Hacker News Reader

Comment by sfblah | Hacker News Reader

There aren't that many, but here's one:

https://www.sciencedirect.com/science/article/abs/pii/S01602...

It's a twin study. Here's a quote: "Mean IQ scores were modestly higher among those from higher SES backgrounds, but the magnitude of genetic influences on IQ was uniformly high across the range of SES." SES means "socio-economic status".

Honestly, the evidence for heritability of IQ is very strong. Arguing against it, in my opinion, borders on arguing that vaccines don't work. I actually understand why people do it (there's a lot of very bad history in how IQ data has been used). But I just think it's more important in the long run to tell the truth and find solutions based on the truth.

View on HN

> Arguing against it, in my opinion, borders on arguing that vaccines don't work.

This is an interesting opinion. So vaccine research has produced a number of successful trials, the research is keeps furthering more knowledge, we develop new techniques, new models, new successful predictions, and new products which yield more successful trials. I consider this a healthy field of science and engineering.

Now how does IQ research fair next to vaccine research. The first IQ test was authored in 1908, not to measure intelligence among populations, but to assess learning disability. Since then IQ researchers have been trying to prove that IQ differences exists among populations. They have created and standardized new types of IQ tests that are supposed to show this difference exists. They have spent the past 100 years doing this, using fancy statistical methods, but have so far failed to convince the other fields of science (especially other fields within psychology), nor have they convinced policy makers (though they were pretty close during the eugenics craze). They have no theory of behavior other then this hypothetical general intelligence. They have no models that predict behavior, they have no products other then their intelligence tests, the tests are constantly criticized, have no successful double blind trials to show for them selves. The techniques are the same as in the 1970s, just plain old pen and paper tests with some factor analysis to spot the correlation they were hoping to spot (some would say, preconditioned or biased to spot).

If I were to be generous I would categorize IQ research among string theory as a scientific dead end. They had a theory, it didn’t go anywhere, and now the rest of the field has moved on. However given the history of IQ research and their ties to the eugenics movement, I’m not gonna give them this benefits. I believe IQ research, with the exception of Alfred Binet him self, was always about putting people in racial categories to show that one group was superior to another, with a made up construct they call intelligence.

IQ testing has numerous applications. For example:

* SAT tests largely amount to IQ tests and are used extensively for college admissions in the US. College degrees from top US universities remain the most sought after degrees in the world.

* Magnet schools in urban areas regularly utilize IQ tests or their equivalents to elevate students into advanced coursework. This enables students to rise above their socioeconomic status.

* European secondary schools regularly use tests that correlate with IQ tests to track students into programs leading to universities versus technical schools. These tests help coordinate society and deploy human capital where it is most needed.

* Even the US military uses IQ tests to avoid accepting recruits who don't meet a certain standard. This aids in ensuring the security of western nations.

IQ testing could be used in additional areas, such as:

* Encouraging higher IQ parents to have more children through subsidies, thereby increasing the likelihood of future technological innovations assisting all mankind.

* Assisting in identifying individuals who need more support from society.

* Better understanding the causes behind social ills like poverty, drug addiction, homelessness and the like. Understanding that an individual's IQ is a stronger determinant of economic outcome than other variables, such as race, is a step toward a society with more solidarity and less friction.

I appreciate that you have a bias against intelligence testing, and I understand its genesis (your mention of the "eugenics craze" makes that clear). I'm suggesting society aim toward an enlightened view of these things where people's IQs don't identify them as "greater" or "lesser". Wishing IQ away is unlikely to work, particularly in a world about to experience the arrival of super-intelligent AI systems.

The geocentric model also had numerous applications. A standard's widespread adoption speaks nothing to its utility.

> Encouraging higher IQ parents to have more children I'm no expert but this sounds exactly like eugenics or at least engineering the gene pool.

Do you really view lower IQ individuals as equal if you want them to have less children (at least as a ratio of total births)? The slippery slope argument would be that you're a few steps from sterilizations.

Are we sure that "super-intelligent AI systems" wouldn't displace intelligent individuals? Seems like they'd certainly be the ones wielding a middle-ground (say, GPT-6/7/8) most effectively but if an AI reached 100% coverage of human skills (imo unlikely in our life time) we wouldn't need any theoretical physicists.

Reply on news.ycombinator.com