Why don't these "researchers" understand the importance of self-selection?!?
Believe it or not, they do. In the second and third studies reported in the paper, subjects were randomly assigned to either play a driving game or a neutral game. Random assignment eliminates self-selection effects.
One thing that many seem to be forgetting is that short excerpts are defined as fair-use.
No they're not. The amount of a work that you use is only one of four factors that go into defining something as fair use. Other considerations include transformative value - for example, using a piece of someone else's work in order to comment on it or parody it. Most of the kind of clips you're talking about aren't transformative at all -- the meaning and significance of 2 minutes of Jon Stewart online is largely the same as those same 2 minutes within the context of a 30 minute show.
Legally you're right. Youtube is within the protection of the DMCA and should win the lawsuit.
However, they are exploiting that protection. When somebody uploads Viacom's copyrighted content to YT, YT makes money off of the content until the moment that Viacom discovers it and tells them to take it down. YT gets to keep the money, and Viacom's product is now less valuable because thousands or millions of people have already seen it.
Seems like a potential legislative tweak to "safe harbor" would be to require service providers like Youtube to track ad hits/revenue gained from any particular content source. If Viacom can prove that the content was theirs, then Youtube should then have to turn over that money to the rightful copyright holder.
You want to play in India, you play by their rules.
If your motto is "don't be evil" and India's rules require you to be evil, then you shouldn't want to play in India. Otherwise you're an evil hypocrite.
Well, technically the summary is correct, though it's misleading. The summary says "Dubai's friendly tax laws will add to Halliburton's bottom line." According to this analysis, that's true -- but the savings will be on non-US taxes:
The move to Dubai could save Halliburton (and CEO Dave Lesar) some money on foreign taxes, though. (With operations in 100 countries, Halliburton had to pay out $289 million to foreign governments last year.) The United Arab Emirates government may have sweetened the deal with favorable real-estate terms or other incentives. Dubai's Jebel Ali Free Zone, which already houses more than 5,000 foreign-owned businesses, doesn't impose corporate or personal income taxes and has a robust workforce with no minimum wage.
Is this a tacit admission that the "content over credentials" philosophy doesn't actually work in the real world?
Yes. If you take "content over credentials" to its logical extreme, you have to refuse to accept information from any source with an author.
Suppose you read an assertion on Wikipedia about something in physics. The edit history shows that it was placed there by someone who identifies as an honors physics student, and you decide not to accept his credentials, so you look up the reference he inserted in the article. The reference, let's say, is to a textbook. But why should you trust the textbook authors with their fancy degrees and their publishing contracts -- isn't that just accepting their credentials? So you look up and read the original study cited in the textbook. But how do you know these scientists really knew what they were doing and accurately interpreted their data? Aren't you just taking their word for it as scientists? Maybe you should replicate the experiment yourself.
At a certain point you have to accept a pair of ideas seemingly at odds. One, that argument from authority (i.e., accepting credentials) is flawed. Two, that if you don't accept an argument from authority at some point, you'll face an iterative regress down to collecting and analyzing the raw data yourself, which defeats the purpose of a reference work.
Here's how this stuff works. Step 1, scientist do incremental, meaningful, but boring (to those outside their specialty) work. Step 2, media picks up on story and puts overreaching spin on story. (Alternatively, the scientists, the journal, or the university's PR office puts out a press release supplying overreaching spin to credulous journalists.) Step 3, everybody sits back in wonderment at a finding that essentially establishes what we already knew: that mental processes take place in the physical brain.
Parent poster is right about the special demands of individual prediction. The basic science might be incrementally useful - trying to ultimately understand how future planning/intentions take place in the brain. (And given the breadth of mental operations that could be considered "intentions," there are probably hundreds of more studies that need to be done before that question can begin to be answered.) But going from a scientific explanatory mode, where you have potentially large samples and budgets and cooperative subjects, to prediction of individual behavior is a huge leap. Just look at a much older psychometric approach, the TAT, which is okay for research but lousy for individual prediction. Brain scanning may well turn out to be the next TAT, for precisely the same reasons.
Part of the problem is that a lot of this work is being done by medical researchers and neuroscientists who have no basic training in psychometrics. They're just reinventing old mistakes (but wasting a hell of a lot more money this time around).
Joking aside, subliminal priming is making a comeback in experimental psychology. It was somewhat discredited in the 60s and 70s (i.e., the urban legend about theaters flashing "Drink Coke" on movie screens), but more recent work has uncovered the parameters and boundaries to make it a viable experimental technique. It is typically used in controlled lab situations to study automatic processing of information in isolation from conscious, intentional thought. It's not entirely clear from the research literature whether it would work in this kind of real-life applied context. But it wouldn't be hard for a casino to do the testing to find out.
it isn't my job to try and convicen my doctor to prescribe some drug, it is the doctor's job to know what drugs are available and prescribe them to me.
Actually, both of you have a job (though you're totally correct that yours isn't to act as a drug sales rep). What's insidious about this kind of advertising is that it exploits what should be a good trend. The old model of "doctor knows best, do whatever he tells you" isn't ideal. The best case scenario is when a patient becomes informed about his/her health and makes decisions in consultation with the doctor, drawing on the doctor's expertise.
When I was seriously ill a few years back, my doctors (who were outstanding btw) encouraged me to do research on my own. I scoured medical databases and brought that info to my doctors, who helped me sort through it. I think it elevated the quality of my care, and it made me feel like I retained some control over a scary illness.
Unfortunately, by feeding patients information that is biased or misleading and playing to emotion rather than providing useful information, these ads are probably eroding the kind of doctor-patient relationships that everyone should have.
This will make an interesting comparison to iTunes... iTunes sells music online with DRM that can, in principle, be defeated (or 1 person could buy an un-DRMed CD and upload it to the rest of the world). But by putting just enough hassle in front of the typical consumer, combined with pricing that is generally perceived as reasonable, iTunes has managed to be quite successful. Consumers could engage in piracy, but most choose not to.
What's interesting about identity watermarking is that instead of using a digital control like iTunes, they're using a more social one -- making people feel accountable. (As was pointed out, it's unclear yet whether people will actually be held accountable.) If that is effective, critics of the **AAs could make a more effective argument that DRM, which restricts legitimate fair use, is not necessary.
The radio station came up with the idea, advertised it, provided an incentive to do it, and intended to make money off of it (via the publicity it got them). She agreed to participate in response to their invitation. If her agreement involved making some assumptions that a reasonable person would make, and one of those reasonable assumptions was that the station wouldn't invite her to do something dangerous without warning her, then you can argue they bear a share of the responsibility for what happened.
As you might guess, there is a whole tangle of legal and ethical issues surrounding testing in personnel selection.
My understanding (IANAL etc) is that you are supposed to assess only the skills, aptitude, etc. that you can defend as related to the job. If that happens to be correlated with sex, race, age, etc., the correlation is not a problem, but you cannot use those things as a proxy for what you're really interested in. For example, in a job that requires quick responses, you can test people's reaction times, but you cannot automatically exclude people based on age (even though age may be correlated with reaction time).
More direct assessment is better anyway. Suppose you are hiring for a job that requires math skills, which you believe is correlated with gender, which you believe is correlated with cat ownership. Even if those correlations exist, you'd still get more accurate results measuring the math skills directly rather than measuring cat ownership which is correlated with something that is correlated with what you need.
You are referring to the guilty knowledge test. Even that is controversial. As you imply, it's only relevant to a pretty limited number of situations (crimes where the police have information that only the perp would know). For mass employee screening, it's not relevant.
You have been infected with a rare and deadly disease. A new drug has become available. In clinical trials, if people are left untreated (control group), only 31% survive and the rest die. If they are given the drug, 69% survive. Do you take the drug, or ignore it because its effects are "unimpressive"? Because that example expresses an effect size of r=0.38.
The "anywhere else" you refer to is in areas of science that deal with deterministic phenomena. In many areas of social science, medicine, and other fields, the phenomena are probabilistic, and effect sizes are judged accordingly.
Equipment for the second part (getting the calls to the ground) is already in many planes for the phones you will find built into seats.
Airlines generally charge a lot of money to use those phones, which raises an interesting possibility... Since the airlines own the onboard cell, could they tack on a hefty toll/surcharge (buck or two a minute) for using it? I'd actually favor that, because it would have the effect of preventing everybody from gabbing all flight long, while still opening an avenue for people who really need to make calls with their cellphone.
Although the study was not as inferentially strong as a randomized experiment, it was a prospective design with a number of statistical controls -- so it's a lot better than the ice cream/drowning correlation. Controls included pre-existing antibodies to the virus as well as self-reported health (which researchers usually consider a useful but imperfect proxy for other indicators).
Also, with regard to someone else's comment, they quarantined subjects and measured for objective markers of illness -- the results don't depend on self-reported symptoms.
A randomized trial would be a great idea to give stronger support to this theory, and hopefully the researchers are planning one. However, the study was focused on fairly stable emotional dispositions. "Training in positive emotions" is no easy thing, as anybody who's ever been clinically depressed can tell you. But this study suggests it's probably worth trying.
In a technical (and technological) sense, you're absolutely right. Given the nature of digital information, anybody putting any information online would be well advised to act as though it is going to get back to everybody they know, perhaps through channels that don't even exist at the time you put the information online.
But the more complicated social reality is that in most people's experience, the public-private distinction has usually been one of probability and degrees, not an all-or-nothing proposition. It used to be the case (and still is, though less and less so) that you could go to certain technically public places and still have a practical/probabilistic expectation of privacy. For example, you could go to a political or cultural event for an unpopular group (a gay pride parade, for example) and have a reasonable hope that it wouldn't get back to your employer or family. You might be in a technically public space and you (hopefully) knew you were taking a risk, but the risk was small enough that it was worth it.
The problem raised by this kind of technology is that it is eliminating those kinds of physical and virtual spaces -- the spaces where you can meet and interact with others and have some practical (if not airtight) expectation of privacy. The fact is, there are very few real places you can socialize with lots of other people that have a truly complete expectation of privacy, so the probabilistic expectation is often the best you can hope for. For people with some kind of politically or culturally marginalized interest -- and let's face it, who doesn't have at least one interest that falls into that category -- it's a sad development.
Your list of exams supports my point, not yours, because in those professions, passing an exam is not enough to certify someone as qualified. Pilots are not allowed to fly based solely on passing a standardized exam (and would you fly in a plane with a pilot who passed the FAA exam but had never been subjected to any other form of evaluation?). Surgeons are not allowed to operate just because they've passed an exam (and would you let someone operate on you who had never passed any form of evaluation except their board exam?). Etc.
In other words, is any test absolutely perfect? Of course not. Nothing is. But I believe we were talking about the real world, not Black 'n' White Land were anything not utterly perfect is ipso facto utter garbage. Nowhere did I claim that standardized exams are useless. Of course exams add useful information to an overall assessment; they are simply not enough. I was specifically responding to your claim (repeated in your reply) that you could construct a comprehensive exit exam that covers all relevant skills and knowledge. I am stating that in practice, no exam is up to that task. (Your claim that you would only accept a "good" exam begs the question. My whole point is that your standard of a "good" exam is impossible in practice. Ironically, this is backed up by the list of professions you offered, all of which supplement standardized exams with other forms of evaluation in admitting people to the profession and making hiring decisions.)
A big comprehensive bugger that would "certify" our graduates in a measureable way and in particular specific skills. (This is in a scientific/technical field, by the way, so such skills are easy to define.)
Spoken like someone who has no understanding of educational measurement. Name one standardized exam that provides a comprehensive assessment of everything that needs to be considered to "certify" college graduates as qualified and hireable employees in any field, scientific/technical or otherwise. Does it measure every area of knowledge? Does it calibrate for differing emphases and subspecialities? Does it measure creative, "out-of-the-box" thinking? Initiative to carry creative ideas into workable solutions? Critical analysis of novel problems? Applying concepts and knowledge from one domain to an unexpected, surprising area? Ability to identify what you don't know and teach it to yourself rather than just applying what you already know? Ability to work in teams? Ability to lead others? Does it measure how much all of these are exhibited persistently in a day-to-day environment, rather than in the focused span of a couple of hours in an exam room?
There's a reason that even employers and grad schools in areas with well-established standardized exams (law, medicine, engineering, etc.) are interested in grades, recommendations and references, interviews, internships, performance tests, etc. -- it's because no exit exam could possibly tell them everything they need to know about whether someone can thrive in their field.
Here's what happens if you implement a do-or-die exit exam: learning of any important area of knowledge, skill, or ability that is not on the exam will get worse, because students will shift focus to learning (by rote memorization, if possible) all the things on the exam and they will ignore all the things that aren't.
And bringing this back on topic, if you read the article, you'll see that overreliance on exit exams is at the top of most experts' lists of what has gotten the Indian educational system into this mess in the first place.
What's annoying about the Indians taking the calls is that they pretend to understand when you use words or phrases they don't get, and it quickly becomes apparent as they struggle to troubleshoot a problem they never comprehended in the first place.
And this is different from American customer service how?
There is no comparison between US college education and the middle-tier Indian colleges being discussed. From TFA:
A deeper problem, specialists say, is a classroom environment that treats students like children even if they are in their mid-20's. Teaching emphasizes silent note-taking and discipline at the expense of analysis and debate...
Rote memorization is rife at Indian colleges because students continue to be judged almost solely by exam results. There is scant incentive to widen their horizons -- to read books, found clubs or stage plays.
The problem isn't one of teaching intellectual disciplines versus practical skills. The problem is that Indian colleges are teaching neither.
Slashdot Mirror
Why don't these "researchers" understand the importance of self-selection?!?
Believe it or not, they do. In the second and third studies reported in the paper, subjects were randomly assigned to either play a driving game or a neutral game. Random assignment eliminates self-selection effects.
Here's a link to the actual journal article.
One thing that many seem to be forgetting is that short excerpts are defined as fair-use.
No they're not. The amount of a work that you use is only one of four factors that go into defining something as fair use. Other considerations include transformative value - for example, using a piece of someone else's work in order to comment on it or parody it. Most of the kind of clips you're talking about aren't transformative at all -- the meaning and significance of 2 minutes of Jon Stewart online is largely the same as those same 2 minutes within the context of a 30 minute show.
Legally you're right. Youtube is within the protection of the DMCA and should win the lawsuit.
However, they are exploiting that protection. When somebody uploads Viacom's copyrighted content to YT, YT makes money off of the content until the moment that Viacom discovers it and tells them to take it down. YT gets to keep the money, and Viacom's product is now less valuable because thousands or millions of people have already seen it.
Seems like a potential legislative tweak to "safe harbor" would be to require service providers like Youtube to track ad hits/revenue gained from any particular content source. If Viacom can prove that the content was theirs, then Youtube should then have to turn over that money to the rightful copyright holder.
You want to play in India, you play by their rules.
If your motto is "don't be evil" and India's rules require you to be evil, then you shouldn't want to play in India. Otherwise you're an evil hypocrite.
Yes. If you take "content over credentials" to its logical extreme, you have to refuse to accept information from any source with an author.
Suppose you read an assertion on Wikipedia about something in physics. The edit history shows that it was placed there by someone who identifies as an honors physics student, and you decide not to accept his credentials, so you look up the reference he inserted in the article. The reference, let's say, is to a textbook. But why should you trust the textbook authors with their fancy degrees and their publishing contracts -- isn't that just accepting their credentials? So you look up and read the original study cited in the textbook. But how do you know these scientists really knew what they were doing and accurately interpreted their data? Aren't you just taking their word for it as scientists? Maybe you should replicate the experiment yourself.
At a certain point you have to accept a pair of ideas seemingly at odds. One, that argument from authority (i.e., accepting credentials) is flawed. Two, that if you don't accept an argument from authority at some point, you'll face an iterative regress down to collecting and analyzing the raw data yourself, which defeats the purpose of a reference work.
Here's how this stuff works. Step 1, scientist do incremental, meaningful, but boring (to those outside their specialty) work. Step 2, media picks up on story and puts overreaching spin on story. (Alternatively, the scientists, the journal, or the university's PR office puts out a press release supplying overreaching spin to credulous journalists.) Step 3, everybody sits back in wonderment at a finding that essentially establishes what we already knew: that mental processes take place in the physical brain.
Parent poster is right about the special demands of individual prediction. The basic science might be incrementally useful - trying to ultimately understand how future planning/intentions take place in the brain. (And given the breadth of mental operations that could be considered "intentions," there are probably hundreds of more studies that need to be done before that question can begin to be answered.) But going from a scientific explanatory mode, where you have potentially large samples and budgets and cooperative subjects, to prediction of individual behavior is a huge leap. Just look at a much older psychometric approach, the TAT, which is okay for research but lousy for individual prediction. Brain scanning may well turn out to be the next TAT, for precisely the same reasons.
Part of the problem is that a lot of this work is being done by medical researchers and neuroscientists who have no basic training in psychometrics. They're just reinventing old mistakes (but wasting a hell of a lot more money this time around).
Joking aside, subliminal priming is making a comeback in experimental psychology. It was somewhat discredited in the 60s and 70s (i.e., the urban legend about theaters flashing "Drink Coke" on movie screens), but more recent work has uncovered the parameters and boundaries to make it a viable experimental technique. It is typically used in controlled lab situations to study automatic processing of information in isolation from conscious, intentional thought. It's not entirely clear from the research literature whether it would work in this kind of real-life applied context. But it wouldn't be hard for a casino to do the testing to find out.
That's because drug companies actively recruit cheerleaders to work as reps.
Actually, both of you have a job (though you're totally correct that yours isn't to act as a drug sales rep). What's insidious about this kind of advertising is that it exploits what should be a good trend. The old model of "doctor knows best, do whatever he tells you" isn't ideal. The best case scenario is when a patient becomes informed about his/her health and makes decisions in consultation with the doctor, drawing on the doctor's expertise.
When I was seriously ill a few years back, my doctors (who were outstanding btw) encouraged me to do research on my own. I scoured medical databases and brought that info to my doctors, who helped me sort through it. I think it elevated the quality of my care, and it made me feel like I retained some control over a scary illness.
Unfortunately, by feeding patients information that is biased or misleading and playing to emotion rather than providing useful information, these ads are probably eroding the kind of doctor-patient relationships that everyone should have.
If someone tells you "You're one in a million," there are 6,571 people exactly like you.
This will make an interesting comparison to iTunes... iTunes sells music online with DRM that can, in principle, be defeated (or 1 person could buy an un-DRMed CD and upload it to the rest of the world). But by putting just enough hassle in front of the typical consumer, combined with pricing that is generally perceived as reasonable, iTunes has managed to be quite successful. Consumers could engage in piracy, but most choose not to.
What's interesting about identity watermarking is that instead of using a digital control like iTunes, they're using a more social one -- making people feel accountable. (As was pointed out, it's unclear yet whether people will actually be held accountable.) If that is effective, critics of the **AAs could make a more effective argument that DRM, which restricts legitimate fair use, is not necessary.
The radio station came up with the idea, advertised it, provided an incentive to do it, and intended to make money off of it (via the publicity it got them). She agreed to participate in response to their invitation. If her agreement involved making some assumptions that a reasonable person would make, and one of those reasonable assumptions was that the station wouldn't invite her to do something dangerous without warning her, then you can argue they bear a share of the responsibility for what happened.
Maybe this youtube link will survive a little better?
As you might guess, there is a whole tangle of legal and ethical issues surrounding testing in personnel selection.
My understanding (IANAL etc) is that you are supposed to assess only the skills, aptitude, etc. that you can defend as related to the job. If that happens to be correlated with sex, race, age, etc., the correlation is not a problem, but you cannot use those things as a proxy for what you're really interested in. For example, in a job that requires quick responses, you can test people's reaction times, but you cannot automatically exclude people based on age (even though age may be correlated with reaction time).
More direct assessment is better anyway. Suppose you are hiring for a job that requires math skills, which you believe is correlated with gender, which you believe is correlated with cat ownership. Even if those correlations exist, you'd still get more accurate results measuring the math skills directly rather than measuring cat ownership which is correlated with something that is correlated with what you need.
You are referring to the guilty knowledge test. Even that is controversial. As you imply, it's only relevant to a pretty limited number of situations (crimes where the police have information that only the perp would know). For mass employee screening, it's not relevant.
You have been infected with a rare and deadly disease. A new drug has become available. In clinical trials, if people are left untreated (control group), only 31% survive and the rest die. If they are given the drug, 69% survive. Do you take the drug, or ignore it because its effects are "unimpressive"? Because that example expresses an effect size of r=0.38.
The "anywhere else" you refer to is in areas of science that deal with deterministic phenomena. In many areas of social science, medicine, and other fields, the phenomena are probabilistic, and effect sizes are judged accordingly.
Airlines generally charge a lot of money to use those phones, which raises an interesting possibility... Since the airlines own the onboard cell, could they tack on a hefty toll/surcharge (buck or two a minute) for using it? I'd actually favor that, because it would have the effect of preventing everybody from gabbing all flight long, while still opening an avenue for people who really need to make calls with their cellphone.
Oh, and here's a link to the full text of the original article in case anybody's interested.
Here's a link to the journal abstract.
Although the study was not as inferentially strong as a randomized experiment, it was a prospective design with a number of statistical controls -- so it's a lot better than the ice cream/drowning correlation. Controls included pre-existing antibodies to the virus as well as self-reported health (which researchers usually consider a useful but imperfect proxy for other indicators).
Also, with regard to someone else's comment, they quarantined subjects and measured for objective markers of illness -- the results don't depend on self-reported symptoms.
A randomized trial would be a great idea to give stronger support to this theory, and hopefully the researchers are planning one. However, the study was focused on fairly stable emotional dispositions. "Training in positive emotions" is no easy thing, as anybody who's ever been clinically depressed can tell you. But this study suggests it's probably worth trying.
In a technical (and technological) sense, you're absolutely right. Given the nature of digital information, anybody putting any information online would be well advised to act as though it is going to get back to everybody they know, perhaps through channels that don't even exist at the time you put the information online.
But the more complicated social reality is that in most people's experience, the public-private distinction has usually been one of probability and degrees, not an all-or-nothing proposition. It used to be the case (and still is, though less and less so) that you could go to certain technically public places and still have a practical/probabilistic expectation of privacy. For example, you could go to a political or cultural event for an unpopular group (a gay pride parade, for example) and have a reasonable hope that it wouldn't get back to your employer or family. You might be in a technically public space and you (hopefully) knew you were taking a risk, but the risk was small enough that it was worth it.
The problem raised by this kind of technology is that it is eliminating those kinds of physical and virtual spaces -- the spaces where you can meet and interact with others and have some practical (if not airtight) expectation of privacy. The fact is, there are very few real places you can socialize with lots of other people that have a truly complete expectation of privacy, so the probabilistic expectation is often the best you can hope for. For people with some kind of politically or culturally marginalized interest -- and let's face it, who doesn't have at least one interest that falls into that category -- it's a sad development.
Your list of exams supports my point, not yours, because in those professions, passing an exam is not enough to certify someone as qualified. Pilots are not allowed to fly based solely on passing a standardized exam (and would you fly in a plane with a pilot who passed the FAA exam but had never been subjected to any other form of evaluation?). Surgeons are not allowed to operate just because they've passed an exam (and would you let someone operate on you who had never passed any form of evaluation except their board exam?). Etc.
In other words, is any test absolutely perfect? Of course not. Nothing is. But I believe we were talking about the real world, not Black 'n' White Land were anything not utterly perfect is ipso facto utter garbage. Nowhere did I claim that standardized exams are useless. Of course exams add useful information to an overall assessment; they are simply not enough. I was specifically responding to your claim (repeated in your reply) that you could construct a comprehensive exit exam that covers all relevant skills and knowledge. I am stating that in practice, no exam is up to that task. (Your claim that you would only accept a "good" exam begs the question. My whole point is that your standard of a "good" exam is impossible in practice. Ironically, this is backed up by the list of professions you offered, all of which supplement standardized exams with other forms of evaluation in admitting people to the profession and making hiring decisions.)
Spoken like someone who has no understanding of educational measurement. Name one standardized exam that provides a comprehensive assessment of everything that needs to be considered to "certify" college graduates as qualified and hireable employees in any field, scientific/technical or otherwise. Does it measure every area of knowledge? Does it calibrate for differing emphases and subspecialities? Does it measure creative, "out-of-the-box" thinking? Initiative to carry creative ideas into workable solutions? Critical analysis of novel problems? Applying concepts and knowledge from one domain to an unexpected, surprising area? Ability to identify what you don't know and teach it to yourself rather than just applying what you already know? Ability to work in teams? Ability to lead others? Does it measure how much all of these are exhibited persistently in a day-to-day environment, rather than in the focused span of a couple of hours in an exam room?
There's a reason that even employers and grad schools in areas with well-established standardized exams (law, medicine, engineering, etc.) are interested in grades, recommendations and references, interviews, internships, performance tests, etc. -- it's because no exit exam could possibly tell them everything they need to know about whether someone can thrive in their field.
Here's what happens if you implement a do-or-die exit exam: learning of any important area of knowledge, skill, or ability that is not on the exam will get worse, because students will shift focus to learning (by rote memorization, if possible) all the things on the exam and they will ignore all the things that aren't.
And bringing this back on topic, if you read the article, you'll see that overreliance on exit exams is at the top of most experts' lists of what has gotten the Indian educational system into this mess in the first place.
And this is different from American customer service how?
There is no comparison between US college education and the middle-tier Indian colleges being discussed. From TFA:
The problem isn't one of teaching intellectual disciplines versus practical skills. The problem is that Indian colleges are teaching neither.