3 ms·

I've always been surprised people thought any different, given the content of IQ tests. Is basically just puzzles and you usually get much better at puzzles if

I've always been surprised people thought any different, given the content of IQ tests. Is basically just puzzles and you usually get much better at puzzles if you do them all the time.

cliveowen 13y ago
True, but since virtually no one takes IQ tests on a daily basis, the IQ test is still a valid measure of one's intelligence (or potential, I don't really know what IQ is supposed to measure).
- gohrt 13y ago
  People who score high on IQ tests do take IQ tests (or similar puzles on a daily basis). > I don't really know what IQ is supposed to measure Neither does anyone else.
  - clarkm 13y ago
    They attempt to measure the theoretical concept of generalized intelligence, known by psychometricians as g. They're obviously not perfect, but they're better than anything else we have.
  - Tehnix 13y ago
    >People who score high on IQ tests do take IQ tests (or similar puzles on a daily basis). not necessarily, on an official test, I scored 140 (which I'd consider high), and I don't do any puzzles or similar at all. that said, obviously an IQ test is not perfect, but, I'd argue a high score is a good indication that a person is relatively intelligent. Middle scores and below can be attributed to a lot of other things (distractions, etc etc), but, the high score is still indicative of your brains capabilities. If that made sense?
gojomo 13y ago
Increasing your score on an IQ test is one thing; getting the benefits expected from more "intelligence" may be another. In such matters, keep in mind Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure." In the domain of IQ, we might observe a very high correlation between higher-IQ and certain desirable results, like a longer/healthier life, lower rates of victimization, higher incomes, and so forth. And we might have a strong case that the test is, at least at the outset, measuring some sort of "general intelligence", and that there is a causal link from "general intelligence" to the other correlated beneficial outcomes. But, then you start aiming to increase "general intelligence", and your only measure is the IQ test, so your target for evaluating all interventions is the IQ score. Now you are quite likely finding interventions that only raise the IQ score, and perhaps skip any "general intelligence" benefits. Even if your IQ score goes up, you might not get the other benefits. Meanwhile, others still researching the correlation/causation questions may see the original correlations start to weaken, or the case that IQ measures "general intelligence" weaken. And in fact the originally-valid ideas are weakening! Motivated action around that target has started to thwart whatever signal it once offered. The problem is universal - in regulation, in education, in investment, in semantics. It's why a lot of high-performing organizations keep their true 'measures' secret... to slow their dilution through motivated optimizations.
- moocowduckquack 13y ago
  The other thing with optimising as a strategy is while it likes small changes, it really hates structural ones as they make it very hard to compare what you have now with what went before, so anything that is hard to measure is resisted. But usefulness and ease of measurement are not always particularly aligned.
jessriedel 13y ago
IQ tests are very robust. The correlation between scores is greater than 0.9. These working memory results are usually good for just a few extra IQ points (as in, literally 2 or 3), and they are far and away the largest boosts people have been able to induce through training.

I've always been surprised people thought any different, given the content of IQ tests. Is basically just puzzles and you usually get much better at puzzles if you do them all the time.

cliveowen 13y ago

True, but since virtually no one takes IQ tests on a daily basis, the IQ test is still a valid measure of one's intelligence (or potential, I don't really know what IQ is supposed to measure).

gohrt 13y ago
People who score high on IQ tests do take IQ tests (or similar puzles on a daily basis). > I don't really know what IQ is supposed to measure Neither does anyone else.
- clarkm 13y ago
  They attempt to measure the theoretical concept of generalized intelligence, known by psychometricians as g. They're obviously not perfect, but they're better than anything else we have.
- Tehnix 13y ago
  >People who score high on IQ tests do take IQ tests (or similar puzles on a daily basis). not necessarily, on an official test, I scored 140 (which I'd consider high), and I don't do any puzzles or similar at all. that said, obviously an IQ test is not perfect, but, I'd argue a high score is a good indication that a person is relatively intelligent. Middle scores and below can be attributed to a lot of other things (distractions, etc etc), but, the high score is still indicative of your brains capabilities. If that made sense?

gojomo 13y ago

Increasing your score on an IQ test is one thing; getting the benefits expected from more "intelligence" may be another. In such matters, keep in mind Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure." In the domain of IQ, we might observe a very high correlation between higher-IQ and certain desirable results, like a longer/healthier life, lower rates of victimization, higher incomes, and so forth. And we might have a strong case that the test is, at least at the outset, measuring some sort of "general intelligence", and that there is a causal link from "general intelligence" to the other correlated beneficial outcomes. But, then you start aiming to increase "general intelligence", and your only measure is the IQ test, so your target for evaluating all interventions is the IQ score. Now you are quite likely finding interventions that only raise the IQ score, and perhaps skip any "general intelligence" benefits. Even if your IQ score goes up, you might not get the other benefits. Meanwhile, others still researching the correlation/causation questions may see the original correlations start to weaken, or the case that IQ measures "general intelligence" weaken. And in fact the originally-valid ideas are weakening! Motivated action around that target has started to thwart whatever signal it once offered. The problem is universal - in regulation, in education, in investment, in semantics. It's why a lot of high-performing organizations keep their true 'measures' secret... to slow their dilution through motivated optimizations.

moocowduckquack 13y ago
The other thing with optimising as a strategy is while it likes small changes, it really hates structural ones as they make it very hard to compare what you have now with what went before, so anything that is hard to measure is resisted. But usefulness and ease of measurement are not always particularly aligned.

jessriedel 13y ago

IQ tests are very robust. The correlation between scores is greater than 0.9. These working memory results are usually good for just a few extra IQ points (as in, literally 2 or 3), and they are far and away the largest boosts people have been able to induce through training.