The DC Snipers and the Limits of Law Enforcement Databases

On October 2nd at 5:20 PM, a .223-caliber rifle bullet shot through the window of a Michael’s craft store in Aspen Hill, Maryland.   Forty-five minutes later, the first victim in what would come to be known as the sniper killings fell dead, shot in the parking lot of a supermarket only two miles away.

It was an interesting time to live in Washington.   Shootings occurred nearly every day.   Local newspapers published advice on how to avoid becoming a casualty:   Don’t walk in a straight line, don’t shop near a Michael’s, don’t use gas stations next to highways.   Revenues at most of the gas stations in the area immediately dropped, although a few of them benefitted from the turn of events – one lucky owner with a gas station located under a church building saw his business more than double.

Many people began to duck and weave while walking to their cars.   Folks in our office building were particularly worried:   our high-rise is next to a mall and at the intersection of two major expressways.   That used to be a good thing, but now it was a liability.   Among the many false rumors floating around at the time, it was thought that the snipers operated from a white box truck.   One day a co-worker and I saw just such a truck in the parking lot as we walked back from lunch.   We probably went an extra quarter-mile just to avoid being in the “line of fire”.

A network of traffic cameras had recently been installed in D.C. to catch speeders and cars running red lights.   They were not very popular, particularly among those who got a ticket in the mail, and a number of civil-rights arguments were advanced to get them removed.   Now there was talk of using them to watch all traffic, in the hope of getting clues.   Spy planes were put up over the city.

The snipers were eventually caught, but not by the planes.   Like so many criminals before them, they were undone by their need to brag.   During a conversation with police they referred to a killing in Montgmery, Alabama.   The police started to re-examine evidence from a double shooting there.   In the store where the attack occurred, a fingerprint had been found on a magazine page.   This was flown to FBI headquarters and matched to Lee Malvo through an immigration database.

After the initial jubilation faded, I started to wonder about that fingerprint.   The murder in Alabama had been brutal, and worthy of swift investigation in its own right.   Claudine Parker, 52, was killed, and a co-worker, Kellie Adams, 24, was seriously wounded.   The Alabama police had been interested in the fingerprint since the day of the shooting, on September 21.   And yet, no match was made for over a month.   Why the long delay?

The fingerprint was sent out first to the Alabama State Police crime laboratory.   27 days elapsed before anyone there attempted to match it against the state database.   According to a spokesman, there are only four fingerprint examiners, which led to the backlog.   No match was found.

Matching against a state database – as slow as it may be – is actually considered faster than sending the print to the FBI.   After the local search came up blank, Alabama was ready to send the print national.   Then the snipers mentioned Montgomery, the FBI sent an agent down to pick up the print, and a match was made within 24 hours.   When a case is important enough, things move fast; but in general the system is shockingly slow.   Only 19 states are electronically tied in to the FBI’s Integrated Automatic Fingerprint Identification System (IAFIS).   The FBI claims that hooking up is relatively easy, and according to assistant director Michael Kirkpatrick can be done for “probably less than $5,000.”   The FBI also provides the Remote Fingerprint Editing Software free of charge to members of the IAFIS community.   The sniper incident has at least provoked more interest in the system, but given that it has been available for several years, some skepticism about fast and universal uptake seems reasonable.

The fingerprint delay highlighted a dramatic shortcoming in law enforcement technology.   Unfortunately, it was not the only one.   Fingerprints are at least reasonably complex things.   Names and license plates are not.   Here too a flawed system led to missed opportunities.   A warrant for John Muhammed’s arrest on a misdemeanor shoplifting charge had been filed in February in Tacoma, Washington.   It was never entered into NCIC, because misdemeanors are considered too minor for extradition – so local law enforcement had little incentive to update the database.   Police stopped Muhammed’s blue Caprice three times in the five days before the shootings started, but they found no reason to hold him.

An additional point of failure occurred with the ballistic fingerprinting.   The bullets used in the early crimes of John Muhammed and Lee Malvo were never entered into the national Integrated Ballistic Identification System (IBIS).   Doing so would have alerted authorities that a shooter was moving over state lines.   It might also have led to the dark blue car, which witnesses saw at one of the crime scenes.   Currently, the IBIS database is not perfect even if used properly.  One-shot nationwide searches are several months away.

The irony is that the technology needed to improve these flaws has been available for some time.   The commercial world has been able to use large databases very efficiently.   The government, which has more political restrictions and a different set of motivations, has not duplicated this success.   Events like September 11 and the sniper attacks may spur some improvements.   Until then, serial criminals are advised to operate on the move.   If the snipers had not stayed in the Washington area, they might not have been caught.

Home
Site Map