IQ is a dismal measure of intelligence (2019)
supermemo.guru
supermemo.guru
IQ is not a perfect measure of intelligence, but I think it's better than no measure of intelligence.
IQ tests are only a clincally valid test of extreme mental deficits. They do not correlate with objective measures of above-average performance, and have created a "metrical culture" with severe unethical concequences.
Fit a straight line, find a straight line.
Where IQ scores correlate, it's at the very-low-score end. It doesnt correlate with independent objective performance measures at average-and-above.
The people writing those papers are psychologists, not statisticians, and often the most poorly trained. Most of it is either circular pseudoscience or unreporducible. Indeed, for much of the IQ literature to be "unreporducible" would be a great credit, since it would mean it had a testable thesis -- which, largely, it never does.
The beginning and end of IQ research is, largely, just that straight-line rulers can draw straight lines in any dataset you put them against.
The takeaway is that IQ doesn't matter as much as other attributes (character, learned skills, etc.) in life outcomes.
HN style crows way overvalue IQ style skills when intelligence has multiple dimensions, many of which matter more than IQ.
IQ is effectively just a measure of test-taking skills. That's useful in some areas, but you could see someone like Gengis Khan or Nikola Tesla be incredibly intelligent and a poor test taker.
[1] https://heckmanequation.org/resource/i-q-isnt-everything/
Uhuh.... so this line it's crazy, right?
It's really strange you concede that we have no theory and hence no way of verifying whether an IQ score is a valid and reliable measure of the thing we have no theory of... and yet, you say it is a "proxy for intelligence"
The entire edifice of research which makes these claims is a product of a pseudoscientific use of statistics by amateur statisticians. The vast majority of the results you think exist in the literature are frauds, and often egregious ones at that.
A human being is one of the most complex systems known to exist in nature, and the frauds in psychometrics are very far away from being up to the task of studying them. The greatest minds in applied sciences, matehmatics and statistics would largely throw up their hands at the job of measuring "intelligence" --- that some fools havent, is no good thing.
> The entire edifice of research which makes these claims is a product of a pseudoscientific use of statistics by amateur statisticians.
That's bollocks. It's just a topic that's too complex. Unfortunately, everybody likes it and draws conclusions from it. Even the HR department. If you would say we should ignore IQ measurements in society, I would agree. But it's not pseudo-scientific use of statistics. Nor all done by amateurs. It might even be more scientific than string theory or dark matter.
None have PhDs in statistics. They're all people who got into psychology and were given GUIs by the maths department. Their papers are a symptom of what happens when you press the "ANONVA" button and print the results.
The entire field is an egregious testament to pseudo-intellectual pseudoscience. It's shocking.
There's nothing there. It's unfathomably bad.
A lot of psychology is bad research. There's a reason there's a reproduction crisis. Statisticians are to blame for that, too. They've been teaching and using NHST, even though all arguments point away from it.
But statisticians can't fix a bad design, because they rarely know the factors involved, and they can't fix errors in the methods, because they've got no idea about the subject matter. The statistician in your argument comes across as some kind of totem.
> There's nothing there.
There's something there, but since we don't know what, it's best to take it with a grain of salt. However, if even the bloody HR department looks at IQ scores, it gets a lot of attention. And one explanation is that it predicts better than nothing. If you have to choose between two candidates, and one has an IQ of 80 and the other of 120, you know who (on average) who fits best.
A score of 80 suggests a possible clinical issue with mental retardation. Few such people apply to roles who compete with people at average or above scores.
What is much more the case is people 95-145 "competing" with each other based, effectively, on what star sign they have.
The test probably breaks down at some level since it is hard to test higher intelligence, but it seems to me like it is still valid below about 120 IQ.
Something having an objective measure or not does not invalidate it.
But in any case, what we're talking about is the meter-ruler called "IQ" and its use by the most batshit HRish elements of soceity, and the most pseudo-scientific elements of psychology.
In a world where major employers are using IQ to guide their hiring practices, we're not talking about a quiz on the back of cosmo.
We're talking about a world where business (, uni,...) has outsourced its obligations to fairly consider the merits of individuals to the cult of psychometricians extremely happy to provide astrological justifications for ruining people's lives.
An entire medical establishment has been created by the IQ test, which exists across (at least the west) as a filter on entrants. What strange, bizarre, and often poorly sutied creatures have found themsleves in positions within this establishment?
When one looks around at the society the IQ test has created, one should be aghast. Why, so often, are nurses better than doctors? (etc. etc.)
What the evidence says is that IQ isn't the only factor, just as height isn't the only factor for NBA players. But just like height, IQ makes things a lot easier.
I'm still waiting for somebody to show me a prolific Math Phd with below average IQ. The only counter example I've heard people use it Feynman whose IQ was "only" 125 putting him smarter than "only" 95% of the population.
It rubs me the wrong way when Hawkins says stuff like "IQ is for losers" but then boasts how his physics undergrad degree was so easy for him he was just bored out of his mind (it sure wasn't easy for me). It's the intelligence equivalent of a billionaire saying you just need to work hard to get there, as if that was ever gonna make a difference if you're an Amazon factory worker.
Note, that when carear-IQ correlations are brought up it is universally in those populations subject to an IQ test filter who are then shown to correlate with IQ scores (ie., those who take, eg., MCAT, LSAT, ...) --- which, if it isnt clear, is circular.
Average and above IQ shows no correlation whatsoever with, eg., wealth.
Am I smarter than the average person? undoubtedly.
But if my IQ is high then I fear I'll avoid doing useful things because they're beneath me. And if my IQ is not high, I fear I'll avoid doing things because they're above me.
There is nothing useful that can come out of knowing your IQ, and as others have said, there's other things that are so much more important. Smart people are unusually good at deluding themselves, for example. They get so used to being right they start to believe if they can rationalize it, it's obviously true and anyone who disagrees isn't as smart.
Having said that, I once made a girl angry in college because I asked the professor if grad school will actually be challenging because undergrad math for me was ridiculously easy.
I think the best way to describe IQ is that it defines your potential (w/i a margin of error), but it's everything else that determines how effectively you use that IQ.
A child with down syndrome will never win a nobel no matter how much they want to or how much effort they put into it. But having an IQ of 160+ will never guarantee you don't end up homeless in a ditch.
Knowing my IQ would let me know if that's reasonable and if I have any kind of shot at it. I know I'm smarter than average, but am I smart enough to be a top scientist? If i'm too stupid to attain my dream and I went for a phd, I would just be wasting my time, my life, my earnings for a dead end. It would be like dedicating your whole life to being a pro basket-ball player while having no shot because you're 5'10" (taller than average, but not tall enough).
The one they settled on was a man in a wheelchair. Through technology they fixed all his ills.
They then put him on a ship with all the others and he was immediately attacked. It turned out there was a hierarchy and it had to do with physical violence. They wouldn't let you die and all injuries would be healed so it was safe, if painful, to do.
Throughout the book you learn the aliens that came to earth are very logical, by the book, run the numbers.
So this human realizes this and realizes he wants more control so he starts working his way thorugh the hierarchy. he gets to the 2nd highest rank, but cannot beat the highest rank alien, but he keeps trying ... and trying ... and trying. Finally the highest rank alien asks him why, he explains, the highest rank alien steps to the side and gives him the highest rank.
Later on in the story the aliens fighting the war decided the statistical odds weren't there and they were going to let all the worlds where they had gathered specimens be overrun. The crew on this ship with the humans rejected it, overtook the ship, and went to fight on their own. They did well enough on their own to tip the numbers just enough that the aliens fighting the war chose to re-engage.
----
The point here is this. You cannot live life by the numbers, you'll be wrong because it doesn't take into account risk. I know you, and many others, will argue against that, but it's true.
Here's an example: Statistically speaking, you should never cross a busy highway. If your child is on the other side of that highway and in danger, you will 100% do it. Statistics can never take into account the reasons for the risk or the worthiness of the risk.
----
The numbers are lying to you. Yes, there's a line at which you'll never be able to do better, but once you're past that line there's a large margin of error. Don't live your life afraid of that margin of error, embrace it.
Find me a below average height person who made it at the NBA or somebody, anybody, with a below average IQ who got a science Nobel prize and you'll have my attention. But if you have to rely on a pure fantasy novel to find a counter-example, it doesn't give me any confidence in your theory.
there, fixed that for you.
As for the rest ...
https://en.wikipedia.org/wiki/List_of_shortest_players_in_Na...
The shortest NBA player was 5'3" and played until 2001.
Isaiah Thomas is the shortest player to play in an all-star game at 5'9".
---
But hey, good luck, I'm sure you're smarter than the rest of us and you'll do just fine without allegories.
Omg, no.
Good test taking skills will help someone score close to their theoretical maximum, but you can’t just train someone in test-taking skills and expect them to score at the top of the scale.
I’ve trained people on test-taking skills to great effect, but there is a limit (often significant) on how much these skills can improve a score.
[all North American cities](https://www.reddit.com/r/MapPorn/comments/ym6ms5/a_map_of_ev...)
When/how is IQ ever used for good? Everyone I have ever talked to about it who talks about others' IQ intends to use those people for their own selfish purposes. And people I've talked to who talk about their own IQ only ever use it to brag, boast, or otherwise elevate themselves. Neither of these are behaviors that should be encouraged.
Generally this is not based on actual IQ test results, but on teachers impressions. And they keep "secret" which is the high IQ class. And if a kid is really motivated to be in the better class, change the kid to the other class (and vice-versa) if at all possible.
Of course this worked better with 15 kids-per-class than with the current 40 kids-per-class education system.
This is doing a lot of good, and has been for decades. Essentially it results in a smarter and more capable society, because people can be educated much closer to the limit of their abilities.
However, it's a politically popular position in some circles in recent years.
Compare Wikipedia in 2015 [1] with Wikipedia today [2].
[1] https://en.wikipedia.org/w/index.php?title=Intelligence_quot...
[2] https://en.wikipedia.org/w/index.php?title=Intelligence_quot...
Subsequently, I watched a long-form breakdown of IQ and my doubts increased. I then "clicked" IQ into place: for most of my adult life ive been complaining about fMRI, pharam, (AI), etc. -- ie., medical misuse of stats --- and somehow Peterson had lulled me into a false sense of security.
Taleb's article was the nail in the coffin, and I read around the references associated with it. I now place psychometrics firmly in that category of pseudoscience which is "absue of statistics by idiots in lab coats with GUIs".
I expect the culture to be following along with me, if at a long long distance. I think in 2022 "the informed elite" are roughly absorbing the obvious issues with AI, Psychology, etc. and that tradition of statisitcal BS which was clear at least a decade ago.
My "pro-IQ" scepticism was initially agaist people worried about racism, etc. It's very lullying for a sceptic to be confronted by people who dont care about reality that muhc. But this is a trap. IQ is actually BS, and its a shame that the people often most against it are also often very poorly informed.
To address IQ research head-on is a little like like addressing old-testament theological scholarship. The level of understanding of both the field, statistics and the philosophy of science required is very high.
It's a field which florishes, as often with theology, from extremely complex arguments with absolutely insane hidden premesies.
https://en.m.wikipedia.org/w/index.php?title=Intelligence_qu...
There is a "pseudoscientific tradition" of statistics wherein one assumes systems are linear, one uses only linear models, and wherein one reports only associations on non-experimental conditions. This tradition was invented by psychologists, for psychology, and it's had horrific ethical consequences. I'd go so far to say that any system or results obtained by this method should now, a priori, be regarded as pseudoscience and should require a massive cross-disciplinary consensus to be regarded otherwise.
People who criticize IQ scores as a proxy for intelligence should publish their own IQ score, no? If the author did so, I missed it, which is certainly possible. The not-intentionally-snarky point is that perhaps the author is not "smart enough" to assess this reliably.
IQ scores are utterly *worthless* as a measure of intelligence because they only measure one's ability to solve a small set of carefully-chosen puzzles under carefully-controlled conditions. They cannot possibly measure anything realistic about intelligence.
Since you want people who criticise IQ scores to publish theirs, I'll tell you mine. Measured at various points across my education and out of five or six scores, they range from 72 to 160. They can't all possibly be right.
Human intellect has the capacity for creativity, empathy, wisdom, self awareness, diplomacy, etc., and those aren't as easily measured on an IQ quiz.
Just my 2c as someone who's not professionally trained to apply IQ tests but have studied them as a side interest, and scored highly on them (probably like many others here). It's just one biased facet of our minds, and we shouldn't give it undue weight or rule people out because their minds work differently.
From the article:
>I did try an IQ test in the past a few times. [..] I will not disclose the results, because I do not want this text to be skewed by a number I am criticizing. If the number was low, skeptics would say I am not smart enough to appreciate the value of IQ or that my claim is just sour grapes.
If IQ actually is reasonably important in some capacity, then it indicates that humans are not fungible and almost everybody would be forced to rewrite all of reality in their heads.
> IQ is popularly perceived as determined by the genes and unchangeable
> IQ tests have a negative effect on test takers
> IQ is often used in "scientific" racism to promote pseudoscientific social theories
1. They say IQ doesn't measure intelligence because according to them "Intelligence is the ability to solve problems". Well IQ at least claims to do that with its with its fluid intelligence component. Maybe creativity is independent of IQ, and also corresponds to better problem solving skills, but that doesn't change the fact IQ is still a predictor of problem solving skill.
2. If you can prove IQ can be boosted, that would be revolutionary. There's the n-back that improves IQ according to some studies, but that's highly controversial. But in any case, I don't see how not being able to change it makes it a bad measure. Me not being able to make myself taller doesn't make height a bad measure.
3. People not handling well knowing their intelligence doesn't have anything to do whatsoever with IQ being a good or bad measure of intelligence.
4. Again, people using metrics to be racist is an argument against measuring intelligence, not against IQ itself. You could make the same argument about literally any other system measuring intelligence.
A point I find interesting is "school improves IQ". What if the reason we can't seemingly improve people's IQ was because the school system already brought them to their full IQ potential? Maybe when we're trying to increase people intelligence, because we're testing on people already at school, it's like conducting studies on how to get muscular on people who are already maxed out gym bros. It's gonna be very hard to get them more muscular/intelligent if they're already reached max allowed by their body through intense daily workout/schooling.
The hell kinda schools do they have around you? I promise you no school I ever attended even tried to do that. Anyone who had potential had to skirt around the school system.
I don't think the school system is necessarily efficient at it, but kids do spend an insane amount of time in school. I imagine people would reach their athletic upper bound too if they spent 8 hours a day in the gym from the age of 5 to the age of 18.
Yeah, the real backlash against IQ is against the implied policy that people think ought to be implemented if IQ is real and determined genetically, as they believe that intelligence is related to moral worth, which raises the question of "what do we do with the worse people if we have a tool that can determine who is a better or worse human."
It's a complex mixup of Equity vs Equality.
All sorts of things goes into a measure.
Central processing speed, memory, memory access speed, cache size, software optimization and cache build up.
I think some people really WANT IQ to be only about central processing speed, but since that's not measurable we're trying to tease out what it is from the end performance metric. But really everything is important to intelligence, and something like software optimization and cache build up is really just education.
IQ measures things that are useful to other people when thinking about a person's as-is work-value, but is fundamentally flawed and useless when applied to a person's potential value or moral value.
But since the former is true it will continue to be a thing that people care about.
Your hypothesis as to what properties minds have, what those properties give rise to, and how IQ tests count as measures *has no scientific backing*. This description of IQ tests is, imv, an extremely unlikely account.
If you want to show IQ scores are measures of these properties, please provide (1) a theory of the mind in which these properties appear as experimental variables; (2) experiments in which these variables are controllable; and (3) evidence that IQ correlates with the expected controlled outcomes.
Note, the above has never been done. Making the assertion that IQ measures intelligence circular. It is simply, asserted.
It is the opposite, IQ correlates with basically every measure and score people have came up with that they think could represent intelligence in some way. This concept is called G.
Not believing in IQ tests (or really in any particular one thing) doesn't make someone intelligent.
I've come to agree with this assessment.
Calvin: If you could have anything in the world right now. What would it be?
Hobbes: Hmmm.
Calvin: Anything at all! Whatever you want!
Hobbes: A sandwich.
Calvin: A sandwich?!? What kind of stupid wish is that?! Talk about a failure of imagination! I'd ask for a trillion billion dollars, my own space shuttle, and a private continent!
Hobbes: I got my wish.IQ tests have their place, but they are grossly and frequently used inappropriately, imho.
Note that actual psychology researchers like Howard Gardner and Robert Sternberg (and many others) have written quite a bit about various aspects of intelligence that the author alludes to. IQ is just one fairly narrow component of intelligence, and I don’t think that this point is debated much by modern researchers.
Let’s take the author’s arguments point by point:
> Problems with IQ
Summary: People use IQ scores in ways that are not appropriate for the measure. Boohoo.
This isn’t a problem with the test, it’s a problem with how the results are used.
> Intelligence cannot be separated from knowledge
IQ test creators acknowledge this, and they try to abstract out knowledge as much as possible. It’s not perfect, but it’s as close as they can get (especially for examinees in “the West”).
The author goes on to complain at length about “the injurious environment we push our kids into”, but this seems to be a different issue.
> IQ is no measure of intelligence
The author starts with “Intelligence is the ability to solve problems”. I would say that it’s the estimate of the capacity to solve problems. As the author mentions, other things come into play with regards to solving real-world problems. IQ tests don’t measure this and don’t claim to.
Interestingly, the author says “IQ tests might be as well replaced with tests of how fast people can count nickels”, and I think that this may not be as bad of a proxy for IQ as he thinks. The problem with nickels specifically is that some people may have done it before and/or may have been taught shortcuts. But if you give a test in which people have to count a variety of things of various shapes, sizes, values, etc. while on a timer, I’m guessing that the results would correlate fairly strongly with IQ.
> IQ is unreliable
Ah, totally agree, but I don’t think that it invalidates the test. This is a theme that I will come back to.
Any single IQ score should be seen as the minimum IQ for that person. There is a theoretical maximum (at least imho) for each examinee that is not always reached. There are so many things that cause people to score below this theoretical maximum, but very few things other than incredibly lucky guessing that will cause them to score above their theoretical maximum (and even this is eliminated when you factor in the confidence interval of a given result).
> Intelligence can be boosted
The theoretical maximum cannot be boosted. All one can do is improve the chances that they score as close to their theoretical maximum as possible (e.g., by getting good sleep, eating well, etc.).
> The psychological harm of IQ
Yes, if people misuse the score in ways that it was not intended (and many do), then it can cause psychological harm.
Any tool can be used in harmful ways, so I don’t think that this is news, and it certainly doesn’t invalidate the test.
> IQ myth
This section is just a weird diatribe of the author. He basically says that IQ doesn’t measure what he wants it to measure. As mentioned in my intro, psych researchers have explored the topic of intelligence writ large quite deeply.
> Child's IQ
Yeah. “Mental age” is an approximation. It’s not a bad one, but it is also misused. If people would just use it as a “directionally correct” measure, then it becomes more functional.
> Schools improve IQ
Same as above. School helps people score closer to their theoretical maximum. It doesn’t improve their theoretical maximum.
> IQ bloopers
Ugh. The author talks about IQ tests that have probably not been validated (at least not recently), and issues with not measuring the theoretical maximum.
Every standardized test used widely enough over time will have these “bloopers”. Every. Single. One.
> New forms of intelligence
Cool story. Arguably not intelligence. Definitely not the part of intelligence that IQ tests claim to measure.
> White supremacy
Very true.
Fwiw, it’s not just white people who think they are superior in terms of intelligence.
I have lived in some countries in which people have told me to my face about how much dumber on average white people are than [their race], using some sort of test score (typically a crude IQ proxy) as evidence. Of course, I’m the exception as a white person according to them, so I get a pass.
This is just another case of people misusing test scores. It’s rampant. It doesn’t make the test scores any less valid for their intended use.
CONCLUSION
This article just seems like a pseudo-scientific rant against IQ tests.
IQ tests aren’t perfect. They have a lot less utility than people think they do. That said, they do have some utility.
IQ tests are pretty good at discovering people at the extremes. One could argue that the folks at the extremes might warrant some different treatment.
IQ tests also do a decent job at identifying large gaps between individuals. This has less proven utility, but there is an idea that people with significantly different IQs perceive the world in different ways. This is a bit of a rabbit hole, but it’s an interesting line of inquiry.
tl;dr - Your IQ score is not you. Don’t worry about your score if it’s not at the extremes or if it’s not vastly different than your partner’s (it’s probably not, and if it is, you probably already expect that this is the case).
I’m glad that my reply was well-received.
Just from spending 20 seconds on his Wiki page.
He is clearly not the original inventor.
IQ is about 4% of what makes up your personality. So whatever your intelligence is within a standard deviation, other factors are much more important.
TL:DR; it's better to be a track star with a big d*k than a fat genius in a basement apartment.
It’s a hotly debated measure of intellect.
It being better to be a big dicked track star vs a fat genius has nothing to do with IQ as a valid measure of intelligence, or even personality as you were talking about. And is frankly really weird you brought that up.
TL:DR; monkey bests fish in tree climbing contest, credits higher IQ.
> it's better to be a track star with a big d*k than a fat genius in a basement apartment.
Why is that? How easy is it to become a well-endowed track star? How easy is to become a genius? How many geniuses are fat? How many live in a basement? This is just distracting from any point related to intelligence and IQ.
Apparently he has a Ph.D. in Counselor Education and Supervision, who primarily spends his time psychoanalyzing pop culture and celebrities on Youtube. I guess in the year 2022 that makes him a qualified professional worth listening to and discussing on this forum regarding his thoughts on IQ?