Vast DNA Bank Pits Policing Vs. Privacy

Opponents of the growing inclusion of people arrested note that a large proportion of charges (fully half for felony assaults) are eventually dismissed. Blood specimens are not destroyed automatically when charges are dropped, they note, and the procedures for getting them expunged are not simple.

Even more controversial are DNA dragnets, which snare many people for whom there is no evidence of guilt. Given questions about whether such sweeps can be truly voluntary -- "You know that whoever doesn't participate is going to become a 'person of interest,' " said Rose of the ACLU -- some think they violate the Fourth Amendment.

Civil liberties issues aside, the sweeps rarely pay off, according to a September 2004 study by Samuel Walker, a criminology professor at the University of Nebraska. Of the 18 U.S. DNA dragnets he documented since 1990, including one in which police tested 2,300 people, only one identified the offender. And that one was limited to 25 men known to have had access to the victim, who was attacked while incapacitated in a nursing home.

Dragnets, Walker concluded, "are highly unproductive" and "possibly unconstitutional."

Familial searches of the blood relatives of known offenders raise similar issues. The method can work: In a recent British case, police retrieved DNA from a brick that was thrown from an overpass and smashed through a windshield, killing the driver. A near-match of that DNA with someone in Britain's criminal database led police to investigate that offender's relatives, one of whom confessed when confronted with the evidence.

Not investigating such leads "would be like getting a partial license plate number on a getaway car and saying, 'Well, you didn't get the whole plate so we're not going to investigate the crime,' " said Frederick Bieber, a Harvard geneticist who studies familial profiling.

But such profiling stands to exacerbate already serious racial inequities in the U.S. criminal justice system, said Troy Duster, a sociologist at New York University.

"Incarceration rates are eight times higher for blacks than they are for whites," he said, so any technique that focuses on relatives of people in the FBI database will just expand that trend.

A Universal Database?

That's a concern that many in law enforcement raise, too -- as an argument in favor of creating a universal DNA database of all Americans. The system would make everyone a suspect of sorts in every crime, they acknowledge. But every criminal, regardless of race, would be equally likely to get caught.

Opponents cite a litany of potential problems, including the billions it would cost to profile so many people and the lack of lab capacity to handle the specimens.

Backlogs are already severe, they note. The National Institute of Justice estimated in 2003 that more than 350,000 DNA samples from rape and homicide cases were waiting to be processed nationwide. As of the end of last year, more than 250,000 samples were backlogged in California alone.

And delays can matter. In 2004, police in Indiana arrested a man after his DNA matched samples from dozens of rapes -- the last 13 of which were committed during the two years it took for the sample to get through the backlog.