A first-ever Justice Department study of deaths in police custody raises more questions than it answers.

The study, “Arrest-Related Deaths in the United States, 2003-2005,” released Oct. 11 by the Bureau of Justice Statistics (BJS), covers people who died in the custody of state and local police. It is based on reports by state governments supplemented by federal data. Although state cooperation was required by congressional mandate as a condition for getting federal corrections funds, three states — Georgia, Montana and Maryland — did not participate. District of Columbia figures are included.

As this is the first time such a report has been compiled, there is no easy way to compare its statistics with those of previous periods. So it is not possible to tell from this whether deaths in police custody are increasing or decreasing.

The study deals with deaths from the moment police attempt to stop or arrest someone to the point that the person is placed in holding cells. Deaths include shootings on the street and also persons who turn up dead in the cell. Not included are deaths related to federal custody, or deaths in prison beyond the point that a person is booked.

According to the report, 2,002 people died in police custody from 2003 through 2005. The overwhelming majority were men. African Americans, 12 percent of the U.S. population, represented 31.9 percent of deaths in custody, while 20.2 percent were Latinos and 43.9 percent were white. In 54.7 percent of the cases, the cause of death is listed as “homicide” by police, meaning the police shot or otherwise killed the individual. Thirteen percent of deaths in custody resulted from alcohol or drug intoxication and 12 percent were suicides. Smaller numbers died from accidents, illnesses or other causes.

In about 55 percent of the cases, the police were arresting the person for a violent offense.

When broken down by race, the rates of “homicide” deaths were about the same as the rates for all causes. But the proportion of Blacks who died in custody from causes listed as “accidental,” “illness” and “other” was sharply higher.

In 80 percent of the homicides, the police claimed to have been threatened by the individual. The report does not say whether such claims were sustained by police review boards or the courts. Thus one cannot tell from this report whether the police officers killed the individual without provocation, or in fact were defending their own lives.

The number of people killed by police tasers jumped from 3 in 2003 to 24 in 2005. Recent highly publicized cases have demonstrated that police may be using tasers, not as a substitute for deadly force as designed, but in circumstances where force would formerly not have been used at all.

The study does not attempt to explain why there is so much violence between police and, in particular, lower-income and minority working-class communities. Such an analysis is not likely to be forthcoming from the Bush Justice Department.

A review of cases in which police killed civilians over the last decade reveals patterns not addressed by the DOJ study. These include:

• police killings of mentally ill individuals,

• police killings of individuals who made a sudden movement which police suspected was to reach for a weapon, but which turned out to be an innocent action such as reaching for a wallet or cell phone,

• killings carried out by police in plain clothes who did not identify themselves, and

• killings following persistent police harassment of an individual.