Skip to comments.What if Airport Screeners Checked Your Square Root?
Posted on 11/20/2010 4:27:54 PM PST by Diana in Wisconsin
(A computer scientist argues that racial profiling is less effective than random searches, but says simple math could offer an even better solution.)
What if airport screeners looked at would-be passengers and instead of assessing the color of their skin, asked: "What's the square root of your likelihood of being a terrorist?"
Such is the world imagined by an American computer scientist who argues that racial profiling to root out potential terrorists is actually less effective than random searches, but says some simple math could offer a better solution.
"When you have any profiling at all, it quickly becomes less effective than random sampling," said University of Texas professor William Press, whose paper appears Wednesday in the journal Significance, a publication of Britain's Royal Statistical Society.
Profiling does not work because "you end up screening the same innocent people over and over again, just because they happen to be in a profiled group," Press said.
Previous studies have shown that any apparent rise in success due to racial profiling is actually due to increased levels of law enforcement. More police focusing on one group will catch more criminals since fewer police and resources are focused on other groups.
"It is simply better to do uniform random sampling, which means everyone who shows up at the airport should have the same chance of being screened in the same way," said Press, who has written on the topic before for the Proceedings of the National Academy of Sciences.
But he has come up with an idea that just might be even better.
"It is this thing called square root sampling," he told AFP.
That way, screeners would approach a given group deemed to be, say 100 times more likely to be harmful, and then check them the square root of that number, or 10 times, more often.
"That actually would be better than uniform (random) sampling. The trouble is there is no good way to do that."
Press teaches university-level statistics and uses the example for his students, who do not argue with his mathematical formulas but do puzzle over practical ways to solve the problem in real life.
"One could imagine a system in which people's risk factors are evaluated and as you show up in airport you know, in some computerized automatic way the computer flashes either red or green and does this square root business which would be some form of optimal profiling," he said.
"But I don't know anyone who actually thinks you could make such a system work."
And when it comes to the latest controversy roiling U.S. airport travelers systems that can peer through clothing and show bodily details Press has just one hope in mind for any new screening technology.
"That it not slow down the lines," he said.
Treating every flier as though they have no history is acting stupidly.
Isn't that like dividing by zero???????????
Yep, part of the weirdness that is statistics is concerned with how certain one wants to be that the answer is the correct one. Stopping all the terrorists 99% of the time just isn’t good enough.
The solution is smimple; Don’t let muslims fly.
Better yet — Americans in general aren’t trying to kill us.
The American people make up the bulk of travelers within this country. It’s extremely silly to focus the majority of our security efforts on the very group that isn’t trying to do harm to our country.
Profile foreigners and those within our country with questionable backgrounds. Leave the rest of us alone. That really is the only sensible course of action on this matter.
Indeed. There's also the matter of how to weigh that history, but then we find ourselves in the neighborhood of the Israeli model.
America has stupidly become dependent on its technology, and now EVERY problem seeks a technological solution (something about if all you have is a hammer, every problem looks like a nail). Israel profiles and interviews. They never let the bad guys get on the plane and they can do it without looking at race or other stupid factors. This is a human problem, and no tech will ever solve it.
As if groping my junk wasn't bad enough, now they want to look-up my square root.
I get profiled all the time when I wear this.
It's very difficult to tell who is a terrorist, but it turns out to be trivially easy to figure out who is probably not a terrorist. If I look you up in a public-records database and a credit bureau, and I find out that you own a $300,000 house, you've had a Visa card with the same bank since 1983, you've had a succession of car loans going back 20 years, you graduated from high school in Rhode Island... I'm wasting time to even mess with you. A computer program could "score" such data in well under a second, resulting in 75% of the travelers being waved right on through.
The trick is to keep the algorithm a secret so that The Bad Guys can't recruit people who would be known to score well on "probably not a terrorist."
This lib idiot is a prof @ UT Austin. miles left of Berkely. Talking out is a$$ cause his mouth knows better.
Which turns out to be a mistake ~ but TSA and other federal agencies are tied to a belief in "random sampling" that assumes uniformity in the flow of passengers is the normal condition.
W. Edwards Deming once observed that over a long enough period of time ANY sampling scheme will tend toward randomness ~ so you don't need to say "Hey, let's hit every n'th item, or person". Instead, you can say "I feel like it's time to hit someone" and do it, and with enough customers you'd end up with something that's arguably "random".
I prefer to think of air carrier safety being a QUALITY CONTROL question ~ not at all a safety question. After all these are airplanes, they're going way high in the air, they are full of all sorts of volatile fuels, even their engines could blow up under NORMAL CONDITIONS.
They are not inherently safe, and bomb carrying passengers may well NOT increase anyone's level of risk enough to measure ~ fur shur 9/11 didn't cause all of us to give up commercial air travel.
As a quality control question we can look at the customers as something other than IDENTICAL ITEMS that we need to polish, buff and gussy up to make them fit right in the airplanes. We can, in fact, view them as nothing more than different items whose relationship to air travel is simply one of volume (affecting absolute capacity limits within the airframe) and weight (affecting absolute lifting capability of the aircraft).
Fur Shur, TSA is not measuring size and weight of the full passenger load for the purpose of detecting a FALSE TRAVELER ~ that is, someone carrying a bomb. They are instead selecting PLACE IN LINE, which on the face of it is a factor of absolutely no consequence to the quality of the coming flight ~ except under special circumstances.
For example, I was once standing toward the end of a line at the Frankfort airport waiting to get on a shuttle to London's Heathrow ~ and a bus to the downtown terminal. A steward comes up and asks me if I'd give up my place in line for this other passenger who needed to jump on that shuttle to link up with a further overseas connection. He assured me I could get on the next segment in 15 minutes.
So, I said "Yeah, sure, whatever".
Wouldn't you know it the first segment crashed and burned at Heathrow. My flight landed. You could look over at the wreck already pushed off the runway.
That's how "place in line" works ~ but rarely. TSA's current approach would have retroactively blamed the crash on my having changed my place in line for air service.
Weight and size can be estimated electronically for each and every passenger with in-motion equipment now widely available. Profiling by skin tone, shape of nose, shape of ears, hair characteristics, and skeletal proportions can be as easily determined by existing modern equipment then matched to weight and size tables to determine if "somebody is amiss" or in a physical category known to be highly correlated with the presence of explosives, bombs or bad attitudes.
Instead of selecting people on "place in line" shifting to a physical profiling system that evaluates weight, volume, reflectivity, and shape of external organs, we can probably reduce the working samples to a very low number of folks for whom we can reserve even BODY CAVITY SEARCHES. That's when we'll start discovering the modern bombs.
LOL! Very creative! Me likes!
Just have the first TSA agent have his “Bomb Sniffing Pot Bellied Pig” with him.
The second agent has his “Bomb Licking Dog” with him.
Neither is anywhere near as offensive as the “grope gauntlet” they have now.
If we simply took any individual actually carrying an explosive device and shot him or her on the spot at the airport and hung the body up for public viewing, we could probably get away with a less than 1% detection rate.
Alas, that would generate additional inefficiencies in the system that would make air travel about as inconvenient as it is now.
Check my earlier post where I advocate moving away from mind-numbingly uniform evaluations of place in line to actual evaluation of statistically correlatable physical charateristics to identify probable terrorists.
“Profiling” has never been “racial” profiling, or “fixed set of attributes” profiling for what’s worth. You take your samples, see what’s common among them and dynamically work with it. It’s just that race, for several reasons, comes up often as a red flag.
Your reward for doing the right thing. Or good luck.
Wait, .... I got to go to the kitchen, I'll be right back.
I’m always polite, and if you want my seat on the plane, go to it.