KILLINGS BY POLICE UNDER-REPORTED

Special Report: Deadly But Legal

Day 2 Of A 2-Part Series

Area agencies not required to give Justice their numbers

DAY 1: At least 81 times in the past six years, officers in Central
Florida have shot or shot at suspects, killing 37 of them. More than
40 percent of shootings by Central Florida police involved suspects
who were unarmed, and officers in those cases were almost always
cleared by prosecutors and internal investigators. Though equal
numbers of black and white suspects were shot, black suspects were
twice as likely to be unarmed.

DAY 2: Because federal statistics on the use of deadly force by police
are incomplete, experts said they can't accurately estimate the number
of suspects killed by officers. When an officer and suspect confront
each other in dangerous circumstances, both are making life-and-death
decisions in a heartbeat. Though most officers are cleared in
shootings, two Central Florida officers have faced
prosecution.

You can find reams of statistics for practically every aspect of law
enforcement in the United States, from the kind of holsters officers
are wearing when they're assaulted to the number of murders committed
every year using blunt objects.

But there's one statistic you won't find among all those numbers: an
accurate accounting of how often police officers use deadly force.

The available nationwide statistics are so inaccurate that the numbers
for Central Florida police agencies from 1999 through 2002 reported
only a quarter of the actual fatal shootings by police, a review by
the Orlando Sentinel and WESH-NewsChannel 2 has found.

In other words, three out of four fatal shootings by Central Florida
officers did not show up in federal reports for those four years.

Critics said there are no reliable figures on killings by police
because some police agencies worry that a full accounting of deaths by
officers would be embarrassing.

Some of that criticism comes from within law enforcement.

"I've been ranting about it for 30 years," said James J. Fyfe, deputy
commissioner of police for training for New York City's Police Department.

"I like to say that we live in a democracy where we don't know how
often the people we pay to protect us kill us," Fyfe said.

The U.S. Justice Department gathers data on "justifiable homicides" by
police and by private citizens -- killings ruled to be legal by
authorities -- and that information is available to the public.

But the accuracy of the data has long been questioned because it's
voluntary for police agencies to file reports on their cases, and many
don't. A detailed report of 20 years of that federal data, released in
2001, included two pages of cautionary notes about errors and
omissions in the data.

In 2003, a group of criminal-justice experts studied the available
national statistics on police shootings, and the group's findings are
summarized by the title of the study: "Underreporting of Justifiable
Homicides Committed by Police Officers in the United States, 1976-1998."

The experts examined the Justice Department data, which reported 8,658
killings by police in that time, and separate figures from the
National Vital Statistics system, which reported 6,686 deaths by
police shootings based on death certificates.

The study found that both are wrong, but for different reasons, said
study co-author Brian Wiersema, a researcher with the Violence
Research Group at the University of Maryland.

The Vital Statistics reports don't reflect the actual number of
killings by police officers because many death certificates simply
fail to note that the death was caused by police, Wiersema said.

And many police agencies either file incomplete or inaccurate reports
to the Justice Department about police shootings, or file none at all,
Wiersema said.

Wiersema's study, published in the American Journal of Public Health,
concluded that "at present, reliable estimates of the number of
justifiable homicides committed by police officers in the United
States do not exist."

Accurate data would be valuable

The benefit of accurate data wouldn't be only academic, Wiersema said,
because how and when police use deadly force is a contentious issue in
many communities.

As his study noted, "Almost every major civil insurrection that
occurred in the United States in the past century was initiated or
accelerated by the perception that the police had misused their right
to use deadly force."

Accurate data would show whether a particular police agency has used
deadly force more or less often than other agencies of the same size,
or whether use of deadly force has declined, as many in law
enforcement contend, with wider use of nonlethal weapons such as Taser
stun guns.

But the available data can't be trusted for those purposes, Wiersema
said.

"It's simply not accurate," he said.

Just how far off those statistics are can be seen by the figures
reported for Orange, Osceola, Seminole, Brevard, Volusia and Lake
counties for 1999 to 2002, the last full year for which the figures
are available.

According to the Justice Department data, which are sent by Florida
police agencies to the Florida Department of Law Enforcement, officers
in those six counties reported seven justifiable homicides by officers.

But a review by the Sentinel and WESH found at least 28 fatal
shootings by officers in those counties in that time, all ruled
justifiable by prosecutors. Those were among 81 cases reviewed in
which Central Florida officers had used deadly force since late 1998,
killing at least 37 people.

Because there is so much missing data nationwide, Wiersema said, it's
not possible to estimate how many people are killed by officers each
year. But, in some states and large metropolitan areas, studies of the
federal data have shown it to be fairly accurate, he said. Nationally,
the percentage of unreported killings by police is probably lower than
among agencies in Central Florida, Wiersema said.

Reporting is voluntary

Why are the numbers wrong?

One explanation might be that the very phrase used to describe
killings by police, "justifiable homicides," means a police killing
doesn't necessarily fit that description until it has been
investigated and prosecutors have concluded it was a legal use of
deadly force.

If those decisions are delayed for months, as they often have been in
Orange County, it's possible the killings weren't classified as
justifiable until after the reporting deadline, said Randy Luttrell, a
senior manager for FDLE's statistics program.

But the most likely explanation is simply that police agencies aren't
required to file data on their killings, and so they put little
emphasis on doing so, experts including Fyfe and Wiersema said.

As Wiersema's study found, statistics on killings by police "are known
to be biased downward because some agencies do not file the necessary
forms."

And if they don't, no one will know.

"This is a voluntary program," Luttrell of FDLE said. "If they're not
reporting them, we'd really have no way of knowing."

Police officials said they weren't aware the data are inaccurate.

Orange County Sheriff's officers fatally shot six people from 1999 to
2002. But data provided by FDLE listed just a single fatal shooting in
that period.

Sheriff Kevin Beary said he didn't know the statistics were incorrect,
but he said they should be accurate.

"Law enforcement shouldn't have anything to hide," Beary said.

Luttrell said he wasn't aware the data for Central Florida police
shootings were inaccurate, and he couldn't say whether the statistics
for other areas of the state also are wrong.

"Not looking at it, I just don't know," Luttrell said.

He said he doesn't recall the state agency's ever reviewing the
police-shooting data for accuracy, which he said it has done with
other crime statistics.

Expert: Bad data easy to spot

But in other crime statistics, it's generally easy to spot incomplete
data when they are filed by agencies, Luttrell said.

For example, if a large police agency reports 5,000 burglaries a year
for several consecutive years and then reports only a few hundred, the
change is obvious enough that the state will check with the agency to
see whether there's a mistake, he said.

But the number of killings by most police agencies is small enough,
and varies enough from year to year, that changes in the numbers
aren't a tip-off that the numbers are bad, Luttrell said.

For example, an agency that hasn't had a fatal shooting in years might
have several in one year. And even large agencies, such as the Orange
County Sheriff's Office, sometimes have no fatal shootings in the
course of a year.

"The numbers are so small, those kinds of change don't really stand
out" the way they do with other police statistics, Luttrell said.

Statistics difficult to compare

Other problems in the data make it tricky to compare Florida's
statistics with other states'. For example, the federal statistics
from 1988 to 1998 include six years in which there is no data from
Florida at all.

Luttrell said that's because Florida officials compiled the data for
those years in a format that included more information than the
federal statistics, but it didn't mesh with the way the federal
numbers were organized.

"Our data on this just did not match," Luttrell said. "And when we
submitted it, they didn't accept it."

The result of the holes in the data is that it's almost impossible to
draw meaningful comparisons about how often different police agencies
use deadly force, the experts said. And another flaw in the data, they
said, is that it includes only fatal shootings by police, and not
cases in which suspects are wounded but survive.

Critics: Agencies avoid exposure

Critics said law-enforcement agencies don't want accurate data
available because it will make it easy to spot agencies that use
deadly force most often.

"I suspect that's information that some law-enforcement agencies don't
want out there because they know it would show they're killing too
many people," said state Sen. Gary Siplin, D-Orlando, a critic of the
use of deadly force by some police agencies.

But the head of one of Central Florida's largest police unions said he
supports compiling more-accurate data, which he said would show that
area police use deadly force only when necessary and no more often
than officers elsewhere.

John Park, an Orange County sheriff's corporal and president of the
Central Florida Police Benevolent Association, a union that represents
about 900 officers, said accurate information about police use of
force would help ease criticism that officers use excessive force.

"I think accuracy is what creates an atmosphere of credibility in the
community," Park said.

Lessons from New York City

In New York, Fyfe credits an accurate accounting of the use of deadly
force with changes that led to fewer shootings.

When officials began compiling data on shootings by New York City
officers in 1971, they learned that police had shot 314 people that
year, killing 93, Fyfe said.

"It was a shock," Fyfe said. "No one could believe it."

The number was embarrassing to the department, Fyfe said, but also
proved the old adage that "what is measured is managed."

What followed, Fyfe said, were years of changes in training and
procedures, and policies that have reduced the officers' use of deadly
force without putting the officers at greater risk.

By 2001, the number had fallen to 26 shootings, 10 of them fatal.

"In recent years, we've been down around 10 or 11" fatal shootings by
officers a year, Fyfe said.

If accurate data were kept nationwide on police shootings, Fyfe said,
it would embarrass some police agencies.

"But you embarrass them in a way that compels them to change," he
said.