SoylentNews is people
SoylentNews
At last weekend’s annual meeting of the Association for Psychological Science (APS) in Boston, Cornell University psychologist Robert Sternberg sounded an alarm about the influence of standardized tests on American society. Sternberg, who has studied intelligence and intelligence testing for decades, is well known for his “triarchic theory of intelligence,” which identifies three kinds of smarts: the analytic type reflected in IQ scores; practical intelligence, which is more relevant for real-life problem solving; and creativity. Sternberg offered his views in a lecture associated with receiving a William James Fellow Award from the APS for his lifetime contributions to psychology. He explained his concerns to Scientific American.
[An edited transcript of the interview follows.]
What I argue is that intelligence that’s not modulated and moderated by creativity, common sense and wisdom is not such a positive thing to have. What it leads to is people who are very good at advancing themselves, often at other people’s expense. We may not just be selecting the wrong people, we may be developing an incomplete set of skills—and we need to look at things that will make the world a better place.
-- submitted from IRC
Related:
AI Researchers Develop Curiosity Algorithm
(Score: 3, Touché) by VLM on Monday June 05 2017, @12:29PM (12 children)
Usually a dog whistle for anti-white racism. Go ahead, just watch how its implemented.
Yeah, about that... I'm familiar with Linda Gottfredson's critique of this cranks theory, she is honest which results in appearing to be absolutely brutal, sure, dude's theory sounds really nice, but the "real" theory correlates with a zillion measurable outcomes indicating success.
Dude is basically trying to sell a nice smelling non-scientific variant of alchemy to a population that's already implemented organic chemistry and chemical engineering. It don't matter how nice it sounds, it predicts nothing and what we have works pretty well and its not going to sell well. Good luck convincing the chief engineer of a successful oil refinery to replace his hydrocracker that eats H2 and outputs lighter "cracked" products that he should replace that piece of hardware with an interpretive bongo drum and dance circle and that'll produce just as much profit and gasoline...
Dude's theory is kind of the "String Theory" of psych. He's (probably) a nice guy. The theory sounds highly entertaining, and is not morally or ethically inherently offensive superficially. You can get normies to believe anything if you state it authoritatively enough, so it sells well to journalists and readers. It makes no measurable predictions about reality, and existing theory predicts the future far better than the proposed theory. As a religious belief there's nothing really wrong with it other than trying to talk about it as if it were a scientific theory, which it is not.
The nicest possible way to interpret Gottfredson WRT Sternberg is that Sternberg's nice sounding religious belief might be on to some if he could produce good numbers out of real data, but the dude's produced numbers at a quality and rate beneath existing theories.
I'm just saying this is kind of like SN reporting on a evolutionary biology conference and not mentioning the guest of honor is a creationist.
I know about Gottfredson vs Sternberg because about a decade ago (this debate goes back awhile ...) there was a pretty interesting AI topic lecture that got recorded and can be watched on youtube and to make a long lecture very short, the Gottfredson way of looking at the world is very traditional and successful way to write neural network training software, you know, make a numeric relevant success metric, grade the result, "punish" or "reward" the neural net, repeat until you get great results, whereas the Sternberg way of looking at intelligence WRT writing software is WTF is "be creative" mean, insert a lot of calls to a random number generator? But that's just simulated annealing which is a whole nother topic. Sternberg-ism at least as of 10-15 years ago was explicitly non-scientific. And because AI tends to be cross discipline, you can get pointy-haired-boss types of a liberal arts bent writing papers and asking questions of why programmers can't write a Sternberg-ist-ic program. They can't because its bunk. Its like blaming programmers for not working hard enough to write "warp drive" navigation software.
I can't find the link to the lecture mentioned above and its early in the morning so I'd rather be rick rolling anyway, but whatever. Its an old lecture I don't remember the conference name or maybe it was a classroom lecture I don't remember much other than Gottfredson powned Sternberg fairly brutally to a level I've not seen much in academia, which in itself is kinda entertaining.
I have my own theory which seems to match reality, that STEM people like Gottfredson-ism because its all predictive statistical models just like engineering an airplane wing (kinda) whereas liberal arts people seem to flock to Sternberg because he tells a nice religious belief of a story or fable that is math free.
(Score: 2) by c0lo on Monday June 05 2017, @01:42PM (4 children)
The number doesn't bear any indication of the truth value. If it did, we wouldn't be today still in the rabbit hole we sunk after zillions were saying "House prices never fall"
But of course... in psychology, the "quantum effect of measuring" is strong. Tell me how you measure me and I'll tell you how I behave.
Measurements confirm the theory... are you sure you've eliminated the experimenter’s influence?
Look, I'm not saying the guy is right, but neither I'm prepared to drink the cool-aid of "Learning to the test is the summit of education and produces the best results evah".
Granted, it produces measurable things... but is measurement the goal?
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by VLM on Monday June 05 2017, @02:48PM (3 children)
Eh kind of good point but it inserts short term time into something timeless.
How about a predictive model that given coordinates of city centers (which a clustering algo could squirt out based on enough population data) then the model predicts that generally price increases as distance to the nearest city center decreases.
Generally. Well enough to model transportation infrastructure and capital financing to build housing and infrastructure and predict financial markets and stuff.
The opponents like to point out that model only applies maybe 99% of the time meaning its "broken" because sometimes recreational small lakefront properties have off the chart property values despite being far away from the city center. Or slums nearby the city center have amazingly low property values. Or white flight means the highest value property is now 20 minute drive away not right downtown. Whatever.
But the model works well enough almost all the time to be useful in practice. So tossing out the model because it fails 1% of the time to replace it with a new model where property values are made based on aesops fables or the story of the three little pigs and the wolf which makes no accurate predictions ever but is a heck of an entertaining story is not necessarily a net gain to the entire system.
Its kind of like that with models of intelligence. One is so accurate and real world that you can apply its philosophy cross discipline into computer science AI work and it actually works to train artificial neural networks. The other is an entertaining fairy tale, but you can't use it to write workable source code, because, well, its basically just a funny story.
In hard sciences models of the world should be replaced with a model that more accurately predicts the world, not a story that sounds more "fun".
(Score: 2) by c0lo on Tuesday June 06 2017, @10:51AM (2 children)
Dismissing what you can't measure goes only that far.
Education is not only a science, it's also an art. A good enough dose of feeling, intuition, acting (creating the emotional connection with the kids, or between the kids and the subject). Tell me what "hard science" can model what makes a good actor?
Look at the Finland's education system and see they integrated the "sciency" part with lotsa "common sense" and creating opportunities for the things to "just happen".
Then think again how's "just happen" treated by the "hard science" - if you can't reproduce, eliminate variability, maintain control, right? This doesn't work too well with education, you'll get a generation of drones (just look around and see where control and taming gets the children when they grow up)
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by VLM on Tuesday June 06 2017, @03:14PM (1 child)
Eh I don't like that style of engineering. Intuition and gut feeling is for when you don't have hard numbers and results that work. Never toss out hard number that actually work and replace with feels.
(Score: 2) by c0lo on Tuesday June 06 2017, @09:05PM
Mate, education is NOT engineering. If one takes it like that, of course one is going to fail.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 1, Interesting) by Anonymous Coward on Monday June 05 2017, @01:47PM
100% accurate. SAT's were introduced to promote meritocracy and overcome class differences. Sternberg's crap is about ensuring equality of outcome and to do that there must be something we are missing because Orientals and Jews cannot be (on average) more intelligent and successful than Caucasians who are, in turn, more intelligent and successful than blacks. Like every piece of leftist claptrap we must promote the least capable and effective individuals just to make sure they're not being oppressed. The exact same method that has always ended in real oppression whenever and wherever implemented.
(Score: 2) by AthanasiusKircher on Monday June 05 2017, @02:40PM (3 children)
Well, reading TFA, you encounter the following information:
There are links in TFA with more details about the "Rainbow Project," etc., though I don't have the time to wade through long academic articles dealing with the stats of standardized tests right now. But one of the reasons why we use standardized tests like the SAT is because they seem to correlate well with college success. If this guy has assessments that can increase that predictive ability, then I'd say he at least has a start at "predicting the future" BETTER than "existing theory." From a quick glance, the proposed assessments sound a lot more laborious both to take and to administer, and I personally would be interested in more long-term follow-ups than just performance in first year of college.
But it at least sounds like he's trying to measure real impact for his theories, contrary to your post which claims he makes no "measurable predictions about reality."
(Score: 2) by VLM on Monday June 05 2017, @02:54PM (2 children)
Yeah fair enough. Science is a race (competition) so my wimpy disclaimer holds.
At least as of 10-15 years ago, when I was introduced to this dude, he was considered a punchline by the hard scientists and statisticians in his field. Maybe 100 more years of data he'll turn out to be right by producing more and better data than the competitors. That seems a realistically possible factual outcome.
My opinion is I'm not betting on it. Too vague and fluffy.
(Score: 2) by AthanasiusKircher on Tuesday June 06 2017, @01:17AM (1 child)
I agree some of this theory sounds pretty vague. On the other hand, what he says has a lot in common with stuff other psychologists have proposed about intelligence. The "g factor" theory of IQ tests has been widely criticized, even among psychometricians. Tests like the SAT used to be heavily g-loaded, but they've shifted further from IQ tests in recent decades -- though arguably in ways that make them more like the stereotypes this guy brings up, and involving less "creative" or "practical" intelligence. Gone are the days of antonyms, analogies, even "quantitative comparisons" in the math section, all of which required different types of thinking than the streamlined SAT of today, which seemed to have given up even on moderately advanced vocabulary.
(Score: 2) by VLM on Tuesday June 06 2017, @04:01PM
Sarcastically speaking slightly off topic, the standardized admissions test of the future is going to be the student's parent's credit report and probably not much else.
(Score: 0) by Anonymous Coward on Monday June 05 2017, @04:09PM (1 child)
So studying the effects of standardized testing and wanting to change things so as to improve the world, this you call a racist dog whistle? I think you're just bent out of shape by the civil rights movement and the attempts to correct years of abuses. You're like the minorities that see everything as racist, except you're a special pure white snowflake. Oh, and one who needs to label themselves ad a white supremacist but that only means you like whites MORE not that you like others LESS. Mmmhmmm
(Score: 0) by Anonymous Coward on Monday June 05 2017, @11:36PM
Anonymous got his virtue signalling in for the day! Congrats, the two minute hate was successful. You've labeled somebody a racist on the internet. You're a hero! Here's a gold star for your Facebook page. Now back to the discussion about how tests must be wrong because stupid people keep failing at them.