Hacker News Reader

Comment by altruios | Hacker News Reader

I would say (it has been a while), the common complaint that I've heard is the tests for IQ are not standardized - there are a million snake-oil pitches for free and paid online IQ tests - or that the real tests have a bunch of cultural bias that might skew the results... Or that IQ is inadequate as a measure of general intelligence (if IQ measures pattern matching effectively and it misses other kinds of reasoning entirely)...

lots of complaints I've heard - can't say one way or the other if there is any salt to them.

I haven't heard the complaint before that IQ is imaginary, that's a new one... I think through reading the exchange I see where the semantics shifted between people's thoughts, I doubt that's what the the poster above you meant - likely they meant a combination of the many complaints to dismiss the general usefulness of IQ tests as a predictor of intelligence.

View on HN

That sounds like a complaint that online IQ tests are bad, which is obviously correct but also not really relevant to IQ as studied in psychometrics.

The complaints I hear about IQ are completely opposite! People dislike standardized exams, including IQ tests, because of their standardization! That is, the need to assess everyone the same way precludes assessment specific to the individual being assessed. Detractors of IQ, insofar as they would like assessment at all, would prefer a more holistic procedure where the characteristics and capabilities of the assessee are analysed with due attention paid to the specific life history of said assessee.

There are certainly a lot of valid complaints about IQ. The lack of standardization is obviously a problem, and the cultural understanding of it is severely lacking. But study after study have shown that IQ is highly correlated with financial success, and at this point the confounders have been controlled well enough to say it’s hard for me to deny that pattern recognition, which is what well made iq tests measure, is a cause of success.

I am not sure what is defined as "financial success," since it appears to be something one can be born into regardless of IQ.

In regards to income, I briefly skimmed the Wikipedia section for IQ and Income, which can be found here:

https://en.wikipedia.org/wiki/Intelligence_quotient#Income

"Highly correlated" would not be the choice of words I use to describe correlations between (0.23 - 0.5).

We, as people, should cherish the contributions of one's intelligence instead of obsessing over an arbitrary, qualitative value painted as a qualitative value like one's potential intelligence.

Maybe you’re not familiar with how correlation is reported, but 0.23-0.5 is very high.

I think I am hung up on semantics.

If 0.23-0.5 correlation is "very high" then what is 0.7-0.9? Would r >= 0.7 be literal causation at that point for psychologist? (I'm kidding, of course)

Again, I am not saying that a correlation of 0.23-0.5 is statistically insignificant by any means. I just think I needed "very high" to be coupled with "for psychology."

> Highly correlated" would not be the choice of words I use to describe correlations between (0.23 - 0.5).

lol, this range is 0.5 to 0.7 variance explained (R squared) which is an insane value. in typical social studies a correlation of > 5% is considered a major improvement. In finance a correlation or > 1% is considered amazing.

You are either a hyper-intelectual drinking Taleb's kool-aid or simply don't know what you are talking about :D

If the correlation is 0.2 then it explains 4% of the variance. If it's 0.5, then 25%. I'm not sure how you got your numbers, can you clarify?

Also 10% variance explained in social science is considered pretty good (which is depressing, but such is life).

> a hyper-intelectual

Very far from it. I'm just an idiot with Internet access.

> drinking Taleb's kool-aid

I do agree with some of what he has said about IQ, but I am not well-versed in any of his other writings.

> simply don't know what you are talking about

Perhaps, but I am still not convinced.

Pattern recognition is an interesting way to describe what is being measured. I've always heard that IQ tests are really good at predicting the ability to understand complexity, or perform complex tasks; such as in the way the military has traditionally used intelligence tests to match recruits. I wonder if being really good at advanced math just means 'really good at pattern recognition'.

Anyway, I saw an interesting theory once that suggested IQ simply measures motivation and highly motivated people will tend to do better on standardized tests as well as more generally in life.

Reply on news.ycombinator.com