix I.
Appendix III
Bureau of Justice Statistics Police Public Contact
Survey
Page 39 GAO/GGD-00-41 Racial Profiling
Appendix IV
Federal Law Enforcement Data Collection
Page 45 GAO/GGD-00-41 Racial Profiling
Presidential Directive
President Clinton directed the Attorney General,
Secretary of the Treasury, and Secretary of the
Interior in a June 9, 1999, memorandum to design
and implement a system to collect and report
statistics relating to race, ethnicity, and gender
for law enforcement activities in their
departments. Within 120 days of the directive, in
consultation with the Attorney General, the
departments were to develop proposals for
collecting the data; and within 60 days of
finalizing the proposals, the departments were to
implement a 1-year field test. This appendix
presents the field locations and data elements
that the Attorney General's October 1999 proposal
indicated would be collected during the field
test.
Locations of Field Testing
Five agencies in three federal departments
are to be involved in collecting data on
individuals who are stopped or searched by law
enforcement. The agencies include the Department
of Justice's Drug Enforcement Administration and
the Immigration and Naturalization Service; the
Department of the Interior's National Park
Service; and the Department of the Treasury's U.S.
Customs Service and uniformed division of the
Secret Service.
Department of Justice
Between six and nine of the following Drug
Enforcement Administration Operation Jetway1 sites
are to be included in the field test:
� Detroit Metropolitan Airport;
� Newark International Airport;
� Chicago-O'Hare International Airport;
� George Bush Intercontinental Airport
(Houston);
� Miami International Airport;
� Charleston, SC, bus station;
� Cleveland, OH, train station;
� Albuquerque, NM, train station; and
� Sacramento, CA, bus station.
The following Immigration and Naturalization sites
are to be included in the field test:
� John F. Kennedy International Airport (New
York City);
� George Bush Intercontinental Airport
(Houston);
� Seattle/Tacoma Airport;
� El Cajon, CA, Station;
� Yuma, AZ, Station;
� El Paso, TX, Station; and
� Del Rio, TX, land-border crossing.
Department of the Interior
The National Park Service was the only agency
identified by the Department of the Interior with
regular public contact. The following Park Service
sites are to be included in the field test.
� Lake Mead National Recreation Area (Nevada
and Arizona);
� Yosemite National Park (California);
� Grand Canyon National Park (Arizona);
� Glen Canyon National Recreation Area (Arizona
and Utah);
� National Expansion Memorial Park (Missouri);
� Indiana Dunes National Lake Shore (Indiana);
� Natchez Trace Parkway (Mississippi and
Tennessee);
� Blue Ridge Parkway (Virginia and North
Carolina);
� Valley Forge National Historical Park
(Pennsylvania);
� Delaware Water Gap National Recreation Area
(Pennsylvania and New Jersey); and
� Baltimore Washington Parkway (Washington,
D.C., and Maryland).
Department of the Treasury
The Department of the Treasury identified the
U.S. Customs Service and the uniformed division of
the Secret Service as the agencies with regular
public contact. The following sites are to be
included in the field test:
U.S. Customs:
� Chicago O'Hare International Airport;
� JFK International Airport (New York City);
� Newark International Airport;
� Miami International Airport; and
� Los Angeles International Airport.
The Secret Service uniformed division will collect
data in Washington D.C..
Data Elements
Agencies are to collect data describing
demographic characteristics, such as gender, race,
ethnicity, national origin, and date of birth
based on agent's observation, or from official
documents such as drivers' license when available.
All participating agencies are to collect a core
set of data elements, but they may collect
additional data as they deem appropriate.
Following is a core set of data elements contained
in the data collection proposal:
� date of encounter,
� start time of contact,
� motorist's gender,
� motorist's race and ethnicity,
� motorist's national origin,
� location of contact,
� motorist's suspected criminal activity,
� reason for contact,
� external sources of information on person
contacted,
� law enforcement action taken, and
� end time of contact.
_______________________________
1 Operation Jetway is a drug interdiction program.
Appendix V
State Legislation and Proposed State Legislation
to Collect Traffic Stop Data: Elements to be
Collected
Page 49 GAO/GGD-00-41 Racial Profiling
Arkansas CaliforniConnectic
Proposed data elements HB 1261 a uta
SB 78 P.L. 99-
198
Race or ethnicity
Age
Gender
Reason for
stop/violation
Search conducted
Who, what searched
Legal basis of search
Oral warning or
citation Issued
Arrest made
Contraband; type,
amount
Property seized
Resistance to arrest
Officer use of force
Resulting injuries
Location, time of stop
Investigation led to
stop
Officer demographics
Passenger demographics
Auto description,
license number
Number of Individuals
stopped for routine
traffic violations
10 6 8
Total number of data
elements to be
collected
aData collection under Public Law 99-108 is to
begin January 1, 2000.
Florida IllinoMaryla Massachuse North Ohio Oklahom Pennsylva Rhode South
HB 177 is nd tts Carolin HB 363 a nia Island Carolin
HB SB 430 Sb 1854 ab SB 590 HB 873 SB 131 a
1503 S.L.199 SB 778
9-26
c
9 10 10 16 15 12 10 12 16
11
bData collection under Session Law 1999-26 is to
begin January 1, 2000.
cIncluding nature of offense for which arrest was
made, whether felony or misdemeanor, and whether
occupants checked for prior criminal record,
outstanding warrants, or other criminal charges.
Sources: Federal and State Proposals on Racial
Profiling, Professor David Harris, University of
Toledo College of Law; California State
Legislature Web site.
Appendix VI
Four Localities Data Collection Plans
Page 53 GAO/GGD-00-41 Racial Profiling
San Diego Police Department
The San Diego Police Department initiated its
program to collect vehicle stop data as a result
of concerns about police racial profiling that
were expressed by community groups, such as the
Urban League and the National Association for the
Advancement of Colored People. Beginning January
1, 2000, San Diego's police force, with 1,300
patrol and 60 motor officers, is to begin using
forms to manually collect stop data. Later, plans
are to use laptop or hand-held computers to
collect information that would be sent to a
department database via a new wireless system.1
Initial officer concerns about the data collection
effort were addressed through departmental
assurances that data would be collected in the
aggregate, keeping officers' and motorists' names
anonymous. In addition, the new data collection
system is to track when a stop was initiated for a
special assignment, such as when targeting African
American gang members. For each stop, officers are
to capture the following information: motorist's
race/ethnicity; motorist's age; motorist's gender;
reason for the stop; whether a search was
conducted and whom/what was searched; legal basis
for the search; whether a consent form was
obtained; whether an oral warning or citation was
issued; whether an arrest was made; whether
property was seized; whether contraband was found;
and whether the officer was on special assignment.
San Diego police officials said that they
plan to enlist the assistance of a statistical
expert in analyzing the data. They hope to obtain
an initial analysis after the first 6 months of
data collection. The department is also working
with community-based organizations to address
questions they have about the project and how data
will be interpreted. San Diego has no plans to
validate data submitted by officers. However,
officials noted that actions by officers could
always be reviewed and scrutinized by their
supervisors.
San Jose Police Department
The San Jose Police Department also began its
program to collect traffic stop data in response
to community concerns about racial profiling by
police. According to police officials, the data
collection will allow them to learn more about the
types of stops being made and to demonstrate the
department's commitment to working with all
members of the community. In addition, if analysis
of the data reveals a pattern suggesting that race
was a factor in motorist stops, then additional
training and supervision will be considered to
ensure fair treatment for all.
San Jose began collecting motorist stop data on
June 1, 1999, and plans to continue the effort
until May 31, 2000. For each stop, officers are to
capture the following information: motorist's
race/ethnicity; motorist's age; motorist's gender;
reason for the stop; and what action was taken
during the stop, for example whether a citation
was issued or whether an arrest was made.
Identities of the officer and motorist involved in
each stop will be kept anonymous and not included
in any reports.
San Jose officers call in traffic stop information
by police radio to a radio dispatcher or by keying
the information into a mobile computer terminal
located in patrol cars. Dispatchers enter the
radioed information into the computer-aided
dispatch (CAD) system, and information entered
into the mobile terminal is automatically entered
into the CAD system. Officers use single digit
alpha codes to identify traffic stop data
elements. San Jose's code system has been in place
since the 1970s; however, what is new is the
addition of three new data elements to the
existing code system. In addition to gender and
traffic stop disposition, San Jose now collects
reason for stop, race, and age information. The
hardware and software cost to implement the data
collection system was less than $10,000. According
to a police official, costs were minimal because
the department was able to make modifications to
its existing automated system, thereby avoiding
the need to design a new, potentially costly, one.
The department's Crime Analysis Unit is to compile
the statistics and prepare two formal reports; one
summarizing results for the first 6 months of data
collection, and the other summarizing results for
the full year. An initial review of the data from
July 1, 1999, to September 30, 1999, was released
by the San Jose Police Department in December,
1999. Aggregate figures indicate that Hispanic
citizens in particular were stopped at a rate
above their representation in the population. A
spokesman for the department stated that the
results do not support this conclusion when the
figures are disaggregated by police district,
although population figures by police district are
not available. The official explained that more
officers are assigned to areas with higher calls
for service, and thus more stops are made in these
areas, which tend to have higher minority
populations. More analysis will be forthcoming.
If results suggest that race may be a factor in
motorist stops, the department may decide to
collect data beyond 1 year. San Jose does not plan
to check the validity of the data being submitted
by officers, except to see if officers have
entered the correct number of codes. However, a
police official told us that supervisors have
access to data submitted by officers, and they can
"stop-in" on an officer call at any time.
Alameda Police Department
According to Alameda Police Department officials,
most of Alameda County's police departments began
to voluntarily collect motorist stop data in
anticipation of state and federal legislation
requiring the collection of such data.2 The
Alameda Police began collecting motorist stop data
on October 1, 1999.
Alameda police officials told us that stop data
are recorded on written or automated citations, if
issued. For all noncitation stops, such as
warnings or arrests, officers use the CAD system
to call in each of the required data elements. For
each stop, officers are to capture the following
information: motorist's race/ethnicity, motorist's
age, motorist's gender, reason for the stop,
who/what was searched, whether an oral warning was
given, and whether an arrest was made.
Alameda police officials said that information
patrol officers write on citations will be keyed
into an automated citations database. In addition,
motorcycle officers have hand-held computers that
they use to input and store traffic stop
information. These data will be printed out and
keyed into the automated citations database as
well. A separate database is to contain the CAD-
collected data for noncitation stops.
Although officers' and motorists' information will
be captured in the data system, the department has
no plans to generate any reports from the data
collected. According to Alameda police officials,
the police department does not plan to analyze,
validate, or publish its data. They said that the
data would be made available to the public if
requested.
Piedmont Police Department
The Piedmont Police Department, located in Alameda
County, began voluntary collection of motorist
stop data in anticipation of pending state and
federal legislation. Piedmont began collecting
motorist stop data on October 1, 1999.
According to a Piedmont police official, Piedmont
is a small department with 21 officers who record
motorist stop data manually. For each traffic
stop, the officer is to fill out an index card
that contains data fields for recording the
motorist's race, sex, and age. At the bottom of
the card, the officer is to record the reason for
stop, whether the vehicle was searched, whether an
oral warning or citation was issued, and whether
an arrest was made. No officer or motorist names
will be included on the cards. A department
official indicated that she expects a volume of no
more than 400 cards per month. Information from
the cards is to be input into an Excel spreadsheet
for analysis, and results are to be tallied on a
monthly basis.
The department reportedly has no planned effort to
validate the information that officers record on
the cards. Piedmont police officials said that the
watch commander can monitor the activity of
officers by listening to interactions between the
officers and motorists over the dispatch system.
The watch commander can then compare the
information overheard on the dispatch system with
that recorded on the index cards submitted by the
officers.
_______________________________
1 The department's move to a wireless system is
part of an overall updating of technology for the
agency. As of January, 2000, some technical flaws
in the system were still unresolved.
2 As noted in table 1 of our report, the governor
of California vetoed legislation proposing the
collection of motorist stop data. A federal bill
(H.R. 118) requiring that the Department of
Justice conduct a study of racial profiling was
referred to the Senate, but no action has been
taken.
Appendix VII
Comments From U.S. Department of Justice
Page 57 GAO/GGD-00-41 Racial Profiling
Appendix VIII
GAO Contacts and Staff Acknowledgments
Page 58 GAO/GGD-00-41 Racial Profiling
GAO Contacts
Laurie E. Ekstrand (202-512-8777)
Evi. L. Rezmovic (202-512-8777)
Acknowledgements
In addition to those named above, David P.
Alexander, Carla D. Brown, Ann H. Finley, Monica
Kelly, Anne K. Rhodes-Kline, Jan B. Montgomery,
and Douglas M. Sloane made key contributions to
this report.
*** End of Document ***