Thursday, July 28, 2005

Avoiding racial profiling at all costs?

This editorial is from Tuesday's New York Times, but the attitude it expresses could have been found in almost any liberal newspaper.

Subsequent to the London subway bombings, New York has had to decide how to increase security on public transit, and so there have been random searches of backpacks and the like. The Times editorial explains:

The police officers must be careful not to give the impression that every rider who looks Arab or South Asian is automatically a subject of suspicion. They will naturally choose to search the bags of those people who appear suspicious, like those wearing bulky clothes in warm weather. But those who are selected simply because they are carrying packages should be chosen in a way that does not raise fears of racial profiling - by, for example, searching every 5th or 12th person, with the exact sequence chosen at random.

This need to avoid racial profiling at all costs, or even the appearance of racial profiling or fears of racial profiling, is puzzling. Yes, of course, it's true that not all terrorists and suicide bombers fit the profile of a young adult Arab Moslem male--there's Richard Reid, and the recent Somalian of the second group of London subway bombings. But there is no doubt that most do indeed fit that profile.

And of course no one is suggesting that all who do fit that profile are bombers; the vast majority are not. And yet it is folly, if not insanity, to deny that young adult Arab Moslem males are far more likely than anyone else to be suicide bombers, and that, in a more logical world, they should be subject to greater scrutiny.

In regular police work, usually a crime has already been committed when the police are doing a search. If there is also a witness who has given a description of the subject--let's say, for example, a young white male of medium height wearing a black hooded sweatshirt and having a huge scar on his forehead--the police are not going to search every fifth person driving through a roadblock. No, they are going to look for young white males fitting the description, or else we would consider our tax money poorly spent and our police hopelessly incompetent. But such a search is not considered racial profiling, because it is based on an actual description of an actual suspect by an actual witness to an actual crime.

Racial profiling is usually done in the context of stopping cars based on the race of the driver and searching them for drugs. Even if members of certain races are statistically somewhat more likely to be dealing drugs than others, it's not as though drug dealing and race are inextricably linked. There are no races or subsets of races, for example, that have as part of their ideology that members must deal drugs; the racial factors in drug dealing are statistical--social and perhaps economic, but hardly ideological in nature. And the need to prevent drug dealing cannot even remotely be compared to the need to prevent suicide bombings.

Which brings us to the other operative word, prevent. With backpack searches in mass transit, we are not looking to solve a crime by finding a criminal ex post facto. We are looking to prevent a crime that has the potential for mass murder. If this doesn't justify a bit of racial profiling that acknowledges what type of person is most likely to be a perpetrator, what would?

It's not that random searches have no place in the prevention of suicide bombings--they do. As terrorists turn more and more to those who don't fit the traditional profile (Somalis, as in London; women, as in Palestine/Israel), the random search becomes more and more valuable, especially as a possible deterrent. It's unlikely that the police would happen upon an actual bomber by chance in a random search, but the knowledge that such searches are occurring might act to keep any non-Arab bombers from feeling safe, and could perhaps discourage some of them from even trying an attack. So I would not be one to argue against random searches--they are one more weapon in an arsenal that needs to be varied.

But it's not an either/or proposition. There's no reason not to do the most reasonable thing: random searches plus racial (or, in this case, religious and even gender) profiling. It's the only thing that makes any sense.

But political correctness often trumps common sense nowadays. The NY Times doesn't seem to feel, in the quoted editorial, that there is even a need to explain why being careful to avoid any hint of racial profiling is more important than the need to protect people who ride public transit against a terrorist attack. The authors of the editiorial seem to think the reasons are self-evident. I, for one, would like to hear their argument, because those reasons are certainly not evident to me.

And yes, I understand the need to protect our civil rights. But we are not talking about throwing young Moslem males of Arab origin into detention camps like those that housed Japanese Americans during WWII. We are talking about subjecting them to greater scrutiny in backpack searches; not exactly the most terrible violation of a person's rights. I like to think that, if I were a young Moslem Arab male in the US who was not a terrorist, I would understand and freely cooperate with such a need on the part of law enforcement.


At 10:59 AM, July 28, 2005, Blogger Baron Bodissey said...

Neo, I'm with you -- the unexamined premise about "profiling" certainly bears examining.

But I can see why NY would want to follow the PC procedures -- I just read today that they're already facing 20 lawsuits from people claiming that they were "racially profiled".

Until judges start throwing these things out the moment they're filed, municipal authorities will remain hobbled.

At 11:02 AM, July 28, 2005, Anonymous Anonymous said...

Oh, Neo, as usual, you make sooo much sense. Frankly, whenever I talk to people in my circle of acquaintance, this is exactly our attitude and point of view. It is reasonable, logical, makes sense if a person has self-preservation instincts. Ironically, people like myself are not nearly as much at risk since our lives do not involve mass transportation on a daily basis. I just don't get the mindset that results in that kind of editorial position.

At 11:18 AM, July 28, 2005, Blogger Alex said...

Only one point didn't make sense to me: this bit about ideology. Why should this matter when you make the decision whether or not to profile Either way, profiling is exploiting the correlation between likelihood of crime (which is unobservable) and some observable physical characteristic. It shouldn't matter whether that correlation is caused by ideology or by some other factor. The far more important thing is the strength of the correlation (i.e. how good a predictor is the observable characteristic).

At 11:46 AM, July 28, 2005, Blogger goesh said...

The underlying premise is the notion that if rules are bent or even broken, then they can never be fixed, that we might somehow become a police state because we profiled people coming from regions that produce islamic terrorists. Young ME males on visas, sounds famaliar somehow....

At 2:10 PM, July 28, 2005, Anonymous Anonymous said...

Neo - we like your site. If you have a moment, feel free to check out our site:

At 3:17 PM, July 28, 2005, Anonymous Anonymous said...

I know it may seem strange but try to look at it practically from the perspective of the Police, the City and the ME/SA/"non-white" community.

How are the police going to do random searches *plus* racial profiling as you suggest? How would you outline the policy?

I'm not trying to be argumentative, I'm just really curious as to how you think it should work, because I can't think of any good way to do it.

Is the random sample pool going to be drawn from the general population or just from the ME/SA population? Should they stop every 10th "white" person but EVERY ME/SA male? Or just stop every 7th, let's be blunt, "brown" person? Or maybe have a "white" entrance and a separate "brown" entrance with metal detectors and dogs?

And sure, we could have a whole line of cooperative middle eastern men happily lining up to have their bags searched, but then wouldn't that just turn into a colossal waste of resources and manpower? How likely are any of them to have a bomb in the first place?

Then there is the practical effect on the "brown" population. You're from NY. If you are late for work one hot summer morning and get down in that steaming hot subway tunnel and you hear the train pulling in and a cop stops you and puts you in a line of brown people to search your bag just because you have brown skin, while all the white people (including the Hasidim in their heavy coats!) get to rush on through and catch their nice air-conditioned train. "Not exactly the most terrible violation of a person's rights," perhaps, but it could get a bit disturbing to you after the third or fourth time, no?

I have to acknowledge that on some level though these random searches make me feel safer, even though I know in my head I'm really not safer at all: The policy now is that anyone can refuse to be searched and is allowed to leave the station, no questions asked. So a determined bomber can just walk a few blocks to the next station where they aren't searching, or even catch a bus.

So what should we do about that, start making enforced searches in the subways? But then why stop at the subway? After all, a middle eastern male with a backpack anywhere in the city is a potential mass murderer, right? After all, we're just trying to prevent a disaster here.

Maybe there is a better way to do it, and I just can't think of it.

At 3:45 PM, July 28, 2005, Blogger Alex said...

I see no conflict in simultaneously conducting random searches and doing profiling. It could work like this:

- Search every 7th (or whatever) person regardless of who they are.

- Police have the power to search a Middle Eastern-looking guy in a bulky jacket, or whomever they find particularly suspicious, even if that person does not happen to be 7th in line.

Alternatively, could do a subtler blend of these strategies by using a random selector, like they have in many international customs areas. In this case, as the subject passes the selector would flash green with, say, a probability of 6/7 and red with a probability of 1/7. Reds get searched, but there is no fixed pattern as to when they appear. Then under the desk you have an override button that ensures that, if pressed by a guard, a given person will flash red. This saves a bit of face for the person, which is nice since in 99.99% of the cases that person will be innocent -- it won't be obvious to everyone that they were singled out, since it might have been random. Of course they were, but it's this sort of thing that might make the plan a little more palatable.

At 3:48 PM, July 28, 2005, Blogger Alex said...

One more advantage to the random method I forgot to mention... So long as there's an identifiable pattern in who gets searched (every seventh person or even something more complex) then people can figure out that pattern and try to position themselves in line to avoid search. A random system makes that impossible.

At 4:09 PM, July 28, 2005, Anonymous Anonymous said...

UB Food for thought, thank you.
One way to help the police would be to allow them some flexibility.
Like the examples Alex brought up, I firmly believe that we must do more. I guess I am just so apprehensive at the thought that this will happen again if we do not take draconian steps now. We spent so much time, money and energy to make flying safer but this only constitutes a small fraction of the public. People that use mass transit deserve as much protection as the air travelers.

At 4:41 PM, July 28, 2005, Anonymous Anonymous said...

Hi Alex, your first scenario is essentially what is happening now. Note that the editorial states:

[The Police] will naturally choose to search the bags of those people who appear suspicious, like those wearing bulky clothes in warm weather.

The key being I guess that they have the power to search anybody in a big coat in the summer, not just the middle-eastern types.

As for implementing an airport-type single-file (or even multiple-file) redlight/greenlight checkpoint in the subways, I dunno, have you ever seen New Yorkers enter a subway station at rush hour? I guess the turnstiles could be rigged up to do something like that, but in a place like Herald Square, where you maybe have a row of 10 turnstiles, and zillions of people zooming through in both directions, I just don't see it happening...

The thing is, unless these searches are done daily at *every* stop (there are 468 of them), there is really no practical net effect. And even then, there are over 4.5 million riders on an average weekday. What percentage would have to be searched to make a statistical dent in the likelihood of an attack?

At 5:02 PM, July 28, 2005, Anonymous Anonymous said...

UB: Maybe there is a better way to do it, and I just can't think of it.

The "better way" is simply to give police the mandate to use their own judgement and common sense. In that way, they would continue to conduct more or less random searches, but would pay special attention to people who fit the visual profile of the vast majority of the suicide bombers: male, young, swarthy complexion, close-cropped hair. If, in addition, they're carrying packs of any sort, are alone, and are acting "oddly", they should get immediate and close attention. The point is that such "profiling" isn't complicated or difficult at all -- it's obvious, and something that everyone does in any case. What's created the problem for liberal sensitivities is that it looks like that old irrational prejudice, rascism -- even though in this case it has a perfectly rational basis in fact.

Now, I'll agree that this would generate some mild frustration for the vast majority of people who fit this profile but who, of course, are not suicide bombers. But I would also suggest that that same "vast majority" should understand very well the seriousness of the situation that has given rise to such policies, and should take out their frustration on the minority amongst them that has caused it.

At 5:10 PM, July 28, 2005, Blogger Alex said...

The Unknown Blogger makes some good points...

No matter what system you set up, the sheer volume of riders and number of points of entry makes the proposition of any type of search extremely expensive and time-consuming. Still, there may be fixes.

I remember reading a while back about those machines that were the equivalent of x-ray specs. They bounced some low-energy beam off people and could see through their clothes, revealing objects, like for instance bombs, beneath. (I remember people got in a huff about these since they could be seen nearly naked.) Anyway, it's possible that machines like these could be rigged to scan people entering. The price of the machines themselves might be high, but if they were coupled with good software that could identify suspicious objects then the whole system might operate with relatively few guards and with little to no time delay. (Salaries and wasted time of course being two of the most costly parts of normal search.)

Another, lower-tech alternative is, sadly, to modify the ratio of people searched randomly to people searched by profiling. Though random searches might deter the very odd case, I think (maybe?) most people agree that terrorists are more likely to be deterred (or caught) through the profiling end of things. If attacks continue, there may be more political will to jettison the random searches, which more serve to make us feel fair than to keep us safe, in favor of a more heavily profiling-based system. But the closer we get to pure profiling, the more people (including myself) who would become uncomfortable with it.

(Why? It is, after all, a difference in degree, not in kind. Yet somehow I just feel it's wrong. And it certainly would get more people angry.)

At 5:21 PM, July 28, 2005, Anonymous Anonymous said...

Physical searches on a mass scale are impractical. The only possible way would be to whittle the numbers down to a managable few by the use of technology. Sensitive screening devices that not only scan visually but sniff for suspect chemical traces. Obviously, these devices need to scan fast and be reliable enough not to produce too many false positives. I suspect that these devices are available but little interest has been shown by Homeland Security due to the enormity of the problem. I can predict with some certainty that if, God forbid, a subway attack happens here in the US, there will be a rush to set these devices up pronto. Alas, there need to be victims first then action.

At 2:51 AM, July 29, 2005, Blogger camojack said...

It's a fairly simple formula:

OK, maybe not quite that simple, but it's true a lot of the time...

At 4:05 AM, July 29, 2005, Anonymous Anonymous said...

" I suspect that these devices are available but little interest has been shown by Homeland Security due to the enormity of the problem."

I don't think this is right. It's not my field of expertise, though a few years back when I worked with people whose field it was, they felt it would be far into the future when it happened and would be VERY expensive.

The problem is that what appears to be possible, and what is possible, may be two totally different things.

Lets take my field of expertese for example: computer science. Lets say we have a program, any will do. Lets say we run it - will it ever finish? That is, given a specific input can I determine if it will run forever or quit? Obviously one can - it should be simple, no? Since this is true, shouldn't deadlocks or livelocks never happen (and they do all the time)?

As it turns out, not only is *not simple*, it is *impossible*. You can not tell if any program out there will run forever or not. You can tell if a *specific* program will - but not the general case. It's called "The Halting Problem" (do a search - lots of info out there) and is pretty famous in Comp Sci. Essentially you can set up a program where if it says it stops it runs forever, if it runs forever it actually stops. Since that can not be - there is a case where it fails. There are many problems that fall into this class, though many are hard to see. The last place I worked was a CS research institution and we actually got a paper out of proving that one of our ideas met the halting problem.

Maybe this case doesn't, but given the amount of money involved, my knowledge of how people are fighting over it (if you think for a second you will come to a pretty close idea - govt isn't that different from industry except that a publication is a "success"), if it was *that* easy it would be out there now. Heck, our project was eventually billed as a "Homeland Security Project" and that was a *real* stretch of the imagination.

At 6:35 AM, July 29, 2005, Anonymous Anonymous said...

Fact: The great majority of bombings have been done by people of Arab descent. If a person looks suspicious check them out !! Enough of this legal correctness!

At 7:52 AM, July 29, 2005, Anonymous Anonymous said...

You know, I avoid political involvement as if it were cursed.

Because it is.

I'm just about the same age as the folks that frequent this blog. Used to be a leftist, too. Used to be involved.

The "I've got to change the world" mentality of the Boomers needs to be discarded. If for no other reason than that it is a pompous bore.

Obviously, I read sites like this. I do it to pass the time and to marvel at the self-importance of everybody involved.

Time for the era of social and political activism to come to an end. Everything has a shelf life. It's been 60 years of obsessing about this stuff.

We elected some people. Let them deal with it. A few years from now, if it ain't going so well, elect some other people. Drink a beer have a glass of wine.

At 8:06 AM, July 29, 2005, Blogger Alex said...

We elected some people. Let them deal with it. A few years from now, if it ain't going so well, elect some other people.[emph added]

Ah, but that's the hard part, isn't it? How do we tell if it ain't going so well. Say there's been a major terrorist attack. Is that evidence of a failure (they didn't prevent it!) or a victory (if they hadn't worked so hard there would have been many more!)? I believe that to make an intelligent evaluation of our leaders, to be able to differentiate a good job from a bad job, we need to at the very least be involved in this way -- reading the best information available and debating ideas.

Notice that I haven't said "hey strpcy and camojack, let's go to NYC and have a direct action where we install scanning devices ourselves!" Or even "let's picket grand central station until they put them in!" No, it's more just a matter of thinking a little about what might be the right course of action so that one can have an intelligent opinion about it come election time. And yeah, I'm self-important enough to feel I should have an intelligent opinion.

At 9:46 AM, July 29, 2005, Anonymous Anonymous said...

"Time for the era of social and political activism to come to an end. Everything has a shelf life. It's been 60 years of obsessing about this stuff."
What is that all about? Stephen, do you understand/comprehend the words that you are ascribing to yourself. You want us to stop exchanging ideas which may improve our lives because it bores you? Talk about being POMPOUS. Have another beer.


Post a Comment

<< Home

Powered by Blogger