/sci/ - Science & Math

lol the cope

>>10571754
>be ultra rich banker hedge fund guy
>born to literal ottoman aristocracy and Lebanese political royalty, with every other ancestors being a governor or judge or something going back centuries.
>taught greek, latin, french, english, as well as his native arabic from a young age
He is the definition of ultra privileged.

>>10571767
Apart from how privileged he is, he is a really mediocre man. If any of us were born into his shoes we could probably have achieved as much as him.

Taleb is right. If you've ever actually read about what an IQ test actually is, this is pretty much self evident. I was really surprised that this became so controversial.

Lets look at the standard IQ test, Wechsler Adult Intelligence Scale (WAIS).

First of all, we start with the basic foundation of the WAIS test, Verbal Comprehension, Perceptual Reasoning, Working Memory and Processing Speed. These are combined to produce a single "IQ Score". So the 4 categories are completely unfounded from any scientific basis. Quite literally these 4 things were chosen because the psychologists who came up with these tests thought they "made sense". We are already off the a bad start. We have an abstract quantity called an "IQ Score", and the factors that make up this abstract quantity are basically just appeals to common sense.

So then we have the same problem repeated, because we need to take measurements for each of the subcategories. So we need to make tests to measure each subcategory. Again we just have unfounded appeals to common sense. Lets look at a test for "Perceptual Reasoning".

>Matrix Reasoning - View an array of pictures with one missing square, and select the picture that fits the array from five options.

So this random test someone came up with is supposed to measure "Perceptual Reasoning". Whatever that's supposed to mean. This stuff is literally the same tier as Myers–Briggs personality types. Just random shit some psychologist came up with the fit an unfounded model they came up with.

>>10571823
Except intelligence tests have progressed in the decades since WAIS was formulated, and now operate in a testing environment completely independent of cultural setting.
When most people say "IQ" they mean "generalized G factor" from modern intelligence tests.
Whatever the G factor is, it correlates with relative success in life better than anything else, including wealth and upbringing (with the exception of catastrophic upbringing)

To say that IQ doesn't exist because we now have better methods is just dishonest.

>>10571823
So lets ignore all of that for now, and give the benefit of the doubt that we can actually calculate the abstract notion of "IQ". To this point "IQ" is a purely abstract number.

Now we need to do correlations into real world data to figure out if it actually means anything.

So for nearly the past century, Psychology grad students have gone out and collected any dataset they can get their hands on and did linear regressions against IQ. Things like Educational achievement, Income levels etc.

Off the bat we have an obvious issue here. Pretty much ever IQ correlation study ever uses linear models, with the implicit assumption that IQ has a linear relationship to some outcome. Unfounded. We are completely blind to any non-linear affects.

Now you factor in that the people doing these correlation studies can just p hack until they get something worthy of a publishable paper, and we then realize that we have an entire body of literature that is pseudoscience non-sense.

Ok so maybe lets give benefit of the doubt again, and assume the studies were done properly. We now have to consider that these result only hold for populations of subjects, and we can't really use them to make predictions about individuals. Why? Because IQ score itself is an extremely noisey estimate, sample size of 1. Really we should have collected the results of multiple IQ tests for each individual and used them to compute an IQ estimate. But there is a huge problem here. IQ tests, and all of the scoring and normalization used have an implicit assumption in them. That assumption is that this is the first time you have ever taken an IQ test. So we can't administer multiple tests even if we wanted to, because every additional IQ test is invalid according to the assumptions we made at the very start.

Theres plenty of more bullshit I can point out, like the inherent cultural biases in IQ tests, or the fact that IQ scores are age normalized but I think I got my point across.

>>10571842
>To say that IQ doesn't exist because we now have better methods is just dishonest.

No one is saying "IQ doesn't exist". Neither I or Taleb. The point is IQ tests measure IQ and IQ only, which is an abstract quantity developed by wonky psychology models from the 1930s.

Strictly speaking we cannot say that IQ == Intelligence. This is basically the bulk of Talebs point if you actually cared to read it. Basically once you pass 100 IQ, all of the correlations against IQ start to rapidly fall apart and become statistical noise.

I will grant, and Taleb makes this point as well, that IQ actually does seem to be a good predict of stupidity. IQ < 100 actually does hold up statistically pretty well.

>>10571846

Also as far as success goes, the role of personality should be taken into account. If you have someone of high intelligence (IQ) but they are low in conscientiousness odds are they will not be as successful as someone of average/slightly above average intelligence and highly conscientious and creative. I highly doubt everyone in MENSA has patents and millions in the bank.

>another IQ thread

>>10571823
>working memory
>unbased
Nigger i did not write a fucking thesis on working memoey for nothing. If anything it tells us of the limitations of the processing retaining of information on people.
It is also heavilly affected by peoples habits, interests and proffessions.

>>10571877
>If anything it tells us of the limitations of the processing retaining of information on people.

Sure, it tells you that...

now make a well founded argument about how it is a factor of IQ... hard mode: you can't use the word correlation

>>10571877

>He did his thesis on working memory

Pic related it's you

>>10571854
IQ or G factor are statistical measures that correlate with a person's proclivity for success in sophisticated environments and complicated mental tasks. Just because it is not a perfect indicator, that does not mean "it predicts only itself." that is purely a lie.
IQ and G factor analysis are the second most predictive traits discovered by scientists in the psychological and sociological domains. (the first being the correlation between income inequality and violence in a small geographic area).

If IQ and G factor analysis are not predictive enough to be scientific for you, then not a single other claim made psychologists, sociologists and psychiatrists is. Including claims made about the chemical nature of the brain.

>>10571860
IQ correlates with success far more than does personality

>>10571965
>If IQ and G factor analysis are not predictive enough to be scientific for you, then not a single other claim made psychologists, sociologists and psychiatrists is
true. real scientists have higher standards, you know like controlling systematic uncertainties and biases. same reason why economists can't predict shit worth a damn

>>10571967

But you cannot discount the role of personality.

>>10571965
>If IQ and G factor analysis are not predictive enough to be scientific for you, then not a single other claim made psychologists, sociologists and psychiatrists is.

That kind if of summarizes it nice. These disciplines at psuedo science and most of what they produce is garbage.

>Including claims made about the chemical nature of the brain.
What? This is just a weird nonsequitar

The only meaningful patterns are found in mathematics.

>>10571976
Literally no one says that personality doesn't have some effect.

>>10571967
No it doesn't.
It correlates with "lack of success"

The correlations hold for IQ < 100. Once you go above 100 it dissolves into complete noise.

And this is exactly what we see from correlation studies. They usually report about a 0.5 correlation. The bottom half for IQ < 100 fits almost perfectly but the top half for IQ > 100 is pure noise. Intuitively you'd expect this to give you a correlation of 0.5 and that's exactly what the empirical data shows.

>>10571767
Yet, he became a statistician instead of a mathematician.
I have a pretty good idea where his aversion to IQ tests come from.

>>10571754
/pol/ btfo!!! This is irrefutable proof that niggers are just as smart (if not smarter!) than white people.

>>10571754
>Is he right?
No
>>10571823
>First of all, we start with the basic foundation of the WAIS test, Verbal Comprehension, Perceptual Reasoning, Working Memory and Processing Speed. These are combined to produce a single "IQ Score". So the 4 categories are completely unfounded from any scientific basis. Quite literally these 4 things were chosen because the psychologists who came up with these tests thought they "made sense". We are already off the a bad start. We have an abstract quantity called an "IQ Score", and the factors that make up this abstract quantity are basically just appeals to common sense.
>
So this random test someone came up with is supposed to measure "Perceptual Reasoning". Whatever that's supposed to mean. This stuff is literally the same tier as Myers–Briggs personality types. Just random shit some psychologist came up with the fit an unfounded model they came up with.
Wrong
Pic related

How can an "arbitrary" test of intelligence be the greatest predictor your brain's ability to recover from a injury? More so than education btw.

>>10571823
>First of all, we start with the basic foundation of the WAIS test, Verbal Comprehension, Perceptual Reasoning, Working Memory and Processing Speed.

What is all this stuff? Every IQ test I've tried is just nothing but abstract pattern recognition, i.e. a bunch of symbols arranged in a certain way and you have to determine the next in the sequence.

>>10572978
Just because IQ is not a good metric for intelligence doesn't mean there is no difference in average intelligence across races.

>>10573007
That's what a real IQ test looks like. You probably only took online tests.

>>10573009
where i can do the real one?

>>10572978
Plus lets not forget the black lady mathematician who was responsible for putting a man on the moon. Katherine Johnson.

>>10573008
Yet that is inevitably the conclusion that is drawn by our intellectual class. The null hypothesis is that race, and racial differences by extension, is a social construct.

>>10573013
I did mine with a neuropsychologist.

>>10573018
wow, someone is super dedicated to this issue

>>10573021
I had to do mine as part of my psychiatric treatment.

>>10573013
uni/psychiatrist or intelligence study

>>10571754
I love how he could not even be assed to write a proper text and just copypasted his random twitter posts.

>>10571860
Creativity aka openness correlates more with IQ than other personality traits.

>>10571754
It primarily measures the sensitivity to patterns. Most of the time high IQ people believe in more complex nonsense than low IQ people, but are not in any way more rational.

>>10571967
taleb literally shows that it only does for the lower quadrant of iq. otherwise, it is useless

>>10573082
you can't say you are measuring openness or creativity with iq from this however. intelligence is extremely complicated. for instance, there is a negative correlation between creativity and working memory

>>10573086
i've heard this claim that iq tests measures pattern recognition but it seems like common sense that is vaguely grounded in reality. for one, if we are talking about crystallized intelligence (the accumulation of knowledge over time as measured by vocabulary and other culturally biased measures), no it does not. however if we are talking about fluid intelligence (the ability to solve novel problems), then there is some merit to the intuition.

even then, i'd argue that it more so measures compositionality (breaking problems into smaller pieces). actually i recall reading an article analyzing this in better detail.

https://www.pnas.org/content/early/2017/04/25/1621147114

and outside of that, working memory (which strongly correlates with fluid intelligence as well).

>>10572978
studies in racial differences of intelligence abuse statistics even more than does iq

http://racialreality.blogspot.com/2011/08/devastating-criticism-of-richard-lynn.html

>We found that Lynn and Meisenberg's assessment of the samples' representativeness is not associated with any of the objective sampling characteristics, but rather with the average IQ in the sample. This suggests that Lynn and Meisenberg excluded samples of Africans who average IQs above 75 because they deemed these samples unrepresentative on the basis of the samples' relatively high IQs. We conclude that Lynn and Meisenberg's unsystematic methods are questionable and their results untrustworthy.

>>10573016
race being a social construct is not a null hypothesis fucking retard

>>10572979
this isn't an argument. iq predicts how iq measures are effected after brain injury. ok?

>>10572979
Wow bro look at their "data"...

weird that the controls had an even stronger correlation.... hmmm....

Also decile regression is a fucking shoddy analysis. just makes your analysis more lossy. By definition IQ should be normally distributed, so there is literally no reason to use quantiles here.

>>10572958
Having experiencing Levantine mental hijinks, I can tell you that he probably didn’t get into the math program and only the stat program so he invented this whole justification for why stars is superior anyway. If he had gotten into math, he’d talk about the importance essential truths timeless and shit.

>>10572979
>Any Taleb thread
> Endless ad hominems
The onions levels are out of control

Get your press up to BW for five reps.

>>10573511
I could do it if I went to an actual gym.

>>10573457
>weird that the controls had an even stronger correlation.... hmmm....
By stronger you mean a difference of 0.03
Let me break down data first:
Control had a correlation of 0.845/1.0 of their IQ scores and their AFQT scores.
Head injured had a correlation had a correlation of 0.816/1.0 of their IQ scores and their AFQT scores
Your argument:
H-HEYY!!! The correlation was stronger by 0.03 in the control!!!
Is there a point you're trying to make?

>Also decile regression is a fucking shoddy analysis. just makes your analysis more lossy.
>I don't like deciles.
Solid argument.
> By definition IQ should be normally distributed, so there is literally no reason to use quantiles here.
Across the population not within a specific field or profession. Please tell me you are not really this stupid.

>>10573545
Your argument:
>H-HEYY!!! The correlation was stronger by 0.03 in the control!!!

(Fixed)

>>10573466
Where is the ad hominem?

>>10573142
>this isn't an argument. iq predicts how iq measures are effected after brain injury. ok?
Do you have reading comprehension problems?
>Greatest predictor of cognitive decline is premorbid IQ.
Do you know what cognition means? It means the ability to acquire knowledge and comprehension.

>>10572192

Some retards on here do. Genetics is not everything.

>>10573142
>taleb literally shows that it only does for the lower quadrant of iq
imagine being this bad at math

can I study physics with an IQ of a mere 124?
However, I do have ADHD and I think my spatial reasoning score got dragged down significantly due to extremely poor working memory.

>>10573596
>can I study physics with an IQ of a mere 124?
Yes, no one is gonna stop you because of that.

>>10573596
Also people tend to use the most inefficient methods to study, like re-reading stuff.
If you can study more efficiently you might get better results than high IQ people.

>>10573626
High IQ people mostly study efficiently because they understand things quicker anyways.

Also re-reading is fine.

>>10573545
>AFQT
Are you retarded. He isn't talking about AFQT/IQ correlation (IQ tests correlate with each other wow). He is talking about his graph.

>>10573596
Of course. You have the same IQ as Feynman. As discussed in thread, IQ>100 is hard to interpret. Follow your gut instinct. Forget that test, dude. I probably have an IQ around yours, and I can do higher math just fine. I am extremely creative compared to my fasttttt peers in grad school. Cognitive complexity is sometimes way more interesting than speed. A fair portion of IQ is speed.

>>10571754
iq is about data retention but is lacking in some regards because its only approved data. iq does not actually dictate how well some one uses spatial reasoning or many other things

some iq test have embedded logic test components but not all. some parts of a logic test can be both worked out in your mind using spatial reasoning or you can memorize the answer. a bike tire has 34 spokes. how many spaces in between does the bike tire have ? the number of spaces = the number of spokes. but you could have also simplified to 3 spokes and seen there were 3 spaces and got the answer. and your SAT score can be used to determine a approximate value for your iq

>>10573779
Okay but how are you going to compete with someone faster AND more creative? Math research (especially higher levels) is filled with these people. Hard work just doesn’t cut it.

>>10573545
Can you read? You responded to nothing that I posted

>>10573790
Its a good point. I find I still usually have more interesting ideas than most. I am slow, but usually the result is very good. I rarely ever meet people more creative than me. Even in math departments.

t. sperg

>>10571754
>Is he right?

Yes. The different 10 points of IQ makes depends on where on the curve it is.

For instance a there is a substantial different in life outcomes between a person with an IQ of 80 and an IQ of 90. However there are negligible quantifiable differences between a person with an IQ of 120 and 130, despite 130 IQ being so much rarer.

In other words IQ is more of a test to detect mental deficiencies, rather than to detect genius. Genius is more likely the result of hard work and determination by people with "enough" IQ.

Take one look at the mediocrity of mensa or triple 9 members if you want to see this in action. High IQ, but relatively normal mostly underwhealming, dull people.

>>10573572
how are they measuring cognition in the study if not iq? they are correlating iq's relation to structural damages in the brain, no?

>>10573585
not an argument >>10572941

>>10573596
yes. 124 is more than enough if we are just going by what iq tells us

brain bulls today

brain bulls forever

>>10573790
>Okay but how are you going to compete with someone faster AND more creative?

You're one serious brainlet for someone posting in an IQ thread, ya know? I bet you're Indian/Asian.

A lot of you Take deniers seem to have major misunderstanding.

Taleb is not denying that "Intelligence" exists. The fact that some people are intelligent and others are not is blatantly obvious. This isn't some kind of egalitarian "we're" all equal argument.

The claim is specifically that IQ scores are not a good model to measure intelligence.

I'm proposing a new model for a binary intelligence metric. If you didn't understand this distinction, you are not intelligent.

>>10571754
>>Is he right?
I can't speak about his statistical rendition of IQ.
I only know that it was originally developed by a child psychologist as a something part of a toolkit to diagnose learning disabilities. Researchers in the US extrapolated the concept to include measures of superior cognitive abilities which went hand-in-hand with institutional racism. This sounds like a lot of SJW shit-talking points but there's some sort of truth to this.
https://en.wikipedia.org/wiki/Alfred_Binet
https://en.wikipedia.org/wiki/Lewis_Terman#Legacy

>>10573779
>Of course. You have the same IQ as Feynman.
>actually believing in the "124 IQ feynman"-meme
Retard.
>As discussed in thread, IQ>100 is hard to interpret.
This claim is evidence of you never having interacted with people from different strata of society.

>>10573947
>Researchers in the US extrapolated the concept to include measures of superior cognitive abilities which went hand-in-hand with institutional racism.
What point are you trying to make? The fact that all "races" don´t achieve the same median IQ-score does not mean that the test is racist.

Furthermore, the US military, intelligence community and academia has used IQ-tests (ASVAB/SAT/ACT) for over 70 years, and their data unequivocally proves that the most mentally demanding tasks (doctorates and graduate degrees in STEM; flag officer positions in the branches of the military; senior positions in the intelligence community) are occupied by persons who score exceptionally well on IQ-tests.

From the point of view of official statistics, the odds of acquiring a doctorate in mathematics are much larger for people with a 130+ IQ, than it is for people with an IQ of 120-125 - those few percentiles make a large difference. In addition, people with an IQ below 110-115 (i.e. the 70th-80th percentile) practically never acquire a doctorate in science, technology, engineering or mathematics - i´ll leave the pondering about why that is to you.

>"IQ tests are stupid!"
- anybody wanting normie cred

They have a specific use that they are pretty good for. Maybe not perfect, but they are a reliable indicator for further tests in another direction if you desire. Maybe somebody scores low, but they are a savant, or somebody scores high but they study sociology.

You're supposed to do them intermittently and rarely. If you kept taking them all day for a few weeks, you would eventually score way above where you actually are.
Just make sure you are well rested when you take one.

>>10574236
None of that is actually verified.
> In addition, people with an IQ below 110-115 (i.e. the 70th-80th percentile) practically never acquire a doctorate in science, technology, engineering or mathematics
You don't even need statistics to see that it's bullshit.

>>10571754
>Modern leftish philosophers tend to dismiss reason, science, objective reality and to insist that everything is culturally relative. It is true that one can ask serious questions about the foundations of scientific knowledge and about how, if at all, the concept of objective reality can be defined. But it is obvious that modern leftish philosophers are not simply cool-headed logicians systematically analyzing the foundations of knowledge. They are deeply involved emotionally in their attack on truth and reality. They attack these concepts because of their own psychological needs. For one thing, their attack is an outlet for hostility, and, to the extent that it is successful, it satisfies the drive for power. More importantly, the leftist hates science and rationality because they classify certain beliefs as true (i.e., successful, superior) and other beliefs as false (i.e., failed, inferior). The leftist’s feelings of inferiority run so deep that he cannot tolerate any classification of some things as successful or superior and other things as failed or inferior. This also underlies the rejection by many leftists of the concept of mental illness and of the utility of IQ tests. Leftists are antagonistic to genetic explanations of human abilities or behavior because such explanations tend to make some persons appear superior or inferior to others. Leftists prefer to give society the credit or blame for an individual’s ability or lack of it. Thus if a person is “inferior” it is not his fault, but society’s, because he has not been brought up properly.
Is Ted right?

>>10574274
Of course.
Show us the part where he talks about "oversocialization" or whatever it was, that one hurt.

>>10571842
>a testing environment completely independent of cultural setting

>>10574274
What the hell is this glorified straw man? What kind of pseud can believe this shit?

>>10574266
http://www.religjournal.com/pdf/ijrr10001.pdf

No, it has been confirmed and I can guarantee if you went into any math department and sought out graduate students and had them tested they'd far outperform the business undegrads, the people working at your local retail stores and nearly everyone working in nursing or doing clerical work at a hospital.

>>10574318
>Study does not state whether subjects have Ph.D.s but merely that they are academic staff. Considerable range overlaping IQ(e.g.,112–132 for social scientists and 112–136 for physicists). Scientists not ranked “eminent,”as in Roe’s study, but working at Cambridge University implies a certain degree of eminence (Simonton 2002) and it is possible (though not detailed in the study) that they have higher IQs than average among Ph.D.holders.
There you go. 115IQ isn't practically nonexistent. And don't fucking quote the other study done on some 64 anonymous eminent scientists, cause it's a category more specific than doctorate, and the test seems to be knowledge based, see physicists not doing math section because it would've been too easy for them, verbal section would probably be something like SAT. It was even some custom made test for the purpose of the experiment, and I doubt its validity very much, especially since before you would also need to take a general population sample that's large enough.

>>10571754
Yes also fuck niggers

>>10574318
It's more proof that math students are better at solving abstract problems rather than anything else. But whatever, go jerk yourself off to your IQ number.

>>10573646
>He is talking about his graph.
What correlation is he talking about then??

>>10573793
>Can you read? You responded to nothing that I posted
You haven't said anything with your posts other than vaguely expressing some opinions.
Here is the reasoning behing your post:
>"I can't really discredit anything but if I express a vague negative opinion, people will think my argument is so obvious that doesn't even need to be said explicitly, and will agree with me."

>>10573844
>how are they measuring cognition in the study if not iq? they are correlating iq's relation to structural damages in the brain, no?
Even if you don't agree that IQ is an indicator of intelligence, the metadata speaks for itself.
If you had a lower IQ, according to the test the decline in the measured functions was greater than if you had a higher IQ. Also they found genetic markers that where significantly correlated with the decline.
Nevertheless, pic related is how they measure cognition. If you don't think this is a proper way to measure it, then you might as well exclude the word cognition from your vocabulary.

>>10571767

Social IQ is the real measure of likelihood of societal success. It's wjy Jews run everything. They have extremely high social IQ.