4 ms·

For some jobs, it matters more than others. I wonder if the data also records profession categories and differences become more clear. Also IQ is a sub- compon

For some jobs, it matters more than others. I wonder if the data also records profession categories and differences become more clear. Also IQ is a sub- component for general intelligence. There are so many other factors. Claiming it's useless or useful is pointless. It probably depends on how relevant that sub component is for the role and task at hand and therefore requires a case-based understanding. It's like claiming that good body strength is useless. For a programmer that's probably true, but for a professional home mover or pyramid builder, results will be different.

x87678r 6y ago
> Also IQ is a sub- component for general intelligence. This is the bit that is more important to me and something I dont see in the IQ debates. Some "smart" people are fast learners, others quick witted, some have great memory, others can distill lots of conflicting data, some are great communicators, others just work really hard. One IQ number is supposed to reflect all of that?
- blfr 6y ago
  Yes, this is one of the surprising things about human intelligence, most intellectual/cognitive abilities are positively correlated. Hence the g factor.
  - vsskanth 6y ago
    Serious Q: I remember taking an IQ test long back and it was a bunch of pattern matching questions. How did they figure out that good performance in this correlates with cognitive abilities that translates to real life intellect at tasks ?
    - disgruntledphd2 6y ago
      This is a great question! You probably took one of the Raven's matrices, which are a culture-blind test of g (the presumed construct underlying IQ). Essentially, each new test is normed against the previous tests, so the Stanford-Binet test would have been the first, and then when a new measure is developed, we'd administer both tests to a sample (hopefully many samples) and use the correlation between the tests to determine whether or not or new test worked. It's not really a great method, but it's the best we've got for most individual difference variables.
      - vsskanth 6y ago
        Thank you for your explanation
      - vllab 6y ago
        The Raven tests are so boring that I'm not surprised that people do better under stimulants (mentioned elsewhere in this thread). I also tend to go off on inner rants like Feynman about round manholes, because the tests do not always seem to have unambiguous solutions.
        disgruntledphd2 6y ago
        Really? They don't take that long though, I vaguely remember doing the APM in about 30-45 minutes for an assignment. They do a much better job of getting estimates of IQ which are robust to culture, which was a real problem for the Stanford-Binet ones (and lead to some belief that the IQ of poor black children was lower because they weren't familiar with some of the analogies).
      - philipkglass 6y ago
        I would like to see some intelligence challenges that do not depend on the identity of the test taker. By that I mean that the results would be accepted as equally valid for people, non-human animals, and machines. If DeepMind developed a system that could ace Raven's Progressive Matrices questions, I don't think that we would say "machines are now as smart as human geniuses." Turing's Imitation Game still seems like a valid challenge to me but is perhaps too stringent. Someone who is not literate in English would fail to convince the questioner even if otherwise intelligent. "General" human intelligence appears to have a very complicated and particular shape. If there is such a thing as universal intelligence (in the sense that there are universal constants), the boundaries of human intelligence in this space look like a gerrymandered voting district more than a symmetrical geometric object.
        username90 6y ago
        The real Turing test is "Get a job as a remote developer and keep it for a year while getting paid". An AI that can reliably do that is what I'd consider general intelligence. If it can't learn how to interact with all the API's and programs you need to complete your tasks such as email, chat, code etc then it is not a general intelligence.
    - brabel 6y ago
      They did not figure it out, the fact that IQ tests as we know today could predict performance in a wide array of tasks was basically an observation coming out of the data. Many other types of tests were tried for this, because being able to predict people's performances has extremely important implications, but the ones that did not predict anything were forgotten, while the ones that worked, like the current generation of IQ tests, became widely accepted (despite being attacked every now and then by a few researches, as in this case - which is a good thing as we need to make sure our current theories are as good as they can be).

x87678r 6y ago

> Also IQ is a sub- component for general intelligence. This is the bit that is more important to me and something I dont see in the IQ debates. Some "smart" people are fast learners, others quick witted, some have great memory, others can distill lots of conflicting data, some are great communicators, others just work really hard. One IQ number is supposed to reflect all of that?

blfr 6y ago
Yes, this is one of the surprising things about human intelligence, most intellectual/cognitive abilities are positively correlated. Hence the g factor.
- vsskanth 6y ago
  Serious Q: I remember taking an IQ test long back and it was a bunch of pattern matching questions. How did they figure out that good performance in this correlates with cognitive abilities that translates to real life intellect at tasks ?
  - disgruntledphd2 6y ago
    This is a great question! You probably took one of the Raven's matrices, which are a culture-blind test of g (the presumed construct underlying IQ). Essentially, each new test is normed against the previous tests, so the Stanford-Binet test would have been the first, and then when a new measure is developed, we'd administer both tests to a sample (hopefully many samples) and use the correlation between the tests to determine whether or not or new test worked. It's not really a great method, but it's the best we've got for most individual difference variables.
    - vsskanth 6y ago
      Thank you for your explanation
    - vllab 6y ago
      The Raven tests are so boring that I'm not surprised that people do better under stimulants (mentioned elsewhere in this thread). I also tend to go off on inner rants like Feynman about round manholes, because the tests do not always seem to have unambiguous solutions.
      - disgruntledphd2 6y ago
        Really? They don't take that long though, I vaguely remember doing the APM in about 30-45 minutes for an assignment. They do a much better job of getting estimates of IQ which are robust to culture, which was a real problem for the Stanford-Binet ones (and lead to some belief that the IQ of poor black children was lower because they weren't familiar with some of the analogies).
    - philipkglass 6y ago
      I would like to see some intelligence challenges that do not depend on the identity of the test taker. By that I mean that the results would be accepted as equally valid for people, non-human animals, and machines. If DeepMind developed a system that could ace Raven's Progressive Matrices questions, I don't think that we would say "machines are now as smart as human geniuses." Turing's Imitation Game still seems like a valid challenge to me but is perhaps too stringent. Someone who is not literate in English would fail to convince the questioner even if otherwise intelligent. "General" human intelligence appears to have a very complicated and particular shape. If there is such a thing as universal intelligence (in the sense that there are universal constants), the boundaries of human intelligence in this space look like a gerrymandered voting district more than a symmetrical geometric object.
      - username90 6y ago
        The real Turing test is "Get a job as a remote developer and keep it for a year while getting paid". An AI that can reliably do that is what I'd consider general intelligence. If it can't learn how to interact with all the API's and programs you need to complete your tasks such as email, chat, code etc then it is not a general intelligence.
  - brabel 6y ago
    They did not figure it out, the fact that IQ tests as we know today could predict performance in a wide array of tasks was basically an observation coming out of the data. Many other types of tests were tried for this, because being able to predict people's performances has extremely important implications, but the ones that did not predict anything were forgotten, while the ones that worked, like the current generation of IQ tests, became widely accepted (despite being attacked every now and then by a few researches, as in this case - which is a good thing as we need to make sure our current theories are as good as they can be).