IQ Is 'Real'

‘IQ isn’t real’ is a claim I’ve argued against too many times, so I’m going to write it down my opinion once and for all. I claim zero novel contributions to this topic.

If you already think IQ tests do an imperfect but decent job measuring the thing we casually refer to as ‘being smart’ then you can skip this post.

B: IQ only measures your score on one test. How could that possibly be useful?

A: Scientists first proposed IQ (back then ‘g’) because they found that mathematical reasoning correlates with verbal reasoning correlates with reaction time, etc. The IQ test is engineered to be maximally informative about all these related attributes in one short test.

B: But there are other ways of being smart. Isn’t one score close to useless?

A: Name your way of being ‘smart’ that doesn’t correlate positively with IQ.

B: Being emotionally intelligent. Being wise. Being virtuous.

A: Emotional intelligence is positively correlated with IQ. $r=.8$.

I suspect you would define wise in a way that isn’t measurable. If this argument ends in us agreeing “all measurable traits which one would describe as ‘smart’ are positively correlated with IQ” then one should have the prior that the non-measurable ones correlate positively too.

Being virtuous is definitely not the same thing as being smart. I accuse you of trying to paint me into the corner of ‘all good human traits are positively correlated,’ which I did not claim.

B: IQ only measures your upbringing and your education.

A: IQ is most strongly correlated with your environment when you’re younger and gradually decouples. Reasonable estimates for what % of IQ variance in adults is caused by education / upbringing vary from 0% to 20%. Reasonable estimates for what % is genetically determined range from 60% to 80%.

Words can be useful descriptors even if they entirely result from our environment.

B: IQ is a test that can be gamed.

A: If you take a second IQ test, your score improves by about 5 points. After three tests, your score stops improving with practice, for a total boost of 7-8 points. An extra year of schooling improves your score by 1-5 points. It is weakly gamable.

Even if it was totally gamable: who cares! One can still make apples-to-apples comparisons, eg two kids who take the test with similar amounts of prep. In practice, one can make reasonable comparisons between different test-takers because the phenomenon is so well-studied.

B: If IQ is real, you’ve wrecked growth mindset.

A: Growth mindset works with effect size $\approx .1$. To the extent that believing IQ is heritable inhibits growth mindset, I’m sad.

Believing certain true things is sometimes detrimental. IQ does not fall in this bucket. Knowing that IQ is real lets one make better choices, by, eg, acknowledging that tracking makes sense and implementing it.⁠^[1]

B: If IQ is real, then some people are better than others.

A: I disagree. Some humans are taller than others. Heightists value tall humans more than short humans, even though height is genetic. Morality issue on behalf of the heightists. I choose to value everyone equally, regardless of their IQ or height.

B: My IQ is low, and I feel inferior.

A: You shouldn’t! Read The Parable of the Talents (SSC). In short: we are compassionate to people who are overweight or balding or unathletic, because these things are largely genetically determined and there’s no point in shaming people for things outside their control. The same should apply to intelligence.