No one cares about the ability to score well on IQ tests. They care about the things an IQ test can predict.

They are good at predicting the ability to do IQ tests!

Seriously though - aren't they still fairly controversial? I have strong doubts on an intuitive level.

Here's a nice short book which cites a lot of literature: https://www.amazon.com/Intelligence-That-Matters-Stuart-Ritc...

There's even a vox explainer on it, hard to believe they would publish this today: https://www.vox.com/2016/5/24/11723182/iq-test-intelligence

1. It seems fairly widely reported that you can practice at IQ tests and get better. Are there any studies confirming or denying this? I presume you'd agree that if true it rather kills the validity of IQ tests dead.

2. I have ADHD and I'm acutely aware of how environmental factors affect my performance in many tasks. How do you square this with the idea that they measure an objective innate quality?

3. How do you account for the Flynn Effect? https://en.wikipedia.org/wiki/Flynn_effect - are people getting smarter?

4. What's your thoughts on the argument that it's impossible to create genuinely culture free tests? This was something I was taught at college and it made sense to me then. Has it been debunked? Conclusively?

5. Steven Jay Gould wrote an entire book critiquing IQ tests : https://en.wikipedia.org/wiki/The_Mismeasure_of_Man - do any of the points he raises stand up to scrutiny?

2. Environmental factors affect everything. You try to standardise them as much as possible for an operational definition of the thing you're measuring. You will never succeed completely. Nevertheless IQ tests time correlate with performance on a wide range of cognitive tasks.

3. Not sure anybody really understands what's going on with the Flynn effect. Nevertheless IQ tests correlate with performance on a wide range of cognitive tasks.

4. I'm sure it is in fact impossible on some level to create genuinely culture free tests. If we're going to be anal about this, this is a consequence of the various no-free-lunch theorems in statistical inference. For any culture/person/algorithm there are problems where they'll do better relative to others, and problems where they'll do worse. Nevertheless IQ tests correlate with performance on a wide range of cognitive tasks.

5. I'm sure they do, but I haven't read the book and the summary in the Wikipedia article is too vague to tell.

None of these objections actually engage with the parent comment's assertion that "Ability to solve these puzzles correlates with ability in a diverse range of cognitive abilities".

(3) is very interesting and there is not yet any agreement as to what is causing it. Flynn seems to have thought our culture is slowly becoming better at training us to think abstractly. There's some evidence for this: one Australian study found that over 20 years the students in a set of schools became better at a vocabulary test but did not improve at all in processing speed. This is not what you would expect to see if the improvements were due to better nutrition or less environmental lead.

For (5), haven't read it. He's stunningly erudite and a careful author though, it's worth taking him seriously. Do you have some examples of points he raised?

IQ tests, or any standardized tests for that matter, blow up this assumption as do aggregated data showing differentials across groups or even nations, and that intelligence as even crudely measured by IQ tests which is just a proxy for something that isn't so precise, still correlates with general positive outcomes in life such as earning potential, health, career success, national achievement in science etc. It is also not politically correct that this dynamic is largely a biological phenomenon, that intelligence like many other physical characteristics is heritable.

This is why generalized IQ tests were banned from job interviews because not everyone scored the same, which is politically incorrect and unacceptable in our current political climate.

Basically for the liberal "left" which dominates western culture, "equality" is almost a religious belief, in opposition to any acknowledgement or even promotion of hierarchies, or recognition of inequality as an objective reality of human existence. Anything that undermines "equality" and points to natural hierarchy is "controversial" not allowed.[0]

[0] A random example, but there are many, of unacceptable differentiation when it comes to group cognitive outcomes. And elimination of standardized tests or advanced classes in the name of "equality" and against any instances of "inequality"

https://www.washingtonpost.com/education/2021/06/04/californ...

They are not controversial at all amongst psychologists. The notion that IQ tests have predictive power is about as settled as the notion that "people get sad sometimes".

/s

For what? (Correlations are interesting scientifically of course, but for everyone else?)

> No one cares about the ability to score well on IQ tests.

Except for Mensa and their applicants.

I don't mind IQ research. I do mind exclusive clubs for the sake of exclusivity, and especially those that are (supposedly and explicitly) genetically determined entry criteria. It smells bad.

Similar tests, like the SAT and GRE, are designed to work even with practice. They have other problems.

What I'm skeptical of is applying stereotyping - and that's what this is, it's s predictive model that's based on averages, not the individual - to any kind of real decision that's any more weighty that letting you into a club. Like education decisions for example.

It is also aversion or even policy regulation against even giving tests and acting on the results towards individuals, as this entails inevitable discrimination when the natural and unavoidable hierarchy of human abilities is observed.

IQ is very much so not genetically determined. See for instance the flynn effect which is happening far too fast to be the result of genetics. IQ is only ~50% heritable.

? what is your stance on, e.g., competitive sports teams ?

The speed at which a person can recognize patterns correlates strongly with their ability to learn in general.

The reason they use shapes is because it's a more universal measurement. Using words introduces issues with a person's ability to speak that language.

How well would you do on any test written in Japanese? Assuming you never learned Japanese. Probably poorly. But if you knew a little bit of Japanese, you might do fine on a math test. And that's the idea with shape-based tests: trying to minimize the influence that language comprehension has on the results because you want the results to be neutral.

Pattern recognition is a large part of that. A person who is able to discern patterns more easily, is usually able to reason better. More able to do second and third order thinking.

IQ "tests" that use words or numbers are also testing knowledge to some degree, which isn't something you typically want when you want to test raw intelligence. Being able to do calculus doesn't necessarily prove you're intelligent, it proves you know calculus.

So that leaves abstract concepts. And that's another facet of intelligence, being able to think abstractly. There should also be some "die folding" questions in there. Where it shows you six connected boxes and you pick the cube that it makes. That one is also about spatial relations as you have to model the cube in your head in some fashion.

All that being said: intelligence is nothing on its own. It's a force multiplier. Intelligence allows you to use the knowledge and skills you do have better. It also makes the acquisition of knowledge and skills easier. People with higher IQs have better life outcomes on average than those with lower IQs among every cohort. What that means, if you take all the poor people, those poor people with higher IQs are going to be doing better overall. Still poor, but able to make more of what they have. Tall people? People with higher IQs will be doing better overall than those with lower IQs. Doctors? Same. Welders? Same.

Now, this doesn't mean that every person with a higher IQ is doing better. Success is very multi-faceted and dependent on a lot of things, some of which are outside of our control entirely. But it helps.

It has a time limit, so it does measure how quick you are.

So it probably bakes in a range. Like if you complete this test in the allotted time, you will be in this range. But if you completed it in half the time, you'd probably be higher, but they haven't calculated the range you'd be in. They'd just say, "Yeah, you're in the top range".

There are awards, patents, accomplishments, etc.

There are positive correlations between IQ and positive outcomes in general. Basically, take a metric of success and group people into it and you'll find people with higher IQs in the groups with better outcomes.

Even if you then slice those groups by other factors.

Second, why would we even have to cut off the tail portion?

Next, you didn't cover that contingency. You keep saying it's all income levels, but it's not. IQ and positive outcomes are positively correlated. Pick a thing, pick a measure of success, you'll find that people with higher IQs will tend towards the successful group.

And this is also a "all doctors are tall, not all tall people are doctors" kind of statistic. Higher IQ people will wind up in the more successful groups more often, but being in the more successful group doesn't mean you have a high IQ.

To be fair, most people don’t understand how paracetamol helps reduce pain. But it does and we are able to measure its effect, so we have scientific studies that prove it. It doesn’t matter if you understand how it works, it does.

Same with IQ tests. It’s not about the pattern matching. It’s about the fact that a huge mountain of scientific literature have found these measures of pattern-matching abilities to be correlated with a wide range of other abilities and accomplishments. If a persons ability to stand on one leg for extended periods of time was a better predictor, then that is what we would use instead. It doesn’t matter if we understand why that would be the case. But ofcause everyone who scores less than they feel that they should end up trying to discredit the entire concept rather than just accept that their poor ability to stand on one leg for extended periods of time does not define them, or impose any new limitations in their life.

I swear every discussion about IQ ends up feeling like a room full of short men debating if measuring tape is even reliable and debating if it’s even fair to say that tall men can reach thing on top shelves more easily since we live in a society with chairs, and there’s obviously something wrong because they have tons of arguments for why they are 6 feet tall, even though the measuring tape doesn’t agree. No one is saying that you can’t have an absolutely great life, or reach the tallest of shelves, being a short man. The fact that we can measure peoples height, and that certain things are different base on how tall you are doesn’t have to offend people.

Last patterns were boolean logic, for example 3 or 4 tests were just XOR stuff, another was AND, another was summing, and so on...

I ended with a score lower than my usual, because I wasn't really paying attention... and bored.

But I agree that it is unfortunate to base the future of a person on a one-time test. That includes the finals in school or a day full of exams after a long semester.