Suspected failed car bombing in Times Square

[Master of Ossus]

There's always the recent fubar in Arizona with perfectly lawful individuals being harassed because they look Mexican, but I thought that was too obvious an example on why racial profiling is a bad idea.

That is a totally different system of racial profiling since (at least my understanding of it) is that it's an ad hoc procedure evaluated by officers on the ground as opposed to a top-down model that systematicaly identifies combinations of characteristics that are associated with terrorism.

Moreover, the problem of illegal immigration is nowhere near as serious as the problem of terrorism (e.g., no one's life is at stake if some guys sneak across the border and forge some green cards), and does not warrant the same sorts of measures that airline security may require.

[General Zod]

There's always the recent fubar in Arizona with perfectly lawful individuals being harassed because they look Mexican, but I thought that was too obvious an example on why racial profiling is a bad idea.

That is a totally different system of racial profiling since (at least my understanding of it) is that it's an ad hoc procedure evaluated by officers on the ground as opposed to a top-down model that systematicaly identifies combinations of characteristics that are associated with terrorism.

Moreover, the problem of illegal immigration is nowhere near as serious as the problem of terrorism (e.g., no one's life is at stake if some guys sneak across the border and forge some green cards), and does not warrant the same sorts of measures that airline security may require.

The point was that it's easily abused by racist shitheads. If they wanted to make someone's life hell, anything tied to airport security (like how they pick out who to hassle) is most likely going to be classified and there isn't much anyone can do to file a complaint against it. In any case, this is getting somewhat off-track. Profiling someone at the airport isn't going to really help much in preventing someone from blowing shit up on the ground if they get all their supplies locally, unless you want to include having the FBI trail them as well. (In which case you should probably already have some pretty solid evidence for arresting them).

[Master of Ossus]

The point was that it's easily abused by racist shitheads. If they wanted to make someone's life hell, anything tied to airport security (like how they pick out who to hassle) is most likely going to be classified and there isn't much anyone can do to file a complaint against it.

How are the "racist shitheads" going to achieve this if they don't have input into whom they search, or if their input could be evaluated later for patterns of abuse? Moreover, if they did try something like this, presumably the system would have a record of which passengers it selected to be searched (since that would obviously be fed back into it for future screenings), so it would be very easy to identify problems like the one you seem so preoccupied with.

[General Zod]

The point was that it's easily abused by racist shitheads. If they wanted to make someone's life hell, anything tied to airport security (like how they pick out who to hassle) is most likely going to be classified and there isn't much anyone can do to file a complaint against it.

How are the "racist shitheads" going to achieve this if they don't have input into whom they search, or if their input could be evaluated later for patterns of abuse? Moreover, if they did try something like this, presumably the system would have a record of which passengers it selected to be searched (since that would obviously be fed back into it for future screenings), so it would be very easy to identify problems like the one you seem so preoccupied with.

So let's assume your system somehow works perfectly. How will this prevent someone from purchasing all his supplies locally and blowing shit up? Once he's gone through the airport and shown that he's not carrying any bombs, he can just go to some hardware stores, pick up a few propane tanks and rig up a halfway competent detonator. I don't see how harassing people at an airport even more than usual is supposed to prevent this.

[Master of Ossus]

So let's assume your system somehow works perfectly. How will this prevent someone from purchasing all his supplies locally and blowing shit up? Once he's gone through the airport and shown that he's not carrying any bombs, he can just go to some hardware stores, pick up a few propane tanks and rig up a halfway competent detonator.

Uh... the system doesn't stop screening people after they pass one inspection and take one flight. It simply scans them at a different rate when they take many flights. Indeed, no group is subjected to zero searches. There is always a possibility that you will be searched, no matter what the information on the system says about you. You are simply searched at a different rate depending on that data.

I don't see how harassing people at an airport even more than usual is supposed to prevent this.

Okay, I'm growing tired of these inane questions. We've been through... what... 3 pages and you have made absolutely no effort to understand the basic subject matter of the discussion (to the point where you quoted a lengthy article describing precisely the sort of system that I've been talking about, indicated positive treatment of it, and somehow believed that it supported your argument against profiling!). The system I'm proposing harasses people less than the current system. It uses fewer resources and (probably MANY) fewer searches to generate comparable or superior results. The distribution of those resources is different than the current system because it dynamically allocates those resources in a more efficient manner by taking into consideration factors that you object to using, but that could easily make the system significantly more efficient. That is the difference.

I understand that there's a visceral reaction against using things like race and religion as even a single of many factors in government selection criteria, but surely you recognize that there must be some sort of balancing act between safety and our concept of racial and religious equality. As I've said before in this thread, I think reasonable people can reach different conclusions, but I've been very disappointed by the failure of profiling opponents to engage in a fair evaluation of the costs and benefits of the system, and this thread is no exception. Frankly, with the exception of the possibility you posited that groups who would be more carefully scrutinized by a statistical system are a few disproportionate inconveniences away from becoming suicide bombers, every single one of your criticisms has completely missed the point. I can completely understand why people are uncomfortable with such an idea, but being uncomfortable with it is no reason not to try to understand its pros and cons.

[General Zod]

Okay, I'm growing tired of these inane questions. We've been through... what... 3 pages and you have made absolutely no effort to understand the basic subject matter of the discussion (to the point where you quoted a lengthy article describing precisely the sort of system that I've been talking about, indicated positive treatment of it, and somehow believed that it supported your argument against profiling!). The system I'm proposing harasses people less than the current system. It uses fewer resources and (probably MANY) fewer searches to generate comparable or superior results. The distribution of those resources is different than the current system because it dynamically allocates those resources in a more efficient manner by taking into consideration factors that you object to using, but that could easily make the system significantly more efficient. That is the difference.

I understand that there's a visceral reaction against using things like race and religion as even a single of many factors in government selection criteria, but surely you recognize that there must be some sort of balancing act between safety and our concept of racial and religious equality. As I've said before in this thread, I think reasonable people can reach different conclusions, but I've been very disappointed by the failure of profiling opponents to engage in a fair evaluation of the costs and benefits of the system, and this thread is no exception.

The problem is you seem to be using some sort of nebulous definition of "safety" that I don't get. It can't be "prevent people from getting killed" because people can still do whatever they want once they leave the airport tarmac under your system or any other system, including building bombs.

[Master of Ossus]

The problem is you seem to be using some sort of nebulous definition of "safety" that I don't get. It can't be "prevent people from getting killed" because people can still do whatever they want once they leave the airport tarmac under your system or any other system, including building bombs.

Fair enough. The simple answer is that no system of security can provide 100% safety. The goal of a security system should be to minimize the number of successfully executed terrorist attacks, and the amount of damage that those inflict, with a finite expenditure of resources (that is "safety"--a 100% safe system would be able to guarantee that no attacks ever take place). Improvements in safety, therefore, require either an increased expenditure in resources or an improved ability to use the same resources currently inputed into the system and using them in a more efficient way to either detect and stop or deter attacks. This is a game theory problem.

Let's go over a simple hypothetical example to see the general idea of how this works. You have 5 people trying to board a plane to New York from Chicago on a Monday: a young black mother with her 4-year-old daughter who's staying the week and over the weekend in New York and flying back Sunday evening, a white man with a round-trip ticket returning late that night, a young Arab woman with a one-way ticket and an elderly Hispanic man who's returning to Chicago on Friday. A real system would probably have much, much more information about all of these passengers (e.g., it could see how often they had flown in the past, detect simple patterns in their flights, see where they've traveled in the past, etc.). In any case, you can search 2 of them (limited resources). A system without any profiling whatsoever would simply pick two of them randomly.

A more sophisticated system, like the one I think most advocates of profiling would argue in favor of, tries to take into account as many variables as possible to try and determine which people are most likely to be carrying bombs. In all likelihood, each person has some non-zero chance of carrying a bomb onto the airplane, but I think you'll agree that some are more likely than others to be carrying the bomb. The four-year-old, for example, seems intuitively unlikely to be carrying a bomb. Let us say that the system evaluates that there is a .0000000000001% chance that any of the five of them are carrying a bomb. You still get to search two of them, but how to do this? A profiling system evaluates the independent likelihood that each person is carrying a bomb, based on the profile data, and then compares that vis-a-vis each other.

Based on that probability, it then assigns each person a probability of being searched. Let's say that, given everything the system knows about these people, it is able to determine that the white man and Arab woman are equally likely to be carrying a bomb, the elderly Hispanic guy is ~80% as likely as the white man to be carrying a bomb, and the young woman roughly 70% as likely to be carrying a bomb; her daughter 30% as likely. Thus, the system determines that the white man and the Arab woman are each given a 50% chance of being searched, the elderly Hispanic guy has a 40% chance of being searched, the young woman has a 35% chance of being searched, and the daughter has a 15% chance of being searched. The exact probabilities are basically irrelevant. The concept is important, though: because not everyone has an equal probability of carrying a bomb (if there is one), it's actually very inefficient to assign them all equal probabilities of being searched--doing so makes it much more likely that, if one of those people is carrying the bomb, they will not be searched and be able to sneak it onto the plane. That means that, if the system has few enough resources that searching every person is impossible, the system can be made safer by use of profiling. If someone's profile suggested that, vis-a-vis his fellow passengers, he was 99 times as likely to be carrying a bomb, he would be searched 99 times as often as any one of his fellow passengers using the profile system. Using the system without any profiling at all, he would be searched only as often as they are. This is clearly an "unsafe" result.

In practice, this result is much more complicated: the "search" or "don't search" decision provides many, many more options (different layers of screening), and so the real choice isn't between "search" or "don't search," but rather to select an appropriate level of screening, none of which are 100% likely to detect a bomb carrier (if they are carrying a bomb) but some of which are better than others. (So, for instance, the screening they make everyone goes through subjectively doesn't seem terribly effective to my untrained eye, but they do it to everyone--would it be better for them to let 20% of people who are identified as being trustworthy, and subject the remaining 80% to more intensive searches? Not clear--it's an empirical question).

In addition, and this gets more to the point that several people have talked about, decisions on safety not only affect the likelihood of detecting an individal bomb carrier but can affect the probability that any of those individuals is carrying a bomb. Perhaps every white guy realizes that, under this system, he's likely to be searched, and rather than carrying bombs into airplanes the crazy white guys of the world decide to shoot up their workplaces, instead. That seems like a bad result, but depending on the relative costs of them shooting up a workplace as opposed to blowing p an airplane, it may actually be a good one. If the costs of shooting up a workplace, for instance, are so expensive that even if every white guy who would have blown up a plane decides to shoot up their boss, instead, only 1/10 as many attacks will occur and those attacks will on average do only half as much damage, then that could actually be a very successful system (as some people have suggested, the terrorists will recruit other types of people--that's actually a partial victory for the system since it seems pretty easy for them to recruit 17 year-old Arab boys; it's much harder for them to recruit 75 year-old Asian women, and so if they have to switch over entirely to that group it will reduce the number of attacks).

There's some complicated math that would have to be done about this, depending on things like the substitution rate, the magnitude of the improvements in efficiency, etc., but it's basically a systems analysis problem, and systems analysis problems are solveable, and are solveable even without perfect information about many of the variables. The point is that that is a better system if your goal is to reduce the number of terrorist attacks and the amount of damage that they do (improve safety) with finite expenditure. I don't think it's unreasonable to say that we should look into it.

[Tritonic]

That is a totally different system of racial profiling since (at least my understanding of it) is that it's an ad hoc procedure evaluated by officers on the ground as opposed to a top-down model that systematicaly identifies combinations of characteristics that are associated with terrorism. <snip>

How are the "racist shitheads" going to achieve this if they don't have input into whom they search, or if their input could be evaluated later for patterns of abuse? <snip>

Let's go over a simple hypothetical example to see the general idea of how this works. You have 5 people trying to board a plane to New York from Chicago on a Monday <snip>

A system without any profiling whatsoever would simply pick two of them randomly.

A more sophisticated system, like the one I think most advocates of profiling would argue in favor of, tries to take into account as many variables as possible to try and determine which people are most likely to be carrying bombs. <snip>

Based on that probability, it then assigns each person a probability of being searched. Let's say that, given everything the system knows about these people, it is able to determine that the white man and Arab woman are equally likely to be carrying a bomb, the elderly Hispanic guy is ~80% as likely as the white man to be carrying a bomb, and the young woman roughly 70% as likely to be carrying a bomb; her daughter 30% as likely. Thus, the system determines that the white man and the Arab woman are each given a 50% chance of being searched, the elderly Hispanic guy has a 40% chance of being searched, the young woman has a 35% chance of being searched, and the daughter has a 15% chance of being searched. The exact probabilities are basically irrelevant. The concept is important, though: because not everyone has an equal probability of carrying a bomb (if there is one), it's actually very inefficient to assign them all equal probabilities of being searched--doing so makes it much more likely that, if one of those people is carrying the bomb, they will not be searched and be able to sneak it onto the plane.

I see what you're proposing, and it is different from (better than) what people tend to traditionally assume when they think of "profiling." Traditionally, the top concern about profiling is if it could become an excuse for too uncontrolled racism by the officers on the ground. However, if the TSA agents were mostly simply taking instructions from the computer network (more so than their own decisions alone, although obviously there'd need to be exceptions for exceptional cases where the guy on the ground notices something), that would help, especially in combination with how the system you describe would have simply weighted probabilities.

So, then, for instance, although a high-risk traveler (like a 19 year-old guy who paid via cash traveling one-way by himself from Saudi Arabia) would have a greater chance of receiving expenditure of limited resources on searches than someone low-risk (like a 79 year-old grandmother from Japan traveling round-trip with her family), everyone would sometimes be searched. If carefully designed, with the risk factors and weighting clearly based on real past experience, not biased more than reality is biased, with national origin being simply one of many factors taken into account by the risk-researching computer system, I think possibly many people could understand the rationality of it. For instance, I'd have a bit higher chance of being searched traveling single by myself than someone traveling with young children and their grandparents, but that's only logical, so I wouldn't object.

[Master of Ossus]

I see what you're proposing, and it is different from (better than) what people tend to traditionally assume when they think of "profiling."

Which is a real shame, because it's what I think that most people who seriously advocate profiling are pushing for, and the counter-arguments are always these inane ideological talking points rather than a reasoned discussion of what the costs and benefits would be.

Traditionally, the top concern about profiling is if it could become an excuse for too uncontrolled racism by the officers on the ground. However, if the TSA agents were mostly simply taking instructions from the computer network (more so than their own decisions alone, although obviously there'd need to be exceptions for exceptional cases where the guy on the ground notices something), that would help, especially in combination with how the system you describe would have simply weighted probabilities.

So, then, for instance, although a high-risk traveler (like a 19 year-old guy who paid via cash traveling one-way by himself from Saudi Arabia) would have a greater chance of receiving expenditure of limited resources on searches than someone low-risk (like a 79 year-old grandmother from Japan traveling round-trip with her family), everyone would sometimes be searched. If carefully designed, with the risk factors and weighting clearly based on real past experience, not biased more than reality is biased, with national origin being simply one of many factors taken into account by the risk-researching computer system, I think possibly many people could understand the rationality of it. For instance, I'd have a bit higher chance of being searched traveling single by myself than someone traveling with young children and their grandparents, but that's only logical, so I wouldn't object.

Right. I'm not sure we should do it, but I am positive that we should have a conversation about it. I mean, what does airline security cost this country, these days? I'd estimate it's something on the order of ten billion dollars per year. So if you could make a system that costs $7bb/year, instead of $10bb, and would make airlines safer, then you have an issue where our ideals of equal treatment regardless of race are directly costing us $3bb annually and are endangering lives and property and inconveniencing many people who wouldn't be as inconvenienced if we were willing to profile. Is it worth it? Maybe, but that's a lot of money--reasonable people could come down on either side, there, but we never get to hear the discussion because people always jump in with "But... wouldn't stop all attacks," or "They'd just hit other targets," or "no profile for terrorists!"

[Gil Hamilton]

It seems to me, MoO, that any system you have is doomed to failure in the same way Arizona's new racial profiling law is. That is, it, by defintion, is actually carried out by people. Wait, check that, it's carried out by TSA guards. I'm trying to figure this out, but I can't see where "a top-down model that systematicaly identifies combinations of characteristics that are associated with terrorism" doesn't involve the actual decision making process be made by TSA guards who make the same ad hoc judgement about who fits the profile as a Maricopa county cop does.

Given the history and performance of the TSA, unless the goal of your program is the increased separation of airline passangers from their valuables, your system boils down to Arizona's. To your average pure blue uniformed TSA nitwit, "characteristicd that are associated with terrorism" might include "person that spoke to me with the wrong tone of voice" or "muslim looking guy (actually Sikh, but they are all the same, right?) whom I feel like hassling".

You argue:

How are the "racist shitheads" going to achieve this if they don't have input into whom they search, or if their input could be evaluated later for patterns of abuse? Moreover, if they did try something like this, presumably the system would have a record of which passengers it selected to be searched (since that would obviously be fed back into it for future screenings), so it would be very easy to identify problems like the one you seem so preoccupied with.

Assuming a magically competant system, this might count for something. However, members of the TSA aren't supposed to "confiscate" valuables from peoples luggage and keep them, something you'd think could be easily checked, but the amount theft at airports has absolutely skyrocketed since the implementation of the TSA. If it's gotten to the point that people are mailing their luggage to their destination due to reasonable fear that they will be looted by security and the powers that be have been unable to stop that, how are they going to analyze and stop abuse of passengers in the manner you describe?

[Sarevok]

Why is it acceptable to expend disproportional amounts of resources to save Americans from death by unlikely causes ? Its ok if you die from disease, we wont pay a cent to see you cured, but heaven forbid should a terrorist bomb kill you. We will move the heaven and Earth to save you !

[Simon_Jester]

Why is it acceptable to expend disproportional amounts of resources to save Americans from death by unlikely causes ? Its ok if you die from disease, we wont pay a cent to see you cured, but heaven forbid should a terrorist bomb kill you. We will move the heaven and Earth to save you !

This isn't new.

There's a long tradition of states doing much more to prevent their citizens being killed by enemy attacks than they do to protect their citizens from disease or poverty. Traditionally the state does have a right to fight external enemies, and terrorists arguably count as such. Only in the past 100 to 150 years have we seen it become commonplace for the state to start protecting citizens from internal problems like disease.

That's still a new idea because it wasn't really practical for government to do that until recently. Back in the day, the king might generously contribute to help keep a famine from getting out of hand, but there was only so much they could do to make things easier on the masses. Whereas there was a lot they could do about an invading army or a bandit troop; that was a simpler problem to solve.

The idea that government should move heaven and earth to fight terrorists while skimping on cancer research is basically the same thing.

[Master of Ossus]

It seems to me, MoO, that any system you have is doomed to failure in the same way Arizona's new racial profiling law is. That is, it, by defintion, is actually carried out by people. Wait, check that, it's carried out by TSA guards. I'm trying to figure this out, but I can't see where "a top-down model that systematicaly identifies combinations of characteristics that are associated with terrorism" doesn't involve the actual decision making process be made by TSA guards who make the same ad hoc judgement about who fits the profile as a Maricopa county cop does.

You can do things like have people identify themselves when they get boarding passes and have a computer tell screeners who to subject to various levels of searches.

Given the history and performance of the TSA, unless the goal of your program is the increased separation of airline passangers from their valuables, your system boils down to Arizona's. To your average pure blue uniformed TSA nitwit, "characteristicd that are associated with terrorism" might include "person that spoke to me with the wrong tone of voice" or "muslim looking guy (actually Sikh, but they are all the same, right?) whom I feel like hassling".

You argue:

How are the "racist shitheads" going to achieve this if they don't have input into whom they search, or if their input could be evaluated later for patterns of abuse? Moreover, if they did try something like this, presumably the system would have a record of which passengers it selected to be searched (since that would obviously be fed back into it for future screenings), so it would be very easy to identify problems like the one you seem so preoccupied with.

Assuming a magically competant system, this might count for something. However, members of the TSA aren't supposed to "confiscate" valuables from peoples luggage and keep them, something you'd think could be easily checked, but the amount theft at airports has absolutely skyrocketed since the implementation of the TSA. If it's gotten to the point that people are mailing their luggage to their destination due to reasonable fear that they will be looted by security and the powers that be have been unable to stop that, how are they going to analyze and stop abuse of passengers in the manner you describe?

To the extent that this is a problem, it is at least as bad in the current system as it would be under the sort of hypothetical screening process I envisage. Moreover, racism could actually be reduced in such circumstances because it allows the government to realistically reassure "guards" that groups who have historically been dangerous are receiving increased scrutiny. However, the increased scrutiny will be at a rational level as opposed to whatever the screeners decide to do that day, and will be reasonably related to the improvement of security as opposed to stupid measures that the screeners might otherwise take.

[Master of Ossus]

Why is it acceptable to expend disproportional amounts of resources to save Americans from death by unlikely causes ? Its ok if you die from disease, we wont pay a cent to see you cured, but heaven forbid should a terrorist bomb kill you. We will move the heaven and Earth to save you !

I don't know that this is acceptable, but it's no criticism of the sort of profiling that's being described. We already spend huge sums of money, manpower, and expertise on airport security. By providing these services more efficiently, at least we could saver on a portion of that and redistribute those resources to other purposes.

[Gil Hamilton]

You can do things like have people identify themselves when they get boarding passes and have a computer tell screeners who to subject to various levels of searches.

Putting aside the clusterfuck that would entail from adding an additional layer at the airport where you've got passangers filling out a survey so that your computer has the necessary information to actually make a determination as to who gets what level of scrutiny, you haven't removed the human element, you've just removed responsibility from nitwits when incidents happen.

They aren't ever going to remove individual initiative from TSA guards, so they are still going to use their discretion on individual passangers. Only now if they get abusive, they can point to your system and say "The government says these people are more dangerous!" if they are being a jackass. Or, more likely, we'll see a huge spike in the amount of theft from people on your checklist, but little more protection than we do now.

To the extent that this is a problem, it is at least as bad in the current system as it would be under the sort of hypothetical screening process I envisage. Moreover, racism could actually be reduced in such circumstances because it allows the government to realistically reassure "guards" that groups who have historically been dangerous are receiving increased scrutiny. However, the increased scrutiny will be at a rational level as opposed to whatever the screeners decide to do that day, and will be reasonably related to the improvement of security as opposed to stupid measures that the screeners might otherwise take.

That doesn't make sense. Bigots are far more likely to act on their behavior when they believe they have back up, this is why many incidents of "hate crimes" happen with groups of offenders, rather than lone criminals. You want to tell them that what they are thinking is not only supported but rational and you think this will make them less likely to act on their behavior? They certainly aren't going think think "Why, the government is takin' care of things, mah work is done", you have just made it their patriot duty to "help out". After all, they can always point to your system and say that that is how it is structured, that these people are terrorists.

People who turn out for security jobs aren't known for the nuance you are counting on. They don't think "individuals with this set of characteristics statistically speaking are more likely to commit terrorist acts". They think "people who look like this are terrorists". Example: homosexual men are not generally allowed to donate blood due to a small but statistically significant higher risk of blood borne illness. You get people who actually use this to support the notion that all gay people in general are plague bearers. The nuance that homosexual women aren't selected against, while anyone who has been to prison in recent history or have gone to a huge laundry list of countries are also selected against (a tiny portion of the survey you take at the beginning when you donate blood), is lost.

I get that you are trying to make a statistical argument, which is fine, I get it, but that fails on implementation because the system may not actually make people any safer while it certainly WILL cause an increase in incidents of abuse. You are trusting this to an organization that the other day had a guard arrested for assaulting another guard with a baton because the guards wouldn't stop making fun of his tiny penis that was revealed on a Whole Body Imager. I would be somewhat more supportive of your argument (though I'd still eliminate race as a criteria) if they organization that would implement it wasn't a failure.

[Master of Ossus]

Putting aside the clusterfuck that would entail from adding an additional layer at the airport where you've got passangers filling out a survey so that your computer has the necessary information to actually make a determination as to who gets what level of scrutiny,

A survey that's conducted in the airport? That's the best way you can think of to ascertain this sort of information?

you haven't removed the human element, you've just removed responsibility from nitwits when incidents happen.

They aren't ever going to remove individual initiative from TSA guards, so they are still going to use their discretion on individual passangers. Only now if they get abusive, they can point to your system and say "The government says these people are more dangerous!" if they are being a jackass.

No, they can't, because the system will tell them what level of scrutiny they're to apply to various passengers. If they constantly exceed that level, their excuse will fall apart completely. Compare with the current system, which appears to give complete discretion to individual screeners.

Or, more likely, we'll see a huge spike in the amount of theft from people on your checklist, but little more protection than we do now.

What are you talking about? Do you really think these baggage screeners are saying, "Well... god... usually I wouldn't take things like this, but the passenger who dropped this bag off was Mexican, so it's mine, now?"

That doesn't make sense. Bigots are far more likely to act on their behavior when they believe they have back up, this is why many incidents of "hate crimes" happen with groups of offenders, rather than lone criminals. You want to tell them that what they are thinking is not only supported but rational and you think this will make them less likely to act on their behavior? They certainly aren't going think think "Why, the government is takin' care of things, mah work is done", you have just made it their patriot duty to "help out". After all, they can always point to your system and say that that is how it is structured, that these people are terrorists.

A system that tells them specifically which level of scrutiny they're to apply to passengers is better than saying, "Check each passenger," when you're trying to get people to buy into it. Moreover, what planet are you living on where people don't think that "these people are terrorists?" It's not like everyone simply refuses to acknowledge that incidents of Islamic terrorism were what prompted the existence of TSA in the first place. At least my system distinguishes people like Sikhs from Muslims.

People who turn out for security jobs aren't known for the nuance you are counting on. They don't think "individuals with this set of characteristics statistically speaking are more likely to commit terrorist acts". They think "people who look like this are terrorists". Example: homosexual men are not generally allowed to donate blood due to a small but statistically significant higher risk of blood borne illness. You get people who actually use this to support the notion that all gay people in general are plague bearers. The nuance that homosexual women aren't selected against, while anyone who has been to prison in recent history or have gone to a huge laundry list of countries are also selected against (a tiny portion of the survey you take at the beginning when you donate blood), is lost.

First of all, I've never heard of anyone trying to make such claims (although, concededly, I also believed that the restrictions on blood donors had been phased out). But if what you say is true, then our ideal of "no separate treatment" is already being violated in blood donations because the public safety would otherwise be jeopardized. What makes it so much worse, in airline security?

I get that you are trying to make a statistical argument, which is fine, I get it, but that fails on implementation because the system may not actually make people any safer while it certainly WILL cause an increase in incidents of abuse.

HOW? How can you possibly say that with any level of certainty? Moreover, why wouldn't it make people "any safer?" Are you just saying we would spend less money on the system? That might be a good outcome, in and of itself--it would require a value judgment as to how much we're willing to spend for safety and how much we're willing to spend for our ideals of equal treatment.

You are trusting this to an organization that the other day had a guard arrested for assaulting another guard with a baton because the guards wouldn't stop making fun of his tiny penis that was revealed on a Whole Body Imager. I would be somewhat more supportive of your argument (though I'd still eliminate race as a criteria) if they organization that would implement it wasn't a failure.

Would you similarly have given up on NASA after Apollo 1? And, really, you don't see a problem with saying, "The current system sucks, therefore significant reforms should not attempted."

[Gil Hamilton]

A survey that's conducted in the airport? That's the best way you can think of to ascertain this sort of information?

Computers don't have the technology yet to go look through a CCD and say "Hey, that guy's an Arab". If you want a complete profile on someone, you can't magic a complete set of data from the ether, you've got to ask them.

No, they can't, because the system will tell them what level of scrutiny they're to apply to various passengers. If they constantly exceed that level, their excuse will fall apart completely. Compare with the current system, which appears to give complete discretion to individual screeners.

The TSA would never do that, because it, theoretically, would prevent them from stopping people who looked suspicious but that your computer system missed. If Timothy McVeigh walked through your system it would go "White, veteran of the US Army, American Citizen of so many years" and then clear him. If McVeigh was trying to blow up an airliner rather than a building in OKC, your system would have missed him. Thus, any system would still have built in individual initiative amongst individual agents. In fact, it wouldn't have caught Richard Reid either, who was initially stopped by French screeners not because of his background, but observation by screeners of an extremely disheveled appearance, dodging some questions, and the fact he didn't book luggage for a transatlantic flight. Only the last thing might flag in a computer as suspicious.

While there are a plethora of technologies come on that are exquisitely sensitive at detecting the tools of terrorism (you'd be shocked at how far away the chemical trace of some explosives can be detected), computers can't look at a scene and detect context. That requires a person. Unfortunately, the people who showed up for the job are TSA agents.

What are you talking about? Do you really think these baggage screeners are saying, "Well... god... usually I wouldn't take things like this, but the passenger who dropped this bag off was Mexican, so it's mine, now?"

You miss the point. Since the TSAs inception, theft from luggage has become rampant. If you have a computer system specifically telling TSA agents to turn a specific groups luggage inside out during an inspection, theft amongst that specific group will skyrocket because they are handling more of their stuff. Double this with specifically assuring the TSA agents that that they are correct and rational to discriminate against that particular group.

A system that tells them specifically which level of scrutiny they're to apply to passengers is better than saying, "Check each passenger," when you're trying to get people to buy into it. Moreover, what planet are you living on where people don't think that "these people are terrorists?" It's not like everyone simply refuses to acknowledge that incidents of Islamic terrorism were what prompted the existence of TSA in the first place. At least my system distinguishes people like Sikhs from Muslims.

I know people think that, and I also know that bigots are far more likely to act on their thoughts when they believe they have back up. Your system makes that back up official policy.

Besides, if we are talking about bigots, how many of them are going to know what a Sikh is? After 9/11, there was an increase in assaults on Sikhs in the United States because asshats thought they were funny looking Muslims.

First of all, I've never heard of anyone trying to make such claims (although, concededly, I also believed that the restrictions on blood donors had been phased out). But if what you say is true, then our ideal of "no separate treatment" is already being violated in blood donations because the public safety would otherwise be jeopardized. What makes it so much worse, in airline security?

Blood banks separate people into a mere two groups. Your system specifically targets them on statistics based on nuance that almost certainly is beyond the average person who shows up for a security job. At the blood bank, your interview is conducted by a RN (not merely one the people drawing blood). At the airport, the person actually interacting with the passengers isn't remotely that trained.

If TSA guards were as well trained in critical thinking and security as RNs are at nursing, then I'd give your system a bit credence. However, you are working with the material you are working with.

HOW? How can you possibly say that with any level of certainty? Moreover, why wouldn't it make people "any safer?" Are you just saying we would spend less money on the system? That might be a good outcome, in and of itself--it would require a value judgment as to how much we're willing to spend for safety and how much we're willing to spend for our ideals of equal treatment.

I can say it will increase abuse because of what you said, that part of your system is specifically assuring people who are racist already that the government is supporting them and that their views are rational.

As for not making people any safer, it would take someone pants as a hat stupid not to be able sneak something past TSA agents, even when they search your stuff. You can't tell me that when you've been going through security you haven't done a gedaken exercise in your head about how you'd move an object quietly through security. The TSA exists to provide the illusion of security. Someone who was bound and determined to blow up a plane and had some organizational skills would walk right through them.

Would you similarly have given up on NASA after Apollo 1? And, really, you don't see a problem with saying, "The current system sucks, therefore significant reforms should not attempted."

If Mercury and Gemini were rife with exploding rockets and then Apollo 1 blew up? Yes, I would give up on NASA. Fortunately, Apollo 1 wasn't part of a trend. The fact that the TSA is a colossal bonehead of an organization and this is another data point that falls on the trendline makes this different. You want to implement your system in an organization of exploding rockets.

[MKSheppard]

It's official. The Administration has come out and said:

Attorney General Eric H. Holder Jr. said investigators had “developed evidence that shows the Pakistani Taliban was behind the attack,” a sharp escalation from the initial assessment that Faisal Shahzad had acted alone and without sophisticated training.

Okay, here are my worries.

1.) We're not the only ones who can do post-mortem analysis and monday morning quarterbacking. A lot of information has been released on how the bomb was arranged, how it was set up, and why it failed in the media and internet; as well as on Shahzad's movements before and after the attack. The taliban can look at it and go "aha, this is where we screwed up with our training!" and alter their doctrinal training correspondingly.

2.) I'm even more disturbed by the fact that this fool went right through our comprehensive layers of security that we put in place post 9-11.

For example, did you know that Shazhad's father is none other than Pakistani Air Vice Marshal (retd) Baharul Haq?

Or that he had taken 12 trips to Pakistan between 1999 and now?

That during many of these trips, he spent significant amounts of time in the North West Frontier Province?

That during his trips, upon returning to the US, he had declared about $80,000 in cash on his person to US Customs?

That's about $6,600 each trip average.

There were so many warning signs in here -- family is from NWF Province? Makes a trip once a year on average? Spends large amounts of time during those trips in NWF? Comes back with large amounts of cash each time?

All those add up to him being "a person of interest" that we'd like to keep an eye on, just in case.

But no; our idea of security is to pat down people like me who wear implants.

[Samuel]

1.) We're not the only ones who can do post-mortem analysis and monday morning quarterbacking. A lot of information has been released on how the bomb was arranged, how it was set up, and why it failed in the media and internet; as well as on Shahzad's movements before and after the attack. The taliban can look at it and go "aha, this is where we screwed up with our training!" and alter their doctrinal training correspondingly.

Actually they always do- the reason they probably don't get a higher success ratio is that we do the same thing and plug the obvious holes.

Of course, when it is terrorism against public locations, you have to fix the current machinary that he slipped through as you pointed out. You can't guard everything.

2.) I'm even more disturbed by the fact that this fool went right through our comprehensive layers of security that we put in place post 9-11.

Call me cynical, but I don't think that the security was as comprehensive of we like to believe. The most obvious part is that the terrorists concentrate on Iraq and Afghanistan, so it doesn't get constantly tested and it was used as a political tool as well as protecting the public which probably messed with its effectiveness.

All those add up to him being "a person of interest" that we'd like to keep an eye on, just in case.

As 9/11 showed putting warning signs together is something the government is really bad at. Especially glarringly obvious ones like "from region sympathetic to Taliban" or "brings large amount of cash to US".

But no; our idea of security is to pat down people like me who wear implants.

It makes them feel like they are making a difference by enforcing petty rules.