Hacker News

points by sjducb 17 hours ago | 0 comments

We care about racial differences in outcomes that are correlated to IQ, such as employment and imprisonment.

What if some of the observed racial differences were caused by a lack of vitimin D, due to darker skin photosynthesising less of it? Vitamin D is essential for brain development. Black people in northern climates get less of it than evolution intended. This lack of vitamin D might be causing reduced brain development in black populations. If that is true, we could fix the prison crisis and racial IQ differences with a pill!

Sadly we don’t know because you can’t do the research to test the hypothesis.

If we block all research into a problem then we can’t solve it.

If you actually care about fixing a problem then we should bring all available tools to fix and solve it.

[–] vintermann 15 hours ago

It's a good tool if you want to rank people, it's a bad tool if you want to help people. IQ tests couldn't (and can't) even detect dyslexia, a specific disability which you can get help with, which was known even in Terman's time. IQ tests by design throw out all information which could complicate ranking people. If your goal is to help, why would you measure people with a tool which can't tell the difference between dyslexia and concussion?

IQ tests have been revised several times to address things like that, like that they "unfairly" penalize dyslexia, or that they had lots of questions which were easier for people of specific social backgrounds. Every time they were, they naturally became less "predictive", since both dyslexia and not belonging to the California white upper middle class give you pretty obvious disadvantages in life outcomes.

So don't waste time on IQ, and don't waste time on other academic culture war BS either.

[–] tptacek 12 hours ago

Wait, this is the opposite of reality. IQ is in fact a useful tool (in a battery of other tools) for helping diagnose cognitive deficits and learning disabilities. That's what it's good for: as a clinical tool. It's in fact terrible for ranking people.

[–] vintermann 10 hours ago

Well that's what it was used for! Do your history homework, mr. clinician!

Tell me then, if someone scored low on a text-based IQ test and that's all you know, how do you tell if it was caused by dyslexia or concussion? I guess you break out that "battery of other tools"? What does IQ tests give you that those other tools don't give better?

Oh, I know one. An excuse to write someone off as simply stupid. An excuse to _stop looking_ for the things which might have helped them. Which is what it was historically used for.

[–] sjducb 15 hours ago

If you can’t study IQ how do you demonstrate that lead poisoning causes reduced intelligence?

The research to ban lead in fuel relied on measuring IQ and measuring lead exposure, then showing a correlation between the two.

At the population level IQ is a very useful tool.

[–] jcranmer 15 hours ago

IQ is a very bad tool for measuring populations. That's because IQ is definitionally designed to force a population to adhere a particular metric: a bell curve distribution, with the median at 100 and the standard deviation of 10. Whatever the thing you're measuring is--and there's good reason to think it's not "intelligence"--you're crunching and distorting that measurement to get the number at the end of it to force the population to adhere to a standard distribution.

We don't have a good model of what intelligence is. But since IQ studies are fairly common, there's a lot of data collected that uses IQ, and since it IQ looks vaguely like "intelligence," a lot of researchers who are motivated to find reasons to declare particular people intelligent or unintelligent are more than happy to use that data to support their claims.

[–] sjducb 14 hours ago

Did you even read my comment?

Answer this question:

> If you can’t study IQ how do you demonstrate that lead poisoning causes reduced intelligence?

[–] jcranmer 12 hours ago

The answer to that question should be blindingly obvious: you don't.

It should be clearly evident that trying to use IQ to establish that "<insert ethnic group here> are stupid" is bad science. It's the same sort of science to establish that "<insert environmental factor here> causes stupidity," so we also need to reject it even if the environmental factor is clearly something problematic. The problem isn't the conclusion, it's the methodology.

[–] sjducb 11 hours ago

So do you reject that lead poisoning causes stupidity because it is unethical to measure intelligence?

It’s a good thing that measuring IQ enabled us to discover that trace amounts of lead in the environment causes reduced intelligence. That was a key piece of data for banning lead in fuel. The loss of IQ points happens to all kids living in cities, rich and poor. Rich powerful people didn’t want their kids made stupid by lead in the air.

I think that removing lead from paint and fuel has greatly improved the lives of billions of people.

[–] vintermann 10 hours ago

No, we knew lead was bad. We knew what heavy metal poisoning looked like, we knew lead caused heavy metal poisoning. We knew of a hell of a lot more concrete effects than average lower scores on IQ tests.

> The loss of IQ points happens to all kids living in cities

No, it didn't. The reduced test scores happened _on average_.

I won't rule out that what you said is correct, in that it may well be the fear that "my child may lose 1-5 IQ points!" helped get lead banned, but if so, that's an indictment of society.

[–] tptacek 11 hours ago

I agree with half of this argument but think you're taking it too far when you reject IQ testing as an epidemiological tool. It's not dispositive. It's statistically ill fitted as a ranking tool, for the reasons you stated, and the fact that it's not reliable across its whole range, and test-test reliability issues, &c. But like, "you can use it as an indicator for environmental causes of cognitive harm" is actually a valid argument, I think?

[–] ogogmad 16 hours ago

> correlated to

Correlation measures whether Y=mX+c holds with small error. I sometimes think that many people unknowingly assume that all dependencies are somehow straight-line ones. This is distinct however from "linear approximations" in physics, which can at least be justified sometimes.

I also suspect that correlations of 0.5 or lower are often just statistical noise.

[–] sjducb 15 hours ago

I don’t see how your reply relates to my comment. At no point do I say that any correlations are linear, nor does my argument depend on linear correlations.

Fundamentally my argument is maybe science can solve a problem, let’s at least try.

[–] mjanx123 2 hours ago

The US never truly abolished slavery, it turned it into a punishment, and started sending black people into prison en masse. There is no prison crisis, in fact the prison industry is flourishing like never before!