Feynman seemed to take glee in sharing his "low" IQ of only a 125. He was one of the smartest men on the planet in the last 100 years. Surely, we should defer to his opinion that the concept IQ is sort of a joke. ;)
I always smile at people putting that much weight into IQ numbers..
IQ is not intended to represent 'all that intelligence is'. It is simply one way to measure a cross section of particular capacities that are valuable in a modern society and are easy to test.
E.g It doesn't attempt to measure the ability to track an animal, or remember the face of a person you met once last year.
This test is of interest because it is strongly predictive of other things we do care about, like educational attainment, criminal behavior, economic outcomes, etc.
As far as I've read about it (not an expert; I've just seen this debate more than once before and tried to educate myself on the basics), that's because every time someone proposes an actual objective test for one of the "other intelligences", it turns out to either be primarily dependent on training (i.e. measuring conditioned behaviors) or to correlate so strongly with IQ that the test is effectively an IQ test.
Equally flawed, taking one outlier as evidence of something inherent.
The last time I had this argument in a big way with people telling me I had no clue what I was talking about, some forum I belonged to was having a heated argument because they had the habit of posting ridiculous tests, like "What kind of dog are you?", and someone had posted an online IQ test and things were getting pretty ugly (in part because a lot of people felt it was scoring them too low).
I stated up front that it likely only went up to 140, then took the test to shut people up and scored a 140. I had training in test taking in my high school gifted classes, so I tend to do well on such tests.
But, whatever.
Plus, this study doesn't care if IQ tests are a good way to rate intelligence, people who score well on them also are more likely to suffer disorders.
If we are going to discuss some correlation between supposed high IQ and various disorders, is it meaningless to talk about how one ascertains IQ and what on earth that even means?
I think it is not terribly different from asking "What color is your aura?" but I am quite sure that if I went around asking that on HN (or, you know, posting articles about auras, chakras, etc), it would be flagged to death. Yet, IQ is equally invisible. You cannot look at someone and say "brown hair, 6 feet tall, IQ of 145." That just isn't a thing.
The correlation is between people who do well on the tests and disorders. It could be that more preparation for the tests, the SAT classes for example, leads to both of these results.
Auras have no repeatable way of being measured. IQ tests show something, even if labeling that intelligence is wrong.
There are people who claim to see auras. How is this different from people trained in administering IQ tests? The average person is just taking the word of some test taker that it means thus and such.
At some point, you simply have to trust the person telling you something. Even when people use machines -- and IQ tests are not measurable by machines like MRIs -- the person has to read the output. How do we know they weren't high on LSD that day?
I mean, this is veering into territory of "Zhuang Zhou dreamed he was a butterfly," * but I am being sincere here. I intend to step away from this particular thread because I don't think people think I am posting in good faith, or being sincere. So there is probably no good that can come from trying to further engage in discussion of a matter that I have thought a lot about and think is a serious issue.
Repeatability, you can do the test multiple times and expect similar results. Not identical, and outliers exist, but similar.
I am not a fan of current IQ tests as a measure of intelligence. I am not trying to change your mind, nor do I doubt you believe these things. This article is careful to distinguish intelligence from IQ test results, and the link is interesting without trusting IQ tests.
What bothered me the most was using the Feynman example without anything scientific in your argument.
Stephen Hsu argued there that Feynmans school iq test must have had a low ceiling. From what we know about the consistently high IQ scores of other incredibly bright theoretical scientists that sounds likely.
If the test administered to Feynman only went to 130, as some do, then a test score anywhere close to that, such as 125, can be assumed to have "ceilinged the test." It can be viewed as a floor for the person's IQ, a minimum score suggesting it is at least that high.
Thus, Stephen Hsu's assertion that it must have just been a bad test with a low ceiling suggests to me that Mr. Hsu knows a great deal less about IQ tests than I know, so I see no reason to defer to his opinion.
Yet, also, IQ tests are strongly influenced by cultural knowledge, good scores in kids strongly correlate to kids coming from upper middle class families with well educated parents, etc. So I think there is quite a lot of room for saying that if you take any child and feed them excellently well, dote on them, invest a lot of parental and societal time, effort and other resources in them, they will tend to perform better than average, both on IQ tests and in life generally.
I am not just screwing with people here. I have studied the space and I think there are serious problems with talking about IQ like it is concrete, unchangeable, primarily inherited with minimal influence from environmental inputs, etc.
Thus, Stephen Hsu's assertion that it must have just been a bad test with a low ceiling suggests to me that Mr. Hsu knows a great deal less about IQ tests than I know, so I see no reason to defer to his opinion.
Maybe there is some nuance I'm missing, but those two paragraphs seem to be arguing the same point - that a 125 score was probably not accurate, but hitting the ceiling of the test? I.e. you and Hsu agree. Here is Hsu's full quotation (never said "bad").
To understand this score we have to remember that typical IQ tests (e.g., administered to public school children) tend to have low ceilings. They are not of the kind that Roe used in her study. One can imagine that the ceiling on Feynman's exam was roughly 135 (say, 99th percentile). If Feynman received the highest score on the mathematical portion, and a modest score of 115 on the verbal, we can easily understand the resulting average of 124.
I elaborated some on my mental model in this paragraph:
Yet, also, IQ tests are strongly influenced by cultural knowledge, good scores in kids strongly correlate to kids coming from upper middle class families with well educated parents, etc. So I think there is quite a lot of room for saying that if you take any child and feed them excellently well, dote on them, invest a lot of parental and societal time, effort and other resources in them, they will tend to perform better than average, both on IQ tests and in life generally.
Part of what I am arguing is that we make up a measurement that validates that our "best" people -- like well educated, well heeled white males -- are best because they have this innate quality called high IQ and then we channel the "smart kids" from "good homes" into classes that cause them to test well.
The mental model that some kids are inherently more deserving of investment at school because they are smarter and do better than other kids winds up being self-fulfilling prophesy. Sort of like redlining of Black neighborhoods was self-fulfilling prophesy, causing Blacks to remain poor and their neighborhoods to remain not worth investing in by banks.
So, hopefully that clarifies it for you because I already said I am stepping away from this discussion.
Adieu.
And I imagine I will regret saying that, but I can't see letting that stand and saying nothing.
Hsu said there was a ceiling on the test, you said he was wrong, there was a ceiling on the test, and IQ is flawed.
The only trouble with your theory is that IQ actually does a very good job of predicting how well members of various minorities will do in life. Half a century you could argue the cultural bias hypothesis but it no longer flies. The tests have been updated and taking "culturally biased" questions out actually made minority scores go down (possibly due to lower G loading on those questions).
> The mental model that some kids are inherently more deserving of investment at school because they are smarter
Of course the test would be "what is the marginal benefit of education to this person" a very different question from what is their IQ level.
You just seem to be making stuff up to fit your ideological preconceptions.
To add to all the other criticisms of IQ score, it’s defined to be mean-100-σ-15* — so nobody has an IQ of over 197, and only newborn anenephalactics ever score as low az IQ 3… and perhaps not even then if they score the same as the comatose.
* or σ-16, which changes the subsequent statements a little.
https://en.m.wikipedia.org/wiki/Fat-tailed_distribution
I regret I can’t find anything on Bing but the below but Google’s blocked where I am.
https://blogs.sas.com/content/iml/2014/10/13/fat-tailed-and-...
There are some other cases where it is generally thought that the distribution is normal, but it is not (in real samples). IQ has too many people in both far ends and has a different shaped tail on the right and left (the maximum is 200 or so, but the minimum is about 40). Weight of adults also has too many in the right tail and is not fully symmetric, but its right tail is also long (unlike that for IQ).
> When current IQ tests were developed, the median raw score of the norming sample is defined as IQ 100 and scores each standard deviation (SD) up or down are defined as 15 IQ points greater or less,
IQ itself — not intelligence or even score on an IQ test — is by definition not fat tailed. If a particular IQ test doesn’t produce a normal distribution, that test is wrong.
At least, wrong by the current standard definition of IQ; there have been others, and the very fact that it is flawed means there probably will be a new one in future.
Raven Progressive Matrices, I believe the most used IQ test today is not influenced by cultural knowledge. IQ has an amazingly high correlation with academic success and work success, see https://www.youtube.com/watch?v=jSo5v5t4OQM it is worthwhile
Wikipedia says that people with autism and asperger's outperform other people on Raven's tests. Most people view ASD individuals as seriously cognitively impaired and they are often seen as low IQ, so I have trouble with your assertion. It suggests that people the world sees as highly impaired whose lives typically are harder than average are smarter than the rest of us. This completely flies in the face of every definition I have ever seen for IQ as a general predictor of competence.
Wikipedia also says the answers for Raven's matrices have been leaked, making them useless as an assessment tool.
Popular culture stereotypes autists as being highly intelligent.
The very existence of autism calls into question our historical conceptualization of intelligence and IQ testing.
I have two ASD sons. They are very bright. They also have some serious personal struggles.
So, which is it? Does IQ measure general competence or does it measure something else? Because this latest twist in the conversation seems just ridiculous.
>Does IQ measure general competence or does it measure something else?
It's an Intelligence Quotient. It measures intelligence. No one said anything about "general competence", whatever that even means. I imagine intelligence is an important part of "general competence" though.
You are contradicting yourself.
We cannot measure intelligence, since we cant even define it properly (in a way that we attach a measurement unit ot it).
Why do you assume that's cause and not effect?
I'm not sure I'd say flawed so much as I'd say myopic. Which is absolutely still a flaw, but a very specific one that allows you to contextualize the results of an IQ test rather than discard them as worthless. IQ measures a subset of human intelligence pretty effectively, but misses the bigger picture. There's large parts of human intelligence that can't be measured using a test like that. A test can only provide stimuli for the mind and measure its reaction over the short period of time while taking the test. Among other things, it can't measure long-term recall or the ability to rehearse/learn information and skills over time, both qualities people would naturally associate with intelligence.
It's our need to have a single spectrum from mental retardation to genius-level that's led us to try to create a test which assigns a number to intelligence. But intelligence is multidimensional and an IQ test just fills in some of those dimensions.
Then for baseball players, everyhing changes.
Intelligence or smart are low-information, high-emotional-signalling terms. It means more about social status than anything else. Everyone has a mix of abilities. Some of the best engineers are terrible writers. Some are incompetent fathers (Steve Jobs). Some people are great at consuming and regurgitating a lot of info, but bad at original/critical thought. Some who would thrive in one life circumstance may struggle in another.
Academically, IQ is supposed to correlate with g, general intelligence. In other words，we're just creatures making up nonsense terms to, at face value, signal about ability, but it's often more of a social construct.
The vocabulary battery would seem to measure long term memorization ability. It actually correlates with IQ better than any other test. Also SAT tests are often used as a proxy for IQ because they correlate so well with it.
>It's our need to have a single spectrum from mental retardation to genius-level that's led us to try to create a test which assigns a number to intelligence. But intelligence is multidimensional and an IQ test just fills in some of those dimensions.
No, it's not. Please learn what you are talking about. It was once believed that intelligence was multidimensional. But then scientists realized all those different dimensions correlated amazingly well. E.g. people with better memory were much more likely to be better at puzzle solving. And a single factor, g, could explain most of the variance. IQ tests try to measure g, which is very predictive of general cognitive ability.
That understates it. The SAT is an IQ test, strenuously as the College Board May deny it. The test was originally the scholastic aptitude test and results from it correlate with other IQ tests just as well as results from IQ tests correlate with each other.
If anybody reading wants to know more about the broad field of which IQ is a small part the Wikipedia article on psychometrics is quite good.
IQ definitely matters. The average IQ of physics majors is like 2 standard deviations above average for instance. Suggesting some fields are very selective of IQ and low IQ people are much less likely to be able to do them.
1. There will be a number of comments saying IQ is outright meaningless.
2. There will be a comment saying Feynman only had an IQ of whatever it was.
Never fails.
Anyway, as I've said here before, Feynman's books are primarily collections of stories. What's presented in them was selected to maximize the impact and entertainment value of the story. And Feynman was very, very good at storytelling.
I am very leery of the power of IQ tests to justify entrenching The Haves and Have Nots. Our mental models have scary amounts of power to shape outcomes and this mental model seems to justify classism as a kind of manifest destiny.
One of the videos linked here in reply to me has some smart authority figure type asserting that university students all have IQs above average. That seems pretty well tautological to me. One would hope that those getting more schooling would do better on tests. But that fails to genuinely prove that what these tests measure is some fundamental personal trait unrelated to the fabric of your life experience.
I was homeless for nearly six years. A lot of people wanted to see me as stupid and incompetent during that time. Those classist perceptions helped deny me opportunities which helped keep me poor which helped deny me opportunities. I can never seem to get that across to well off people that their tautologies simply entrench class based outcomes and really say nothing about personal competence.
When we design a world that closes doors to people arbitrarily, we design a world that actively wastes human potential. This makes all of us the poorer.
Look, no one is saying your individual IQ test is the only thing that matters. There is random error on tests, and it's not like intelligence is the only thing that matters in a person. But if you go around denying that it exists at all, you are just spreading pseudoscience.
I am trying to do some things to bring down those barriers. I was able to start making some earned income online while homeless. My lack of nice clothes or a recent shower did not matter. I am trying to foster that model as a thing that can help protect against sliding into homelessness, help you scrape by while on the street and be something that can help build a future that won't be taken away because you got better. Plus, it is not a homeless service, is open to anyone interested.
http://independentdigitalworkers.blogspot.com/
I was one of the top students of my graduating high school class. I got a lot of brainwashing that I was destined for greatness etc. It didn't go down that way. So I think these mental models are mostly propaganda and we need to guard against that.
With regard to your last statement, here's a (very) vaguely related post from Penelope Trunk who also writes about aspberger's and life. http://blog.penelopetrunk.com/2011/12/01/living-up-to-your-p...
Saying that there's lots of great chess players with low IQ doesn't mean the two things are not correlated
For some actual numbers, see https://www.researchgate.net/publication/307874653_The_relat...
This means that, given high IQ, it's not necessary to have good chess skills to be in the elite group, whereas given low IQ, it is necessary to have good chess skills to be in the elite group. This will establish a negative correlation between "chess skills" and IQ within the elite group even though they are positively correlated overall.
Similarly, SAT math and verbal scores tend to be negatively correlated within the student population of any particular university, because universities will reject applicants whose total scores are low, despite the fact that the math and verbal scores are very strongly positively correlated in the larger population.
> Restriction of range, such as floor and ceiling effects or selection effects, reduce the power of the experiment, and increase the chance of a type II error.[5] This is because correlations are attenuated (weakened) by reduced variability
If you want to try this for yourself, try generating some fake SAT scores: (a,b) pairs where a and b are normally distributed with mean 500 and sd 100, and where a and b correlate at about 0.8.
Then, from your unbiased data set, select all the pairs for which 1600 >= a + b >= 1300, and check the correlation of a with b in the new, smaller data set.
Just examining this as a thought experiment, you can see that someone with an 800 math score might have a verbal score anywhere from 500 to 800. If we lower the math score, our expectation for the verbal score rises -- someone with a 600 math score will have a verbal score between 700 and 800. That is the basis of the negative correlation in the subgroup, but it's purely an artifact of the way we chose to label data points as being in or out of the subgroup.