The Reality Of Statistical Profiling
Wednesday - August 11, 2005
About 10 years ago, New Jersey State Troopers were brought to task about the criteria they were using to stop motorists on the New Jersey turnpike. The turnpike had become an acknowledged thoroughfare for drug traffickers and, over time, the troopers developed a profile for high probability vehicles: the make, type, year and condition of the vehicle, speed, time of day/night, origin of license plates, and any erratic driving patterns. This profile, when used, maximized the troopers’ efficiency in stopping the right vehicles. Once the vehicle was stopped, then another set of criteria could be pursued: driver behavior, IDs of driver and passengers, relationship of passengers to the driver and to one another, inconsistent stories about their destination, reason for trip, scratch marks around usual drug hiding places, etc. The profile did not include race of the driver. Realistically, how could it be used to stop a vehicle at night anyway?
This profiling system netted more drug dealers and mules (transporters of drugs) than would have otherwise been possible — a plus, right? Wrong! The problem was that statistically the system was netting a higher percentage of minority criminals than their percentage of turnpike drivers, therefore the troopers “must be” using race as a criterion for stopping vehicles. Suddenly, “racial profiling” was now on everybody’s radar screen. And under the mantra of political correctness, trooper heads began to roll, and the turnpike drug traffickers got a green light.
A few years later, then New York Mayor Giulliani, in a Harlem community meeting, was inundated with irate complaints about the prevalence of drug dealers on the streets. The profile was known: certain popular street corners, cars stopping at curbside, times of day or night, dress and demeanor of the dealers, etc.
The mayor increased the police presence and the targeted crackdown almost eliminated the trafficking. Good, right? Wrong! Turned out, of course, all the Harlem drug criminals arrested were of the same minority race. Some of the same irate Harlem citizens now charged, “Had to be racial profiling!” So the trafficking resumed to the detriment of all of Harlem’s citizens.
What had been charged as racial profiling in these two cases was really nothing more than statistical profiling. By building a bank of statistical data, the New Jersey troopers were able to narrow down the characteristics of most likely drug vehicles, and then with another set of statistical characteristics determine if a vehicle search was appropriate/legal. The New York police in Harlem, by using statistics to narrow down the characteristics of a street drug deal, were able to effectively crack down on those deals. The race of the guilty emerged only after the objective statistical data was applied.
Today we continue to face the challenges of drug trafficking, but now the mother of all challenges has emerged: stopping mass murder. With the single exception of Oklahoma City bomber Timothy McVeigh — an aberration in the context of modern-day terrorism — every one of the nearly two dozen terrorist attacks on Americans in the past 25 years has been perpetrated by Middle Eastern Arabs of the Muslim faith. Our statistical “data bank” puts that at 100 percent, a profoundly significant indicator of the source of future terrorist attacks on Americans. More recently, scores of innocents — not just Americans but from many nations, Muslim included — have been murdered. It turns out that all but Arab Muslim jihadists are vulnerable.
With the expansion of terrorism into our most mundane transportation systems — commuter trains and buses — and with the difficulty in identifying suicide bombers before the explosion, “random” searches at turnstiles and bus stops, to be effective, will paralyze those systems. I’m suggesting the time may be near for statistical profiling. Profiling based upon a combination of statistical factors is not subjectively or emotionally biased, and not racist. It’s by the numbers. Another attack on the level of 9-11-01 will make it inevitable.
With statistical profiling it will turn out, of course, that a certain minority will be more inconvenienced — even though for their own safety as well. But if the inconvenience sparks “outrage” anyway, perhaps instead of running their outrage to the ACLU, they could run their outrage back to their own ethnic community and demand intolerance and condemnation of those who skew the statistics against them.
E-mail this story | Print this page | Comments (0) | Archive | RSS Comments (0) |
Most Recent Comment(s):