Can pad the watch list. Witness this story in the Washington Post, where the state of Maryland law enforcement added peace activists to the terrorist watch list because there was no other category for them in the database (um, people exercising their first amendment rights?)

http://www.washingtonpost.com/wp-dyn/content/article/2008/10/07/AR2008100703245.html

Better recheck that. It's either 1 out of 300 or 1 out of 1000 (more or less).

I agree with your article, but I just want to point out an error. 1 million people is not 1 out of every 10 Americans. (It's more like 1/300.)

The list has 1 million names, not 1 million social security numbers. Assuming an average of about 30 people walking around with any given name on the list and you get the 1/10 figure.

Timely. Has a definite perspective. Informative.

It even has something some a principal in the story that is not publicly available. Not a scoop, but a real original quote.

Farivar is looking better this week.

However, it is important to note that 400,000 is about one out of every 763 people in America, not one out of ten. More importantly, this is not a list of suspected American terrorists, but rather of suspected terrorists worldwide. That makes it one out 16,500 people.

Is it possible that one out every 16,500 people in the world would do the USA harm? Yeah, for various reasons, that is actually possible.

Quick description of data mining: If you lay on your back and stare at the clouds long enough, you'll see a cloud that looks like an elephant. That doesn't mean you're looking at a real goddam elephant.

I did some computations a few years back regarding the CAPPS II thing which I believe was rejected eventually. Using conservative estimates (conservative in a scientific sense, not politically) and Bayes theorem, I computed that using such a system, the probability that a person identified as a terrorist ACTUALLY IS a terrorist is about 1 in 50,000. Imagine being a poor airport security officer, knowing that each time a red flag comes up there is practically no chance that the suspected flyer is a terrorist. Utterly demoralizing, and it would completely defeat the intentions of the system.

It's really weird that it takes the concentrated effort of a team of experts to convince the government of something which is just common sense to the average person.

... I think they would do well to concentrate their attention on Wall Street.

This is really stupid. Our government goes totally overboard watching regular people for the slightest sign of "terrorist" activity, but they completely miss the financial disaster of the century brewing right under their noses.

As a social scientist who works with statistics and analysis of verbal content, I promise you that this won't work. We've been unable to get to good reliability with basic motivational content, let alone subtle markers of terrorism. It would be nice to be able to target someone through patterns, but buying the same brand of popcorn ten times in a row is light-years away from a constellation of words, syntax, and meaning that add up to terrorism plans. Read it the right (wrong) way, and someone describing a video game would get nailed. Personality is very basic (there are only about five major trait dimensions); people add their training, upbringing, values, skills, and experiences on top of that, let alone the fantastic complexity of how people actually think and behave. If you really think about the complexity demanded, it becomes obvious how futile this effort is.

This is like the FBI's "profiling" myth -- the level of precision they claim is not really based on science at all, and nowhere near as useful as they would like people to believe. Similarly for lie detectors, which have a massive fudge factor. It's useful to threaten a guilty party who believes the nonsense, but it doesn't work for a smart, educated person who knows what they are doing -- the very kind of people we should fear most as terrorists.

But this myth, however appealing and even useful in daily police work, is dangerous when it justifies all sorts of privacy invasions because maybe it might work someday. Let's get the technology first, and then perhaps we can try applying it broadly.

This is not a list used by the local police to keep an eye on someone who has been noted to be doing "suspicious" things (whatever the hell they are. The gov is worried about the Uigers because they've had weapons training. Hell, I've had weapons training courtesy of the NRA and the US Air Force. There's a whole lotta people in this country who have had weapons training.) Theoretically, the whole point behind the bullshit at the airport is to ensure that no one can get on the aircraft with a weapon (regardless of how ridiculous the definition of a "weapon"). Once everybody's disarmed, what do you care about their politics?

But that's the point, isn't it? Keep us in fear (the definition of assault) and meek?

I take issue with this point (quoted, not written by the author):

"From my point of view, a program that relies on personal data but then fails to detect or prevent terrorism necessarily constitutes an unwarranted invasion of privacy. In short, no invasion of privacy is justified if the program doesn’t work."

This assumes that if the program actually worked then it would be justified. That's like saying that a cop can break into your house for no reason as long as he or she finds something incriminating as a result. The whole idea behind the rule of law is that no one is above the law - including those who enforce it. We can't say that it's fine to bend the rules as long as it works out in our favor in the end, otherwise the rules have no meaning in the first place. The price of freedom is the risk that someone will abuse it. Frankly I'd prefer living in a society that placed a value on the privacy of the individual but included some risk to one that was completely safe but allowed the authorities into all aspects of our lives.