What Medical Tests Should Teach Us About the NSA Surveillance Program

First time accepted submitter Davak writes "In many ways finding the small amount of terrorists within the United States is like screening a population of people for a rare disease. A physician explains why collecting excessive data is actually dangerous. Each time a test is run, the number of people incorrectly identified quickly dwarfs the correct matches. Just like in medicine, being incorrectly labelled has serious consequences."

Comments

The crazy is strong in this one

[davebarnes]

You, sir, are a nutjob.

Re:The crazy is strong in this one

[Anonymous Coward]

Not so fast. There is precedence. The Gulf of Tonkin Incident for one.

Re:The crazy is strong in this one

[Anonymous Coward]

Democracy exists because for every group of people that unquestioningly support the leadership, there are others that without any reason support it. And it's because of those "nutjobs" that the rest have to make transparency possible.

When all people start to agree with the government and paranoids disappear or go silent, that's when you'll know democracy is dead.

Just because he says conspiracies exist at government level and sounds like a nutjob, doesn't mean he isn't right, but without proof, it doesn't mean he isn't wrong either (or insane, although, considering the definitions for the various mental afflictions change every year, I'd go for the safe description of "eccentric").

Re:The crazy is strong in this one

[Artifakt]

Does that matter, in re. the parent post?

        Start with the assumption that terrorism is NOT just a bunch of false-flag state operations from some group such as the CIA.
If generating these huge lists creates a cloud of false positives that will make us actually less safe, then we (we meaning citizens of the country making the lists) want the state not to waste that money. We ought to oppose wasteful, counterproductive ways of fighting terrorism.

          Now try the counter-assumption - Terrorism originates within our government: If the government actually is comitting terrorist acts, still, the government gets money to accumulate these big lists of potential terrorists. In fact, one big motive for false flag terrorism would be to get lots of taxes devoted to creating such huge lists, and thus make a big profit for some contractors. We don't have open proof that this is happening, but we can prove the part about waste. If we oppose wasting that money on counterproductive list-making, and it turns out the government really is behind false flag operations, then the government has failed to get something it wants by using these false flag operations and should logically try something else to get money and create contractor opportunities. Maybe the government will even try something more ethical.

          xiando thinks the government is doing false flag operations. xiando appears to think this is bad. xiando appears to want the government to stop doing false flag operations. But, this article is giving a good reason for the government to stop doing something that, if xiando is right, is a consequence of the thing xiando apparently wants stopped. Yet xiando wants to lump this article in with government propaganda for doing the very thing the article says to stop doing. There's a gear slipping there, somewhere. Other people want at least some of the same goals xiando presumably wants (shutting down the programs that may have been enabled by false flag operations, if xiando is right). But because they aren't advocating other things xiando presumably also wants (like exposing the government secret program, or maybe putting a bunch of people in the government on trial for war crimes), they get lumped in with xiando's great enemy.

Re:The crazy is strong in this one

[Anonymous Coward]

Does that matter, in re. the parent post?

It demonstrates that there are black swans. Having established that black swans exist, you must now observe your swan to determine if it is, in fact, white.

Re:The crazy is strong in this one

[Livius]

The fact that the NSA is not really trying find terrorists does not mean that other terrorists do not exist.

Then what do you do then?

[Anonymous Coward]

So you have no test and just let the virus spread? You will end up with an epidemic. If there are better followup tests and second opinions in order to determine who is sick, then that is the way to do it. Just telling everyone, including those who are violent, what you are doing so they can get around the tests won't help anything. There will be plenty more false positives now once people who are seen as avoiding the system seem guilty of something and need to be followedup on...

Violent extremism is a ideological disease that spreads in the same way as a infection through society.

Re:Then what do you do then?

[BrokenHalo]

So you have no test and just let the virus spread?

You extend the analogy too far. In fact, the analogy in TFA, while interesting, has limited relevance. Yes, the danger and destructive effects of false positives are important in both medicine and national security, but where TFA mentions (almost in passing) that "The balance between privacy and security is always difficult", it sidesteps the simple fact that this surveillance is about neither. It is about control.

Let us not fool ourselves that the US (or any other) government is actually likely to prevent all (or any) acts of terrorism with these efforts. We have recent proof otherwise. Our various governments have simply seized on this supposed threat as a means to exert control - for no other reason than because they can.

Re:Then what do you do then?

[jc42]

Let us not fool ourselves that the US (or any other) government is actually likely to prevent all (or any) acts of terrorism with these efforts. We have recent proof otherwise.

It's no longer "recent" by media standards, but the (second ;-) attack on the World Trade Center is an excellent example. Much of the news coverage of the event is still available online, and if you dig it up and look at it, you'll see that several things stand out. One is that the US authorities were totally taken by surprise, and didn't have any idea what was happening until after the second tower was hit. However, it became clear in the first several hours that they'd decided who to blame. The reports from everywhere were full of "Al Qaeda" and "Osama bin Laden" (often badly mispronounced ;-), despite the obvious fact that they couldn't have collected the evidence in such a short time.

Over the following weeks and months, it also became clear that their ignorance was pretty much self-imposed. They had been warned about the specific perps by various other countries' security folks, and chose to ignore the information. This was in part due to a serious shortage of Arabic-speaking translators in the US military/security agencies. This was in turn due to their mistreatment of Arabic speakers, which the US has millions of. If you look into this, you'd probably also conclude that anyone fluent in Arabic would have to be really stupid (or suicidal) to volunteer for a translator job in those agencies.

The most parsimonious theory explaining this is that the US government isn't particularly interested in finding and blocking terrorists; they are mostly interested in using such things as a way of instilling fear in the general population. With this understanding, the government's "anti-terrorist" activities make a lot of sense.

(And, of course, treating the US government as some sort of unified, monolithic entity is a major mistake. There are lots of people in various government agencies who understand the situation pretty well. But they're generally not the ones in charge. Or if they are, they also understand that it's all to their own personal benefit. Or they keep quiet because they understand how "whistle blowers" are treated, and don't want that to happen to them. But we may hear from them after they retire. ;-)

Re:Then what do you do then?

[zippthorne]

Yeah the medical issue is a problem with matching test results to actions taken. The argument that we should do less tests, because the tests might find something and then we must take action always seemed a little specious to me. Sure, you might end up doing extra procedures based on the results of the test, but that just means that we need to do one of:

1. suck up the cost of the extra procedures (if there's little to no health risk and the cost is reasonable)
2. improve the accuracy of the tests themselves
3. research or use follow up tests to filter out issues that are benign
4. Change the threshold for which liability is incurred if results are not followed up on to one which makes sense and results in the best overall outcome.

If the problem is that lawyers make it difficult to gather information because they will require action where none should be taken, the solution isn't to cut back on information. It's to cut back on lawyers.

Re:Then what do you do then?

[tylikcat]

That's not the precise argument, at least in a medical context. If the tests themselves and the responses to false positive have no significant medical downside, that's one thing.

But, let's say we starting giving all women yearly mammograms at age 16. Now, while this might reveal a very tiny number of additional breast abnormalities (many of which won't be cancerous) it's going to expose a lot of women to increased amounts of radiation, and while that amount of radiation is slight, that is likely to lead to a measurable increase in rates of cancer. If you're causing more cancer than you're catching, it's a stupid test, right?*

In addition, the response to false positives needs to be taken into question. Further procedures have their own medical costs. If you have a high rate of false positives leading to painful and hazardous procedures, that cost, too, has to be weighed against the value of catching those cancers early. ... and I will stop here as the breast cancer analogy in particular is one I can babble on about for a very long time. (My mother is a breast cancer survivor, and was diagnosed fairly young, which puts me in a high risk category.)

* One could make the argument that this is a very tight analogy, as if surveillance is increasing hostility towards the government by US citizens, and towards the country overall abroad, we could be creating a worse situation than we're addressing. I think this is a pretty strong argument applied to some of our foreign wars.

Re:Then what do you do then?

[swalve]

As always, one size does not fit all. That yearly mammograms cause more cancer via radiation than it catches might raise the costs and the prevalence of all cancers in the population, it doesn't change that there will be many, many people whose cancer is caught early and treated successfully. Plus, one has to imagine that of the small group of people who conceivably get cancer from the xrays, most of them were probably going to get cancer sooner rather than later. It's herd risk versus individual risk. And a signal to noise ratio problem. If you increase the signal, you are going to increase the noise. But you know it is happening and can account for it.

Not to mention, at least in the medical world, most tests have data about what their false positive versus false negative rates are. There is no perfect test, but you want one that gives false positives rather than false negatives. I mean, I can claim that I can detect cancer by pinching bottoms, and be very nearly 100% right in telling every person I see that they are cancer free.

Re:Then what do you do then?

[Dunbal]

Violent extremism is a ideological disease that spreads in the same way as a infection through society.

[citation needed] Violent extremism is not a "virus" at all. It's a product of poverty and lack of education, making the masses exploitable by unscrupulous individuals who seek quick wealth and power. Educate your people and extremism disappears. Do you think it's because of the TSA and the DHS that the US suffers relatively little "terrorism" (both before and after 9/11)?

As for specificity and sensitivity of the "tests" for terrorism, I mentioned this problem oh, 10 years or so ago. I guess it's about time people start catching on.

Re:Then what do you do then?

[jkflying]

You're ignoring the side effects from treatment of people who didn't have the condition, and the suffering they go through. In the breast cancer analogy, chemotherapy is terrible: it causes your hair to fall out, you lose months or years of life to something that wasn't necessary. The alternatives are radiation therapy and mastectomy, which are worse. So how many people wrongly getting their breasts removed, or getting chemo, is worth saving a single person's life?

This is true in military and intelligence situations as well. If law enforcement starts having negative side effects (think TSA nude scanners and groping, SWAT teams being called as pranks etc) then the negative effects on society are worse than the actual problems they would be preventing. Not only that, but if they aren't seen as helping, people will become less cooperative to law enforcement officials, which will further break down social peace.

Re:Then what do you do then?

[russotto]

Violent extremism is a ideological disease that spreads in the same way as a infection through society.

Really? Most people, when presented with extremism (even non-violent) tend to back away slowly and treat said extremist as a nutjob.

Re:Then what do you do then?

[tylikcat]

One size doesn't fit all. That's why I cheerfully began annual mammograms when I was 33 - because I am in a high risk group, and for me there is a clear benefit.

And for others, there isn't, and generally speaking what has been considered to be a clear benefit from these tests is being reconsidered as the cost of having the test run becomes better known. Breast cancer is, in fact, one of the more involved and contentious ones (as well as the one that impacts me the most directly.) The cost benefit analysis (and again, I'm talking health costs, not financial costs) aren't trivial, but they're also not that difficult. And it only gets more involved from there, because current evidence suggests pretty strongly that a lot of breast cancers, particularly those found in routine screenings, most likely including my mother's, are neither quick growing or dire, and all the early interventions are not decreasing the incidence of later stage cancers. If getting them early means they never turn into later stage cancers, one would not expect that.

When I was in my early teens, I took care of my mother while she went through a fairly mild course of chemo, and her hair fell out and she lost damn near a quarter of her body weight, and entered menopause early from the experience. It's easy to talk about being safe is always the best course... but this isn't just about being safe. Radiation treatment carries a substantial risk of causing cancers downstream, and no, not just cancers that would have happened anyway. This is much more true of the kind of doses used when treating a late stage aggressive cancer - so if you're otherwise looking at weeks or months to live, it probably seems like a decent trade. But you damn well don't want to undertake this kind of treatment by mistake.

Re:Then what do you do then?

[tnk1]

Using the US Government's attitude to pre-9/11 is not very fair. They ignored the evidence because they figured that perhaps there would be some hijacking or some fairly minor bombing. They were caught by surprise because they were complacent, but that doesn't mean that they lacked the information to know exactly who was responsible.

I'm not at all surprised they had the information that fact. There were likely people in the CIA and FBI trying to get someone to listen in the upper management levels for years. Now those executives wanted answers and they finally listened.

The only thing that happened with 9/11 is that the government got an attitude adjustment. George W. Bush wanted to completely ignore the Middle East, back in the day. Then the Middle East came to him.

Why is it that people want to see intricate plots in something that can easily be explained by heading down to the DMV and checking out the average initiative level of a standard government worker? There are no "plots", there is no "campaign of fear". There is only crass incompetence. Don't kid yourself. 9/11 was a tragedy of bureaucracy and political tunnel vision. Believing in some sort of fiendish plot is giving them far too much credit.

Re:Then what do you do then?

[kgskgs]

There is one key difference in medical testing and NSA surveillance that no one seems to be talking about.

When I go to the doctor, it is politically and morally correct if the doctor looks at my race, age, gender, etc. and decides what is high risk to me and then narrow down the tests specifically for me. In short "discriminate" based on my age, sex, race, etc.

When a security enforcement person in a broad sense (NSA, cops, everyone combined), looks at you, they are not allowed to discriminate based on your age sex gender whether you should be monitored more or not, whether you should be kicked off the airplane or not. That makes them search needle in the haystack.

Don't get me wrong. I am not for all out surveillance. Merely pointing out a key difference between the medical and NSA world.

K

Re:Then what do you do then?

[crbowman]

Mohamed Atta the ring leader of the September 11th attacks was not uneducated, he had a degree in architecture and was pursuing a graduate degree. He came from a wealthy family. It is dangerous when you make the claim that violent extremism is a product of poverty and lack of education. While they may be correlated, correlation is not causation and the statement leads people to believe that we can eliminate violent extremism simply by eliminating poverty and providing education.

Re:Then what do you do then?

[ppanon]

If you have a genetic condition, then your ethnicity may have some relevance. However in the case of nearly all germ infections, what you did in the last week has much more relevance. But yes, women do have an extremely low incidence of testicular or prostate cancer.

Re:Then what do you do then?

[hackwrench]

There is no such thing as a "standard" government worker.and it is just s ludicris if not more so given recent and not so recent historical events to state definitively that there are no plots or a campaign of fear as it is to state definitively that there are. I don't know, maybe it's all a coincidence or maybe collateral damage that the government actions result in no little amount of fear in the general populace

Re:Then what do you do then?

[Dunbal]

What I mean is that poverty and lack of education makes for many people being willing to blindly follow these idiots and be bullied by "peer pressure" into staying quiet and accepting extremism. For exactly the same reasons that religion is very much on the decline in modern, educated countries. When you're educated you tend to take fanaticism with a pinch of salt. Of course there are highly educated, manipulative people who are more than willing to take advantage of their less fortunate bretheren. It's called politics.

Re:Then what do you do then?

[gzuckier]

The point being, that if the only test you have produces more false positives than true positives by orders of magnitude, it's worthless for actually finding the true positives anyway, so if there's going to be an epidemic, your lousy test isn't going to help at all.

Gedanken experiment, Reductio ad absurdum, etc.: Consider the ultimate in such tests: Just flag everybody who gets the test as a positive result. You will absolutely be guaranteed to catch every single true positive case, and you will be absolutely guaranteed to be no further ahead than you were without the test.

In order to see if the test has any advantage at all, you need to evaluate the diagnostic efficiency (sensitivity * specificity)
This would eliminate such items as the "flag everybody" test, which would have an efficiency of 1*0=0.
But for a more sophisticated look at a test where the penalties for false positives and false negatives are different, you should probably look at the Receiver Operator Characteristic, which despite the abstruse name is basically just a graph of the sensitivity versus selectivity for a given test as you vary the threshold at which you decide negative vs positive result. You can then balance the advantage the test gives you versus the differing penalties for either kind of error.

Re:Then what do you do then?

[gzuckier]

Vaguely related to my previous post above, having all the data in the world can actually impede your ability to find what you are looking for. Consider the frequent TV lawyer show of responding to the other side's request for disclosure by handing them a truckload of documents related to the case, since they're not going to be able to find the one smoking gun document they are looking for.

That's what happened in 9/11 in part; the government had tons of data, but without having a focus on al qaeda, bin laden, etc. they just didn't have the data mining capacity to see the links in the data and pull them out. And the Cheney administration was still focused on fighting the damn commies and russkies, and had dropped what they considered Clinton's bizarre obsession with Al Qaeda and bin Laden. (None other than Paul Bremer told the Washington Post in 2000 that the Clinton administration had been "correctly focused on bin Laden". Now that is ironic.) And of course brushed off the famous "Bin Laden determined to strike in US" memo that could probably have netted some info from the pile of data if they had taken it seriously and gone on a hunt in their data. If you're looking for needles in your haystack, you're not going to find a toothpick.

Re:Then what do you do then?

[gzuckier]

Generally spot on, but I have to point out that the standard government worker is no more useless than the standard corporate worker. Corporations are routinely damaged or destroyed because some executive won't listen to the lower level guy who happens to be immersed daily in the reality of the business situation, telling him that his brilliant idea for a new formula for Coke isn't going to work.

After all, those people in the CIA and FBI trying to get someone to listen in the upper management levels were also standard government workers; the guys who tossed the memo "bin Laden Determined to Strike in UN" into the trash can were the corporate titans who came in to government to run it like a business. Which it appears they did.

Re:Then what do you do then?

[zippthorne]

But the solution to "acting on the data is burdensome" isn't to collect less data. It's to be smarter about how you act on it.

In the case of a screening test, sure, you want to set the frequency of tests to the optimal one for "health problems successfully detected and mitigated" vs. "health risk increase from the test itself".

"mamograms increase cancer risk" is a valid argument for controlling the number of mammogram tests. "Mamogram test results must be acted upon" is not, because you have the choice at each step along the way as to whether or not to continue to the next step. Each step should be evaluated on its own, based on the available information.

If the followup action has its own risks, they should be factored in to the decision process for that step, not the preceding step.

Ignorance for the sake of ignorance isn't a plan.

Re:Then what do you do then?

[jkflying]

Sometimes we have no way of knowing that the test was a false positive, that's the definition of a false positive. Not only that, measuring less often doesn't mean that you won't catch the cancer, it means you'll catch it later. If the negative effects from the number of acted-upon false positives outweighs the benefits of the true-positives being found earlier, it means it is better to measure less frequently. Like GP, you're making the assumption that there is no such thing as a false positive, and/or that false positives never get acted upon, both assumptions which are provably false.

The social analogy is how soon we should stop a 'terrorist group' into their evil plans for destruction. Should we wait until they call a friend about making a bomb, should we wait until they get the plans, or should we nip it in the bud and just wait until they make friends? Should we wait until they buy supplies for the bomb, should we wait until they have a working bomb, or should we just take them out with a drone the moment they start talking about making one? Or should we wait until they're about to set it off, ie. the only point when it is actually dangerous.

At some point, the intrusion to society outweighs the additional risk of the terrorists being successful, because you start pulling in high-school students doing science projects and farmers buying fertilizer for their crops. Measuring less often means the number of people affected when your measurements are wrong is lower.

Of course, if you think you can design a cheap cancer measurement system with 0% false positives, be my guest ;-)

Re:Then what do you do then?

[zippthorne]

While you can't know if any given result is a false positive on its own, you can eventually determine the false positive rate using the results of follow up (or non-follow up outcomes), and use that to inform the decision about whether to take additional steps.

Further, I somewhat doubt that it's a simple yes/no on the false positives. Surely things like imaging tests will have varying degrees of confidence in the results depending on how well resolved the detected structures are, or even the shape or location of said structures.

If a doctor sees something that has a 90% chance of being a false positive, should the follow up test or procedure be performed? I propose that it should depend on the nature of the follow up. Perhaps a second test might be in order (assuming that a repeated test is capable of improving the confidence or disproving the hit) if the tests are cheap and low risk and other measures are not.

This is an optimization problem, and more information shouldn't lead to less favorable results unless we're handling that information poorly.

well duh.

[Gravis Zero]

the NSA is not concerned about infringing on people's rights and civil liberties. if we are going with medical analogies, i think the NSA would rather amputate than treat an infection.

Re:well duh.

[Anonymous Coward]

"The NSA" would not "rather" do anything, because "the NSA" doesn't make up missions for itself; it only responds to Information Needs (INs) levied upon it. "The NSA" is only interested responding to the INs as aggressively as possible under the law. If it's not doing everything possible under the law, I'm not sure what it should be doing.

Of course, this analogy is all wrong, too, because everyone assumes that NSA is "mining" the phone call metadata, when in reality it is only collecting it so that it may be queried for specific targets later. (I realize people believe it is all being mined with no proof of this and in contravention of everything in the leaks and numerous on- and off-the-record statements from current and former officials and other experts. I also realize people think all internet traffic is being collected and mined, even though there is no proof of this, either, and a program that was collecting internet metadata for later searching was terminated in 2011.)

Re:well duh.

[Anonymous Coward]

"The NSA" would not "rather" do anything, because "the NSA" doesn't make up missions for itself; it only responds to Information Needs (INs) levied upon it.

Really? So everything the NSA does is decided explicitly by politicians. I find that hard to believe. Care to provide any citation backing up this claim?

"The NSA" is only interested responding to the INs as aggressively as possible under the law.

However, this claim is false because we now know that the NSA violated the US constitution.

If it's not doing everything possible under the law, I'm not sure what it should be doing.

Gathering a reasonable amount of intelligence under adequate judicial oversight? A reasonable cost/benefit tradeoff like everywhere else and without violating the constitition? Just some crazy ideas, I know, I know...

Of course, this analogy is all wrong, too, because everyone assumes that NSA is "mining" the phone call metadata, when in reality it is only collecting it so that it may be queried for specific targets later. (I realize people believe it is all being mined with no proof of this and in contravention of everything in the leaks and numerous on- and off-the-record statements from current and former officials and other experts. I also realize people think all internet traffic is being collected and mined, even though there is no proof of this, either, and a program that was collecting internet metadata for later searching was terminated in 2011.

These claims are mostly false. However, independently of whether they are false or not, one can also demand of an agency as powerful as the NSA to show a certain amount of maturity and social responsibility in constitutional matters crucial to democrarcy, rather than the "hey, you don't have conclusive proof, so you're wrong and fuck you" approach of little children they (and you) are illustrating. That and not being caught in obvious lies like James Clapper would be helpful to the credibility of the NSA. Another question is how it is possible that the director of the world's largest intelligence agency can remain the director after having comitted perjury in front of Congress. In a truly democratic country this would be unthinkable.

Re:well duh.

[hackwrench]

There is no proof that all internet traffic and phone metadata is being collected and mined? All anyone has to do to determine that that is a blatant falsehood is google nsa hdoop mine. Even if that were not the case, you are making the assumption the NSA is not. That is av far more dangerous assumption to make.

No time to read TFA

[Sponge Bath]

After glancing over the summary, I'm fine with doctors are experimenting on terrorist dwarves.

Re:No time to read TFA

[Anonymous Coward]

Also, bomb the population from the orbit because it's the only way to be sure.

Re:No time to read TFA

[intermodal]

I'm not sure I like that bald-faced attack on dwarves. It ends up siding against dwarves while equating the NSA with the Uruk-hai, and I don't think that's fair to the Uruk-hai.

Re:No time to read TFA

[gzuckier]

On behalf of the Holy Alliance Against Little People With Facial Hair, I must protest.

After the attack

[Anonymous Coward]

It could be the "blue screen of death" or "Abend"?

How do you find out what went wrong? You look at all the activity that led up to the event. That is why you log events, why you save memory.
The is a lot more to saving information than predicting events.
If saving this information is a good idea is a political question.

It seems likely

[The Real Dr John]

That the NSA is not specifically looking for terrorists, although that is the convenient excuse. They are looking for all sorts of things, and that is why they are collecting everything. They are listening in on foreign diplomats to see what they are up to, they are eavesdropping on foreign corporations to give US companies an advantage in trade deals, they are digging up dirt on political enemies and protesters, and they are checking up on reporters to help keep them in line, and they are especially looking for whistle blowers who might throw some light on what they are doing with our tax dollars. All of these activities have been reported, so it doesn't take much imagination to realize they are collecting everything they can on purpose and for numerous reasons, most of which are not to the benefit of the American people. If the intention was to help the American people, they would be putting all that computing power into bioinformatics to cure cancer and other diseases that kill half a million Americans a year.

Re:It seems likely

[Livius]

If the intention was to help the American people

The goal is to benefit the bank accounts of a small set of the American people.

Sociopaths will flatter themselves that they got it close enough.

Re:It seems likely

[gringer]

Of course. Pat Buchanan talked about this recently at NetHui in New Zealand:

Terrorism is a fig leaf placed on the intelligence business to justify what they do. Terrorism is not the bulk of what intelligence agencies do. The bulk of what they do (to include the GCSB) is traditional state-to-state espionage, Increasingly cyber in manifestation. But 90% of what intelligence agencies do, in this country and elsewhere, is spy on other states (perhaps spy on commercial entities connected to a state). But terrorism is the buzzword that western intelligence agencies use to justify all sorts of sins.

Video link here, which also explains how the GCSB gets around not being able to spy on NZ citizens via contracting their staff out to other agencies. Also, 80-90% of intelligence is gathered from freely available sources (e.g. facebook, twitter), so terrorism is a 10% of 10% sort of thing in terms of surveillance laws.

Re:It seems likely

[lightknight]

Indeed. In much the same sense as the warrantless wiretaps got the nod because they would primarily be used on terrorists, and have since been found to be used almost exclusively on non-terrorist related crimes, there is a fair chance that the NSA is not watching only foreigners engaged in acts of sabotage, but is 'helping out' other agencies as well with their daily chores. That's on a broad basis -> on an individual basis, there are, no doubt, petty individual requests / favors being carried out, much like the rest of the government, with a lack of evidence being the sole reason that the agency itself has not been cleansed (well insulated is well insulated). If the NSA director told us that he knew for a fact that no one under his watch was engaged in acts of personal favoritism, I'd want to check it myself: as we've seen elsewhere, people of these personality types cannot help abusing the power that they've acquired...and I somehow doubt the NSA is stocked only with angels.

Re:Slashdot Propaganda Machine Working Overtime

[slick7]

Truth and innocence are the first casualties of any conflict. The real truth, is that the government should fear its people, not the other way around. Government should nurture the population and not herd them to slaughter. When you use fear and sanctions against the people, a time will come when the people have nothing left to lose and nothing left to fear, this is when the people become the most dangerous.

Re:hmm...doctors just don't worfk as hard

[uglyduckling]

No, because the point is that the false positive results lead to more invasive tests (which in themselves may do harm), over-interpretation of other physical signs, worry etc.. The parallel with terrorism is that people end up on no-fly lists, get invasively searched and questioned, might get turned down for jobs or credit etc.. The uselessness of screening tests for low prevalence diseases is well known in the medical world, which is why tests need to be targeted to a high-risk population to have any value.

Flawed Analogy

[the eric conspiracy]

Correctly done, Medical testing is made more accurate by gathering additional data.

Basic tests are generally inexpensive but have a pretty high false positive rate. The key here is to have a very low false negative rate first and then minimize the false positive rate with additional tests.

If a positive result is obtained additional data is gathered using different tests aimed at eliminating the false positives. This additional testing is often more invasive and expensive, however it drastically reduces the number of false positives.

The premise this article is based on is just repeating the initial screening over and over. That's not what happens.

Re:Flawed Analogy

[Davak]

When you screen huge masses of people needlessly, almost all to all of your hits are going to be incorrect. Additional testing of these false positives are harmful. Biopsies, radiation, no-fly lists -- harmful.

Nobody is saying that we should never wiretap if we have evidence. That's testing a small population. The problem here is that we are wiretapping everybody to attempt to find evidence.

Re:Flawed Analogy

[sjames]

Unfortunately, in the case of the NSA, that's exactly what happens. If you show positive on one of the screening tests for terrorism, boom! you're on the no-fly list. It's like immediately giving full chemo and whole body radiation if the nurse thinks a mole looks a bit suspicious.

Even with properly used screening tests, if the followup test is expensive or invasive, it's often better to just skip it if you have no particular reason to suspect you have a given condition.

Re:Flawed Analogy

[Alomex]

If a positive result is obtained additional data is gathered using different tests aimed at eliminating the false positives.

If the additional test is too expensive, then at least for the case of illnesses you wait for the first indicator to be something else like "doc, I'm not feeling ok". By now if the test comes positive you have two indicators of a possible illness which now makes the expensive test worthwhile.

Re:Flawed Analogy

[AHuxley]

Re The premise this article is based on is just repeating the initial screening over and over. That's not what happens.
We do have some historical pointers. The CIA, MI6, NGO, faith based support for protest movements sent into "sealed" 1980's Eastern Europe.
Printing equipment (small and large scale), tv/radio broadcast efforts, books, Bibles, capturing images of life under house arrest.
Every container, bag, box, person, car, van, truck would have to be searched entering - no fun if you want hard currency, export friendly factories and have new loans to repay.
The police and security forces could not politically afford to incorrectly “diagnose” people as dissidents. Moscow wanted daily reports too.
So you watched anyone of interest for a really long time no matter the cost, lack of any progress or hardware limits.
The police and security forces offered the leadership a deal - time and new equipment to map all foreign involvement - UK/US NGO, diplomats, spies and locals all caught for that perfect show trial.
For every year spent mapping out/infiltrating/trying to move up the ranks of protesters, creating fake charismatic protest leaders - the printing equipment runs, tv/radio broadcast spread -books, Bibles seem to be pouring in.
In the end the State blinks, years wasted on 24/7 surveillance of a family wanting an exit visa, suicides in jail with sealed coffins and extra "guests" with visible video equipment at the funeral, torture and liquidation in a very public way of prominent dissidents, liquidation of escaped dissidents in the West.
Once the bureaucracy starts with a file the actions and results can become self-fulfilling and budgets run away.
What can the USA do with all its massive global data mining results? Pass on the results on and pay an informant to get close?
A plot is suggested, arrests, press conferences and convictions are handed out by courts with very good odds for not understanding terms like "entrapment".
Initial screening over and over pays off for any State and selected private contractors- you get to follow people and shape the future in any way you want. Eliminating the false positives can also get very tricky with so many infiltrating/freedom fighter efforts. Is that really bad group cover for your best freedom fighters from two wars ago and been protected for the next small war? Or one of your best informants?

Re: Flawed Analogy

[catchblue22]

Some doctors have a name for what happens when modern hi-res imaging causes doctors to think benign irregularities may be harmful: getting "VOMIT"ed on...ie "Victim Of Medical Imaging Technology".

Re:Flawed Analogy

[Anonymous Coward]

The premise this article is based on is ignoring statistical inference.
This guy is a terrible doctor.

Re:Flawed Analogy

[Anonymous Coward]

The doctor in the story is an idiot. The problem is not with collecting too much data. The problem is with the incorrect interpretation of that data.

Re:Flawed Analogy

[anagama]

The problem here is that we are wiretapping everybody to attempt to find evidence.

Honestly, I think the Feds know that collecting huge amounts of random data makes the job of finding bad people harder, not easier. But the point of the program isn't about finding bad guys, it is mainly to create a repository of information that can be accessed whenever they want to silence critics.

They don't care if they send you to prison because of your activism itself, they just want you in prison. This data collection coupled with a Federal code base so vast and vague as to be unknowable, basically ensures that everyone is a criminal and makes it trivial to suppress dissent simply by rummaging through the data store, finding some random bit of nonsense, and charging that person with 50 years worth of bullshit. Or as Snowden would say, it's "turnkey tyranny."

Re:Flawed Analogy

[the eric conspiracy]

No-fly lists aren't a test. It's more like a vaccine which we don't know what the positive effects of. The negatives are well publicized though.

Re:Flawed Analogy

[the eric conspiracy]

Either way you are gathering more data, not less.

The analogy this article is based on is not valid.

Re:Flawed Analogy

[AthanasiusKircher]

When you screen huge masses of people needlessly, almost all to all of your hits are going to be incorrect.

Yes, this is something that apparently even most doctors don't understand. Suppose who had a simple problem like this:

1% of women at age forty who participate in routine screening have breast cancer. 80% of women with breast cancer will get positive mammographies. 9.6% of women without breast cancer will also get positive mammographies. A woman in this age group had a positive mammography in a routine screening. What is the probability that she actually has breast cancer?

The correct answer (calculated from Bayes' Theorem, or simple logic) is 7.8%. Most doctors cannot do this problem, and that not only get the answer wrong, but they often get it wildly off -- estimating the answer to be much greater than 50% (often 70% or so, probably from simply subtracting the two numbers).

If you don't believe me, have a look at this link. As the author says there:

usually, only around 15% of doctors get it right. ("Really? 15%? Is that a real number, or an urban legend based on an Internet poll?" It's a real number. See Casscells, Schoenberger, and Grayboys 1978; Eddy 1982; Gigerenzer and Hoffrage 1995; and many other studies. It's a surprising result which is easy to replicate, so it's been extensively replicated.)

The author here is being generous. I looked at these studies years ago, and many of them show only 5-10% getting the answer to such problems correct.

And if this is true of physicians, it's probably true of just about anyone else who encounters a lot of false positives and isn't used to thinking statistically. That means most people are very likely to draw incorrect conclusions about the prevalence of something when the false-positive rate is high... making those using the methodology assume that (1) their methodology is better than it is, and (2) that with more "assumed positives" from incorrect logic, the incidence of whatever they're looking for in the population is higher than it is.

Re:Flawed Analogy

[Alomex]

You gather less data as compared to the test everyone option.

Re:Flawed Analogy

[Anonymous Coward]

yes biomedical science of the last few decades has been crap. Name one unequivocal success...

Re:Flawed Analogy

[DMUTPeregrine]

Let's say there are about 1000 terrorists in the US at any given time (likely a vast overestimate.) There are about 300,000,000 non-terrorist people in the US (319 million total US population as of the last census.) Let's assume a hypothetical test for being a terrorist is given to everyone in the US, which detects 99.9% of all terrorists, and gets false positives 0.001% of the time.
999 terrorists will be caught by the test. 1 will go free.
3000 innocent people will be caught as terrorists.

That's part of why blanket testing/surveillance/TSA style airport security and similar mechanisms don't work well: they lose the real results in a flood of false positives. And that's with an unrealistically low false-positive rate, and an unrealistically high number of actual terrorists in the US.

Re:Flawed Analogy

[gringer]

human genome project

Re:Flawed Analogy

[Anonymous Coward]

But where's Snowden's years of bullshit?

The US didn't even bother to give very good reasons for why Snowden is such a high priority and worse Spain, France etc didn't even require very good reasons to block an ambassador's plane just because Snowden might be on it.

They don't even do this sort of thing for known murderers.

Re:Flawed Analogy

[hackwrench]

No, the problem is confirmation bias and vast amounts of historical data being ignored and effectively being treated as stale in favor of scant scraps of more recent data, then in turn being both intentionally and unintentionally misinterpreted,and used to justify redidulous expenditures to collect as much data as possible,which again is both intentionally and unintentionally misinterpreted and secretly dispersed with the intention to benifit a small minority of people with the effect of harming the whole, ironicly even the minority that was ssupposed to benefit. The moster's hunger is never sated by gorging on that which is external to it; it must appese its hunger by devouring itself as well.

Re:Flawed Analogy

[gzuckier]

viagra

cure for cancer?

[Anonymous Coward]

Surely you jest... Curing cancer would be a financial disaster that would kill the US financial system. The actuarials are built on an 80 year life span average, now extend that to 120 + years and see what happens to the investment funds that back insurance and retirement plans. A stock market dive.

Re:cure for cancer?

[Anonymous Coward]

A small price to pay.

good for the Affordable Care Act (ACA) as without

[Joe_Dragon]

The databases become a nice preexisting conditions and Recissions list.

Re:hmm...doctors just don't worfk as hard

[Davak]

If only I had mod points to give... thanks.

Let's not be rational about this ...

[Alain Williams]

As your elected representative let me enlighten you as to why you voted for me rather than the other guy:

* I made good, powerful speaches. I went to some classes to help with this, it is more important that I dress in a good suit and have a strong voice than what I say makes sense.

* I avoided checking facts when making opinions. If you know the facts you realise that things are not black & white, but to express that makes people think that you are a ditherer, that you don't know what you stand for. Who wants a politician who, when asked a question instead of saying ''yes'' or ''no'' says something long and boring that starts with ''It depends'' ?

* Most of you don't look at the facts, you work on gut feeling and gross extrapolation. You remember that story in the local press last week about the thief from out of town who had green eyes, blond hair and a limp ? Yes: you are quite right to know that everyone from out of town with blond hair & a limp is a good for nothing crook and we don't want people like that round here!

* You people just want to be safe. You don't care what happens to out of townies, how hard we make it for them; or even foreigners -- some of who have a skin of a funny colour. They just don't matter!

* You don't really know what safe means, but are happy if you can still watch TV and drink beer when supporting your team. My predecessor did not do anything to make you realise that you can do something else, neither will I --so you will vote for me next time.

* In order to get on the short list for election I had to sign up to what the party says. They won't listen to a newbie like me, if I ask questions there are plenty of others to choose from who do what the party bosses say.

* Do you know how much I got in ''research grants'' and travel ''expences'' from the large corpotations? To say nothing about my fee for 2 days work a year as a consultant. I must not upset them by saying something that upsets them. All that money buys a lot of publicity as well as letting me buy that new yacht..

* I have a good friend who knows people, (I don't want to know why they are), but I got warnings of the other guy's plans and it was mighty useful when his campaign manager was caught in bed with that young ... that no one had seen before

So you see, I would be really silly if I upset the status quo and made you think for yourself.

Snowden For President 2016

[Anonymous Coward]

You're also missing a much bigger problem with his choice. When you collect data on a disease, the disease doesn't take steps to avoid being in the data. It's not a sentient being like terrorists. The terrorists aren't in the dataset he's collecting, because he's collecting the low hanging fruit in the easy to process data formats.

General Alexander decided simply to store it all, and so that is what happened. Terrorism is just the excuse. IT WAS NEVER HIS CHOICE TO MAKE. He was never given the power to flip a political system from 'innocent by default' to 'suspect-of-terrorism' by default. I don't believe the FISA court had that power even. It was a political decision that needed a change of constitution. A political decision that never happened because what he's doing is not legal within the constitution and nobody has the balls to fire him.

I think you need a president who will restore the constitution, and that ain't Obama. He doesn't even seem to be in charge at this point, just some sort of PA who makes the phone calls.

No, you want a President to fix things? Vote Snowden 2016.

Re:Snowden For President 2016

[gweihir]

Hehehe, nice sentiment. Unfortunately, the opposite could happen, namely a president could be elected that thinks all this surveillance is a good idea and to start using it against dissenters and "undesirables" in the population. And political opponents as well. The date collected is ideally suited to these purposes. If that happens, the US will make Nazi Germany look tame.

Ultimately, the same motivation

[Rambo Tribble]

Whether excessive medical tests or excessive surveillance, the minions happily promote it to ensure their job security. If the patient or the society suffers, well, that's okay. Perhaps a bit regrettable, but okay.

Ultimately, a society that strenuously promotes competition also engenders a mercenary attitude. So, you see, the excesses of Wall Street are not that far removed from the excesses of the NSA, or Microsoft, to pick but a very few examples.

Re:Ultimately, the same motivation

[Anonymous Coward]

I wonder if that's just due to promoting competition, or more specifically rewarding being right vs punishing being wrong? If Wall Street had been punished as they should have been, if the government agents and departments had been punished for their transgressions appropriately (I don't think it will ever happen, hence my wording) would people learn a different lesson? I think some personality types probably wouldn't, but overall maybe the different industries would?

True positives versus False positives

[Anonymous Coward]

"Each time a test is run, the number of people incorrectly identified quickly dwarfs the correct matches"

This is simply not true as a universal statement. It is only true if the false positive rate of a test exceeds the true positive rate. While screening tests are selected for their high sensitivity and may suffer from lower specificity and therefore higher false positive rates, no one is labeled as having a condition until a confirmatory test with higher specificity is positive. I can't speak to how terrorists are labeled, but please don't drag medicine down into that morass.

Re:True positives versus False positives

[jc42]

..., no one is labeled as having a condition until a confirmatory test with higher specificity is positive. I can't speak to how terrorists are labeled, but please don't drag medicine down into that morass.

Here in the US, we're still suffering from a case of just such labeling around 20 years ago. That was when we started seeing widespread suggestions that people should avoid being tested for AIDS. The explanations was that the first two widespread tests had false positive rates of about 10% and 5%. The actual incidence at the time was somewhere around 1 per million. So, it was explained, if the entire US population were given the first test, it would catch around 300 people with AIDS, and would also finger 30 million people with a false positive, which would be a disaster for most of them.

Now, it was probably true that the medical people wouldn't take any of the positives seriously, but would just note it in their records, and proceed with the second test, which would find maybe 30 more true positives and 1.5 million false positives. They would then do further tests. But this didn't matter to most of us.

The problem was that all these positives went into a record system that wasn't at all secure, and there were no legal repercussions against people like employers and insurance agencies who found out about the positives. As a result, there was a growing epidemic of job losses, insurance cancellations, etc. Those false positives followed people around, and many of them are still suffering from the results although they never had the disease.

True, this isn't the fault of the medical system. But being "put on a list" has a history of meaning the end of employment and all sorts of other punishments for the rest of one's life. Look up the history of the "Red Scare" of the 1940s through the 1970s for details and many examples.

The only real answer is to impose serious legal penalties on companies, agencies, and others who act up on such positive tests without verifying their validity. So far, the US has very little in place to do this, for either false medical reports or false terrorism reports. So any of us can easily become a victim of a false positive, regardless of our actual status.

(I was tempted to add "unless we're rich and powerful enough", but the story of Ted Kennedy being blocked at an airport because he was on a no-fly list says that that might not even protect you. ;-)

Re:Slashdot Propaganda Machine Working Overtime

[ericloewe]

Where's the "-1: conspiracy theorist" option?

Bruce Schneier pointed this out 6 years ago

[Anonymous Coward]

This is a pretty classic statistics problem called the base rate fallacy. I've seen it come up pretty often when doing Bayes theorem exercises. Bruce Schneier pointed out the pitfalls of data mining back when the Bush Administration was pushing the Total Information Awareness program:
http://www.schneier.com/blog/archives/2006/03/data_mining_for.html

Re:Bruce Schneier pointed this out 6 years ago

[gweihir]

Indeed. The only rational explanation is that the NSA surveillance is not about fighting terrorism, but about identifying dissenters, independent thinkers, etc. These people are undesirable in a police state and that seems to be what all this is aimed at.

This analogy makes no sense

[Anonymous Coward]

Of course more data makes it more likely that they'll catch more terrorists. The question is if that is worth it. If nothing else, they will have a large amount of information about a newly discovered terrorist immediately already on file, so they can instantly track down any co-conspirators who might think they can't be found and who are planning something nefarious.

Data-mining should also be valuable. If they investigate 10 000 people per year, then more information used properly will increase the probability that terrorists will actually be among those 10.000 people. The flaw in the analogy is that the doctor here is assuming that everything mildly suspicious will be investigated, as with medical matters, but the capacity for investigation is independent of the amount of information. All that will happen with more information is that the selection of who to investigate will be better.

This NSA overreach and spying is terrible for the country because, but that's mainly because of the possibility of misuse of the data and the international repercussions. The issue pointed out in this story isn't really an issue at all in this context. Criminal investigations are not done the same way as medical decisions, mainly because the criminals who go free will not sue the police department if they are guilty but not investigated. Huge difference.

Re:hmm...doctors just don't worfk as hard

[Anonymous Coward]

And this in itself will create more "terrorist" because some of the people who end up on a no-fly list, lose their job and credit, wife and kids probably too; will be pretty mad and decide to do crazy things. Then more surveillance is needed and more control. The more control the more people will become "terrorists" true or not and repeat until either the terrorists on one side win or the other terrorists.

Haven't people seen enough science fiction movies by now to realize everyone is a terrorist if you push them hard enough.

https://en.wikiquote.org/wiki/WarGames

Joshua/WOPR: Greetings, Professor Falken.
        Stephen Falken: Hello, Joshua.
        Joshua/WOPR: A strange game. The only winning move is not to play. How about a nice game of chess?

Here's the same argument about drug testing

[nbauman]

Here's another doctor who made the same argument about testing for illegal drugs. Be sure to catch the distinction between screening tests and diagnostic tests.

http://www.kevinmd.com/blog/2013/07/drug-testing-considered-screening-tests.html
Should drug testing be considered screening tests?
Chris Rangel, MD | Conditions | July 12, 2013

... The problem of a false positive test is frequently encountered in the practice of medicine. Depending on the clinical circumstances and the nature of the initial test, follow up evaluation with more expensive and possibly more invasive testing is often required in order to verify the results. For example, an abnormality found on the chest x-ray of a smoker with a bad cough requires further evaluation. A CT scan of the chest, bronchoscopy, and even a needle biopsy to obtain a tissue sample for analysis are required before making a diagnosis of lung cancer and starting treatment.

However, the possibility of a false positive drug screen and the need for further testing and evaluation is rarely considered outside the context of clinical practice. Employers, school administrators, government agencies, and law enforcement can and do consider a positive drug test to be perfectly equivalent to an admission of illicit drug use. This frequently results in the administration of some form of punishment or corrective action being delivered without giving the accused the right to defend themselves in any way. Essentially, drug testing is an effective way to violate a person’s right to due process since most drug screening is managed by lay people in non-clinical roles who believe that drug testing is 100% reliable. But this would be the same absurdity as giving chemotherapy to the smoker with the abnormal chest x-ray without first trying to verify the diagnosis with further evaluation (due process).

The other problem comes from the mass drug testing of large numbers of people (either random or at the initial point of contact). The interpretation of the results of a medical test are never as simple as positive or negative. The statistical probability of a false positive or a false negative result must be considered in concert with the pretest probability....

Re:Slashdot Propaganda Machine Working Overtime

[wmac1]

Yes, specially with the revealings of the recent weeks, it is impossible to think about any other wrong doing of governments and spying agencies (i.e. conspiracies).

Annual physical exams are problematical too

[Paul Fernhout]

Both for false positives and ineffective treatments: http://www.drmcdougall.com/misc/2005nl/july/050700physical.htm
" The annual physical exam is an intensive, well-orchestrated, experience designed to make apparently well people, sick (with good intentions). You walk into the doctor's office as George or Francine and you leave as a breast cancer, prostate cancer or heart-disease victim. The initial exams commonly lead to more tests â" some of which are painful, disfiguring, and dangerous, such as mammograms, breast/prostate biopsies, colonoscopies, and angiograms. Ultimately, the costs of all this meddling can make you homeless and take away your life savings.
    The annual physical is supposed to be a means of prolonging your life â" and it could have been, except for the fact that the treatments that follow the initial exam are at best useless, and at worst, dangerous. Let me give you two fundamental reasons why the annual physical is doomed to failure, and because of lack of real life benefits all major health organizations have recommended against it: ...
The goal of every patient should be to remain out of the health care system. This is accomplished by staying healthy. This highly desirable state is not simply a matter of good luck, but rather a result of your behaviors; more specifically, following a low fat, plant-food based diet, getting moderate exercise and having clean habits. ..."

However, note that such advice is also in the context of teaching people how to avoid most disease through better nutrition.
"3 Biggest Mistakes People Make in Their Diets - Dr. John McDougall"
http://www.youtube.com/watch?v=OF7Uanr-lYA

See also on why we don't change because we're invested in a belief system (cognitive dissonance):
"Mistakes Were Made (But Not by Me): Why We Justify Foolish Beliefs, Bad Decisions, and Hurtful Acts"
http://www.amazon.com/Mistakes-Were-Made-But-Not/dp/0151010986

Re:Annual physical exams are problematical too

[Anonymous Coward]

Obamacare is making annual physicals free, thereby increasing our medical costs and doctor profits.

Wrong analogy.

[will_die]

If you want a medical to NSA the analogy would be they test someone and they have a sexual transmitted disease and it shows positive. The person says I had sex with person X so the medical team says we need to check with person X and see if they actually had sex with the person and if they did if they got the disease. If so then person X needs to be handled as required.

Re:Slashdot Propaganda Machine Working Overtime

[anagama]

Maybe it should be "+1 conspiracy theorist"

I've had my fun at the expense of foil hatters in my day, but recently I kinda wonder.

Bad on purpose

[Alomex]

The NSA is full of really smart people. There is not much we can come up with here they haven't thought of. The problem is that they are not being evaluated by how many attacks they stop (see the Boston bombings). They get measured by how active and busy they look.

Political ignoramuses consider a short, narrow targeted no-fly list a failure (picture Bush Jr in the oval office: "you've only found 100 people after I gave you 10Gigadollars???") while they are very impressed with a 100K long no-fly list ("you are catching so many of them! good job! here's another 3Gbucks").

Re: Bad on purpose

[Anonymous Coward]

You really don't understand what the NSA does. They provide information to people who make decisions, that's what intelligence boils down to. They are doing their job well, or they'd be shaken up and we'd hear about new agencies being formed or other agencies starting to do SIGINT themselves.

They are not judged by how many terrorist plots they thwart, speaking frankly, only people like you judge them in that way.

Thanks

[Anonymous Coward]

Thanks!

I use the exact same analogy now for over 10 years. Does anybody listen? No.

Is far worse than that

[gmuslera]

When you weaponize computers everything could look as a disease. Running a trojan (or worse, removing a software that is a government trojan), receiving spam message, doing a "funny comment", or just someone else playing social engineering could put you in the enemy of the state list. The fake version of the disease is the one viral, not the disease per se (even if the government is trying very hard to have sick people to justify what they are doing)

In medical terms, what is being perpetrated is creating a new disease, forbidding doctors to make a vaccine, intentionally release in all the world, and asking foreigners to not bring that disease here, because we won't develop any cure even if we want because that cure could be used elsewhere.

The NSA does not care anout terrorists

[gweihir]

Unless they are terminally stupid, they do not. Surveillance does not actually help against terrorism, and the NSA does know that very well. Terrorism is just a convenient pretext (i.e. lie) to justify the surveillance. What they are really interested in is profiling every person they can get data on and identify dissenters, independent thinkers, etc. as these can threaten a police state, as the US is more and more becoming.

What they are overlooking is that this is extremely dangerous. Just have one president go off the deep end, and the US will make Nazi Germany look tame. All the surveillance and population control mechanisms are already in place. The police is already used to shoot citizens as a matter of routine. Prisons are in ample supply. The only thing missing is the madman at the top. It will be just a question of time before that one is found.

Liability determines the motivations

[mounthood]

When doctors say it's bad to collect too much information, they're talking about medicine not liability. Liability determines the motivations, and tells us how both doctors and the NSA will act:

If Doctors or the NSA don't identify someone: Major liability (although doctors only have to ID patients they encounter, not everyone in the general population)
If Doctors or the NSA have false positives: No liability (because it was an honest mistake, by people doing their best)
If Doctors or the NSA don't treat/investigate someone who's identified: Major liability
If Doctors or the NSA treat/investigate someone correctly: No liability (even if the patient dies or the person becomes a terrorist)
If Doctors or the NSA are negligent when treating/investigating: Major liability (NB: negligence is NOT determined by outcome, but by professional standards)

When politicians hold agency witch-hunts after disasters they make the liability unlimited, and we get the NSA breaking laws (even the constitution!) to adjust.

Wrong.

[Anonymous Coward]

They're not concerned about any rules at all since they have enough loopholes to do whatever they like, down to largely making their own rules. But they're not in the business of treating anything.

The NSA will take a picture, possibly put it on the (secret, internal) wall as an interesting example, maybe push a copy to some analyst who may or may not shove it on through to someone in some other agency to eventually take a look at, and move on to the next thing to take a picture of. Just in case someone somewhere might want to look at pictures of interesting medical cases later. They do have the most extensive collection of medical pr0n on the planet, only it's secret. Useful, no?

people incorrectly identified dwarfs

[Anonymous Coward]

Did Steve Ballmer write "people incorrectly identified quickly dwarfs"? While this phrase may be part of a coherent sentence, groaners like this don't pass the eyeball test. Why not use simple, direct wording?

"Each time a test is run, significantly more people are incorrectly identified than are correctly matched." - doesn't use more words or take up more space, but does not have the groaner.

Another great video on this by Dr. McDougall

[Paul Fernhout]

"Avoid Doctors to Protect Your Health!" http://www.youtube.com/watch?v=ahELa5oYrkM

To be clear, at the link in my previous post there are a few specific tests McDougall suggests for early cancer detection (including visual exams of the skin for melanoma).

Glad I'm not his patient

[Anonymous Coward]

So this doctor really refuses to do a test if the false positive rate would be high? The right thing to do when a test comes out positive is to do a second test. The reason why you do the less acurate test first is that it's less expensive and/or less risky, so even if it produces some false positives, it still greatly reduces the number of individuals to undergo the second test that is more expensive or intrusive. The decision what second step to take is still the patient's and the doctor's responsibility. There is no automatism. The more data this decision is based on, the better.

I can't believe this guy thinks complete ignorance is better than some knowledge. Isn't M.D. a scientific degree? What scientist would say: "If the outcome won't give me 100% certainty I'd rather not do this experiment at all"?

Having said that, the analogy with governments spying on people is totally flawed. A medical test is a privacy concern, but the patient consents to being tested, and the test result is (should be) confidential. Government spying is done without the knowledge or consent of the person targetted, and without consent of the so-called sovereign, the people. If there weren't some people with a conscience inside the apparatus, we wouldn't even know about it. The analogy should be: Don't test me without my consent : Don't spy on me.

From a Criminal's Perspective

[LifesABeach]

When Law Enforcement sits on its ass and reads a screen, few things are more useful.

Needle in a haystack?

[Anonymous Coward]

Obviously seems that it would be harder to find a needle in a haystack if you make the haystack larger......

But then again this is almost certainly backed up with some high powererd filtering algorithms that eliminate most of the hay immediately..... so maybe the analogy isn't as appropriate as it seems due to the existence of, you know, fast data processing.

Me, I guess some fairly crude URL list based and keyword based stuff, backed up with some strong human intelligence which is quite successful in identifying the really risky idiots out there is what's really at the heart of all this.

Therefore you can leave your tinfoil hat off when visiting hilariouscatvideos.com - their budgets do not really stretch to employing tens of thousands of staff to monitor most stuff in the detail many people imagine.

Having said all that, you might have just flagged yourself onto the "idiot" scoreboard by doing so, and who knows what consequences that might have the next time someone with access to it leaves their laptop on a train?

 

Re:Slashdot Propaganda Machine Working Overtime

[Anonymous Coward]

the conspiracy nutjobs are suggesting that there is some sort of useful policy in blowing up innocent people. That requires some serious extraordinary proof.

Abdulrahman al-Awlaki

Reliable tests

[manu0601]

TFA has some merits, but also some limits, as some medical test are quite reliable. For instance, I do not see how we are going to make fake positive by screening vitamin D levels.

Mission is to keep money flowing

[Anonymous Coward]

"The NSA" would not "rather" do anything, because "the NSA" doesn't make up missions for itself

The purpose of NSA (and any other government dept.) is to perpetuate itself. So if some politicians want NSA to "catch terrists", they will "catch terrists". Their real existence does not matter.

Back in the days of KGB, politicians in Soviet Union were always telling them to "catch spies". It did not matter that there were no spies to catch. But the KGB did catch spies in their internal security division. You know, local informants saying that their estranged neighbor or whatever was a spy, etc. etc. But numbers are numbers.

Another case would be Florida police force and war on drugs. They catch the "little guys" (users). They they make a deal with the little guys that if they snitch on dealers, and others, they will get reduced sentences. Since users and other small timers don't really know anyone except other drug addicts, they end up just snitching on other users. The police get more arrests. They look good. The little guys go in jail for a long time. Rehabilitation - none.

http://www.bbc.co.uk/news/magazine-21939453

So, as long as NSA provides names to the list(s), they remain relevant. They remain the "doers" and the money keeps flowing. Hell, they even got the late Sen. Kennedy on the list and that was OK.

http://www.washingtonpost.com/wp-dyn/articles/A17073-2004Aug19.html

Bayes Theorem

[pugugly]

I had the good fortune to run across Bayes Theorem (Not by name) in an article about misdiagnosing problems in Discover magazine back in the 80's, and for some reason filed the factoid away as 'Oh, this is *important* and is going to apply to a lot of things' and have never forgotten it.

The fundamental takeaway for me is "It doesn't *matter* how accurate your test is - what matters is how accurate it is compared to how rare the condition you're looking for is.". Random drug tests, random highway stops, the instant you are doing anything that force a 99% accurate test on a population that might only be 1% guilty, you should be fined for a violation of Bayesian logic.

It is one of those universally applicable truths, and we need to hammer it into the brain of every teenager before the get out of High school.

Pug

Sharpshooter Falacy

[IBitOBear]

You know, when people talk about who was warned about what, they completely forget the sharpshooter falacy. Warn everyone about everyone, then when some one does some one thing you can say "you were warned" because, in the huge pile of everything-squared you can find that nedle in the nedle-stack.

Now all the people who pointed at the nedle demand a bigger nedle-stack full of smaller and smaller nedles.

More signal. But more noise. And more noise per each increment in signal.

And more blame to go around.

There was a song, it has a point. "You have to hold-on loosly but don't let go". There was a movie, and it has a point "the more you tighten your grip the more systems will slip through your fingers." It's like there are all these old aphorisms and they came about for having truth within them. The truth of moderation.

More isn't better, it likely never was.

Re:Sharpshooter Falacy

[Thanshin]

*needle.

Re:Slashdot Propaganda Machine Working Overtime

[hackwrench]

I would counter that no one in their right mind is 100% certain that no false flag operation have been done. The question is not whether there is some sort of "useful" policy, or even whether aq givengiven operation is a false flag op, but whether false flag operations serve the ends of people with the capacity to make them happen, and given recent and not so recent historical events its even more irrational to suggest or imply that no publicscrutiny should be made as to whether false flag operationsw have been run.