Hacker News Reader

Comment by teaearlgrey_hot | Hacker News Reader

May I be the first to say that I’m not sure what “IQ” actually measures? Naturally I have the same understanding as most: IQ is supposed to be like Horsepower for brains. “This brain has so and so much more thinking ability than others”. Yet I find a curious phenomenon when spending time /with/ the brains of others through their work, writings, or speech: often my level of engagement or intellectual fulfillment that I gain has little to no correlation with their IQ. What IQ seems to fail in capturing is the human ability to be creative and unique; whereas I believe now one can design a sufficiently advanced AI/ML instance which could learn the patterns in this Mensa test, our ability to assess value and meaning from a perspective of “well they recognize certain types of patterns” seems carelessly limited. But that’s just my perspective.

View on HN

Instead of asking what IQ measures I think it's probably more useful to simply ask what it predicts.

Often discussions regarding IQ seem to get bogged down in debates around the definition and value of intelligence, or the ability of an IQ test to accurately measure it. But whatever your opinion on what IQ tests are measuring, IQ has undoubtably been shown to have a high statistical predictability for many of the positive (and negative) life outcomes we're interested in as a society. For example, if you score higher on an IQ test you're statistically much less likely to go to prison and much more likely to have a higher income.

Personally I think IQ is something worth thinking about in the context of group behaviour and performance. For example, at the level of a nation it's worth measuring mean IQ and being concerned about any persistant annual declines in IQ, or simply to understand why some cohorts might have significantly higher or lower IQs. On the level of the individual I'd question how useful it is.

"For example, at the level of a nation it's worth measuring mean IQ and being concerned about any persistant annual declines in IQ"

Then the nation should filter its immigrants by IQ and promote natality for the most smart stratum of its own people.

As research using twins followed by 30 years in adoptive family claims majority of IQ is genetic.[1]

[1] - https://www.sciencedirect.com/science/article/pii/S016028962...

This particular test is an imitation of Raven's Progressive Matrices, which measures working memory (WM). WM is the amount of information "chunks" an individual can work with in their short term memory. For an average person it's "7+-2" or 5-9 "chunks".

The patterns you see on the test introduce progressively more and more changes, until you are getting beyond 9 changes a piece at 120+ IQ. Let me know if you have any questions.

Interestingly, young chimps outperform humans in working memory tasks...

https://www.newscientist.com/article/dn12993-chimps-outperfo...

But has it been shown that working memory correlates to anything in specific?

> IQ is supposed to be like Horsepower for brains.

I don't think this is a very valuable way to think about IQ. It is what it is, nothing more nothing less. It's the location you fall on in a normal distribution of intelligence as measured by specific tests.

I have heard that IQ measurements correlate very well with general success in society. Peterson says they correlate better than any other measurement in human Psychology. (Paraphrasing and I can't validate it)

Peterson also says that it doesn't matter which questions you use. People with hi scores on one test will typically have high scores on the other tests.

It seems to be a hard thing for intelligent creatures to hear:

There are some smarter that you and some less smart and you are most likely in the middle range somewhere.

I'm not disagreeing at all. I'm just saying that isn't what IQ is. It correlates with all sorts of stuff and understanding those correlations is useful. But IQ isn't its correlations. I definitely believe those correlations exist and are pretty strong.

It isn't "Mental Horsepower" though. If something called "Mental Horsepower" or "General Intelligence" exist which I don't personally buy, IQ probably correlates with them but it isn't them.

IQ is a location on the point of a normalized distribution of scores on a specific set of tests. Nothing more, nothing less.

Correlations don't guarantee things to be true. If IQ correlates strongly with "Mental Horsepower" it could still be very possible for a specific individual to have high IQ with low "Mental Horsepower" although it is statistically less likely.

> There are some smarter that you and some less smart and you are most likely in the middle range somewhere.

This always seemed to me like an important piece of social intelligence. I've noticed that people who are intelligent enough to be smarter than most around them, but who take that to mean nobody else is as smart as they are, end up getting played a lot by other folks who recognize this failure.

We can also test how far most people can jump and their ability to memorize numbers or do arithmetic or algebra quickly. We don’t have a “movement quotient” or a “calculation quotient” so what exactly does an “intelligence quotient” aim to quantify precisely? It’s not intelligence, I’ll argue.

We absolute have and use these things. This is why the US army makes you run a mile and counts your pull ups. These things are not measures of your "true physical fitness", they are just data points. But they have extremely high correlation on other fitness metrics so we just measure them and assign a score.

You could very reasonably label the outcome of that test a movement quotient.

For the general population, pullups are just measuring bodyweight.

??? Yeah if you average everyone together. You could say this about every fitness activity that exists.

It doesn't mean you aren't going to see big differences in results when you test it which is why they do it.

You can't say the same thing about everything. If you measured absolute strength the correlation with weight would go the other way!

> What IQ seems to fail in capturing is the human ability to be creative and unique

This is completely contrary to the evidence we have. Creativity is correlated with IQ and inversely correlated with age. A lot of famous artists in a variety of fields are recognized to be quite intelligent. Most creative arts programs in universities are highly competitive; perhaps the most competitive of any college in a university. And a surprising number of famous musicians have degrees in other difficult fields, for example the lead singer of the band The Offspring has a Ph.D in biology.

Additionally, it has been observed that papers written by younger mathematicians are much more likely to have novel approaches to solving problems. Suggesting that younger minds are more creative than older ones.

Reply on news.ycombinator.com