Thank you for saying this. The Puritans get a bad rap, in part because many people assume there is an unbroken line from them to the modern religious right. In fact, Puritans would have had deep theological disagreements with the modern evangelical movement, which has its own origins.
The Puritans were a diverse lot, but among them were some of the early heroes of church-state separation. Take Roger Williams. His religious beliefs were so extreme that he thought even most other Puritans were going to hell. But he was also a vehement proponent of church-state separation. He thought that moral behavior should be freely chosen, not compelled by the state; and like many Puritans, he had seen firsthand back in England the corrupting influence that church-state mingling could have on both sides. He founded Rhode Island after a rift with other Puritans, and made it a safe haven for people of many different religions (including one of the first settlements of Jews in North America). That tolerance is as much or more a part of the Puritan legacy as "The Scarlet Letter" and religious prudery.
Not to throw water on the fires of righteous indignation, but... did this incident actually happen?
There are no links in the summary. I tried searching Google for phrases quoted in the summary. I couldn't find anything that wasn't a repost or link back to this Slashdot thread. No sight of the original forum post. Granted, it may not be indexed... but it's a little weird.
(The reason I went looking, BTW, is that it isn't clear from the summary whether this was a college professor, which everybody seems to be assuming, or a high school teacher, which seems more plausible to me. I have trouble imagining a college-level instructor even trying, never mind getting away with this. By contrast, I have little trouble imagining this sort of story being spread without verification.)
For similar reasons, many businesses protect themselves with policies that they do not even look at unsolicited submissions of ideas. That's true in the tech world as well as other creative fields. If you allow yourself to be exposed to somebody else's ideas, and later you come out with something similar, the onus may be on you to prove that you weren't influenced by them.
Why not extend this test to national-security information? I bet if you ask the military, CIA, NSA, etc. whether they would consider their own "meta-data" as less sensitive than the actual content of messages, they'd say no. The history of traffic analysis in military and foreign-policy applications is a pretty good indication of that.
But whether you'll observe consistent responses to stimuli will depend on whether you're (a) measuring the right responses and (b) using the right stimuli. In this case, the stimuli were video images. Other researchers have found personality differences when using real stimuli. Maybe there is something about video stimuli that overwhelms individual differences?
Where is this "scientific fact" that you speak of? The only "emotionally unstable" teens that I've read about...
If you search Google or Google Scholar for "teen brain development" you will find some relevant information. Like this or this or if you've got access, stuff like this.
A lot of scientific attention has focused on charting the ongoing physical maturation of the frontal lobes through adolescence and even into early adulthood. The frontal lobes are involved in executive functioning, which includes things like self-regulation and impulse control. The frontal lobes are also involved in self-monitoring, which interestingly ties back to the grandparent poster's statement:
I'm sure you didn't think you were emotionally unstable, teens generally don't, doesn't mean you weren't.
Do we really want to saddle ourselves with more debt?
The counterintuitive but economically sound answer is yes.
An economic downturn is precisely the wrong time to be balancing the federal budget. On a personal level, yes, right now individuals should absolutely be preparing for difficult times to come. But as a matter of government policy, the government should spend during bad times, to buffer the effects of a sluggish economy. The time to save (pay down the debt) is when things are going well in the private sector. (And if recession-era spending is done wisely -- for example, investing in infrastructure -- then the benefits will carry into later boom times without new money being spent.)
It would be a little easier to stomach spending now if our current president had been following this advice and running surpluses during the boom (like his predecessor did). But that doesn't change the fact that balancing the federal budget right now will make things worse.
What many people don't realize is that detection procedures with very impressive-sounding statistical properties generally do horribly at catching rare events.
Imagine some very impressive numbers. Suppose that this procedure has 99.999% sensitivity -- it catches nearly every wannabe terrorist who tries to board a plane intending to do harm. And suppose it also has 99.999% specificity -- out of 100,000 innocent passengers, 99,999 will be correctly identified as innocent, and only 1 will be a false alarm. Sounds good, right?
Not really. In a given year, only a very small number of passengers are wannabe terrorists -- say, 10 per year. (That's probably high.) On the other hand, there are 1.6 billion air passengers per year (that may be a low estimate, since it's a 2000 number). So if this were implemented worldwide, then in a given year, we can assume that this profiling procedure will flag 160,010 people as terrorists. Only 6 x 10^-5 of those will be actual terrorists.
Of course, those hypothetical sensitivity and specificity numbers are unrealistically, ridiculously good. With more realistic numbers, the problem gets much worse. Even if the detection procedure is very sensitive and very specific -- and I doubt that it is -- the low base-rate of terrorism means that an enormous number of people will be falsely accused of being terrorists.
If you want to show causation, then you must have a model and you must subject it to experiment.
As pointed out earlier in this thread, the linked article is about a randomized experiment.
Many studies take correlation coefficients of 0.5 to be "significant". What a joke.
Imagine that you have been diagnosed with a potentially fatal disease. The research shows that if you go untreated, you have a 75% chance of dying within a year. Your doctor offers you a treatment that, in well-run clinical trials, has been shown to reduce the chance of death to 25%. Do you take it, or dismiss it as "a joke"? Because those probabilities express a correlation of r=.5 between taking the drug (vs. not) and living (vs. dying).
Medical and behavioral science study phenomena that are inherently complex, noisy, and non-deterministic. Your standards for what constitutes a meaningful effect size have to be adjusted accordingly.
To all the people tagging this "correlation is not causation," do you even know what you're talking about? This was a randomized experiment.
I'm not saying this is a perfect study -- there might be plenty of other things wrong with it. But the phrase "correlation is not causation" has an actual meaning. It is not just a synonym for "I had a kneejerk reaction to dislike this study but I can't say why."
And here is a somewhat more technical essay on why free trade is better than the alternatives even if your trading partners have horrible records on the environment and labor rights.
In his popular writing, including his NY Times column, Krugman is a pretty outspoken liberal on most issues. But within his academic expertise -- which is what he won the prize for -- he is very willing to depart from liberal orthodoxy if that's where logic and evidence lead him.
Especially interesting is that the work Krugman won his prize for is about global free trade. Like most economists who've seriously studied the issue, Krugman has concluded that free trade is unabashedly a good thing.
But in the current U.S. political climate, free trade is mostly being touted by conservatives and reviled by liberals. So if you're a conservative and you want to claim liberal bias, you have to account for the fact that Krugman got the prize for work you probably agree with. And if you're a protectionist liberal who wants to boast, you're similarly stuck.
And if you're just generally tired of ideologues crowing about victory or whining about bias when neither is deserved, you can enjoy the whole spectacle of people getting tongue tied when someone wins the Nobel prize on (gasp) the strength of his ideas.
Three reasons this is a bad idea. One, if you suddenly have to swerve to avoid something, you might miss the edge that the power steering would provide Two, you might need to suddenly apply engine power in some emergency situations, like if you hit black ice and start skidding. Three, if you accidentally turn the key too far back, you'll engage the steering wheel lock and won't be able to steer at all.
Absolutely. The summary and linked article get it wrong. The Ig Nobel prizes are not about "irrelevant" research:
The Ig Nobel Prizes honor achievements that first make people laugh, and then make them think. The prizes are intended to celebrate the unusual, honor the imaginative -- and spur people's interest in science, medicine, and technology.
From the Ig Nobel Website
A lot of scientific research seems pointless or silly to people who don't know what it's really about or why it was done. Hence the regular "still no cure for cancer" and "I can't believe my tax dollars fund this" comments. The Ig Nobel prizes acknowledge that science can sometimes seem funny on the surface, but they definitely do not concede that it is irrelevant.
If best means "in theory, the technology is extremely difficult to break when human beings know how to use it correctly and do so" then yeah.
If best means "in practice, it is likely to work effectively when used by ordinary actual human beings in the real world" then um, no.
Human beings are part of the security equation. Gnupg is a great piece of technology, but unless it is wrapped in an interface and set of procedures in such a way that Nina from accounting is able and willing to use it correctly every time, it will not improve security.
Richard Dawkins is an atheist who believes that science and faith are fundamentally incompatible, so it makes sense that he'd think that. But that is not a consensus view among scientists. Plenty of other people think that science is compatible with religion and spirituality because they address different concerns.
Unless the Royal Society is now taking positions on the acceptability of religion, there should be no consideration given, pro or con, to the religious beliefs or affiliations of its officers.
Agreed. If you read what Reiss actually said, it is clear that he was NOT advocating for giving creationism any scientific legitimacy in the classroom. Rather, he was giving some very sound and humane advice for how teachers can respectfully reach students who arrive in the classroom as creationists. Among his suggestions:
Students should be encouraged to voice their doubts so that teachers can deal with them in the open
Educators should see their role as making sure that students know the scientific methods, theories and evidence, even if the students' beliefs conflict with the science
Don't expect your students to abandon their long-held beliefs soon, or even at all. Hold them responsible for what they know, not what they believe
Everything in his essay seems reasonable to me. The fuss arose in part, I think, because attacks on the scientific community have forced scientists into such a bunker mentality that they acted irrationally (i.e., not like scientists)
Interestingly, David Foster Wallace wrote the book Everything & More about the work of mathematician Georg Cantor, who suffered from mental illness late in life. Around the same time as Wallace's book was published, another book came out trying to argue that Cantor's mathematical genius stemmed from his illness. In an interview, Wallace called it "dreck, the very worst kind of appeal to a flabby, unconsidered pop version of what you just now called the 'mad genius' syndrome." Wallace goes on:
...The fact is that it's all but certain that Cantor was bipolar, that his professional insecurities and travails aggravated the illness but didn't cause it, that most of his worst episodes and hospitalizations occurred when he was older and his best work was long behind him. Etc., etc.--some of the unsexy truth gets talked about in [Everything & More].
People like the "mad genius" meme, but it paints an overly simple picture of things. It might be the case (and empirically, the jury is still out) that some of the same predispositions that put people at risk of mental illness in some circumstances can contribute positively to intellectual functioning in others. But that is most definitely NOT the same as saying "with great genius comes great madness." If anything, the madness comes along and puts a sad end to the genius, rather than enhancing it.
I would draw a clear and bright line between what you say about David Foster Wallace ("he took the coward's way out") and others' reactions ("fawning news coverage"). The second point, I totally agree with you on. The first, I vehemently disagree.
We don't know why DFW committed suicide, and we might neverk now. But in the vast majority of cases, "cowardice" isn't even a relevant concept. Depression -- real, deep depression -- is not just about being in a crappy mood. Real depression (and other kinds of serious mental illness) messes you up so deeply that up seems like down and you cannot make rational sense out of yourself or the world. To call someone a "coward" implies that were faced with a choice and, with faculties intact, made a weak decision. Like I said, not a relevant concept for suicide.
And for the exact same reason, all the tributes making this into some sort of penetrating existential act of a man who saw the world too clearly... please! DFW was a brilliant thinker and writer, but his death is a tragedy and a loss. It is not an artistic act.
Interesting point. The linked site doesn't say much about its algorithm. But I'm guessing that it uses some sort of probabilistic matching of your rankings across many preferences, rather than an exact match. After all, they say right on the sight that you shouldn't worry about remembering things exactly. So I bet your preferences don't all have to be perfectly stable for it to work. Rather, enough of your preferences have to be just stable enough for whatever probability threshold they use.
As for guessing preferences from known behaviors like reading slashdot, I bet the predictability is lower than you'd think. Do all slashdot readers like video games? More so than non-slashdot readers? On average maybe so, but I bet there is huge variance in both groups. Additionally, if they ask about multiple preferences that have low joint probabilities, they can make the algorithm more secure. Knowing that you're a slashdot reader may help me guess that you are slightly more likely than the average person to like videogames, but it won't give me much traction on something like your favorite foods.*
* - Assuming they leave Cheetos and Mountain Dew off the list.
"This is no science. It is only the hope of a science."
- William James, The Principles of Psychology, 1890
There, fixed that for you. And seeing as how James is generally credited for establishing the field of psychology in America, I think he can be forgiven for limiting himself to some cautious optimism.
Slashdot Mirror
Before everybody starts tagging this story "correlationisnotcausation," let me just point out that this was a randomized experiment.
Carry on.
Can you opt out of this "service" if you don't want it?
Thank you for saying this. The Puritans get a bad rap, in part because many people assume there is an unbroken line from them to the modern religious right. In fact, Puritans would have had deep theological disagreements with the modern evangelical movement, which has its own origins.
The Puritans were a diverse lot, but among them were some of the early heroes of church-state separation. Take Roger Williams. His religious beliefs were so extreme that he thought even most other Puritans were going to hell. But he was also a vehement proponent of church-state separation. He thought that moral behavior should be freely chosen, not compelled by the state; and like many Puritans, he had seen firsthand back in England the corrupting influence that church-state mingling could have on both sides. He founded Rhode Island after a rift with other Puritans, and made it a safe haven for people of many different religions (including one of the first settlements of Jews in North America). That tolerance is as much or more a part of the Puritan legacy as "The Scarlet Letter" and religious prudery.
Not to throw water on the fires of righteous indignation, but... did this incident actually happen?
There are no links in the summary. I tried searching Google for phrases quoted in the summary. I couldn't find anything that wasn't a repost or link back to this Slashdot thread. No sight of the original forum post. Granted, it may not be indexed... but it's a little weird.
(The reason I went looking, BTW, is that it isn't clear from the summary whether this was a college professor, which everybody seems to be assuming, or a high school teacher, which seems more plausible to me. I have trouble imagining a college-level instructor even trying, never mind getting away with this. By contrast, I have little trouble imagining this sort of story being spread without verification.)
For similar reasons, many businesses protect themselves with policies that they do not even look at unsolicited submissions of ideas. That's true in the tech world as well as other creative fields. If you allow yourself to be exposed to somebody else's ideas, and later you come out with something similar, the onus may be on you to prove that you weren't influenced by them.
Why not extend this test to national-security information? I bet if you ask the military, CIA, NSA, etc. whether they would consider their own "meta-data" as less sensitive than the actual content of messages, they'd say no. The history of traffic analysis in military and foreign-policy applications is a pretty good indication of that.
But whether you'll observe consistent responses to stimuli will depend on whether you're (a) measuring the right responses and (b) using the right stimuli. In this case, the stimuli were video images. Other researchers have found personality differences when using real stimuli. Maybe there is something about video stimuli that overwhelms individual differences?
If you search Google or Google Scholar for "teen brain development" you will find some relevant information. Like this or this or if you've got access, stuff like this.
A lot of scientific attention has focused on charting the ongoing physical maturation of the frontal lobes through adolescence and even into early adulthood. The frontal lobes are involved in executive functioning, which includes things like self-regulation and impulse control. The frontal lobes are also involved in self-monitoring, which interestingly ties back to the grandparent poster's statement:
The counterintuitive but economically sound answer is yes.
An economic downturn is precisely the wrong time to be balancing the federal budget. On a personal level, yes, right now individuals should absolutely be preparing for difficult times to come. But as a matter of government policy, the government should spend during bad times, to buffer the effects of a sluggish economy. The time to save (pay down the debt) is when things are going well in the private sector. (And if recession-era spending is done wisely -- for example, investing in infrastructure -- then the benefits will carry into later boom times without new money being spent.)
It would be a little easier to stomach spending now if our current president had been following this advice and running surpluses during the boom (like his predecessor did). But that doesn't change the fact that balancing the federal budget right now will make things worse.
Actually, let's do...
What many people don't realize is that detection procedures with very impressive-sounding statistical properties generally do horribly at catching rare events.
Imagine some very impressive numbers. Suppose that this procedure has 99.999% sensitivity -- it catches nearly every wannabe terrorist who tries to board a plane intending to do harm. And suppose it also has 99.999% specificity -- out of 100,000 innocent passengers, 99,999 will be correctly identified as innocent, and only 1 will be a false alarm. Sounds good, right?
Not really. In a given year, only a very small number of passengers are wannabe terrorists -- say, 10 per year. (That's probably high.) On the other hand, there are 1.6 billion air passengers per year (that may be a low estimate, since it's a 2000 number). So if this were implemented worldwide, then in a given year, we can assume that this profiling procedure will flag 160,010 people as terrorists. Only 6 x 10^-5 of those will be actual terrorists.
Of course, those hypothetical sensitivity and specificity numbers are unrealistically, ridiculously good. With more realistic numbers, the problem gets much worse. Even if the detection procedure is very sensitive and very specific -- and I doubt that it is -- the low base-rate of terrorism means that an enormous number of people will be falsely accused of being terrorists.
CmdrTaco: "have a serious discussion on the issues surrounding this election"
You must be new here.
As pointed out earlier in this thread, the linked article is about a randomized experiment.
Imagine that you have been diagnosed with a potentially fatal disease. The research shows that if you go untreated, you have a 75% chance of dying within a year. Your doctor offers you a treatment that, in well-run clinical trials, has been shown to reduce the chance of death to 25%. Do you take it, or dismiss it as "a joke"? Because those probabilities express a correlation of r=.5 between taking the drug (vs. not) and living (vs. dying).
Medical and behavioral science study phenomena that are inherently complex, noisy, and non-deterministic. Your standards for what constitutes a meaningful effect size have to be adjusted accordingly.
To all the people tagging this "correlation is not causation," do you even know what you're talking about? This was a randomized experiment.
I'm not saying this is a perfect study -- there might be plenty of other things wrong with it. But the phrase "correlation is not causation" has an actual meaning. It is not just a synonym for "I had a kneejerk reaction to dislike this study but I can't say why."
Some supporting evidence making it hard to fit this prize into an ideological box...
In his popular writing, including his NY Times column, Krugman is a pretty outspoken liberal on most issues. But within his academic expertise -- which is what he won the prize for -- he is very willing to depart from liberal orthodoxy if that's where logic and evidence lead him.
Especially interesting is that the work Krugman won his prize for is about global free trade. Like most economists who've seriously studied the issue, Krugman has concluded that free trade is unabashedly a good thing.
But in the current U.S. political climate, free trade is mostly being touted by conservatives and reviled by liberals. So if you're a conservative and you want to claim liberal bias, you have to account for the fact that Krugman got the prize for work you probably agree with. And if you're a protectionist liberal who wants to boast, you're similarly stuck.
And if you're just generally tired of ideologues crowing about victory or whining about bias when neither is deserved, you can enjoy the whole spectacle of people getting tongue tied when someone wins the Nobel prize on (gasp) the strength of his ideas.
Three reasons this is a bad idea. One, if you suddenly have to swerve to avoid something, you might miss the edge that the power steering would provide Two, you might need to suddenly apply engine power in some emergency situations, like if you hit black ice and start skidding. Three, if you accidentally turn the key too far back, you'll engage the steering wheel lock and won't be able to steer at all.
In short: really bad idea.
Absolutely. The summary and linked article get it wrong. The Ig Nobel prizes are not about "irrelevant" research:
A lot of scientific research seems pointless or silly to people who don't know what it's really about or why it was done. Hence the regular "still no cure for cancer" and "I can't believe my tax dollars fund this" comments. The Ig Nobel prizes acknowledge that science can sometimes seem funny on the surface, but they definitely do not concede that it is irrelevant.
Depends on your definition of "best."
If best means "in theory, the technology is extremely difficult to break when human beings know how to use it correctly and do so" then yeah.
If best means "in practice, it is likely to work effectively when used by ordinary actual human beings in the real world" then um, no.
Human beings are part of the security equation. Gnupg is a great piece of technology, but unless it is wrapped in an interface and set of procedures in such a way that Nina from accounting is able and willing to use it correctly every time, it will not improve security.
Richard Dawkins is an atheist who believes that science and faith are fundamentally incompatible, so it makes sense that he'd think that. But that is not a consensus view among scientists. Plenty of other people think that science is compatible with religion and spirituality because they address different concerns.
Unless the Royal Society is now taking positions on the acceptability of religion, there should be no consideration given, pro or con, to the religious beliefs or affiliations of its officers.
Everything in his essay seems reasonable to me. The fuss arose in part, I think, because attacks on the scientific community have forced scientists into such a bunker mentality that they acted irrationally (i.e., not like scientists)
Interestingly, David Foster Wallace wrote the book Everything & More about the work of mathematician Georg Cantor, who suffered from mental illness late in life. Around the same time as Wallace's book was published, another book came out trying to argue that Cantor's mathematical genius stemmed from his illness. In an interview, Wallace called it "dreck, the very worst kind of appeal to a flabby, unconsidered pop version of what you just now called the 'mad genius' syndrome." Wallace goes on:
People like the "mad genius" meme, but it paints an overly simple picture of things. It might be the case (and empirically, the jury is still out) that some of the same predispositions that put people at risk of mental illness in some circumstances can contribute positively to intellectual functioning in others. But that is most definitely NOT the same as saying "with great genius comes great madness." If anything, the madness comes along and puts a sad end to the genius, rather than enhancing it.
I would draw a clear and bright line between what you say about David Foster Wallace ("he took the coward's way out") and others' reactions ("fawning news coverage"). The second point, I totally agree with you on. The first, I vehemently disagree.
We don't know why DFW committed suicide, and we might neverk now. But in the vast majority of cases, "cowardice" isn't even a relevant concept. Depression -- real, deep depression -- is not just about being in a crappy mood. Real depression (and other kinds of serious mental illness) messes you up so deeply that up seems like down and you cannot make rational sense out of yourself or the world. To call someone a "coward" implies that were faced with a choice and, with faculties intact, made a weak decision. Like I said, not a relevant concept for suicide.
And for the exact same reason, all the tributes making this into some sort of penetrating existential act of a man who saw the world too clearly... please! DFW was a brilliant thinker and writer, but his death is a tragedy and a loss. It is not an artistic act.
Interesting point. The linked site doesn't say much about its algorithm. But I'm guessing that it uses some sort of probabilistic matching of your rankings across many preferences, rather than an exact match. After all, they say right on the sight that you shouldn't worry about remembering things exactly. So I bet your preferences don't all have to be perfectly stable for it to work. Rather, enough of your preferences have to be just stable enough for whatever probability threshold they use.
As for guessing preferences from known behaviors like reading slashdot, I bet the predictability is lower than you'd think. Do all slashdot readers like video games? More so than non-slashdot readers? On average maybe so, but I bet there is huge variance in both groups. Additionally, if they ask about multiple preferences that have low joint probabilities, they can make the algorithm more secure. Knowing that you're a slashdot reader may help me guess that you are slightly more likely than the average person to like videogames, but it won't give me much traction on something like your favorite foods.*
* - Assuming they leave Cheetos and Mountain Dew off the list.
There, fixed that for you. And seeing as how James is generally credited for establishing the field of psychology in America, I think he can be forgiven for limiting himself to some cautious optimism.
I have a very similar theory about the origin of dragon stories.