The New York Times hosts a debate over profiling:
The first contributor says:
The Obama administration has announced that it will subject citizens of 14 countries, including Pakistan, Saudi Arabia and Yemen, to intensive screening when flying to the United States (the rule will also apply to those passing through those countries). This means treating people differently depending on where they come from or what passports they hold.
Does it make sense to concentrate security efforts on more limited populations — through profiling, behavioral or otherwise? Is profiling effective, compared to other strategies?
Bruce Schneier is a security technologist and author of several books on computer security, including “Beyond Fear: Thinking Sensibly About Security in an Uncertain World.”
Terrorists can figure out how to beat any profiling system.
There are two kinds of profiling. There’s behavioral profiling based on how someone acts, and there’s automatic profiling based on name, nationality, method of ticket purchase, and so on. The first one can be effective, but is very hard to do right. The second one makes us all less safe. The problem with automatic profiling is that it doesn’t work.
Terrorists don’t fit a profile and cannot be plucked out of crowds by computers. They’re European, Asian, African, Hispanic, and Middle Eastern, male and female, young and old. Umar Farouk Abdul Mutallab was Nigerian. Richard Reid, the shoe bomber, was British with a Jamaican father. Germaine Lindsay, one of the 7/7 London bombers, was Afro-Caribbean. Dirty bomb suspect Jose Padilla was Hispanic-American. The 2002 Bali terrorists were Indonesian. Timothy McVeigh was a white American. So was the Unabomber. The Chechen terrorists who blew up two Russian planes in 2004 were female. Palestinian terrorists routinely recruit “clean” suicide bombers, and have used unsuspecting Westerners as bomb carriers.
In reality, as sportswriter Damon Runyon said, "The race is not always to the swift, nor the battle to the strong, but that's the way to bet."
Without an accurate profile, the system can be statistically demonstrated to be no more effective than random screening.
Actually, the link says the opposite, as I'll show below.
And, even worse, profiling creates two paths through security: one with less scrutiny and one with more. And once you do that, you invite the terrorists to take the path with less scrutiny. That is, a terrorist group can safely probe any profiling system and figure out how to beat the profile. And once they do, they’re going to get through airport security with the minimum level of screening every time.
Sure, as long as Al-Qaeda can recruit Mexican grandmothers to be suicide bombers as readily as it can recruit young men with Muslim names.
As counterintuitive as it may seem, we’re all more secure when we randomly select people for secondary screening — even if it means occasionally screening wheelchair-bound grandmothers and innocent looking children. And, as an added bonus, it doesn’t needlessly anger the ethnic groups we need on our side if we’re going to be more secure against terrorism.
A recurrent theme of mine is how the demand for denial of average IQ differences spills into seemingly unrelated issues, like airline security, causing widespread intellectual stultification. The modern liberal mind thinks in black-and-white Manichean terms, rendering it unarmed for dealing with a probabilistic universe.
It's hard to deal with liberal arguments because they tend to be so Gladwellian in their mental rigidity. Here we are, more than eight years after 9/11, and this "expert" picked by the NYT for his wisdom can't imagine any profiling system smarter than he is.
Schneier seems to be assuming that profiling means that 100% of attention would be devoted to people in category X and 0% to people in category Y. The weird thing is, that's common among progressives. They really just don't get it. The conventional wisdom is a form of unilateral cognitive disarmament.
He's like a pitching coach who tells a baseball pitcher, "Your fastball is above average, your slider average, and your change-up below average, but if you only throw your fastball, they'll expect it, so you should choose your pitches randomly, throwing one-third of each."
Obviously, when stated in those terms, it's easy to see the fallacy: there are superior methodologies in-between all fastballs and total randomness. If your fastball is relatively more effective than your other pitches, you want to throw relatively more fastballs. But you still want to "mix 'em up," as every pitching coach from Babe Ruth League onward as told pitchers.
Why can't Americans be as smart about public policy as they are about sports?
Thus, if you read the article Schneier links to behind his phrase "statistically demonstrated," you'll find it's merely a debunking of a braindead "100% fastballs" profiling method:
Press then examines the effect of what he terms a strong profiling strategy, one in which a limited set of screening resources is deployed solely based the risk probabilities identified through profiling. It turns out that this also works poorly as the population size goes up. "The reason that this strong profiling strategy is inefficient," Press writes, "is that, on average, it keeps retesting the same innocent individuals who happen to have large pj [risk profile match] values."
The very next paragraph of the article linked to by Schneier explains that non-braindead profiling is the best method:
According to Press, the solution is something that's widely recognized by the statistics community: identify individuals for robust screening based on the square root of their risk value. That gives the profile some weight, but distributes the screening much more broadly through the population, and uses limited resources more effectively. It's so widely used in mathematical circles that Press concludes his paper by writing, "It seems peculiar that the method is not better known."