5 ms·

IQ is a shit measure of intelligence

Proven 8y ago
You're trying to stifle the debate by debating. Good luck with that.
_Schizotypy 8y ago
wow thanks for all the downvotes, apparently nobody is current on psychological research https://www.webmd.com/brain/news/20121218/iq-test-really-measure-intelligence#1 https://www.webmd.com/brain/news/20121218/iq-test-really-mea... https://www.psychologytoday.com/us/blog/beautiful-minds/200910/intelligent-testing https://www.psychologytoday.com/us/blog/beautiful-minds/2009...
commandlinefan 8y ago
Maybe, maybe not, but how would you know? As of now, IQ is the _only_ measure of intelligence, so there's nothing to compare it against except anecdotes. You can say, "so and so has a low IQ but he seems smart to me", but there's nothing objective to compare it against, so there's really no good way to tell if it does measure that ephemeral quality that we attribute to the term "intelligence". It does track pretty closely with every other objective measure of intelligence, though, like academic performance and financial success.
- Junk_Collector 8y ago
  Of course if the IQ test was flawed and just a measure of where we expend resources on people rather than intelligence, it would strongly track with academic performance and financial success.
- _Schizotypy 8y ago
  Just because something is the only game in town doesn't mean you should use it, that logic doesn't make sense. We need to work on developing better systems rather than saying "oh well, it's the only thing we've got!"
  - candiodari 8y ago
    You don't understand. IQ isn't "the only game in town" because there aren't any other games, it's the only game in town because all other games were gradually unmasked to all be the "IQ game". IQ started out being one of about 5 measurements of people. Because of the age, one other measure was "nobility". Well, corrected for IQ, it turned out, that didn't matter. Same happened to many others. You can characterize and predict how humans will perform in many, many games with this single variable. It's the one that works best. So the current thinking is that there is only one measurement of intelligence. They all converge into a single one. It's called IQ because that was the first one that worked well enough, not because there's just one way to measure it. (Technically there is also age. Younger people are dumber than older ones up to a pretty high age. But that's the IQ game: age and "raw iq". Nothing else really matters in predicting if someone can succeed at something. IQ, the measure used, is corrected for age, so it should take that out. But a 10 year old with 100 IQ is, in absolute terms, dumber than a 20 year old with IQ 100) So your criticism is rather a lot beside the point. We have worked, a long and hard and often on developing better systems. We have just never actually succeeded in creating a truly different system (this of course is a "technically incorrect" statement. There are recent research results, there are ... blah blah blah ... there is no halfway decent reason to assume those research directions will lead to anything, but we're still trying. We've been doing that for >100 years, so the odds of the next big thing being in the pipeline right now is small). The basic observation behind IQ is that "all skills converge" (I would argue this has even been observed in machine learning). If you know one skill, whichever one it is, you become faster, better and quicker to learn almost any other skill. Not much, a little. In machine learning there are papers that point out that the ability to distinguish poodle races in pictures help with music composition. Not "much", but the key is what happens over time. It helps something like 2-5% (measured in time to learn a certain proficiency in music composition). The issue is what happens with kids when you give them new skills regularly over a 10 year period. Then they have 100 more skills, or 100 times 2% easier time to learn a new skill compared to someone who learned nothing. 100 times 2% easier ... is (1.02^100)-1, which is an enormous number (which is why IQ is an exponential score, not a linear one. Someone with 120 should be about double as smart as someone scoring 100. By which we mean that if you take a random intellectual skill, the odds of the person with 120 having some proficiency at that skill is about double that of the person with 100). To add insult to injury IQ compounds on itself. The higher your IQ, the easier it is to get it even higher. What people hate about IQ is the underlying theory clearly shows that human brains are designed to amplify differences in intelligence over time, not to minimize them. That is a clear and incontestable fact at this point. This is presumably also not a bad thing, or it would have been selected out. This also won't change because of any amount of changes occurring in psychology or the measurement of it. We can prevent ourselves from knowing about it, but we can't change it.

You're trying to stifle the debate by debating. Good luck with that.

wow thanks for all the downvotes, apparently nobody is current on psychological research https://www.webmd.com/brain/news/20121218/iq-test-really-measure-intelligence#1 https://www.webmd.com/brain/news/20121218/iq-test-really-mea... https://www.psychologytoday.com/us/blog/beautiful-minds/200910/intelligent-testing https://www.psychologytoday.com/us/blog/beautiful-minds/2009...

commandlinefan 8y ago

Maybe, maybe not, but how would you know? As of now, IQ is the _only_ measure of intelligence, so there's nothing to compare it against except anecdotes. You can say, "so and so has a low IQ but he seems smart to me", but there's nothing objective to compare it against, so there's really no good way to tell if it does measure that ephemeral quality that we attribute to the term "intelligence". It does track pretty closely with every other objective measure of intelligence, though, like academic performance and financial success.

Junk_Collector 8y ago
Of course if the IQ test was flawed and just a measure of where we expend resources on people rather than intelligence, it would strongly track with academic performance and financial success.
_Schizotypy 8y ago
Just because something is the only game in town doesn't mean you should use it, that logic doesn't make sense. We need to work on developing better systems rather than saying "oh well, it's the only thing we've got!"
- candiodari 8y ago
  You don't understand. IQ isn't "the only game in town" because there aren't any other games, it's the only game in town because all other games were gradually unmasked to all be the "IQ game". IQ started out being one of about 5 measurements of people. Because of the age, one other measure was "nobility". Well, corrected for IQ, it turned out, that didn't matter. Same happened to many others. You can characterize and predict how humans will perform in many, many games with this single variable. It's the one that works best. So the current thinking is that there is only one measurement of intelligence. They all converge into a single one. It's called IQ because that was the first one that worked well enough, not because there's just one way to measure it. (Technically there is also age. Younger people are dumber than older ones up to a pretty high age. But that's the IQ game: age and "raw iq". Nothing else really matters in predicting if someone can succeed at something. IQ, the measure used, is corrected for age, so it should take that out. But a 10 year old with 100 IQ is, in absolute terms, dumber than a 20 year old with IQ 100) So your criticism is rather a lot beside the point. We have worked, a long and hard and often on developing better systems. We have just never actually succeeded in creating a truly different system (this of course is a "technically incorrect" statement. There are recent research results, there are ... blah blah blah ... there is no halfway decent reason to assume those research directions will lead to anything, but we're still trying. We've been doing that for >100 years, so the odds of the next big thing being in the pipeline right now is small). The basic observation behind IQ is that "all skills converge" (I would argue this has even been observed in machine learning). If you know one skill, whichever one it is, you become faster, better and quicker to learn almost any other skill. Not much, a little. In machine learning there are papers that point out that the ability to distinguish poodle races in pictures help with music composition. Not "much", but the key is what happens over time. It helps something like 2-5% (measured in time to learn a certain proficiency in music composition). The issue is what happens with kids when you give them new skills regularly over a 10 year period. Then they have 100 more skills, or 100 times 2% easier time to learn a new skill compared to someone who learned nothing. 100 times 2% easier ... is (1.02^100)-1, which is an enormous number (which is why IQ is an exponential score, not a linear one. Someone with 120 should be about double as smart as someone scoring 100. By which we mean that if you take a random intellectual skill, the odds of the person with 120 having some proficiency at that skill is about double that of the person with 100). To add insult to injury IQ compounds on itself. The higher your IQ, the easier it is to get it even higher. What people hate about IQ is the underlying theory clearly shows that human brains are designed to amplify differences in intelligence over time, not to minimize them. That is a clear and incontestable fact at this point. This is presumably also not a bad thing, or it would have been selected out. This also won't change because of any amount of changes occurring in psychology or the measurement of it. We can prevent ourselves from knowing about it, but we can't change it.