CITY'S PLAN TO PROBE RACIAL PROFILING FLAWED

Some HPD Officers Ignore Form

In August 1999, Mayor Lee Brown announced that Houston police would begin
documenting every traffic stop they made in an effort to see if any were
using racial profiling.

"We owe it to the community to find out," he said at a news conference
during the National Urban League's convention here.

But police might not be finding out much, if anything. A computer analysis
by the Chronicle shows that after nearly 20 months, the data collected by
the state's largest police force may be seriously flawed.

As the Texas Legislature considers whether to require every police agency
in the state to record the race of everyone it stops, Houston's biggest
contribution might be to illustrate the policy's potential pitfalls:

* While Houston police have made at least 450,000 traffic stops since the
program began -- and likely hundreds of thousands more -- their
racial-profiling database had only 110,000 entries as of a month ago. Some
officers say they simply don't fill out the computerized form.

* The error rate in the "race" box is at least 19 percent. Officers can
choose among "w" for white, "b" for black, "h" for Hispanic, "a" for Asian
and "n" for Native American, but thousands of records contain blanks and typos.

* Officials seem unsure whether they can determine which officers made
which stops.

The department last year told the Chronicle that because of unforeseen
computer problems, the officers' names could be matched to their stops only
by hand-checking roll-call rosters. When the Chronicle asked for copies of
those rosters, the department said it would charge the paper $6,462.90 in
copying and other fees.

But it might not be so complicated after all. Upon receipt of a computer
file of officers' log-in times to their car terminals, the Chronicle last
month was able to match most stops to an officer in less than five minutes
using off-the-shelf software. Still, more than 7,000 stops appeared to be
unmatched with any officer.

Police Chief C.O. Bradford declined a request for an interview. Through a
spokesman, he said the department would be modifying the data-collection
process and would be asking two college professors for their assistance in
analyzing the data.

Meanwhile in Austin, the Legislature is considering bills sponsored by Sen.
Royce West, D-Dallas, and Rep. Senfronia Thompson, D-Houston, that outlaw
the practice of stopping and questioning a citizen because of his or her
race. A new system of either reporting or videotaping all stops would be
mandated.

While civil rights groups strongly back the legislation, some police
associations are wary of it. And Houston's program may shed light on the
problems involved in collecting and using such data.

A police officer sees a young black man in a new, expensive car. He
suspects the man is a drug dealer and follows until he finds a reason to
pull him over.

The man turns out to be an attorney and is let go. He is furious, knowing
he was pulled over because of his race.

This, critics say, is "racial profiling" and is widespread among the
nation's law enforcement agencies. For years, police officials said that
such instances were unverified anecdotes, not common occurrences.

To try to settle the issue, the police department in San Jose, Calif., in
1997 became the first to announce that it would track all of its stops for
racial patterns.

In August 1999, Brown, during a national convention featuring a panel on
police and race, announced that Houston would begin a similar program.

Officials said tracking all officer-initiated stops -- whether a citation
was issued or not -- would send a powerful message to the community that
racial profiling would not be tolerated.

It was supposed to send the same message to police officers. But many of
them complained that the collection process would be an annoying waste of
time and that the data collected could only be used against them.

They cited the case of officers working in all-black or all-Hispanic
neighborhoods whose stops would be overwhelmingly minority and thus
presumably suspect.

Bradford said he didn't believe his officers were making stops based on
race but that the community needed assurance. He also said supervisors
would know which officers worked in minority areas.

As the program kicked into gear, traffic tickets dropped precipitously as
wary officers stopped fewer people. (Other factors also were involved in
the drop, such as changes in court procedures and a temporary loss of grant
money for overtime patrols.)

Ticket volume has crept back up. But though some elected officials in
Austin are citing Houston as a role model, close scrutiny reveals serious
flaws in the program.

From Aug. 24, 1999, to March 8, 2001, the racial-profiling database,
obtained by the Chronicle through the Texas Public Information Act, shows
110,496 officer-initiated contacts.

"Officer-initiated" means the officer makes the decision to stop a person,
like a traffic stop. It does not include responding to an accident or to a
911 call.

The problem is that the data suggest Houston's 5,400 officers initiated
stops on only about 200 people per day. That's an average of one such stop
per day for every 27 officers, an impossibly low number.

According to police records, officers issued about 770,000 traffic
citations in the same period that 110,496 entries were made in the
profiling database.

A portion of that gap can be explained. First, officers can issue multiple
citations in a single traffic stop. According to a Houston Municipal Courts
study, the average is 1.7.

Dividing 770,000 by 1.7 indicates that police made about 450,000 traffic
stops in which a ticket was issued.

Another part of the gap is explained by the fact that some officers are not
required to collect profiling data. Those on a horse, bicycle or motorcycle
do not have the "mobile data terminal" needed to fill out the computerized
form.

But that doesn't close the gap. Motorcycle officers write by far the most
traffic tickets of the three excluded groups, police officials said, and
they account for only 78,000 citations, or perhaps 46,000 stops in the
period examined.

And, of course, the preceding comparison between the officer-initiated
database and traffic-ticket stops excludes thousands of other stops that
should be included in the database -- such as stops in which no ticket was
issued and stops in which someone was arrested on suspicion of driving
while intoxicated or other crimes.

A final way to narrow the gap in the favor of police would be to consider
the fact that not all traffic tickets are written after an
officer-initiated stop. Sometimes dispatchers send officers to an accident
scene, where they then write citations.

However, during the time period studied, Houston had only 75,000 traffic
accidents, according to the Texas Department of Public Safety -- not enough
to fill the gap.

There is a more likely explanation: Officers are not filling out the forms.

Among 10 officers interviewed at a shift change at the Houston Police
Department's central division, two said they thought they were no longer
required to fill out the form.

Most said that they filled it out but weren't sure if others did. One just
laughed. Another said "no way" did his colleagues fill out the forms,
because it was "a pain in the ass."

When the program started, officers could not get to the form they needed to
write a ticket without first entering some data on the racial-profiling
form. But that feature has been disabled, several said.

And apparently no one is checking up on the program.

Under the Texas Public Information Act, the Chronicle picked out 13
officers at random and requested one "work card" filled out by each, from
randomly selected days.

Although the handwritten cards indicated those officers made more than 40
stops that should have been entered into the database, none was found there.

That's unacceptable, says a leading national advocate of collecting
profiling data.

"If you really want something out of this, you have to give people
assurance that it was collected properly," said David Harris, a University
of Toledo law professor. "If the effort is not given a chance, you'd be
better off not doing it at all. Nobody is reassured."

The question is posed to Kevin Begley, chief of the West University Place
police:

It is 3 p.m. on a quiet cul-de-sac in his affluent, nearly all-white
enclave. A young black man in shorts and a T-shirt is walking around. If an
officer stops him to see what the man is up to, is this racial profiling?

"Well, no. Well ... based solely on a black individual on a sidewalk, yeah,
it would be," Begley waffled.

But, he added, a young white man wandering in Houston's predominantly black
Third Ward would also be considered suspicious.

"Because it's out of place," he said, "and that's what police officers are
trained to look for."

Begley agrees in principle that singling someone out solely because of his
race is wrong. "No one should have to worry about driving or walking while
black or brown," he said.

But does it work? Does racial profiling help officers do their jobs,
especially when combined with other aspects of a full "profile" -- for
example, if someone is nervous, wearing certain clothes or driving a
particular kind of car?

Among top executives of larger law enforcement agencies, the debate is
largely over. They now agree that the anger and distrust racial profiling
engenders in minority communities is not worth any extra arrests it
purportedly leads to.

But Begley, as a representative of the Harris County Area Police Chiefs
Association, which comprises 80 smaller departments in the Houston area,
opposes the main racial-profiling bill in Austin, SB 1074.

Already the bill is a product of hard compromise between civil rights
groups and police groups in meetings overseen by West.

It requires all departments to keep track of the race or ethnicity of
people they ticket and make special records of searches. The state already
requires race on all tickets, but Hispanics are technically classified as
"white."

HPD and some other departments have addressed that problem internally with
a "Hispanic" category, but some have not.

Departments would also be required to compile a second database of the race
or ethnicity of all people who are stopped, not just those who are ticketed
or arrested. Harris and others say it is even more important to document
those stops where there has been no violation of law.

But the legislation includes a huge exception to its reporting
requirements. If every patrol vehicle is equipped with a video camera that
records all stops, and the tapes are saved for three months, a department
doesn't have to keep a database of all its stops, just those that result in
a ticket.

West's chief of staff said this will reduce racial profiling because
citizens will have video to verify any claim of mistreatment. The bill
requires departments to have a well-publicized complaint procedure.

Begley testified last month that data entry will take too much of officers'
time and cause them to avoid contact with citizens. He also said the video
requirements are an "unfunded mandate" to local governments.

Rep. Thompson likened such complaints to Chicken Little protesting that the
sky was falling. But the points raised by Begley and others were not
dismissed. Although the bill has been passed by the Senate and cleared a
House committee, it is on hold while West and Thompson scramble to find $35
million to cover the costs being forced on departments.

The biggest boosters of profiling statistics admit there's still a major
problem. Once the statistics have been compiled, what do they show?

Are they compared with a city's racial makeup as determined by the
nationwide census? Are they compared with licensed drivers living in the
city's jurisdiction? Are they compared with the racial composition of the
drivers on the roads, if that could be determined?

In West University Place, surrounded by Houston, that's a major issue. The
residents are mostly white, but drivers passing through are of the
variegated colors of greater Houston.

"That really is the million-dollar question. The analysis (of data) has
been done really poorly," said Amy Farrell, a senior research associate at
Northeastern University and co-author of a U.S. Department of Justice study
titled, "A Resource Guide on Racial Profiling Data Collection Systems."

A professor at Temple University in Philadelphia, out to prove that New
Jersey state troopers target blacks, went so far as to have his assistants
drive at 5 mph over the speed limit on the New Jersey Turnpike and record
the race of everyone who passed them.

In that way, they had the rough racial composition of all drivers who
deserved to be pulled over. (He found blacks and whites sped in equal
proportion.)

A look at Houston's meager racial-profiling database illustrates the
problem. Eliminating the 19 percent of entries that have a blank or typo in
the race category, the database shows:

. Whites, who constitute 31.5 percent of the city's population in the 2000
Census, were pulled over 37.8 percent of the time.

. Hispanics, 37.4 percent of the city's population, were pulled over only
27.9 percent of the time.

. Blacks, 25.3 percent of the population, were pulled over 30.7 percent of
the time.

The data would appear to indicate that police pick on white people even
more than black people, and pick on Hispanics the least.

That's a surprising conclusion, and probably inaccurate because of the
skimpiness of the database and other factors. For example, Hispanics may be
37 percent of the city's population, but as Houston's youngest demographic
group, they probably aren't 37 percent of the drivers.

The decision to simply eliminate the 19 percent of entries that were blank
or typos may also be a problem. It assumes that they were mistakes made
randomly. But what if officers were carefully filling in their "white"
stops but leaving their "black" stops blank?

"Officers that do racially profile will never fill out that form or get on
the computer showing they've made a traffic stop," said Hans Marticiuc,
president of the Houston Police Officers Union.

"As a 21-year police officer, I can tell you, if someone is out there doing
illegal activity, they're not going to tell you," he said.

But what if the statistics do show a tendency to pull over more minorities,
as they have in other cities? Bradford has offered explanations for why a
high stop count for minorities might not indicate a racial-profiling problem.

In a December 2000 memo to City Council member Annise Parker, who had
inquired about the program, he wrote that "many of the calls for police
service in minority populated areas include criminal activity occurring on
the public streets (i.e., loud noise, open drug activity, trespass).
Therefore, police officers responding to these requests initiate a greater
number of contacts."

Bradford also said that more low-income people tend to drive cars that have
problems than do higher-income people, problems that would lead to a police
stop.

He could have added that some middle-class areas contract with county
constables to patrol their subdivisions, while Houston police focus on
areas with high numbers of apartments. Unless 911 has been called, HPD
rarely enters some largely white neighborhoods.

"How much can (data collection) do? I'm not sure. It kind of sensitizes
people" to the issue, said West, author of the legislation.

"It is very demeaning to be stopped because of the color of your skin, in
front of your children," he said.

Professor Harris said data collection is only one tool, requiring the
support of the community and the police. Departments should not throw
together a "slapdash" system just to meet minimal requirements, he said.

"You wouldn't approach crime-fighting that way," he said.

West's chief of staff, Janna Burleson, said his legislation is more than
just a data-collection measure, citing the videotaping requirement and the
fact that it includes a forthright declaration by the state that racial
profiling is forbidden.

"You don't have to look far to find a minority who it's happened to," she said.