Despite widespread public interest and a provision in the 1994 Crime Control Act requiring the Attorney General to collect the data and publish an annual report on them, statistics on police shootings and use of nondeadly force continue to be piecemeal products of spotty collection, and are dependent on the cooperation of local police departments. No comprehensive accounting for all of the nation's 17,000 police department exists.

The lack of good data is a national scandal.

Geoffrey Alpert University of South Carolina

The major reasons for the vacuum, the experts agree, are twofold. The lack of information on police shootings is attributable to the failure of police departments in many cities to keep and report accurate figures that distinguish between what the police see as "justifiable" shootings � those in which the suspect posed a serious threat � and incidents where an officer may have unlawfully fired at an unarmed civilian.

The International Chiefs of Police, a police organization, tried in the 1980's to collect such information, but "the figures were very embarrassing to a lot of police departments," said James Fyfe, a professor of criminal justice at Temple University who is a former New York City police lieutenant. The results, he said, varied wildly. New Orleans had 10 times as many shootings per 100 officers as Newark. Long Beach had twice as many as neighboring Los Angeles, which in turn had three times more than New York.

Some cities did not provide data at all, Professor Fyfe said, but the results, such as they were, showed that "the rates of deadly force are all over the lot," meaning that some cities appear to be much better and some much worse at managing their police forces.

As for the lack of figures on the use of nondeadly force, the situation is even murkier because there are no uniform definitions of force and no standard reporting requirements from one police department to another.

The Bureau of Justice Statistics, the statistical arm of the Justice Department, has tried to fill in some of the blanks on police behavior, issuing a number of surveys and reports on the topic. Most recently, the bureau quietly released a report, "Policing and Homicide, 1976- 1998." But the report itself underscores the continued problems in knowing what is really happening.

On its cover, for example, the report refers to all the victims of police shootings as "felons justifiably killed by police," a categorization that Samuel Walker, a professor of criminal justice at the University of Nebraska at Omaha, termed "deeply offensive and legally incorrect." In fact, a Justice Department official said the bureau was so embarrassed by the term, and the lack of distinction between justifiable police shootings and murders, that it did not send out its usual promotional material announcing the report.

BUT, the official said, the bureau was trapped because it depends on local police departments to report their figures on police shootings to the Federal Bureau of Investigation, and "felons" is the term that police departments insist on using when they do so.

Making matters worse, some police departments fail to report their shootings at all, and for some years, figures from entire states are missing. Although the 1994 crime act ordered the Justice Department to collect such data, there is no law requiring local police departments to provide it, Janet Reno, the former attorney general, acknowledged in a 1999 speech.

Based on the data available, this most recent report suggests that the number of "justifiable" police killings has not increased since 1976, averaging 373 a year, despite a growth in both the population and the number of police officers. And while the rate at which blacks are killed by the police still far surpasses the rate at which whites are shot and killed, it has dropped to four times the white rate in 1998 compared to eight times in 1976.

But the report also acknowledges its own limitations. "One statistic that is impossible to obtain" from the Justice Department's database, it said, "or from any other currently existing database, is the number of murders by police," because in reporting their shooting figures, the police don't distinguish between justified and unjustified killings. The report also fails to break down the number of police shootings by city, unlike other Justice Department reports on crime, making it impossible to compare police performance.

Viewed on a case-by-case basis, things do appear to be improving in some cities. Minneapolis, Boston, Miami-Dade County, Tampa, Phoenix, Seattle and Portland, Ore., are among cities that have implemented improved systems to report all use of force and shootings by officers, better systems for keeping track of civilian complaints and a program police departments are calling an early warning system.

"The early warning systems collect data on citizen complaints and look for patterns by computer to flag officers with problems," Professor Walker said, adding that this "creates more accountability."

Still, the efforts of individual cities often get lost amid the drumbeat and drama of each new instance in which a police officer shoots a civilian, partly because there is no national store of numbers to provide a sense of proportion. And until more comprehensive data are collected, those looking for the national trends behind each local incident are likely to be frustrated.

The lack of good data "is a national scandal," said Geoffrey Alpert, a professor of criminology and criminal justice at the University of South Carolina and a leading authority on police use of force. "It's a scandal in the sense that these are public servants who work for us and are paid to protect us."