[Related to: Attitude vs. Altitude]
I.
I write a lot about the importance of IQ research, and I try to debunk pseudoscientific claims that IQ “isn’t real” or “doesn’t matter” or “just shows how well you do on a test”. IQ is one of the best-studied ideas in psychology, one of our best predictors of job performance, future income, and various other forms of success, etc.
But every so often, I get comments/emails saying something like “Help! I just took an IQ test and learned that my IQ is x! This is much lower than I thought, and so obviously I will be a failure in everything I do in life. Can you direct me to the best cliff to jump off of?”
So I want to clarify: IQ is very useful and powerful for research purposes. It’s not nearly as interesting for you personally.
How can this be?
Consider something like income inequality: kids from rich families are at an advantage in life; kids from poor families are at a disadvantage.
From a research point of view, it’s really important to understand this is true. A scientific establishment in denial that having wealthy parents gave you a leg up in life would be an intellectual disgrace. Knowing that wealth runs in families is vital for even a minimal understanding of society, and anybody forced to deny that for political reasons would end up so hopelessly confused that they might as well just give up on having a coherent world-view.
From an personal point of view, coming from a poor family probably isn’t great but shouldn’t be infinitely discouraging. It doesn’t suggest that some kid should think to herself “I come from a family that only makes $30,000 per year, guess that means I’m doomed to be a failure forever, might as well not even try”. A poor kid is certainly at a disadvantage relative to a rich kid, but probably she knew that already long before any scientist came around to tell her. If she took the scientific study of intergenerational income transmission as something more official and final than her general sense that life was hard – if she obsessively recorded every raise and bonus her parents got on the grounds that it determined her own hope for the future – she would be giving the science more weight than it deserves.
So to the people who write me heartfelt letters complaining about their low IQs, I want to make two important points. First, we’re not that good at measuring individual IQs. Second, individual IQs aren’t that good at predicting things.
II.
Start with the measurement problems. People who complain about low IQs (not to mention people who boast about high IQs) are often wildly off about the number.
According to the official studies, IQ tests are rarely wrong. The standard error of measurement is somewhere between 3-7 points (1, 2, 3). Call it 5, and that means your tested IQ will only be off by 5+ points 32% of the time. It’ll only be off by 10+ points 5% of the time, and really big errors should be near impossible.
In reality, I constantly hear about people getting IQ scores that don’t make any sense.
Here’s a pretty standard entry in the “help my IQ is so low” genre – Grappling With The Reality Of Having A Below Average IQ:
When I was 16, as a part of an educational assessment, I took both the WAIS-IV and Woodcock Johnson Cognitive Batteries. My mother was curious as to why I struggled in certain subjects throughout my educational career, particularly in mathematical areas like geometry.
I never got a chance to have a discussion with the psychologist about the results, so I was left to interpret them with me, myself, and the big I known as the Internet – a dangerous activity, I know. This meant two years to date of armchair research, and subsequently, an incessant fear of the implications of my below-average IQ, which stands at a pitiful 94…I still struggle in certain areas of comprehension. I received a score of 1070 on the SAT, (540 Reading & 530 Math), and am barely scraping by in my college algebra class. Honestly, I would be ashamed if any of my coworkers knew I barely could do high school-level algebra.
This person thinks they’re reinforcing their point by listing two different tests, but actually a 1070 on the SAT corresponds to about 104, a full ten points higher. Based on other things in their post – their correct use of big words and complicated sentence structure, their mention that they work a successful job in cybersecurity, the fact that they read a philosophy/psychology subreddit for fun – I’m guessing the 104 is closer to the truth.
From the comments on the same Reddit thread:
Interesting, I hope more people who have an avg. or low IQ post. Personally I had an IQ of 90 or so, but the day of the test I stayed up almost the entire night, slept maybe two hours and as a naive caffeine user I had around 500 mg caffeine. Maybe low IQ people do that.
I did IQTest.dk Raven’s test on impulse after seeing a video of Peterson’s regarding the importance of IQ, not in a very focused mode, almost ADHD like with rumination and I scored 108, but many claim low scores by around 0.5-1 SD, so that would put me in 115-123. I also am vegan, so creatine might increase my IQ by a few points. I think I am in the 120’s, but low IQ people tend to overestimate their IQ, but at least I am certainly 108 non-verbally, which is pretty average and low.
The commenter is right that IQtest.dk usually underestimates scores compared to other tests. But even if we take it at face value, his first score was almost twenty points off. By the official numbers, that should only happen once in every 15,000 people. In reality, someone posts a thread about it on Reddit and another person immediately shows up to say “Yeah, that happened to me”.
Nobel-winning physicist Richard Feynman famously scored “only” 124 on an IQ test in school – still bright, but nowhere near what you would expect of a Nobelist. Some people point out that it might have been biased towards measuring verbal rather than math abilities – then again, Feynman’s autobiography (admittedly edited and stitched together by a ghostwriter) sold 500,000 copies and made the New York Times bestseller list. So either his tested IQ was off by at least 30 points (supposed chance of this happening: 1/505 million), or IQ isn’t real and all of the studies showing that it is are made up by lizardmen to confuse us. In either case, you should be less concerned if your own school IQ tests seem kind of low.
I don’t know why there’s such a discrepancy between the official reliability numbers and the ones that anecdotally make sense. My guess is that the official studies give the tests better somehow. They use professional test administrators instead of overworked school counselors. They give them at a specific time of day instead of while the testee is half-asleep. They don’t let people take a bunch of caffeine before the test. They actually write the result down in a spreadsheet they have right there instead of trusting the testee to remember it accurately.
In my own field, official studies diagnose psychiatric diseases through beautiful Structured Clinical Interviews performed to exacting guidelines. Then real doctors diagnose them through checklists that say “DO NOT USE FOR DIAGNOSIS” in big letters on the top. If psychometrics is at all similar, the clashing numbers aren’t much of a mystery.
But two other points that might also be involved.
First, on a population level IQ is very stable with age. Over a study of 87,498 Scottish children, age 11 IQ and adult IQ correlated at 0.66, about as strong and impressive a correlation as you’ll ever find in the social sciences. But “correlation of 0.66” is also known as “only predicts 44% of the variance”. On an individual level, it is totally possible and not even that surprising to have an IQ of 100 at age 11 but 120 at age 30, or vice versa. Any IQ score you got before high school should be considered a plausible prediction about your adult IQ and nothing more.
Second, the people who get low IQ scores, are shocked, find their whole world tumbling in on themselves, and desperately try to hold on to their dream of being an intellectual – are not a representative sample of the people who get low IQ scores. The average person who gets a low IQ score says “Yup, guess that would explain why I’m failing all my classes”, and then goes back to beating up nerds. When you see someone saying “Help, I got a low IQ score, I’ve double-checked the standard deviation of all of my subscores and found some slight discrepancy but I’m not sure if that counts as Bayesian evidence that the global value is erroneous”, then, well – look, I wouldn’t be making fun of these people if I didn’t constantly come across them. You know who you are.
Just for fun, I analyzed the lowest IQ scores in my collection of SSC/LW surveys. I was only able to find three people who claimed to have an IQ ≤ 100 plus gave SAT results. All three had SAT scores corresponding to IQs in the 120s.
I conclude that at least among the kind of people I encounter and who tend to send me these emails, IQ estimates are pretty terrible.
This is absolutely consistent with population averages of thousands of IQ estimates still being valuable and useful research tools. It just means you shouldn’t use it on yourself. Statistics is what tells us that almost everybody feels stimulated on amphetamines. Reality is my patient who consistently goes to sleep every time she takes Adderall. Neither the statistics nor the lived experience are wrong – but if you use one when you need the other, you’re going to have a bad time.
III.
The second problem is that even if you avoid the problems mentioned above and measure IQ 100% correctly, it’s just not that usefully predictive.
Isn’t that heresy?! Isn’t IQ the most predictive thing we have? Doesn’t it affect every life outcome as proven again and again in well-replicated experiments?
Yes! I’m not denying any of that. I’m saying that things that are statistically true aren’t always true for any individual.
Once again, consider the analogy to family transmission of income. Your parents’ socioeconomic status correlates with your own at about r = 0.2 to 0.3, depending on how you define “socioeconomic status”. By coincidence, this is pretty much the same correlation that Strenze (2006) found for IQ and socioeconomic status. Everyone knows that having rich parents is pretty useful if you want to succeed. But everyone also knows that rich parents aren’t the only thing that goes into success. Someone from a poor family who tries really hard and gets a lot of other advantages still has a chance to make it. A sociologist or economist should be very interested in parent-child success correlations; the average person trying to get ahead should just shrug, realize things are going to be a little easier/harder than they would have been otherwise, and get on with their life.
And this isn’t just about gaining success by becoming an athlete or musician or some other less-intellectual pursuit. Chess talent is correlated with IQ at 0.24, about the same as income. IQ is some complicated central phenomenon that contributes a little to every cognitive skill, but it doesn’t entirely determine any cognitive skill. It’s not just that you can have an average IQ and still be a great chess player if you work hard enough – that’s true, but it’s not just that. It’s that you can have an average IQ and still have high levels of innate talent in chess. It’s not quite as likely as if you have a high IQ, but it’s very much in the range of possibility. And then you add in the effects of working hard enough, and then you’re getting somewhere.
Here is a table of professions by IQ, a couple of decades out of date but probably not too far off (cf. discussion here):
I don’t know how better to demonstrate this idea of “statistically solid, individually shaky”. On a population level, we see that the average doctor is 30 IQ points higher than the average janitor, that college professors are overwhelmingly high-IQ, and we think yeah, this is about what we would hope for from a statistic measuring intelligence. But on an individual level, we see that below-average IQ people sometimes become scientists, professors, engineers, and almost anything else you could hope for.
IV.
I’m kind of annoyed I have to write this post. After investing so much work debunking IQ denialists, I feel like this is really – I don’t know – diluting the brand.
But I actually think it’s not as contradictory as it looks, that there’s some common thread between my posts arguing that no, IQ isn’t fake, and this one.
If you really understand the idea of a statistical predictor – if you have that gear in your brain at a fundamental level – then social science isn’t scary. You can read about IQ, or heredity, or stereotypes, or gender differences, or whatever, and you can say – ah, there’s a slight tendency for one thing to correlate with another thing. Then you can go have dinner.
If you don’t get that, then the world is terrifying. Someone’s said that IQ “correlates with” life outcomes? What the heck is “correlate with”? Did they say that only high-IQ people can be successful? That you’re doomed if you don’t get the right score on a test?
And then you can either resist that with every breath you have – deny all the data, picket the labs where it’s studied, make up silly theories about “emotional intelligence” and “grit” and what have you. Or you can surrender to the darkness, at least have the comfort of knowing that you accept the grim reality as it is.
Imagine an American who somehow gets it into his head that the Communists are about to invade with overwhelming force. He might buy a bunch of guns, turn his house into a bunker, start agitating that Communist sympathizers be imprisoned to prevent them from betraying the country when the time came. Or he might hang a red flag from his house, wear a WELCOME COMMUNIST OVERLORDS tshirt, and start learning Russian. These seem like opposite responses, but they both come from the same fundamental misconception. A lot of the culture war – on both sides – seems like this. I don’t know how to solve this except to try, again and again, to install the necessary gear and convince people that correlations are neither meaningless nor always exactly 1.0.
So please: study the science of IQ. Use IQ to explain and predict social phenomena. Work on figuring out how to raise IQ. Assume that raising IQ will have far-ranging and powerful effects on a wide variety of social problems. Just don’t expect it to predict a single person’s individual achievement with any kind of reliability. Especially not yourself.
Nice try, but waaaaaay too late Scott. The psychosis has already established itself. You’ve done too much damage to too many people’s self image already, and one half-assed little article saying that you never really meant all that stuff you said isn’t putting this genie back in the bottle.
I know, because I’m one of those people. Thanks bro.
A: Why don’t you guys invite me over to dinner any more? It’s already been months.
B: Well, after the last time we had you over, we noticed that some of our silverware went missing.
A: What?! You don’t think I stole it, do you?
B: No. We eventually found it, but the antipathy had already set in.
If you are told a lie and made distraught by the truth, the blame lies with the liars. And with you.
How well does incorrectly interpreting small but definite correlations between IQ and life outcomes as predictive edicts correlate with IQ?
Sadly, the human brain really, really want to cast to boolean.
There’s either a 100% correlation or there’s 0% correlation. Article after article has persuaded many that it’s not 0%, so now their brain is continuously whispering to them that the correlation is 100%. Sure, they understand that IQ is not destiny on an individual basis, but the problem is that their brain *knows* it is.
And facts and some logic is not going to stop their brain from performing what is was built to do. And in doing so, destroy them.
Oh Scott, maybe you should consult with a psychologist from time to time.
There are a thousand reasons why someone might end up with an IQ underestimate, but really no reason for an overestimate (except cheating or Clever Hans administration, I suppose). A decent interpretation of a battery should include caveats and explanations, point out significant intra- and intersubtest scatter, significant observations about test behavior etc. Too often they don’t, but we’re all overworked these days. Composite variables are always better predictors than discrete variables (of course), because of the magic of canceling error terms, thus verbal subtests will be more stable predictors than nonverbal tests, especially in areas like engineering (where ceiling effects are also at play).
And maybe you shouldn’t reconsider emphasizing IQ research quite so much. Those studies do play well with people’s just-world fallacy.
Well, except (as you’ve noticed) when they don’t.
Where did Scott talk about overestimates?
I don’t think he did, but he is talking about measurement error. Did my point not seem relevant to you?
Frankly, no. I’m probably being stupid or missing something.
He mentioned the likelihood of underestimates, and my point is that these are common, while overestimates are really not an issue at all. He’s referring to individual tests, but it turns out there are analogous problems with group comparisons, which is something that ideologically-driven IQ researchers fail to appreciate, be it willfully or unwittingly.
Overestimation is possible if the test has been written in a way that accidentally aligns with whatever biases are in your brain. Or if you are having an unusual good day and your cognitive function is higher than your baseline. Or if you just randomly guessed some answers and got them right by sheer luck.
I have an untested belief that IQ tests select for mild ADD, if you’re lucky.
It’s being fed a bunch of little puzzles/problems, and if they’re the sort of questions you like answering, you don’t have to stay with any of them long enough to get bored.
vV_Vv
An IQ test that produces variable results on the basis of idiosyncratic interests has a validity problem. What you’re saying about unusually good days, cognitively, suggests there’s something undermining your performance that maybe you should see your doctor about. If random guesses pay off, your test has a validity problem.
Nancy, ADHD is a poorly understood condition that is independent of IQ. If you’re an intelligent person who finds cognitive problems stimulating, you’re likely to do well in many parts of an IQ test (though intersubtest scatter is likely to be an issue, because you’re unlikely to find all associated tasks interesting). This also can apply well to academic performance, though the same caveat (is this particular task stimulating to you?) applies.
People with ADHD tend to do either very well or very poorly on processing speed tasks, as a function of what you happen to find stimulating. And they either read voraciously (minority), or never.
Where the hell did you get that idea? You’re entirely wrong, of course.
Oh? Feel free to explain.
IQ can overestimate or understimate g-factor. There’s no reason to expect it to only underestimate. And it doesn’t.
Every item on an IQ test is a sampling. What you’re saying implies that the likelihood of a false positive and a false negative are equal. Think about that for a minute, and the implications of multiple sampling.
Any given individual is much more likely to score lower than their “real” IQ than to score higher.
There is almost nothing a normal person can do to *significantly* increase their IQ in the short term. Maybe one of the no nootropic stacks can get you four or five points if you’re lucking, or have focus problems or whatever.
But smoke a bowl before going in. Or have a couple long island Ice teas.
Or heck, influenza or rhinovirus. A room that is too hot, or too noisy. Having someone sitting behind you that smells REALLY bad etc.
Actual measurement error is likely to be +/- about the same, but local conditions can do VERY little to help and a lot to hurt. Which I suspect is what GeneralDisarray is getting at.
Your intuition about IQ testing and testing generally is flawed. There are more ways to get lucky than getting a false positive on one of the test questions, or ‘samples.’ Even on types of tests where it is impossible to get a question right without knowing the answer, overestimation of your abilities is not only possible but will happen about half the time if the test is of appropriate difficulty for your ability. You have some ‘representative error rate’ and you can overshoot or undershoot that by luck. You misremember less than you normally would, you recall quicker than you normally would, or the reverse. You reason more sloppily or less sloppily.
There is no reason to expect that you would systematically undershoot more than you overshoot, unless you are up against a ceiling that is too low. Certainly the idea that you would never overshoot is insane. Nothing makes the best score you could possibly get on a test your ‘true score.’ The run where you get the luckiest, where you get every single question that you could possibly scrounge the knowledge and reasoning ability to get the right answer correct, is one where you’ve gotten a profoundly unrepresentative score. Tests are tools for comparison, and you’re not comparing with other people’s ‘best possible’ scores but their actual scores.
Antistotle: that’s most of it, yes.
Ikarnal: if when you read this thread tomorrow you still don’t understand what I’m referring to, I’d be happy provide you with a concrete example. But I’m afraid you’re mistaken, and also demonstrating an unfortunately overlooked downside to high IQ, which is that it rather magnifies one’s capacity for self-deception.
I don’t think you understand what ‘real’ IQ is. Your ‘real’ IQ is the result that you get. If you do multiple tries and take the best or worst one, that isn’t your ‘real’ IQ. There’s nothing you can do to make your result more ‘real’ and a hell of a lot you can do to make it less ‘real.’ The fuzziness, the noise, is part of the real test. This is what has been given around the world and characterized.
You can have ‘local conditions’ (illness, blood sugar level, test room distractions, itchy legs, what the fuck ever) that are worse than average, better than average, or average. This is part of noise. Everyone suffers from this. You could get more than average or less than average. But if you try the test again because you are convinced you had more than average negative noise, you are creating a systematic bias which is far worse if you want a representative result.
The average test taker was hurt some amount, by ‘local conditions’ and probably much more of other sorts of blind poor luck (not winning dice rolls on ambiguous choices). That’s your unicorn ‘representative sample,’ with no noise pushing them up or down. There’s no reason to expect that there are way more people below this average level of bad luck than above it.
There are some really unlucky people who lost way more dice rolls than chance would indicate, and some really lucky people who won way more. But if you let people go ‘wait, that was unrepresentative, I don’t like my result, I’m going to try for another’ you are replacing noise with systemic error. Which is bad. Even if the only effect is to remove the noise – which isn’t the case – this guy with his de-noised test is failing to represent IQ properly. IQ is a thing in the real world with certain error bars. He’s a dude on the low end of the error bar, of the negative misrepresentation. That’s important. If he goes on to win a Nobel Prize, that g-loaded achievement should be stacked up next to his actual IQ result, and that revealed error is important. If you take 5 different IQ tests and average them, or whatever, the accuracy of that result is not necessarily relevant at all to the guy who took AN IQ test and wants to know how representative that is. We have data on how g-loaded AN IQ test is, we don’t have nearly as broad a dataset of some homebrew composite. We haven’t gone all around the world giving that battery. And also, if you started and stopped this re-test thing because of dissatisfaction/satisfaction with the results, you’ve introduced systemic error.
And it’s possible that your composite wouldn’t reduce error as much as you think. These aren’t really independent samplings. There could be something biasing the results of all these tests. Like you’re just unusually bad or good at taking tests, for example.
There are tests for whether you’ve been exposed to some disease and are a health risk, they detect antigens. Let’s say you have one of these tests for smallpox, and it has a 90% rate of accuracy within some population of a given age or whatever. Meaning 90% of the people who test positive have been exposed to the disease. The 10% error could be entirely people who have been vaccinated, which is unusual for their age group at this particular time, or whatever.
You won’t be able to get rid of this type of error (unlike the fumble fingered technician kind of error) by testing multiple times. It’s not the case that someone who fails twice has a 99% chance of having been exposed to the real live disease. Get it? You don’t get to just assume that you can completely ‘de-noise’ some test with multiple samplings. There is very likely some that you can reduce, but to characterize the reduction you need a significant sample OF these multiple sampling composite tests. And you then have a new result that is not ‘real’ IQ but your SuperSample MegaBlast IQ or whatever you want to call your dubious contribution to psychometrics.
This is worth re-iterating in a brief post on its own. If we are talking about IQ as the useful test/set of tests that have been given round the world with proven predictivity – there is NO ‘REAL’ IQ separate from the result you get when you take an IQ test. That IQ score IS your IQ. It measures your IQ PERFECTLY.
What it measures VERY IMPERFECTLY, with errors that have been characterized, is ‘g,’ which is what gives IQ its value. The general intelligence factor, your success at intellectually straining tasks.
General intelligence is not IQ. IQ is an attempt to measure general intelligence. A pretty damn good attempt, at that. Still very imperfect. What can I say, we live in an imperfect world.
When you say ‘omg this IQ test is wrong it gave me the wrong IQ I need a new test,’ whatever hare-brained scheme you go on you are not changing your IQ. Your IQ is the result that you got. You could in your quixotic quest come up with new procedures and/or a new test that has a 1:1 correlation with g. You can call this test IQ and get your ‘real’ g-percentile tracking score. Guess what – your IQ didn’t change. Your IQ is exactly what it was before. You created something else, which happened to be better (lucky you, very lucky). It’s still something. else.
But more likely you won’t come up with something better. Much more likely you will do some ‘clever hans’ nonsense where you keep trying tests and stop when you get a result that you like.
We have lots of data about what it means when you take a dudester and you give them an IQ test and you look at the result and you look at the dudester. This is what we have data on. Stay in this world. This is the world of information that is of known utility. Don’t do random other shit. You’re just exiting the world where results are known to be meaningful and entering the world where ???maybe they are meaningful maybe they are moderately biased maybe they are super biased who knows????
The distinction here is between:
* Taking one full IQ test where all factors (environment, time of day, tiredness, whether you’re hung over etc.) are nearly constant for the duration of the test, and
* Taking a number of full IQ tests at random times of day in random locations in random states of mind over a period of several months.
In the former case we would expect errors (really, “noise” in this context) to cancel each other out, and produce a measure of the individual’s IQ at that time when they are in whatever state of mind they’re in at that time. In the latter case (as GeneralDisarray and Antistotle are arguing), there are lots of factors that might cause the results to skew downwards (hungover, tired, uncomfortable); but almost none that will cause a symmetrical bias in the positive direction.
Having an amount of ‘negative bias’ or bad luck below the average amount doled out to your fellow participants is a bias in the positive direction. These are two ways of saying the exact same thing.
You seem to assume that ‘skew downwards’ factors are some kind of boolean that has a ten percent change of being flagged ‘true’ or whatever. This is nonsensical. There are gradations to everything, including how hungry or tired or itchy you are. Including how distracting or comfortable the testing environment is. Including how lucky or unlucky you get on the many dice rolls.
@ilkarnal
Okay, obviously, yes, in some sense that’s true. But it still seems to be missing the point of contention here. An IQ test may “perfectly” measure your IQ on a given day, but there are other quantities we could be interested in. One of them might be, I don’t know, let’s call it IQ_ave (your average IQ score after taking the test a number of different times). Obviously IQ_ave is no more a measure of “real” intelligence than any individual IQ score would be. But it’s still a meaningful number, and you could say it’s closer to what people mean when they talk about “IQ” in the first place.
Anyway, the question is then: how well do we expect a given IQ score to reflect someone’s IQ_ave. And that depends very much on the distribution of IQ scores relative to IQ_ave. I think all GeneralDisarray was trying to say was that this distribution is likely to be skewed – it likely has a much longer tail to the left than the right. So getting an IQ score on a particular day that’s, say, 25 points below your IQ_ave, is much more likely than getting a score that’s 25 points above your IQ_ave (because there are very few things that can increase performance as much as, say, being hungover can decrease performance). This is true despite the fact that you should on average expect your IQ score to be neither above nor below IQ_ave. It just means that small overestimates would be relatively more common than small underestimates, and large overestimates would be relatively less common than large underestimates.
So let’s imagine we took you, ikarnal, and you undertook a series of full IQ tests over a period of several months. The conditions, daily timing, location etc. of each test were as identical as we were able to make them. Similarly, you assured us that in each case you felt well-rested, healthy, weren’t drunk, and weren’t hungover etc.
The total number of tests (lets say, 12) were divided equally in two. The only difference between the two groups of tests was that in one I was standing over you with a fork, jabbing you in your abdomen every so often; and in the other group of tests, I was not doing this thing.
Now let’s clarify things. Which of the following are you, ikarnal, claiming:
1. That your results in the fork-abdomen tests would be roughly the same as the non-form-abdomen tests, with no statistically significant difference in your test result each time.
2. That you would accept the outcome of the tests, perhaps averaged over time, as being as accurate a reflection of your true IQ as we are able to achieve.
3. That even if 1. is not correct, and there is a statistically significant difference in your measured IQ in the two groups of tests, there exists some equal-and-opposite effect that will counteract the deliberate attempt to sabotage your test taking (even though we’re at pains to avoid any such thing happening).
4. That even if 1. is not correct, and there is a statistically significant difference in your measured IQ in the two groups of tests, your measured IQ when you’re being struck in the ribs with a fork is in fact *higher* than when you’re not being struck in the ribs with a fork.
Put in statistical terms, what I’m claiming is that the following quotation is correct:
If this *is* correct, we would expect that if *the same individual* took the same full IQ tests over an extended period of time, we would see a skewed distribution, clustering round a particular point (the ‘true’ IQ), with some noise on either side; but with a long-ish tail on the negative side.
This is for the reason stated in the quote: there’s lots of things that can make you perform *worse* than what you might be able to achieve in an obsessively well-controlled series of tests. There are rather fewer things (notwithstanding nootropics) that would have as big an impact on the positive side.
This isn’t about random noise, it’s about the fact than when one person takes one IQ test, and they’re ill or drunk or whatever, the result isn’t an accurate reflection of their IQ.
(I accept that there’s an ambiguity here between the Platonic “true” IQ and the actual measured value in any particular case, but I’m sure people can follow what I mean).
The fact that an individual’s results might (or might not) skew in this way over time is irrelevant to the distribution of the population as a whole, the central limit theorem being what it is.
As you say, the aggregate IQ of a population will automatically correct for the fact that a few people were ill/hungover on the day. But this just means that the measured population IQ as measured is lower than it would be in the case that a minority of people *weren’t* hung over when they took the test, or otherwise experiencing someone sticking a fork in their ribs.
So, I’ve made a specific, falsifiable prediction; and I’d be very grateful to hear from anyone who can point me to any research that disproves it. I know good, longitudinal studies of IQ are pretty rare, so I suspect my hunch won’t be proved or disproved either way.
The last part of this post dealt with that. You are replacing well characterized error bars with uncharacterized error bars, you have no idea whether you’re actually making any appreciable improvement and could get a totally false sense of certainty, and you will introduce systemic bias if you do something like stop when you get a score you like. Or maybe there’s systemic bias in obsessing over IQ results and taking multiple tests, period – after all, those sampled around the world didn’t do that. They took an IQ test. They didn’t obsess over it, they didn’t view this as the key to prophesying their future. They took. A. Damn. Test. Not seven, stopping at seven because the seventh test gave a satisfying result.
You don’t know how much better IQ_ave (I preferred my term, SuperSample MegaBlast IQ) is than IQ, and it could well be worse. Move from the well-characterized to the poorly characterized as you like, but don’t delude yourself into thinking this improves your certainty.
No. It’s not ‘no more,’ it is less. We have characterized IQ, various IQ tests and what they mean. We have not characterized some grab-bag composite with more tests thrown in until some badly characterized condition is satisfied.
Is it meaningful, when compared to no data at all? Definitely. But the most meaningful thing in this domain is the thing that has been done all around the world, for many many decades. Fiddling with the methodology could be interesting, but the idea that this is a way of being more certain about underlying g is nonsense. You would have to actually study this, it cannot be assumed.
Well, you’re wrong about what GeneralDisarray was trying to say. Maybe there will be a significantly fatter tail to the left – that certainly cannot be assumed. But contra GeneralDisarray, there will certainly be a very significant tail to the right, and there’s nothing that makes the right end of the distribution more ‘real.’
Look, how many catastrophically ill people take IQ tests? If you’re in a coma that doesn’t mean you take an IQ test and get a zero, it means you don’t take the IQ test. You will be within some certain range of normality if you’re sitting down and taking a test.
Yeah, some people will feel unusually bad. Some, one presumes, will feel unusually good. Yeah, you can feel way way way worse than you can feel good – like, you’re not ever going to wake up feeling as good as someone spraying shit from one end and vomit from the other feels bad, unless you wake up with the aid of some interesting substances. But mr two-way hose is not going to go take a test. He’s going to sit at home, doing his two-way hose thing.
Feelings probably just aren’t as important as you seem to think. More elemental chance probably overwhelms this as a source of error. This would include things like a certain category of test just systematically overestimating or underestimating your abilities, which we don’t have any reason to think is impossible. What we know is there’s a lot of error, a lot of variance. No way of removing this error has been found, as of now. Go try to find one if you think it is simple. We have, essentially, a moderately tall and wide shotgun scatter blast cluster that defines your ‘g.’ There’s no reason to expect you’ll find a pellet that is guaranteed to be dead center.
And that is very unfortunate, because this almost religious false sense of certainty and meaning is unscientific and will lead people astray. In both directions. It makes a useful tool into something very dubious.
This is quite outside the range of conditions under which people have taken IQ tests, and is as a result useless.
People, who are healthy enough to have gotten out of bed and walk around, go take a test. They complete the test, and hand it in. If someone came in with an axe and chopped the head off the person next to them, this would not result in them getting a worse score. It would result in them getting no score.
If I was experiencing extreme sharp pain in my abdomen, I would go to the bathroom. I would not keep taking the test while muffling my screams. If someone was poking me even lightly, obviously I would deal with that rather than taking the test.
None of those claims bear any relation to mine.
And this is just wrong. Tons of things can have an impact on the positive side, which you can view, if you like, as a less than average negative side. You can say winning all the dice rolls is getting lucky, or you can call it ‘never getting unlucky.’ It doesn’t matter.
It is a perfectly accurate reflection of their IQ – it is their IQ. You take an IQ test, the result is your IQ.
You are using some quasi-religious ideal of IQ. IQ is an attempt to measure general intelligence. It has certain error bars. General intelligence is not IQ. IQ is not general intelligence. Your general intelligence is not your ‘real’ IQ. Your IQ is not your general intelligence. These are two different things that are statistically linked.
If this was the pattern we saw, this would not prove that the cluster round a particular point is “the ‘true’ IQ.” First of all, that’s just a nonsensical way of phrasing this and you really need to stop. Let me phrase this correctly. There is no implication that this cluster is more g-loaded than the first test.
One way of interpreting this result is that some guy took an IQ test, didn’t like the result, bunch of times getting higher and more consistent scores as they remembered more answers or increased proficiency in more categories, until they plateaued (reaching questions or categories they could not answer no matter how much practice they got.) If this interpretation is correct, you would expect the high cluster to be unrepresentative and less g-loaded.
Not only would you have to demonstrate that this pattern exists, you would have to then demonstrate that the high cluster is more g-loaded than the average, or than the first test. I consider it unlikely that the high cluster, should it appear, will be more g-loaded than the first result. Following that, I consider it unlikely that it will be more g-loaded than the average result. There is no obvious reason to remove low scores.
Then, in order for this exercise to be useful, the improvement in g-loading would have to be usefully high. You are spending, at this point, a very significant amount of time.
I consider each of these propositions, which all must be correct, unlikely to be true.
Sorry, yes, I missed that.
Okay, but..why, though?
Like, look, I have no horse in this race. I have no idea what my IQ is, and I don’t really care. I’m not trying to find a way to “game” IQ scores to get a better result, and I certainly wouldn’t do anything stupid like stopping when I got a better result.
But why do we have no idea whether we’re making an appreciable improvement when we average over multiple scores? I’m genuinely asking here, that seems like a very counterintuitive thing to say. Forgetting for a second about whether the distribution is skewed or not, we know that IQ is noisy to some degree, right? Why wouldn’t averaging over a bunch of test results (assuming you don’t do something stupid like stopping when you get a good result) give you some kind of better indicator of your “true” IQ?
Like, you say in your other comment:
But isn’t that exactly what multiple sampling does? Reduce noise? Like, noise is pretty much defined as the thing that goes down once you do multiple samples.
This seems a bit uncharitable. Obviously I agree that SuperSample MegaBlast IQ has not been carefully investigated in the literature. But to call it grab-bag seems strange to me. We’re just talking about a straight-up average – it’s pretty much the most obvious thing you can think of to do with a sample. My prior would be on the average of a bunch of tests being at least as meaningful as the test itself, unless I had some specific reason to think otherwise.
Well sure, I wouldn’t want to assume it, exactly. But doesn’t it seem like a pretty good guess going forward? The two most obvious things that could screw up SuperSample MegaBlast IQ would be, like you said, a selection bias on who decides to do multiple test compared to one, and a biased stopping condition. But those aren’t particularly hard to mitigate.
Sure, but this just seems like quibbling over degrees. I completely agree that if someone feels bad enough they probably won’t take the test. But that still leaves plenty of room for people who feel semi-crappy, and I don’t think they’d have a corresponding population of people who feel semi-awesome. It seems likely to me (as in I would be willing to bet on it) that in a proper study the distribution would turn out to be left-skewed. I agree that we shouldn’t assume it, certainly, but I’d be surprised if it wasn’t the case.
Well look, if you’re just trying to push back against people using IQ in an overly certain/deterministic way, then I’m 100% with you. I don’t have any kind of belief that IQ captures everything there is to know about a person’s intelligence. But given that IQ seems to be predictive and seems to capture something real, I honestly don’t see why a straight-up average of multiple tests (assuming you don’t do something stupid like not determine a stopping condition in advance) wouldn’t be better than a single test.
You can think of it like that, or you can think of it as a random sample from a probability distribution. As I understand it, you’re claiming that the same individual taking a test multiple times (and assuming that the test is a ‘good’ IQ test and doesn’t benefit from practice or remembering past questions) will get a series of results that will cluster around some value, and the distribution of that clustering will be symmetrical.
(I mean, the way you’ve phrased it kind of makes it sound like you’re saying that IQ tests don’t experience noise and the measured IQ result would be perfectly replicated again and again, but that would be absurd.)
The point of my example was to highlight what I took GeneralDisarray’s to be; that is, that the distribution from which an individual IQ test is taken is asymmetrical. There’s a ‘hard limit’ (plus some amount of noise) above which you simply can’t go, then there is a long tail to the left, caused by things ranging from mild illness, distraction, feeling hungover, etc.
The result of this would be an asymmetrical distribution. *Or maybe not*. This is a testable claim, but I don’t think anyone’s done a study where they apply the same (type of) IQ test to someone over regular intervals for an extended period of time. For most people, an IQ is a one-off event, and probably one of only a handful of times in their lives – at most – when they get it measured. If they were ‘slightly off’ on that day for whatever reason it seems strange to claim that this particular measurement should hold so much significance, whatever the shape of the individual distribution.
After reading all this back and forth, I think you’re both right.
My take on what GeneralDisarray is saying is that someone whose ‘true’ IQ (however we measure that) is around, let’s say, 140 can do badly on a test for whatever reason (all the external influences ilkarnal mentions) and so only get a score equivalent to IQ 120.
But for someone whose ‘true’ IQ is 120, doing well enough to score at IQ 140 level would be so much harder even if they felt really good that day and had practiced hard and all the rest of it, that it’s much more likely that most IQ scores would be under-estimates rather than over-estimates.
A sprinter can run slowly and below their best but an 800 metres runner, no matter how good they are, is highly unlikely to be able to run a sub-10 second 100 metres dash.
If for no other reason, then when a test is calibrated by normalizing scores so the average becomes zero etc, the underestimates in the calibration data will skew this process, meaning all the non-underestimated scores become slight overestimates.
But it is truly hard to imagine a test with an element of random error that only errs in one direction.
Below you say
but I don’t think anybody said the likelihood had to be equal, or of equal magnitude. Just that it would be non-zero.
With each successive sampling (item on the battery), the likelihood of an overestimate becomes increasingly trivial. This is why it takes over an hour to administer one (usually), and why some of us took strong issue with the tightened-up discontinuation rules designed to decrease administration time (why we limit test).
Seems trivially wrong if IQ tests still include multiple choice questions.
Most questions are open-ended. Some are multiple choice. Item analysis is of course performed on every item, and there are subtest discontinuation rules that minimize the impact of guessing. Most items are open response, or require fluid reasoning to complete them.
I always wonder if people saying this are just trying to comfort me.
See, I was tested in 8th grade or so, and got told I was kinda gifted but not really genius-level. Just around the cutoff for being seriously bright. My verbal score was, admittedly, way high, but it was offset by a performance/math score that was barely even above average.
Whatever. It was immensely disappointing at the time, because I was a massive nerd. It was basically telling me that I’d never have much hope of being all that good at STEM stuff, no matter how well I could seemingly speak and write. It’s even been depressing later in life, since I really want to be a proper scientist rather than some guy who can talk real good. All this despite my getting pretty good marks in STEM classes in college.
Then it turns out that people go around saying the verbal component is more stable, and that’s the one that’s higher for me. Seems like a really self-serving thing to say, since – me imagining everyone else as copies of myself – I tend to assume everyone will have verbal scores higher than their performance scores.
So, big questions:
1) Do most people actually have higher verbal than performance IQ scores? Is that a trend, or am I projecting myself onto everyone else? Are there people who default to effortless greatness on quantitative problems but can only just about communicate with expert fluency?
2) Conditional on the above questions being answered, “no, yes, and yes”, does the pattern of verbal scores being more stable, and performance scores more variable, still actually hold? Why should any such asymmetry actually occur?
(On the upside for me, the verbal score really was quite good, and does show up in my real performance on verbal/speech tasks. People always compliment my abilities with second languages, even if I’ve never reached literary-level bilingual fluency for lack of practice and opportunity. I can speed-run the verbal sections of standardized tests. It makes the flagrant difficulty with mental computation and pattern-recognition all the more painfully apparent.)
IQ scores are normalized, so no, an equal number of people have imbalances between verbal and performance IQ in each direction.
Verbal scores, particularly vocabulary, are what we call “hold” scores. They’re more robust over time, and if I want to estimate someone’s pre-TBI functioning, I want to look first at what’s most stable. Verbal scores are much better predictors of academic performance in large part because much of academic instruction, even on nonverbal topics, is verbal.
Verbal scores are based less on discrete skills. There are many books on this subject, and despite what some on this thread are espousing, the debate on the nature of the “g” factor has not been fully resolved (it may be a statistical artifact resulting from the narrow environmental scope in which it is measured, and evidence of its impact is assessed, and there are almost certainly epigenetic factors that impact its expression).
Many experts in this area have very little training in IQ testing itself or the statistics involved in test construction, norming and interpretation, and little to no experience in actually administering IQ tests. That lack of context shows in their thinking; IQ can do wonders to bolster one’s overconfidence. (I wish they better appreciated the relationship between heuristic biases, false positives and type-1 errors).
How do verbal scores translate into verbal skill?
My VCI was in the 99.9 percentile (way above working memory or processing speed) but I’m an Aspie and I mumble, ramble and often fail to recognise if the listener is understanding me or, indeed, interested.
If the errors only point one way, then you’re defining someone’s IQ as their peak possible performance. But that isn’t very useful or realistic.
For example, we know effort matters. We usually talk about this as if it underestimates people who don’t try that hard. But it’s equally true that if someone is unusually motivated to take an IQ test, their score is likely an overestimate relative to their normal intellectual performance, since they’re typically not that motivated.
Or if someone is unusually well-rested and relaxed, they might turn in an unusually good performance. But that’s actually an overestimate relative to their “typical” performance, which will be a better predictor of how they perform on everyday g-loaded tasks.
Presumably someone’s cognitive ability fluctuates with mood, rest, diet, circadian rhythm, etc. We typically define IQ as the average of this fluctuating ability, not the peak.
Errors only going one way:
You get to the end of the information subtest on the WAIS. I ask you, “Who wrote Hypnerotomachia Poliphili?” [Not an actual item, but relevant to my point.]
If you know the answer, by damn, you know the answer; even if you know this for idiosyncratic reasons (say, your mother is writing a dissertation on 17th century French literature), you remembered the title and the associated author, which is quite a feat. In order for the error term to vary equally in both direction, you’re requiring the likelihood of a correct answer with a guess on that question.
That’s not the way it works.
If you do happen to know the answer, but you’re tired, anxious, low potassium, dehydrated, hypoglycemic, hungover etc, you might not recall (false negative).
There are thousands of explanations for an underestimate. But there’s really only one explanation for a correct response. Capiche?
Applies across the test. (And no, nootropics don’t really help all that much. Mostly they help modulate arousal for people who are a bit suboptimally low in the testing context, but there are other folks who will then skew suboptimally high.)
You might get lucky on all the questions because they relate to specific interests or memories you just happen to have but the chances of that happening are so rare they make Bangara musicals out of it.
Are you saying that IQ is not Normal ? (badump-bump)
One disagreement here, relating to the SAT scores. Based on that post, that person is in their first year of college, which means that they would have taken it on the new scale. On the new scale, that score corresponds to roughly the 50th percentile, which is much closer to his IQ. The site you linked to dates to 2007 (or maybe even older).
> Based on that post, that person is in their first year of college, which means that they would have taken it on the new scale. On the new scale, that score corresponds to roughly the 50th percentile, which is much closer to his IQ.
A 1070 apparently corresponds with an IQ a hair below ~108, which is pretty much bang-on what Scott wrote in the post (and 14 IQ pts higher than the guy’s original estimate).
EDIT: I see from a comment further down that the original text said 114, which indeed is a lot further away from 94 than 104 is. Never mind..
[1] https://pumpkinperson.com/2015/12/16/revised-chart-converting-sat-scores-to-iq-equivalents/
Your parents’ socioeconomic status correlates with your own at about r = 0.2 to 0.3
Am I the only one to whom this sounds shockingly, even implausibly, low? I tend to roll my eyes at the perennial articles in SWPL outlets claiming that the US has a rigid class society where people from poor families never get a chance, but even I expect parental SES to predict more than 4-9% of their children’s SES, especially since this is everything parents contribute (genes, culture, and money, not just the latter). I suppose this can be partly explained by being (I assume) based on single-year snapshots rather than long-run averages, but that’s still pretty low.
Gregory Clark’s research on surnames suggests that correlations in income and social status are higher between grandparents and grandchildren than you’d expect just from the fairly low parent-child correlations that economists come up with.
One reason is that it’s hard to come up with a single metric to adequately measure what we are interested in overall, which tends to depress parent-child correlations.
Is that Pearson’s correlation or Spearman’s rank correlation? If it’s Pearson then I can easily see why it is low, given that wealth and income are distributed according to Pareto distributions.
For instance, you have people like Bill Gates and Mark Zuckerberg who are way much wealthier than their parents, so they make the linear correlation statistic go down, but their parents were still upper class, hence they make a rank correlation statistic go up.
People do all sorts of transformations. For income, they either top-code or log-transform, often both. Bill Gates isn’t in the sample, anyhow.
But this isn’t about income. This is socio-economic status. There isn’t a numerical ground truth. It’s probably on the basis of 10 bins. But it does seem small for such a metric.
Judging by reading random New York Times marriage announcements, it would appear that trans-generational transmission of social status is higher than what the social scientists estimate. But I can sympathize with the social scientists’ difficulties in adequately quantifying social status in order to calculate correlations.
Mathematician/child-prodigy Terry Tao speaks to these same issues in a weblog essay that addresses the question “Does one have to be a genius to do maths?”
Tao’s essay begins by quoting the philosopher / critic / essayist José Ortega y Gasset, as follows:
Tao concludes “The answer [to the titular question] is an emphatic NO.”
In high school, while taking a state-mandated test of educational development, my own empathic quotient (EQ) was so low, as to deny me an appreciation of the creative vision and moral courage of the girl seated next to me, who marked-up her test-sheet so as to depict a magnificently life-like (albeit pointillist) dog’s head.
This young lady deserved to receive OuLiPo membership on the spot. 🙂
Your link to Tao’s blog actually goes to a comment about Jewish STEM performance.
By the way, one of the reasons I’m always skeptical about this so-called Jewish STEM advantage is that (as noted about myself) as far as I can tell, Jewish brains have a primarily verbal advantage, which is actually what we’d expect if the arranged-marriage selection pressure was for good religious scholars, rather than good scientists hundreds of years ahead of any concept of science.
Given the real history, I expect to see fewer Jews in STEM or FIRE jobs than I actually see, and many more in literature, politics, history, and philosophy than I actually see. Law seems to be the exception, with about the number of Jews I expect.
Maybe the difference comes from an economic pressure towards money-making job fields? But that wouldn’t really explain the amount of Jews (especially Israelis, nowadays) in the academic sciences, where making a secure living is actually quite difficult.
But 1) verbal and mathematical ability are positively correlated, and 2) European Jews have been craftsmen, traders and moneylenders for hundred years, which could have caused sexual selection pressure towards mechanical/mathematical ability. After all, being good at winning Talmudic disputes doesn’t look like something particularly useful to feed your children per se, while being good ad making good deals is.
Being good at making deals sounds primarily like a verbal-IQ thing. Math helps, but particularly in pre-modern markets I think most of the gains are to be found in what you can talk the other party into rather than finding some unrecognized cost or profit hidden in the math.
But the skills in abstract thinking that are useful in maths might well be those that are also useful in spotting an unrecognised cost or profit, even if you never formally lay out said market (because algebra hasn’t been invented yet).
E.g. realising that this material can be made more cheaply over here or the new barge boats make this spot more profitable for a factory.
Gentlemen, you are forgetting the ladies, and in breeding the mother is very important.
From the vague impressions I get, the ideal of the Talmudic scholar means that the wife is the one who deals with the practical side of life – which includes “earning a living to support the husband studying and discussing Torah and the Talmud twelve hours a day”:
Look at the traits of the Ideal Wife in the Book of Proverbs, which includes:
So while selecting for verbal intelligence in the male line, might not selecting for mercantile ability = mathematical talent in the female line?
Very good point, and thank you for the correction. The question then becomes how much mathematics was involved in running a prosperous medieval Jewish household. Certainly enough arithmetic to keep the household finances in the black, which sounds trivial to most of us but may have stretched the limits of 14th-century lay education in math.
Verbal ability 100% helps a person get ahead in scientific academia. Presenting your data in a well-shaped verbal package helps almost as much as having good data in the first place.
I’m not aware of the rules but does it work like marbles where the winner keeps the loser’s Talmud?
I’m skeptical of the idea of a Talmudic dispute being “won” during the life of the disputants.
“But 1) verbal and mathematical ability are positively correlated, and 2) European Jews have been craftsmen, traders and moneylenders for hundred years, which could have caused sexual selection pressure towards mechanical/mathematical ability. After all, being good at winning Talmudic disputes doesn’t look like something particularly useful to feed your children per se, while being good ad making good deals is.”
As I understand it, Jews show up well on verbal and mathematical ability, but not visualization.
It’s probably more accurate to framing it as good at taking part in Talmudic disputes rather than good at winning Talmudic disputes.
Again, as I understand it, the interesting thing was that there were two paths to success for Jewish men. One was making money. The other was being good at Talmud– poor boys who were good at Talmud had a chance of marrying a rich man’s daughter. Wikipedia described this as an unproven theory about Jewish intelligence.
Apologies are extended! The intended link was to Tao’s essay itself, “Does one have to be a genius to do maths?“, rather than to any of the commentary upon it.
The error occurred as the result of a cut-and-paste from a draft essay relating to Geoffrey Cantor’s history Quakers, Jews, and Science: Religious Responses to Modernity and the Sciences in Britain, 1650-1900 (Oxford Press, 2005), John H. Morgan’s article “The Free Quakers: Reaffirming the Legacy of Conscience and Liberty (The Spiritual Journey of a Solitary People)” (Journal for the Study of Religions and Ideologies, 2012), and (for younger readers) Silas Weir Mitchell’s turn-of-the-19th-century proto-Heinlein young adult novel Hugh Wynne, Free Quaker: Sometime Brevet Lieutenant-Colonel on the Staff of His Excellency General Washington (1897, text here, historical context and original documents here).
SSC readers who object to in-depth scholarly references, explicit chains of reasoning, deconstructive critiques of (what can justifiably be called) “Cluster B rationalism”, and conclusions that are heretical from a traditional rationalist perspective sensu stricto, may be discomfited even by mention of IQ-related postulates — which are provisionally slated for eventual posting to (e.g.) the LW/relaunch essay “Why I am not a Quaker (even though it often seems as though I should be)” — to the general effect that:
Less formally, these readings provide historical, philosophical, moral, and even neuropsychological foundations that naturally explain why, for 21st century Enlightened progressivism in particular, “It’s always a beautiful day in the neighborhood.” 🙂
——-
@book{cite-key, Author = {Cantor,
Geoffrey N.}, Publisher = {OUP Oxford},
Series = {Cambridge studies in
nineteenth-century literature and
culture}, Title = {Quakers, Jews, and
Science: Religious Responses to
Modernity and the Sciences in Britain,
1650-1900}, Year = {2005}}
@article{cite-key, Author = {John H.
Morgan}, Journal = {Journal for the
Study of Religions and Ideologies},
Number = {32}, Pages = {288-305}, Title
= {The Free Quakers: Reaffirming the
Legacy of Conscience and Liberty (The
Spiritual Journey of a Solitary
People)}, Volume = {11}, Year = {2012}}
@book{cite-key, Author = {Silas Weir
Mitchell}, Publisher = {Century
Company}, Series = {Works}, Title =
{Hugh Wynne, Free Quaker: Sometime
Brevet Lieutenant-Colonel on the Staff
of His Excellency General Washington},
Year = {1897}}
This would not make me skeptical. A verbal advantage may be very useful for learning… everything, including mathematics. And for lecturing, and writing papers, and convincing people to hire you or give you grant money.
Take two people with similarly high mathematical ability, and I would not be shocked if the one with a higher verbal ability were both a faster learner and more likely to get ahead in whatever technical profession for non-technical reasons.
Here’s my IQ FAQ from a decade ago that makes some similar points:
http://www.vdare.com/articles/why-do-we-keep-writing-about-intelligence-an-iq-faq
My God, here you are bringing up that odious Saletan debacle after all these years. Scott, there are myriad examples of the misuses of IQ research. Please learn from the mistakes of people like Mr Sailer here, who appears committed to a rather unfortunate and destructive form of social engineering.
Can you give me a little history on the “debacle,” please? I’m at work and can’t spend a lot of time looking it up right now, and the odds of my remembering to do it when I get home are slim.
It’s mentioned on Saletan’s Wikipedia page, along with links to rebuttals in other publications. What is no longer accessible is the rather lengthy, boisterous discussion on what was the Slate’s discussion board, the Fray (the presence of which was almost certainly instrumental in Slate’s decision to shutter the Fray). It was a big deal at the time, and many people became acquainted with a new breed of genteel, academic racists, like Mr Sailer here whom we’d never realized were out there. Learning that a psychologist who’d published papers on altruism that I quite liked, had also attempted to verify an inverse correlation between penis size and IQ was rather disheartening. Or learning about the origins and continuing activity of The Pioneer Fund.
But ask Mr. Sailer. He can tell you all about it.
I usually only lurk, but I registered to say that I was disappointed to read your comments in response to Mr. Sailer. I enjoyed reading your civil discussion about IQ in the comment thread above, but here is where you turned from arguments to vague smears and insults. If you have a fact-based disagreement with the posted FAQ, I would enjoy reading it.
Thudbit, Mr Sailer and his associates are a bright and sophisticated bunch of ideologues who are very familiar with the psychology of movements and of persuasion. In fact, if one were inclined to make an argument about the dangers of intellectualism, they’d be a sterling example of the dangers of confirmation bias and premature cognitive closure. Comparisons with the intellectual wings of vilified historical nationalist political movements are apt.
Melvin Konner provides an eloquent warning in his preface (titled Caveat) to the notes section of The Tangled Wing. Those notes are available for free download directly from the following link (on his website): http://www.melvinkonner.com/wp-content/uploads/2009/02/tangledwingnotes.pdf
I’d be happy to discuss Konner’s essay with you. Mr Sailer is s banner carrier for the worst of intellectual society. It bothers me to see him here because he’s a pied Piper, of sorts. We should all be suspicious of those more committed to proselytizing than in the state and fate of actual people.
I’ve always found the latter to be a lot more dangerous. In particular they seem to have a rather restrictive definition of “actual people” which varies but tends to exclude me.
At least with a proselytizer I know what they’re selling up front.
1: if only it were so easy. If you think you always know when you’re being sold to, you either haven’t encountered any sophisticated salesmen, or you didn’t realize it.
2: that is the crux of the problem. I wish people could better remember that by affiliating with exclusionists they and everyone else will always be vulnerable to future exclusion, and will invariably encounter some strange social triangulation occurring as folks attempt to preempt their exclusion. I’d argue there’s something like this going on with ex military families and the NFL right now.
More hate than rationality here …
Philosophical question of the times: are hating and hating hate equivalent? Also, how can we determine one’s intellectual honesty in a scholarly discussion on a socially loaded issue?
Steve, you might need to put a trigger warning on some of your links. Though GeneralDisarray appeared to be more triggered by your name and the offhand reference to Saletan than anything else.
In fairness, what seems to be happening is that race/IQ connections are known to be somewhat taboo to discuss here but no one knows exactly how taboo, partly because Scott seems somewhat conflicted on the issue, so Steve posts a link that comments on the issue and General responds with ad hominems + a link + more ad hominems. But no one wants to enter into an actual argument. And such an argument would probably just be mind-death. So instead you just get some shots across the bow.
Though General, if you want to argue for excluding Steve, you should probably first argue that Scott cease linking West Hunter. I mean, that could be construed as an implicit endorsement! His official stance towards Steve is mere tolerance.
Sorry, I was pointing out that 10 years ago I had made a point similar to the point of Scott’s post:
“Q. So, do IQ tests predict an individual’s fate?
“A. In an absolute sense, not very accurately at all. Indeed, any single person’s destiny is beyond the capability of all the tests ever invented to predict with much accuracy.
“Q. So, if IQ isn’t all that accurate for making predictions about an individual, why even think of using it to compare groups, which are much more complicated?
“A. That sounds sensible, but it’s exactly backwards. The larger the sample size, the more the statistical noise washes out.
“Q. How can that be?
“A. If Adam and Zach take an IQ test and Adam outscores Zach by 15 points, it’s far from impossible that Zach actually has the higher “true” IQ. A hundred random perturbations could have thrown the results off. Maybe if they took the test a dozen times, Zach just might average higher than Adam.
“But for comparing the averages of large groups of people, the chance of error becomes vanishingly small. For example, the largest meta-analysis of American ethnic differences in IQ, Philip L. Roth’s 2001 survey, [Ethnic group differences in cognitive ability in employment and educational settings: a meta-analysis, Personnel Psychology 54, 297–330] aggregated 105 studies of 6,246,729 individuals. That’s what you call a decent sample size.
“Q. So, you’re saying that IQ testing can tell us more about group differences than about individual differences?
“A. If the sample sizes are big enough and all else is equal, a higher IQ group will virtually always outperform a lower IQ group on any behavioral metric.
“One of the very few positive traits not correlated with IQ is musical rhythm—which is a reason high IQ rock stars like Mick Jagger, Pete Townshend, and David Bowie tell Drummer Jokes.”
The consequences of the Saletan / Sailer discourse have been sterile at best, and toxic at worst, substantially because the Saletan / Sailer writings to date (to my knowledge) have shown scant historical, anthropological, or philosophical knowledge of real-world social weighting practices, and no appreciation at all of the toxic sequelae of IQ-obsessed Cluster B rationalism for those practices.
Oh, in academic circles the prevailing opinion about that series and the players involved is definitely “toxic.” The problem is, it’s not that parties have been ignorant about the social implications of the debate; there is a political agenda to foster and magnify social disparities, almost exclusively on the basis of an ethnic covariate.
This is less surprising now than when Saletan’s series appeared, more’s the pity. It’s difficult to be sympathetic to eugenicists when they are without exception inveterate racists.
> but actually a 1070 on the SAT corresponds to an IQ of 114
Is that SAT-to-IQ site accurate? It shows a 1070 at the 82% but the college board lists a 1070
at the 53% for test takers and 61% for the “Nationally Representative Sample”.
Thanks, I think I was using old data. Fixed.
Thanks for posting this. When I first started going to rationalist meetups, I did have somewhat of an inferiority complex / impostor syndrome because the average IQ on LW/SSC is always reported to be like 2 billion, and I am but a mere mortal. But I’ve found that I can hold my own in discussing topics with most members, and even seemingly grasp concepts from readings quicker/better than them from time to time.
Good post.
I think it’s fair to say Feynman once got 124 on an IQ test, but I think saying “Richard Feynman’s IQ was 124” gives a really misleading picture of both Feynman and IQ testing. He is widely acknowledged by other brilliant minds as one of the most dazzling of his generation. That 124 should come with an big-ass asterisk like a home run record in the steroid era.
Apparently writing good test questions is hard and it’s easy to wind up with a question that the smart students in the class get ‘wrong because they see something in it the exam-writer didn’t. That could have happened with Feynman.
Was tested as a seven year old.
“Is it snowing today?”
I know it isn’t. Here. But I know the world has vastly varying climate, and it must snow somewhere! And yet… I don’t know that, do I?
Answer: “I don’t know”
There were a few more like that that they talked to my parents about.
I had a chemistry teacher in high school who kept giving us tests with questions suffering from that sort of fault. I don’t recall any of the exact questions but it took him the better part of a year to stop having arguments with us saying “Well that might be literally true but that’s not what I was asking.”
Your are in good company.
That’s a great story.
I do have one question about it:
What is he referring to here? A quick google search turns up nothing. I get that the one hand up one hand down thing is not how most people would move when asked to put out their hands, but the ‘bloodsucker’ thing seems to imply there’s a second part to the ‘trick’ that he never got to pull off, and I’m curious what it is.
I’m not sure. Maybe some variation of the way to spoil the “right is where the thumb is left” silly advice?
The bloodsucking probaly refers to the blood sampling station in the line of diagnostic booths the draftees have to go through.
The link had only the middle part of the story: “Uncle Sam Doesn’t Need You!”, in: “Surely You’re Joking, Mr. Feynman!”
This is something that bothers me to no end. I’ve seen IQ test items in the psychological literature that for all I can see have two logically possible solutions, one of which is probably the intended one. I’m pretty confident in that diagnosis because example items in scientific papers are unlikely to be so hard that I couldn’t solve them, and because one of the logically possible solutions is usually more subtle.
But still, I found this shocking. Are IQ tests meant to, among other things, test how good you are at predicting what the test writer intends? I really do wonder how prevalent this sort of issue is.
The 5lb Book of GRE Practice Problems has an entire section with hundreds of these, entitled (ironically) “Reading Comprehension”.
The irony is that despite getting a nearly maxed-out score on my Verbal GRE, I can never seen to answer more than 2/3 of those practice problems correctly. The ones I fail at basically always have two answers out of the four that seem very plausibly correct, and try as I might, I cannot figure out how to guess which one. Other people have looked at it over my shoulder, and can’t seem to guess them either.
Not all sources of “practice problems” are reliable. The GRE used to have an “analytical ability” section, which seemed to consist of logic problems of the form “which conclusions follow from this given information.” On the practice exam I took before my actual GRE, it was always possible to use the information given (about who was older than whom, or which house was next to which other house, or whatever) to construct a complete description of the state of affairs, and use that complete description to derive the correct answers to all the questions. On the actual GRE I took, that was never possible (maybe it was in earlier years? Or maybe the practice questions were all rejects). There was enough information to answer all the specific questions asked, of course (and I was able to; that part of the test seemed to be almost designed to inflate philosophers’ scores, so I don’t know that I approve of its removal in 2002) but always left some of the details that weren’t asked about ambiguous.
I bought that particular practice book because reviews promised that it was more difficult than the actual exam.
That definitely happened with me. I remember the school gave me some kind of IQ test in kindergarten and I have very distinct memories of it. I don’t remember all the questions, but I do remember one part of it extremely clearly, at the end:
the tester held up these cards that had a cartoon drawing on them. It would be something like a teddy bear with no head. And then they’d ask me what was missing. I remember it being really easy.
Then they got to the last question: I think everything before was teddy bears or some animals or something. this one is a pen line drawing of a classical ‘house’ (slanted roof, square windows divided into quadrants, some bushes and trees up front, a path leading to the front door, etc. When asked what was missing, I offered tons of options, (e.g. the sun, a garage, (which were not depicted), etc.) eventually the examiner decided I’d tried enough times and wasn’t going to get it. Curious I asked what the answer was, but she refused to tell me, which annoyed the hell out of me (which is probably what caused me to remember the whole thing). I’d offered tons of plausible ideas, and even some that were a stretch.
I’d remembered it and told the story on and off for the rest of my life, still always a little annoyed that I wasn’t told the answer: I was sure that I’d gotten it ‘wrong’ for a dumb reason like that.
20 years later, I was telling the story to my best friend while eating in a cafe. I got to the point where I was describing the drawing, and suddenly I stopped and my jaw hit the floor. I’D FIGURED IT OUT! 20 YEARS LATER! (at least I think so): all the other drawings had the lines filled in with colors, like animation cels, while the house one was just the pen lines. So I’m pretty sure the answer was supposed to be “color”. Which I didn’t say at the time (even though i noticed it wasn’t there) because you woudn’t say a drawing isn’t complete because the artist chose not to use color. there are plenty of examples of drawings that have no color that are considered complete works.
It’s kind of funny too because later in life I became really into drawing and animation, and I don’t use any color in any of my works, because I don’t think it’s that important to what I’m trying to convey. To me it’s all about structure. It’s funny that I had those same proclivities at age 5.
>He is widely acknowledged by other brilliant minds as one of the most dazzling of his generation.
Due to the Flynn effect, shouldn’t that mean is IQ in modern terms would only be 103?
IQ isn’t intertemporally valid. Scores have risen precipitously without a corresponding increase of g. It remains, as far as we can tell, co-temporally valid. Which is nice. But don’t try and put too much weight on that branch. Focus on g, and understand that there’s nothing particularly special about any test that is loaded on g.
Since we can’t measure g directly, how do we know this?
Some theories I’ve heard hold that the Flynn effect is much more prominent on tests that are loaded for fluid intelligence, a measure of abstract reasoning and problem solving that is postulated to be one of the two components of g.
This is wrong. g is just the first factor component in the correlation matrix of the results for a variety of cognitive tests.
If the population’s IQ has changed over time, g will change also.
You seem to be referring to the idea that some proportion of the variance in IQ is attributable to genetic, rather than environmental factors; and this (we assume) has not changed significantly in the last 100 years, whereas IQ famously has. But this is not the same thing as g.
cor·re·spond·ing kôrəˈspändiNG,ˌkärəˈspändiNG/ adjective
1. analogous or equivalent in character, form, or function; comparable. “we discussed our corresponding viewpoints”
Sorry ilkarnal. What are you actually trying to claim here?
g is the first principal component of a factor analysis of the correlation matrix of a bunch of different tests of cognitive ability. Are you trying to claim that the various types of IQ test are inferior to other tests of cognitive ability, because IQ tests exhibit the Flynn effect? This seems decidedly odd, if true; because it raises the question of why psychologists wouldn’t update their IQ tests to include these other, more intertemporally valid cognitive tests.
I suspect Feynman’s cognitive profile was a little different than that of his peers. Feynman’s Regular Guy act was something of an act, but not wholly. Most of the other great physicists had highbrow tastes, but Feynman really was a regular guy — a very smart, original Regular Guy.
I don’t. At least, not unusually unusual. His peers weren’t the most normal people.
Most comparably famous physicists had several highbrow predilections, such as listening to baroque music or reading philosophy. Feynman didn’t. Was he intentionally burying a part of himself under his populist act, or was that just the way he was?
What is more indicative of cognitive abnormality – liking bongos more than violins, or playing around with a barely sub-critical hunk of plutonium one screwdriver slip away from death? How about handing secrets over to the Soviets?
I wouldn’t describe Feynman’s tastes – his art, his music – as ‘populist’ exactly. Idiosyncratic, sure.
It wasn’t that he was turned off by the idea of philosophy so much as he thought the philosophers he ran into were snake-oil salesmen.
Here’s Gell-Mann talking about Feynman:
https://www.youtube.com/watch?v=rnMsgxIIQEE
As idiosyncratic as Feynman’s behavior was, he didn’t really go far beyond the pale. Yes, he was a bit of a play-actor. But he wasn’t exactly play-acting a ‘populist’ or someone who was just ‘normal.’
Being someone who finds highbrow culture much less fun than (admittedly nerdy) lowbrow culture, I can pretty easily believe it was just the way he was.
On the other hand, there’s an urban legend that members of heavy metal bands are actually tilted towards geniuses who gave up being physicists because they enjoyed headlining a band more.
A number of rock stars have gone on to get Ph.D.s in technical subjects: Brian May of Queen, Phil Alvin of the Blasters, and Greg Gaffin of Bad Religion. Dexter Holland of The Offspring hasn’t finished his, last I heard.
Steve Hsu on Feynman’s IQ:
You guys are reading a lot into a vague anecdote about an IQ score. For all we know, 125 was the maximum score on the test, and the test was only intended for the lower IQ ranges.
IQ measures general intelligence. This is because every intellectual activity that can be scored measures general intelligence. That’s kind of inherent in the concept of general intelligence.
Per Jensen, g-loading of IQ is similar to g-loading of vocabulary tests.
What is special about IQ is that it is acultural, you can’t give the same vocabulary test to a Chinese person and an African and a Scotsman, but you can give the same Raven’s Progressive Matrices problems to all three and have direct comparison. Also, the acultural nature of IQ gives it some protection from accusations of ‘cultural bias’ that inevitably issue forth from the left.
There are reasons to be skeptical of IQ, relative to other highly g-loaded tests. It isn’t intertemporally valid, for some reason. That’s enough to make you worry.
There is absolutely no surprise that we’re seeing lots of people with IQ underestimating their g-factor, considering the g-loading of IQ you’re going to have a lot of people in say the top or bottom 5% that have scores that are really not very representative.
This doesn’t mean that IQ tests don’t matter for individuals. The further you are from the average of your chosen arena of competition, the less likely you are to be able to hack it.
Not much below average, according to that table of yours. Lotta those occupations have no grey below 90 IQ. Also, becoming “scientists, professors, engineers, and almost anything else you could hope for” is very very very different than being SUCCESSFUL scientists, professors, engineers &c.
“Desperately try to hold on to their dream of being an intellectual” being “an intellectual,” I would suppose, involves more than being able to plausibly and unproductively blend in. What’s the IQ average of those who actually move things forward? Probably that would be a grey bar substantially further to the right than any on that chart.
Plausibly. However, I think a good part of moving things forward is simply having sufficient nonconformism to come up with your own way of looking at things. Folks have argued that North European cultures punch above their weight in terms of important discoveries, and if so I suspect this is a result of North European cultures being highly individualistic. (This article has some great anecdotes that provide perspective on European individualism.)
> [IQ] isn’t intertemporally valid
You have claimed this twice now in this thread. Why do you think so? I think that’s false.
http://scottbarrykaufman.com/wp-content/uploads/2013/12/Armstrong-Woodley-2014.pdf
https://imgur.com/a/kL32P
Arithmetic gains per decade are a tenth those of Raven’s Progressive Matrices. Not intertemporally valid. Something’s wrong.
Flynn’s What Is Intelligence? covers this stuff.
You are mistaken. A typical IQ test is not just Ravens. A full-spectrum IQ test like WAIS includes extensive verbal testing.
Therefore you can’t give the same test to people from different countries/cultures. You have to translate them. And international IQ comparisons are consequently difficult.
IQ tests are more g-loaded than a vocabulary test, because otherwise psychometricians would just use a vocab test. Vocab tests are part of the WAIS package.
https://en.wikipedia.org/wiki/Wechsler_Adult_Intelligence_Scale
How can you say it’s acultural when you know it’s not intertemporaliy valid? It’s the different culture between now and 50 years ago that caused the +1SD increase in IQ.
By all indications and almost all the tests I ever took my IQ is between 4 and 5 SDs above the mean, except for one test I took in 2nd grade where I scored 133, and that’s the score my school district gave to my teachers at the beginning of each new school year. So measurement error can be quite large, especially among children.
I knew Feynman. He was obviously smarter than me, in his case the IQ measurement was off by at least 3 SDs. (John Conway and a couple of other people I know personally are also smarter than me, but I won’t estimate this for anyone I don’t know personally. And when I say “smarter than me”, I mean “at least as smart as me in every way, and smarter than me in some ways”. Feynman excelled at every kind of thinking, there was nothing narrow about him.)
This strikes me as an obvious fallacy. Imagine you knew a very smart person who had a merely average SAT score, would you conclude that the SAT was wrong? That kind of thinking seems to me to be wrong headed. After all what is your SAT score, other than whatever score you got on the SAT?
Since the only reasonable definition of IQ is the “the score one receives on an IQ test”, I don’t see how, if it was constantly tested at a given level (and in this case that is a big if), it could be wrong, even in principle. You do mention childhood tests as being unreliable, but I was under the impression that the score came from a test he took as an adult.
What you should instead say is that in Feynman’s case, and perhaps in others, IQ was unrepresentative of his overall intellectual abilities; which I think is the point that people telling the 124 anecdote are trying to get across.
I suspect that the truth is that Feynman’s intellectual talents were not evenly distributed. He may have had an incredible mathematical ability, and merely above average verbal ability.
I was under the impression that he fabricated the story to undermine IQ tests.
Or, at minimum, Feynman was a great storyteller and his only-moderately-high-IQ anecdote has made for a great story, considering how many millions of times it has been repeated.
Feynman might wind up in American folklore in a position kind of like Davey Crockett, whom many people don’t realize was a real person.
It’s possible.
It’s kind of hard to track down an original source. The closest I’ve got is a reveiw of “Surly you’re Joking, Mr Feynman!” in people magazine from 1985.
So I revise my impression from adult to high school student. If somebody here has a copy of the book, can you look through it and find the relevant portion?
I’ve also heard that William Shockley, James Watson, and Luis Alvarez had similarly gifted, but not extraordinary, sub 130 scores. At least in Shockley, and Watson’s cases this would be a lot less surprising than Feynman, since their work work was basically experimental, and demanded a lot less in terms of abstract mathematical reasoning.
It is well-documented that Shockley and Alvarez were tested at age ~11 by Termin for his Termites in 1921-1922 and both failed the 130 cutoff. I think both by 1 point, but I’m not sure.
Watson entered college at age 15, so I would be surprised if he scored <130 at, say, 11.
Added: I don’t see anything in Feynman’s book. Faber’s review does kind of imply that she got it from the horse’s mouth, not the book. It’s weird that people cite Faber in Reader’s Digest rather than People.
Actually, Termin’s threshold was 135 (ie 1%), so Shockley and Alvarez could have exceeded 130.
You need to read more carefully. You missed the part where I said I knew Feynman personally. I specifically stated that his intellect was not narrow, that he was good at every kind of thinking. Also, 124 was a single test result, and I had just made the point that individual IQ tests can have a pretty high measurement error.
No, I got you on both points. Though I may have been a little unclear in how I phrased my post, I did qualify my statement about the innate validity of IQ tests.
After a little more digging it seems that the anecdote came from testing that was done during high school. It’s not clear, but I’m guessing that means it was only one test, so make of it what you will. I really have no idea what the confidence interval on a 1930s era IQ test looked like, if you do, please contribute what you know to the conversation.
On the Internet, nobody knows your a dog, so I have absolutely no way of verifying that you knew Feynman; and for that matter no way of estimating the accuracy of your evaluation of his intelligence.
People are subject to cognitive biases, and often attribute positive traits to people who they hold in high esteem for unrelated reasons (E.g. believing that physically attractive people are more intelligent). It’s possible that you were so awed by Feynman’s mathematical abilities that you overestimated his verbal intelligence.
If we don’t think that Feynman was lying about his score, and he may well have been, then we are left with two possibilities. Either someone with a ”merely” gifted IQ did Nobel prize worthy work in theoretical physics, or there was a massive measurement error.
On the one hand Feynman was obviously a genius, his work on quantum electrodynamics proves that better than any test ever could. On the other hand, we have at least one test that seems not to have captured the magnitude of that genius. This was either due to systemic limitations of the test, for instance it might have been very heavily weighted for a type of intelligence in which Feynman was merely above average, or it may have been due to some kind of random error. If three standard deviations worth of error was common on whatever test he took, then that test wasn’t worth very much, and It seems like somebody would have noticed that.
Of course, some combination of both factors could have been at work. Perhaps there was only one standard deviation worth of measurement error, and the test was heavily weighted for verbal IQ. In that case if he took the same test again we might have seen that his IQ was around 140, and was still unrepresentative of his intellectual abilities.
Above you say that your IQ is between four, and five standard deviations above the mean. If that’s true, and the test is generally a fair representation of intellectual ability, then you are somewhere between being on par with one of the top few thousand, and one of the top hundred minds in the country. If that is the case, then it’s likely we could be talking about far more interesting things than psychometrics.
What do you do for a living?
If you knew Feynman, and Conway, then I take it you are a mathematician, or physicist, or perhaps some other sort of natural scientist. What is your primary area of research?
I know a couple of guys who knew Feynman. And they are both really smart.
I don’t see anything terribly implausible about some extremely smart person who knew Feynman commenting on this blog. On the other hand, without giving up your anonymity it’s hard to prove you’re not a dog on the Internet …
@Steve Sailer
You are of course right. There are between a hundred, and several thousand people in this country alone who have the abilities that Polymath is claiming for himself. They must do something with their spare time, and the SSC comment section seems to be as likely a place as any to find them as any.
I would hope that nothing I’ve said would be taken to imply that I disbelieve him, per se. It’s just that if what he says is true then there are probably things he could contribute that
would greatly enrich the SSC discourse.
I personally don’t get a lot of opportunity to talk to world class minds who knew Richard Feynman, and I bet it would be a lot more interesting than the average conversation around here.
I don’t know whether you are using a standard bell curve to estimate the number of people in this bracket, but you should be aware that the actual distribution of IQ scores in the population is fat-tailed; there are quite a lot more people above 160 than a purely statistical projection using an ordinary normal distribution would predict.
@bbartlog
Related to my post below, I don’t see how this is possible. It was my understanding that the normal curve, along with the standard deviation of 15, was built in to the definition of IQ under the newer formulation (i.e. after it stopped purportedly being about mental age).
In other words when the raw score to IQ calculations are being made they are supposed to map the raw scores a normal curve with a SD of 15. So a deviation from that would mean the test designers screwed up rather than something about the population.
If he claims something unlikely (that would have been considered significant in advance) and which there is a plausible reason to lie about, you should update in the direction of not believing him. (although you may still end up thinking it’s more likely than not that he’s telling the truth.)
I don’t understand how this is possible.
4SD above the mean is about 1 in 25,000 and 5 SD is about 1 in 3.3 million. I can’t imagine any exam format norming study has included 25,000 adults, much less 3.3 million 12 year olds. And even those numbers wouldn’t be high enough to have any confidence in the results that far out.
Okay, so you extend the norms by over-weighting the number of extremely bright people in the pool. What does that do for you, really? It isn’t as if you can use some other test to determine the true IQs of your extended norm sample and use that to calibrate your new test, unless the other test itself had been normed with some fantastically large pool of people.
Seeing something like 5sd immediately sets off alarm bells in my head, and I would think it would do likewise for other people familial with statistics. But maybe I’m missing something.
All the simpler tests that maxed out at 150 or so, I maxed out on, except that one 2nd grade test. Later I took adult tests and scored in the high 160’s.I also got 2400 on the GRE back in the 80s when it was really hard. I knew Feynman because his best friend, Ed Fredkin, was one of my professors, and because I worked for a time as a researcher at the MIT Lab for computer science in Fredkin’s group that studied the same kinds of theory that Feynman was working on over at Danny Hillis’s company down the block.
I have been a mathematical consultant or software developer for thirty years or so, have a math Ph.D., an Erdos number of 2, have won several awards for research and writing, taught on the side a lot, and know a few people who are smarter than I am.
My dad’s IQ was measured at 171 back when he was at Bronx Science, which he graduated from 2 years early, and I’m way better at math than he was (he was an attorney, my mother is a physician).
Okay. You’re a really bright guy that knows a lot about math. Those weren’t the questions at hand but perhaps it’ll mean you know the answer to the one that is: Viz. how can those tests forms possibly be giving out numbers in the 160s or 171 given the inherent difficulties in norming that far out?
I think the New York City school system had data on millions of children over the years, including hundreds of thousands of the high IQ subpopulation of Ashkenazi Jews (which my dad was), they could certainly try to give a test that went up to 170 or so.
I agree that the numbers probably aren’t very precise or accurate, and also that there are fat tails due to combining populations with different means. But the GRE is taken by a large enough group of smart people to give pretty good calibration, and even the new GRE was considered by high-IQ societies to have a top of 4 sigma, while the old one that I took had a lot fewer perfect scores and therefore a higher cutoff at the top end.
I think, if you have never done a test where the difficulty of the tasks extended into a range that was clearly too difficult for you, you only know a rough estimate of your IQ and not necessarily an underestimate.
In my experience the difficulty of IQ test tasks usually maxes out somewhere around 130, and where you land in the extrapolated range beyond that depends more on your motivation and your ability to concentrate than on your fundamental limit of understanding.
I think my IQ is realistically around 135 but I have scored above 170 in some tests. I can always tell when I will overscore, because the tasks are kinda hard but still doable. Then, when I concentrate hard enough on doing these doable tasks, the extrapolation takes me to Neverland.
My guess is, that fat tails usually come from extrapolation. Norming a test means squashing the actual distribution into a gaussian shape. If you do it correctly, you will by definition not have a fat tail, even for a multimodal actual distribution.
I don’t see what nyc “having data” on millions of children over the years does for them. Unless the claim is that 1) the same test form was in use over those millions and 2) they were continualy carefully collecting raw score information beyond the then calibration of test and 2b) the questions scaled enough to have meaningful raw scores into the 1 in millions and 3) some statistical expert went back over the carefully collected data to extend the calibration of the test.
And all this assumes they were going for a deviation IQ in the first place, which give how long ago this was may not well be true.
Exactly here is the problem with trying to have discussions about IQ. People’s ego are so tightly tied to these numbers that they are willing to swallow any old snake oil that tells them what they want to hear. Even if they are smart and statistically knowledgable enough to know better.
There are clearly people on this planet who are 1 in a million with respect to intellegence. And 1 in a billion. And there’s a smartest person. There just aren’t any standardized tests capable of identifying these folks. Absent truly enormous expense there can’t be. Which is perfectly okay! This little fact of the matter need not be taken as an insult or an attack on anyone’s intellegence. But for some reason it too often is.
And, just like that, Scott eased the burden of existence for 10,000 nerds with imposter syndrome.
Meanwhile, I have a very high IQ, I’m very good at taking tests, and I’m lousy at things that aren’t like taking tests. So when I consider my life, admittedly impressive for an average citizen but a severe disappointment compared to how I was doing calculus at 8 and I’m supposed to be off curing cancer or proving the Goldbach conjecture by now, I think I’m going to keep clinging to the delusion that IQ “just shows how well you do on a test” a little longer.
Do you mind saying a little more about your experience, and what particularly you think holds you back? Like, what are the things you think you’re lousy at?
I ask both out of curiosity, and in the hope that people here might be able to say things that you would find helpful, either personally or psychologically.
(I suspect a lot of SSC readers might fit into the “high-IQ underachievers” category; maybe that should be a topic unto itself.)
From my perspective if you’re in a school system that doesn’t track well you wind up with massively underdeveloped work ethic and social skills since your raw intelligence is adequate for most pre-puberty challenges where normal kids have more incentives to work hard and make friends. Usually you realize how fucked you are socially when puberty hits and it’s a crapshoot how much you can repair it in high school and college (college is easier since it’s a fresh start so you see a lot of catching up there). Work ethic can take much longer to catch up with you, depending on what path you take academically and variable teacher/professor emphasis on exams vs busy work, but since most jobs, even high-g ones, are mostly busy work, it’ll be obvious by the time you have one full-time.
Yeah, sounds like me. Now, what I said above was no exaggeration – I literally was doing calculus at age 8, and I haven’t the slightest idea how an exceptional case like me could be “tracked well.” In retrospect, I think I’d have been better off holding myself back, staying with my grade level, and developing socially like a normal human, instead of jumping around among schools and tutors to nurture my “genius” while leaving me emotionally stunted.
Long story short, my social skills are still very weak, my work ethic is decent if I can stay focused, but I’m easily distracted if the work isn’t engaging enough, which it often isn’t. My main issue, I think, is that I lack creativity. Set me on a task with a clear answer and I’ll find it, ask me to come up with something myself and I’m stumped.
What makes you think that would have worked? I was in the normal public schools from 2nd grade on and it was hell which did nothing for my social skills except teach me how to be the paraiah.
@BBA:
I hope you don’t mind if I ask a few further questions:
Are you on the Autism/Aspergers spectrum?
Do you have particular unusual mental abilities (e.g. memory, calculation)?
Are your abilities focused in a particular area (e.g. math) at the expense of others?
Do you have difficulty handling verbal ambiguity? Or, what is difficult about interacting with people?
What are your parents like? Siblings?
Did you participate in math competitions or similar activities?
What do you do now? Do you work in math or programming?
It’s late and I don’t want to delve too much into my personal life, so I’m not going to answer all of those questions.
I’ve never been diagnosed with an autism spectrum disorder. My younger brother has, and he shares a lot of my personality quirks, only his are much more severe and interfere with his daily functioning, while mine just make me obsessive and anti-social. So I consider myself a low-functioning neurotypical.
I got a 1570 on my SATs. I’m better at math, but my language skills are perfectly up to snuff. I’ve always hated writing but I’ve also always been told I’m a good writer. Must be doing something right.
I just find social situations exhausting, and hate the feeling of being “trapped” at a noisy, crowded gathering of some sort when I’d much rather be in my room, alone, reading. I’m okay at one-on-one conversations but with more than a few people I find myself listening more than talking and rarely able to get a word in edgewise, which usually suits me.
@Nybbler: yeah, I probably would have been miserable no matter what, but never having the chance to make friends as a youngster has made it so much harder now that I don’t even know where to begin.
Seconded, alas.
Thirded. I think the primary reason I haven’t lived up to all that “99th percentile” grade school hype is that I was really bad at going forth and doing tasks that were not laid out for me by authority figures. Getting a good job in particular, which is to say the combination of choosing a field, seeking out appropriate jobs in that field, and applying for those jobs in such a way that I actually got called for interviews, was existentially terrifying to me as a new college graduate, in much the same way that being told to plan and then embark upon a solo trip backpacking across Europe might be existentially terrifying to someone whose prior travels in life always involved being shepherded around by others.
I was really bad at going forth and doing tasks that were not laid out for me by authority figures.
This sounds familiar to me. Fortunately I’m getting better at this.
Yeah. I think there’s a certain brand of high-IQ nerd that wasn’t properly socialized by our society, since standard socialization methods were such a mismatch for them, and thus missed out on essential character development.
Fortunately, I think it’s possible to bootstrap a lot of that yourself, though it’s a bit painful.
I’m actually starting to think about this from the perspective of a father, rather than a young adult reflecting on my own childhood. If my son turns out to be as much of a nerd as early indications (and his lineage) suggest, I’ll need to find a good way to avoid these pitfalls.
I’m thinking boy scouts and martial arts. Anyone have other ideas? Science-focused summer camps?
Martial arts is a really, really good idea – but it’s important that he want to do it at least as much as you push for it. Easy motivation hack: Expose him to Teenage Mutant Ninja Turtles early.
Science-focused summer camps also good. I would expand that to: help him find IQ peers as soon as possible. Center for Talented Youth (CTY) might have a presence where you are.
I was a Boy Scout. Also probably a good idea, how good depends on the quality of your local leadership.
Boarding school for highly intelligent kids? Social development would be less distorted by peers being more or less distant, IQ-wise. Less nerding-out of socially demanding situations, less jealousy or punishment for being smart, better nourishment of talents from teachers being aware of them.
I only acquired the ability to actually pursue a goal like this at age 39, and only by reading countless guides that broke down how to do every single step from finding leads to preparing questions for the interviewer. I’m not sure what’s different about 21-year-olds who just go out and do the interviews. Maybe they’re used to life being difficult while I was lazy, maybe they’re more used to playing out social scripts while I expected to be me without changing, maybe they don’t react to unsure situations with executive dysfunction. Maybe they have friends and family that encourage them and normalize the process. Maybe they have less complacency that things will work out for them if they don’t make them happen (I never found a mate, either).
At any rate, I spent about ten years driving forklift with my BS in Math and 1480 SAT, then got forced to take a much tougher job in the same factory where I had to adapt to survive. After several years of being forced to develop maturity in this way I started playing D&D, and applied the strategies from work to prepping at home. After that I was able to get into the field of computer programming via self-study and showing up for interviews.
I don’t know if I could ever manage to teach my college self how to be that responsible, or what makes other people able to do it. I had to learn so much about business people’s perspectives by reading their writing on blogs and Reddit, I had to develop so many time management skills. To me it seems like there should be a lot of people like me who are easily able to perform intellectual work but just don’t get invited to by the system.
My brother was completely incapable of getting a job after high school. My mom would pick up applications for him, hold his hand through filling them out, but somehow he never returned them. He had the same problem in college — he had the brains, but he didn’t have the executive function, and without our parents to help, he flunked out. My parents finally put their feet down, afraid he’d never leave home, and said if he didn’t get a job, he had to join the Navy. So … that’s what he did. In the military, *they* try to win *you* over, at least if your ASVAB scores are good enough. (His were good enough to get his choice of jobs.)
Every time his term in the military runs out, he talks ahead of time about leaving the Navy, maybe studying computer science …. and then the clock runs out and he just reenlists because he hasn’t made alternative plans. He’ll probably make a career of it, although it isn’t his ideal job.
Autism runs in our family and I’m fairly certain he has it. Maybe it’s something to look into. Executive dysfunction is a really big sign. I have it to a lesser degree; it takes me a week to make a doctor appointment. Luckily I’m married to an organized person so actually paying the bills isn’t my job.
Remember: a smart person can get an almost arbitrarily low score if he doesn’t care, or deliberately blows the test.
Bill Cosby apparently scored below 500 on his SAT, at a time when you got 400 just for writing your name on the test. Obviously he isn’t in the first percentile intelligence-wise.
Source. This article is about gender differences, and it also has an interesting video midway through about how men learn to let things roll off of them through playing sports and getting casually insulted when they are kids. This fits with an analysis I saw indicating that people who play sports while they are students make more money (sorry I can’t find the link, I swear it was on halfsigma.typepad.com or somewhere like that). And it is giving me a great justification for all of the times I teased my little sister when she was growing up 😛
Here’s the original study; the experiment described in that paragraph is Experiment 4 in the paper, and its results are summarized in Table 1.
“Dramatically” in this case appears to mean “by about a third of a standard deviation”. And it’s significant at “p = 0.05” (not “<", notably).
On the one hand, I want to make an IQ test. On the other, I’m kinda terrified about getting a low score. I think it’s less about finding out my life is screwed rather than personal pride but even if my life was screwed if my IQ was below a certain benchmark, I don’t think I would need to worry less about it since I usually score pretty well on IQ-test-like tests.
I like this post, I think it’s a good reminder of an important principle although for those without any grounding in statistics I think it may not make things too much clearer; but there are surely plenty of places that offer good introductions to statistics to which you could point such people. So none of that is why I’m commenting.
I’m commenting because I was triggered by something said offhand, and I’m just going to try to keep this brief. I don’t think that someone who is fairly familiar with it would consider music to be a subject which can occupy the intellect at a capacity below the average of various other scientific pursuits. I think that there is certainly an anti-intellectual mythology around music; and even more than those who create it, many who critique it believe that a frankly offensive ignorance of its workings is not only acceptable but may be considered an advantage. It is troubling not only that the view is held by some in “authoritative” positions, but also that this represents a sort of baseline view in our culture.
Again, the remark was offhand and simply served a shallow illustrative example in the piece, the purpose of which example was clear. But I will push at any turn against the notion that music is a field in which the intellect is best left out.
Arthur Jensen mentioned musical rhythm as the one talent that didn’t seem to have any correlation, positive or negative, with IQ.
That might be why Mick Jagger, David Bowie, and Pete Townshend all tell Drummer Jokes.
http://www.musicradar.com/news/drums/the-23-best-drummer-jokes-ever-169967
On the other hand, the careers of Jagger, Bowie, and Townshend all suggest that having a few dozen IQ points above average can be useful even in the Rock Star business.
FWIW, I have both a high IQ (well above the nominal genius cutoff) and a strong musical talent, including for rhythm. That makes the fact that I agree with Jensen probably of some interest. These are different things.
However, I also notice that other aspects of musical talent seem to recruit the same pattern-matching ability I use for demanding cognitive tasks like writing software, and thus are probably heavily g-loaded. The “rhythm” part of “musical rhythm” is an important qualifier in Jensen’s negative statement.
It is common folklore among both programmers and mathematicians that music talent is strongly correlated with talent in their fields. This matches my experience
which I have never had any reason to do
What would you say rhythm is, if it isn’t pattern-matching?
Well, in a sense it is. But Jensen’s report, and my experience, suggests it might be handled by a different neural subsystem than other kinds of musical talent. One pergaps less tid to whatever Spearman’s g is measuring.
One of my favourite musician putdowns is that they are so primitive they only know the alphabet up to G (although they can say it forwards and backwards), and can only count to four; and drummers don’t even have the alphabet down.
I can actually speak to the drummer jokes as a professional, though; obviously there are a lot of aspects of music that a drummer gets to just ignore, or it’s perhaps better said that while others are doing quick calculations to get the particulars of things like a chord progression, drummers think at a level of abstraction higher, just understanding the overall trajectory. They essentially read and interpret an abstract but intrinsic property of the piece called harmonic rhythm; the way that the flow of music sets up certain expectations, and how those expectations are resolved and in what time frame, which creates the division of phrases and the arc which makes those phrases, separately and as a whole, intuitively make sense to a listener. So, a drummer reinforces and modulates the music at the level where a listener projects their understanding of how the music “works”, and possibly even what it means. This is not to say that other instrumentalists do not work or think at this level, but again they simply have concerns to address in playing that a drummer does not.
In this way, I might think of drumming as displaying a more emotional intelligence. It is generally true in music that while there may be wrong answers, there is no single right answer, and this is even more the case in drumming; a drummer hears what everyone else is playing, extrapolates to how they are each thinking, combines these impressions to interpret the flow and meaning of the music as a whole, and then makes specific interventions to impose and reinforce that interpretation. But the whole process is generally an unconscious one that you “feel” intuitively.
Neil Peart is a drummer and sure seems to be the smartest Rush member. I mean, he writes all the lyrics and they’re full of references to obscure literaure, science, and philosophy. Between his writing and his drumming talents, I’d be surprised if he weren’t significantly above average in intelligence.
The average person who gets a low IQ score says “Yup, guess that would explain why I’m failing all my classes”, and then goes back to beating up nerds.
I know this is meant in jest, but it seems really uncharitable to the “low IQ” crowd, which for the anxious portions of the SSC audience appears to be around mid-90s. Most of these people are the perfectly productive, reasonably articulate and curious types to whom Reader’s Digest Classic Editions were targeted in decades past. Statements like these suggest that a bit of ingroup/outgroup thinking has crept into our conception of the meaningfulness of the measure. De-individualization and “the soft bigotry of low expectations” (sorry) are among the concerns that I’ve seen the online IQ-fretters express.
I don’t think the IQ-anxious are limited to SSC or the rationalist community, or even the broader nerdy proto-intellectual community. They turn up in fora (both intellectual-leaning and basely popular), clever-sounding and funny and concerned that they or their kids may be at a disadvantage. They may not worry about their ability to contribute an oriel window to the edifice of human knowledge, but they are concerned about their economic and social security, or their ability to be taken seriously as a leader in a hobby community.
Additionally, since this note of reassurance was meant to resonate on the individual level, I’m not sure I would call “emotional intelligence” silly, at least in its popular understanding. Being able to fluidly navigate primate social hierarchies and intuit the desires of others is a useful skill in many contexts. If Dale Carnegie and “game” practitioners are correct, performance on these measures can be improved with practice, unlocking a great deal of value for the individual.
Yeah, I don’t get why Scott is sh*ting on emotional intelligence. It may be kind of hard to measure quantitatively, but I suspect a test for the ability to read people, understand their emotional states, and model their mental process, seems to me like it would be very predictive of future income, at least for people with a certain minimum IQ. If you look at the leaders of fortune five hundred companies, I don’t think what sets them apart is their spatial reasoning skills.
And they’re right to express those concerns! Almost nobody in “the system” actually pushes you as hard as your IQ actually allows you to be pushed. They instead take the IQ number as a signal of, “This is how well you can do without any particular effort on our part or your part”, and let you slouch along. You can almost always go further and do more in intellectual subjects than the people giving you an IQ test say you can.
It’s important to put in the effort, because the people getting extraordinary results aren’t coasting on their IQ’s. Coasting on your high IQ leads to Brilliant But Lazy/Gifted Underachiever Syndrome. Great results come only from having a decent IQ and then working hard as hell.
the anxious portions of the SSC audience appears to be around mid-90s
I dunno, the IQ-anxious strike me as “My score says I’m 110/120, I expected/hoped to be 130/140/150, which is less correct – the test or my expectations?” rather than “Oh god oh god oh god I couldn’t even manage to get the dumb normed-to-be 100 score???? I’m a sub-human!”
*proudly waves 99 score on that Ravens Matrices test: yay for sub-humanity!* 🙂
Wait, you’re average-to-slightly-dumb? I’d always pinned you as really clever.
Excuse you, I are very dumb! 🙂
Mathematically, that is. Pattern matching, spatial awareness/manipulation, the rest of it – at times I still go “It’s over here on the right” *points to the left* “Ah ha ha, other right, I meant!” So any test that relies on “what pattern comes after this jumble of lines and dots?” is going to have me taking off my shoes to count on my toes ‘cos I don’t have enough fingers to count that high.
Words, though, I’ve never had any trouble with them. Which is really fun all through primary school when you’re bad at maths but reasonable at other subjects, so the teachers go “Well, the reason you’re doing poorly in maths can’t be because you’re stupid, else you’d be doing poorly in everything, so you must be lazy and just not working hard or trying instead!”
You’re Irish, aren’t you?
Mark, are you referring to Professor Richard “Ulster Says No!” Lynn’s work? Because I think he’s as trustworthy on this topic as a fox in the henhouse; I do believe he has a lot of political inclinations which lead him to interpret results through a particular lens and if the Wikipedia article on the criticisms of his work is correct, then he used a lot of fudging and guesstimating and “If I take this result from 1920 and extrapolate a bit” for his “list of global IQ”.
I really find it very hard to believe that people five miles on one side of an arbitrary line will have a much better average IQ score than people five lines on the other side of an arbitrary line (as you may see from this graphic, Norn Iron is lumped in with the UK’s average of 100 as against the Republic of Ireland’s average of 92 – or 96, depending on which story you read).
But faix and begorrah, Muster Mark, shure isn’t yourself right about meself being Irish, like! Now excuse me, yer honour, I must go and give the pig in the parlour a feed of stirabout, wirra wirra! 😉
This shower are probably feckin’ liars, because they told me my score was
It’s 20 questions on an untimed (supposedly 10 minute) test (I did it in around the 10 minutes but I didn’t time myself), plus they want you to pay for the result, so I’m very dubious as to the whole honesty of it all, plus they’re freely using Professor Lynn’s name to tout themselves 🙂 (According to himself, Amerikay only scores an average of IQ 98 so you lot are only slightly smarter than us dumb Paddies!)
But there – either I’m a hidden genius who miraculously got way better at pattern matching since I took that online Ravens test, or this lot are hucksters. Let the people decide! 🙂
You don’t write like most people I’ve met with 99 IQ, I can tell you that much.
But the whole point of OP is that you shouldn’t put too much weight on a single number for individual purposes, because all sorts of shit can tweak that number up or down, or make it nonindicative for the stuff you actually do, that’ll average out on the population level. This forum just has trouble accepting that because we’re mostly nerds who compare IQs like a drunken football team compares dick measurements.
I might also say that any online IQ test is only slightly more reliable than tea leaves.
Isn’t this exactly the kind of statement this article is warning against?
“Your IQ is slightly below average” does NOT mean “you will perform at an average or below average level at “. (I mean, Deiseach confirms here that she does perform below average at the first two, but we believe her because she says so, not because of her Ravens Matrices score.)
There aren’t any completely average jobs, really. Individuals should be worrying about which particular things they are good and bad at.
Intelligence doesn’t have the practical value you’d think it would. The same with wealth really. People assume they would be happy if they were super smart and super rich, and that they would have no problems – but generally speaking the complexity of problems assigned to super smart people and/or rich people scale up. Studies show the most competent people in a workplace are assigned the toughest, most important jobs. Which is actually smart when you think about it. And rich people have rich people problems, like trying to line up a cleaning service for their ski chalet in Alsace (better learn French!).
People assume that have seven cars is more fun than having none, or that being a veterinarian is better than being a dog groomer. But when we study happiness material possessions such as lots of cars don’t really correlate. And while a dog groomer and a vet both smell like dogs, only one of them has to euthanize beloved family pets and coach parents through how to grieve with their children.
Especially at the community level, wealth and IQ can solve lots of problems. But there are generous poor people with low IQs who make important contributions to the world too.
Yes, but it’s nice to live around smart people. As Scott’s post suggests, while your personal IQ may or may not do you much good, in the aggregate being around a lot of smart people is better.
For example, a general trend over the last 50 years or so is that house prices have gone up most in high IQ neighborhoods. For example, a typical 2000 square foot house in Palo Alto now costs $3 million.
Intelligence doesn’t have the practical value you’d think it would?
Depends on what you’re doing. I’m a systems programmer and for that it matters a hell of a lot. I planned to be a theoretical mathematician; it would have mattered a lot there, too. I recognize that not all fields are so IQ-loaded, of course.
I don’t know whether this counts as “practical”, but I enjoy being really bright. My universe is a tremendously more interesting place because of all the things I know and my ability to reason about them. Grokking the physics of rainbows gives them a depth of beauty and meaning they wouldn’t have as mere colors in the sky.
Mind, being in the right place at the right time matters a fuckton for systems programming, too. I loved being a firmware engineer, for instance, until they cut the money for our project!
Now I’m just the guy working on a goddamn Rails app who uses engineering practices designed for Linux kernel patch submission.
Wealth might not be a guarantee of happiness, but poverty sure comes with a lot of suffering. And as more and more jobs that don’t require much intelligence get automated, people who have a low IQ have a really good reason to be worried. What if they can’t get any job that pays a living wage? What is that going to do to their chances to live somewhere that’s not a slum, have kids, pay the doctor when they get sick?
That kind of problem probably doesn’t kick in at the 95 IQ level, but ten points lower and there’s likely good reason to worry.
I think that the jobs that are likely to be automated first are those that involve manipulation of information and language.
Translators are going to be automated before hairdressers. Clerks take money (information) and talk to people (language). Automated.
I don’t think it’s freedom from automation that will protect brainy professionals and knowledge workers – it’s social inertia. Look at universities – for undergraduates, there isn’t really any reason for them to exist, and yet they do.
Fund managers?
I think that the jobs that are likely to be automated first are those that involve manipulation of information and language.
Dentists may need to worry, if this story about the Chinese robot is true.
The older I get the more I realize that pretty much the only thing you get for being smart* is a richer and deeper appreciation and understanding of the truth and beauty of the universe. and from that:
1: that’s actually a big deal. The joys of being able to catch some glimpses into the rhythm of the world are immense, and I think those who don’t have that are really missing out.
2: but it doesn’t make this ‘life’ business any easier in terms of getting resources for yourself and dealing with dumb problems.
* to be specific, I think there are ‘practical’ values to more intelligence, but they kind of stop past +1 sd, because that’s all it takes to deal with a lot of the junk in the social/cultural/machine of society landscape. and more than that doesn’t help on that axis
These are good points, I thought it might be worth thinking about the selection of the anecdotal evidence too. Consider the selection process for the other anecdotes:
1. Contacts Scott about their IQ test. These are people who have got a low IQ score. On average they feel that their IQ score is lower than they would have expected. They also have goals and ambitions which are not easy to achieve for people who are not smart. They have enough intellectual discipline to not just decide IQ tests are bullshit in light of this discomforting news.
2. SLC reader. We are drawing from a different distribution than the general population. SLC readers have higher IQ and higher intellectual motivation/interest than the average (also different variance, skew etc.).
The implications of point (1) are as follows.
a) We are selecting people from the bottom of the distribution. There are two reasons to be at the bottom of the distribution, low IQ and a negative error in the test. Consequently the anecdotes are selecting on people with higher expected downward error in the IQ test.
b) We are selecting people who feel like their IQ score is too low, or not compatible with the environment they are in, or is insufficient for achieving their goals. This is almost like selecting on a second test. The people who are reporting low IQs have got this other signal (their expectations, environment or goals) about their IQ and are reporting the mismatch.
These two issues alone are going to lead us to picking out anomalously high errors. The Feynman story is a little like this too. Why did we hear about Feynman’s low IQ score? Because he’s a nobel prize winning Physicist who got a low IQ score. If he was 40 points too high we wouldn’t know
The implications of point (2) are less clear. We are picking from a different distribution and so that means we cannot interpret the anomalies in the same way as if we were drawing from the population. How this impacts things depends on the true distribution of IQ for this group. But suppose, for the sake of argument, that the true selection process was read SLC only if true IQ>120. In this case all scores below 120 would be due to measurement error. Draw a readership of x thousand from above 120 and you will get some really anomalously low scores – moreover when they see these scores they will be really miffed and quite probably want to ask someone about this who seems to know about it. I don’t really believe that this is the selection process but it illustrates the kind of selection worries that we should have.
Finally, we are also selecting on people’s interest in intellectual matters, willingness to read long blog posts that contain statistical analysis and complex arguments. This means that there is going to be another bias to worry about other than test-error. That is the relationship between IQ and positive outcomes. Someone who has a low IQ and is reading SLC must have much higher interest/motivation regarding a set of skills and subjects that typically lead to success. Consequently, even when we are getting the IQ test right we happen to be picking a person with low IQ who is going to outperform people with similar IQs.
How important could this factor be? I’m not sure but it is important to note that while the relationship between IQ and outcomes is significant (in both senses) it only explains a small fraction of the variance in people’s success. In any rigorous analysis of success we can explain around %50 of variance in terms of observables such as IQ, personality, education choices, etc. This leaves a hell of a lot unexplained i.e. there is a lot of room for people’s motivations, interests and choices to determine their outcomes. Moreover, many of the observables (e.g. education, fitness) are partially choice variables themselves – that is, driven by preferences (I think philosophy is interesting) as well as capacities (I find philosophy easy).
So people who are reporting low IQ scores who are reading SLC should both have anomalously high intelligence/success. Except, it’s only really anomalous if we don’t condition on the selection process.
Section I: “From an personal” should probably be “From a personal”
I would put a faint question mark over your assertion at the end that we should study ways to increase average IQ. Can we really be certain that a society with ever higher average IQ would be an improvement? More likely there is an optimal average level and distribution. Given human societies function like ecosystems due to specialisation of roles even just decreasing the standard deviation of a population and keeping the same average could have unexpected negative effects.
The proportions of people in various roles increase and decrease over time, and roles even appear and disappear. I’m not saying it doesn’t create problems (it obviously does), but unless you want to put a stop to all that (difficult, and only appealing if you have a higher opinion of current society than I do) they are problems you need to deal with regardless. Even if raising average IQ increases the amount of such problems (which is not guaranteed; if we don’t devote effort to raising average IQ, we’re likely to devote effort to something else which will also produce change, or just be changed by becoming bigger slackers if we don’t devote effort to anything), it seems not implausible that a higher average IQ might increase our ability to cope with such problems.
IQ is associated with so many good things that it’s hard to see why we wouldn’t try and improve society’s IQ if we can. As for the idea of ‘specialisation’ being a problem? Not really. You just use your higher average IQ to automate more and more unskilled and low-skilled jobs. Any remaining really unpleasant jobs could be done by people who would be sufficiently articulate and aware to ensure they were adequately compensated (as distinct from the status quo, where people in poorly paid, unpleasant jobs are often unable to negotiate themselves into a better position).
Also, brilliant people who prefer to think on their own time while doing a menial job do exist. I know two such people, and I spent a very enjoyable summer in college cleaning houses. I think I’d still rather be a smart person bagging groceries than a dumb person bagging groceries, if those were the only two options.
I agree that an average person with an IQ of 110-130 bagging groceries would probably be an improvement on an average person with an IQ of 70-90. But non-linear things happen when people get over IQ 150 where people can become quite dysfunctional. If the average increased to 110 and the standard deviation stayed the same the percentage of people over 150 would increase significantly.
It is also worth remembering that at least some of the genes that increase intelligence are beneficial when heterozygous but cause diseases like Tay Sachs disease when homozygous. Increasing the frequency of these genes in the population could have all sorts of unexpected effects. The pleiotropy of genes associated with high intelligence is also a big unknown. For example myopia seems to correlate with high IQ. (https://www.ncbi.nlm.nih.gov/pubmed/3417304)
If there are mechanisms that select for the current range and average of IQ in society then it might be worth determining why they are the way they are before we go shifting population averages in the hope of some “obvious” global benefit.
Hasn’t Scott already written a post about this?
Briefly, there isn’t any real reason to think that genius is inseparable from dysfunction. Yes, there are genes related to intelligence which can also result in dysfunction, but there are plenty of genes related to intelligence which don’t. And the main way we could improve intelligence is simply by correcting harmful mutations and genetic load, which would pull the 80’s and 90’s up to what we now consider average without pushing the upper bound of human intelligence any higher. Well, after we first deal with the low-hanging fruit which causes low intelligence — lead poisoning, fetal alcohol syndrome, and so on. Trying to create super-geniuses is another question, and I certainly agree we shouldn’t do that until we know all of what a gene we wish to insert/correct would actually do.
In short, if we can cure mental retardation, we should do so, because while it’s great that a person of 70 IQ can get a job putting grocery carts away, that is no reason to preserve 70 IQ for the purposes of having someone put carts away. Brave New World was wrong in the assumption that you would have to create dumber people to operate the elevators. The same person could do the job with the defect corrected, and also have a happier life in many other areas.
Well, more intelligent people get bored more easily. They are less likely to just keep putting away carts day after day, and more likely to scheme to take your place as the grocery story manager.
Being “overqualified” for jobs is a thing. There are many employers who prefer not to hire PHDs for menial jobs. The employer quite reasonably fears that they will get bored and quit after a while, and the employer will be stuck retraining. They also may be worried that the person would be less likely to follow orders.
There are 7 foot tall humans, and even 8 foot tall humans. But not many of them. There is probably a reason for that. There are Einsteins and Fermis, but not many of them either. There’s probably a reason for that too!
It’s fun to speculate about why human intelligence is not higher than it is. Does the mind get too unstable at high IQ? Or do people get too bored with the daily grind that they don’t pass on their genes as well? Was there just not enough selection pressure? Or is human intelligence just difficult to push much past where it is (diminishing returns).
This is great! Although what this article says should have been obvious, I never really thought about it this way until now.
When I was younger, I took multiple IQ tests. Apparently I got a high IQ. I didn’t care at all. I don’t even know the result of the latest test I took. I only took the tests because I was young and some respectable adults asked me nicely and I believed them.
And this article explains that I was right, and at the same time, the people who took the test and cared about whatever research they did were also right. I was right because my individual IQ score doesn’t matter too much, and because I have more reliable information about what intellectual tasks I am successful at that just the IQ. But at the same point, my school and whatever researchers cared about those tests were also right, because they got some objective measurements out of those tests and they could compare them to the same kinds of measurement done on thousands of other young people. It’s a win-win situation.
I would argue that a very similar article could have been written about discrimination, where you see the same issue that people with certain traits often regard discrimination that clearly happens at the group level as a hard limit on their personal ability to achieve things, rather than ‘merely’ a hindrance where they can still achieve great things. Especially if they are smart and avoid failure modes that have been shown to cause much of the group level differences, I feel that much of the hindrance can often be avoided, especially as a non-trivial part of the failures are caused by people in the group not being taught to make choices that lead to success.
Yes, I could easily envision such an article. However, I don’t think many people would have the chutzpah to write it.
Discrimination certainly doesn’t appear to affect all people equally. Obviously people vary in their natural aptitude for social skills, and those with unusually high aptitude are often able to find ways to “fit in” in groups other than those in which they were raised. As a result, I feel like there is a pattern of at least some of the most successful people in disadvantaged groups (who are most likely to be among those with this aptitude) sometimes being dismissive of the struggles of those who are not so successful; the hindrances didn’t give them (the successful ones) much trouble, so they couldn’t be that big of a deal.
One aspect of how discrimination affects people differently is that if someone is told “You’re the sort of person who is fundamentally inferior”, some people will react with “Fuck that, I’ll show YOU!” and others react with “might as well give up”.
Reality is more complicated because people are also affected by them being told that people exist who discriminate against them and those people then interpreting various experiences with ambiguous causes as discrimination.
For instance, I read a comment by a white person the other day who had been stopped by the police 5 times last year for no clear reason. Because that person is white and there is no narrative of racial profiling against white people, that person didn’t conclude that he was racially profiled. Instead he was unable to point to a reason. Many black people who had the same experience would logically have adopted the explanation that the media tells us is common: racial profiling.
Because we know that cases exists like the white person above, a non-zero percentage of black people will falsely attribute experiences to discrimination. We can expect errors the other way too, where black people will falsely attribute discriminatory experiences to other causes.
We can expect that different black people have different levels of eagerness to classify ambiguous experiences as discrimination. Unlike what Protagoras seems to believe, this eagerness doesn’t necessarily match how much this person is disadvantaged by discrimination. It seems plausible to me that black people in liberal environments will on average be very eager to interpret bad experiences as discrimination, because they live in an environment where it is commonly believed that discrimination is really bad; while that environment will actually have relatively low levels of discrimination, because most of the people oppose discrimination.
Another issue is that in itself it doesn’t necessarily matter whether people correctly judge whether they are discriminated, but that it’s really important whether people respond in ways that make the situation worse for them or better. Nancy correctly argues that different people respond in different ways, although I would argue that the range of options is broader than just working harder or giving up. For example, some of the police-on-black video’s seem to feature black people who angrily confront the police officer(s) about perceived discrimination in ways that doesn’t make them safer or get them treated better.
The conservative argument seems to generally be that it’s better to respond by working as hard as you can, while the progressive argument seems to generally be that the perceived discrimination should be punished, the black person given help, etc. My position is somewhere in the middle with the caveat that I think that progressives often falsely claim a level of group-level discrimination where the scientific evidence doesn’t support their claims.
PS. I suspect that the conservatives also regularly make claims that the scientific evidence doesn’t support, but since they are my fargroup…
One of the engineers in our office is licensed as a structural engineer, even though she’s not actually a structural engineer. Back in the ’80s, her supervisor told her that women weren’t smart enough to pass the SE exam. She took the test so she could show him the score sheet.
I can’t say for sure that the engineer I’m discussing is the smartest person in our engineering division, but she’s certainly on the short list for that distinction.
I think it might be useful to talk about some of the ways that people (of all sorts of intelligence) use to up-level their mental abilities. (And here by ‘mental abilities’ I mean ‘all the things that we do with our brains that smart people are measurably better at.)
Off the top of my head I can think of various practices : setting concrete, achievable goals, making lists of tasks (grocery lists, anyone?) and prioritizing them, doing systemic research (how do I find out where to find something?), practicing skills from spelling and multiplication tables through how to shmooze a receptionist/gate keeper (muscle memory and rote memorization are both useful tools) and using things like checklists and daily planners to keep one on track. Oh, and Big Rest & Clear Fluids, plus Big Exercise.
Sure, if you are one of the fortunate ones with the super brain, and you use these tools, then you’ll be more effective than a middling intellect person who uses these tools. But I have run into so many people – and have myself struggled with so much in my life – where just being *smart* wasn’t nearly as appreciated as, oh, showing up on time every day. Plus, it takes a lot of flashes of brilliance to run past steady level grinding.
I think there is a lot of room for people who aren’t brilliant but who are methodical and who persevere, to do well and lead happy lives that change the world.
Also – it might also be useful to identify those things which we do that “smart people” are not obviously better at than the average human. If we want to acknowledge the worth of all humans regardless of intellect (*) then identifying a positive quality which is not necessarily linked to intelligence might be a start.
(*) Not that everyone does, mind you. But I prefer the sort of human who sees worth in the rest of the species.
> it might be useful to talk about some of the ways that people (of all sorts of intelligence) use to up-level their mental abilities
This sort of research is done under the umbrella of “Cybernetics”, specifically in the sense of how physical systems (e.g. to-do lists) can be used to augment human capability. I haven’t studied it at all myself so I can’t provide reading material, but studies on this topic exist.
I’ll just observe that most of what you’re talking about here are strategies to be more effective at life, not how to actually increase cognitive function.
For the latter, my best guess is there’s no silver bullet, outside of the general observation that we get better at things we do, which transfers to similar tasks. Like, if you do a lot of math and programming, you’ll probably increase your general analytical ability; and if you read a lot of philosophy, you’ll probably increase your reading and logical ability.
Oh, and getting enough sleep.
My dad — a naturally smart guy trying to raise kids who had inherited his intelligence — used to say to us, “You can be dumb or lazy, but you can’t be both.” He also pointed out that we’d never reach our full potential if we counted on being smart and lazy.
But it really is obvious that studying does actually make a huge difference. I’ve always felt the main thing intelligence gives you is the ability to learn stuff faster. If you’re not smart but put in five hours of homework a night, you can pull down the same grades as a genius. Most people don’t feel like putting in that kind of effort, but some care enough to do it, and they succeed.
A framing that I find useful for interpreting social science results is: assuming that the study was perfect, had a large sample size, and had no particular sampling bias, even then its conclusions about the impact of measurable factor X on measurable outcome Y are: “In the absence of any other specific knowledge about the person, we predict that X leads to Y with some frequency”.
It’s important to remember that where one has additional specific data on a person, these are also relevant, perhaps tremendously so.
make up silly theories about “emotional intelligence” and “grit” and what have you.
I guess on this site there is a lot of prior literature about these silly theories. For me I can only take the terms “emotional intelligence” and “grit” as standard English expressions describing personal qualities. To me it seems that both of those personal qualities are virtues that I would expect to be both measurable and correlated with life outcomes in just the same kind of statistical, precise-for-populations-but-not-for-individuals way that I expect for the virtue of (ordinary) intelligence.
Agreed. But when people use those terms, they aren’t referring to them generically, but to actual psychometric measures people have tried to construct, which have generally failed.
My understanding is that, besides IQ, the only psychometric measure that reliably correlates with success is the personality trait “conscientiousness”.
Of course, that we haven’t so far been able to construct a reliable psychometric measure of something doesn’t mean that it doesn’t exist! There’s a lot of variance in personal outcomes left to be explained.
The rule is: all questionnaires measure one or more of the Big 5 plus IQ, either well or badly.
Grit is just Big 5 Conscientiousness. Emotional intelligence, insofar as it can be measured, is mainly some combination of regular IQ, B5 Agreeableness, and B5 Extraversion.
Ever since I was a kid (I’m 35 now) I was regarded as intelligent but I never took an IQ test, probably because I was always worried I would turn out to be “ordinary” and lose the only thing that made me special somehow. After reading this article I took a deep breath and did the IQTest.dk Raven’s test right here in the workplace. I opened it in an incognito window somehow thinking if it turned out to be 95 or something I will just turn everything off and erase/block the memory from my brain. I scored a 138 and I couldn’t answer the last question because I was out of time and if my boss didn’t ask me something in the middle of the test maybe I’d even have time for that.
Then I activated my wp account to write this comment here (long time reader, first time commenter) to thank you for inadvertently encouraging me to take the test and feel genuinely special and not worthless at all.
Thank you.
PS: Sorry for my English if there were any errors, it is not my primary language.
No errors! Congrats on the score. I kind of want to do the same. It doesn’t really matter in any way, but I have to admit, I want the bragging rights.
@sconn, what do you mean “no error”? A score of 138 implies that a few errors were made. I scored 138 the first time; when I redid the test to figure out what I had done wrong, I score 143, and when I did it again I got “your IQ is above 145”.
So it seems that if you make no mistake, the program simply tells you your IQ is above 145 with no more detail (which makes a lot of sense). Maybe one mistake drops your score to 143. A score of 138 means at least two mistakes, probably more.
Sconn was saying “no errors” in response to the OP’s apology for errors in English in their post.
Oops, I completely misunderstood that. Apologies.
Thanks, Rachel. Yes, OP’s English is excellent and it didn’t occur to me they weren’t a native speaker till they said so.
That was a fun test, though the last ~8 questions murdered me. Were there just a bunch more transformations on those, or were they relatively simple after some twist of lateral thinking?
Discussing IQ on a board like this is like discussion sexual conquests in a locker room or in a PUA forum. Inevitably some people will claim to have phenomenally high IQs. Just like there’s always a few 50-year olds who are banging a new hot college co-ed every week and having all of them beg for more.
The real fun happens when we switch the two around.
Please let us not get into the sexual conquests discussion again, we’ve (I’ve) already had the go-round with the guy who was a professional sex-worker saying he had a fool-proof method for getting women and that was to be a bit of a hound.
TL;DR: The world is multi-causal. IQ researchers spend a huge amount of effort eliminating confounding factors to isolate individual causes. That doesn’t mean the research is wrong. You’re not meant to use isolated-cause research to extrapolate individual effects from your multi-causal observation. Deal with it.
Hear hear!
I would just add two comments:
First, I think that people have a very inaccurate understanding of the actual IQ distribution because what you read on the internet is so misleading. People always read inflated IQ claims, and so they take a score of 105 to be absolutely crushing to their ego. But realistically an IQ of 105 just means you’re around average, like most people.
Second, generally individuals know more about their IQ than what an IQ test can tell them. An IQ test gives you a noisy estimate of g, i.e. general mental ability; but a whole lifetime of living in your own brain should give you a much better understanding of your mental abilities. I mean, we spend most of our formative years in school, which is pretty much the most g-loaded thing people do regularly!
So how about this: if you had an easy time learning a wide variety of material in school, can easily read and write complex prose, and are pretty good at math, you have high g. And if an IQ test tells you differently, the test is probably wrong.
Now if you always struggled in school, but persisted in believing that you possessed a True Spark of Genius within, and then you get a low IQ score — well, probably you were just in denial. Or maybe you have a learning disability.
(p.s. I think the online SAT/IQ conversion charts people generally cite are wrong; they seem to assume correlation 1 between IQ/SAT, while it’s more like 0.7).
I certainly hope so. MY SAT equivalent is apparently 155 which 1. nahhhh 2. if true plunges me into a state of gloom about underachieving
So how should it be done? Should we take a random variable as a prior and then adjust for the strength of the evidence?
Or maybe one with an adjusted mean given that not everyone takes the SATs.
There are two issues: SAT takers are not a random sample of the population, and there is a less than perfect correlation. I think you deal with them in that order.
So here is what I would say:
(1) Take your SAT score and look up the corresponding percentile for that year. (e.g. in 2016, a 1450 is the 98th percentile, i.e. top 2%).
(2) Adjust for the fact that not everyone takes the SAT (no perfect way to do this; err on the side of caution.) I couldn’t find the info online, but let’s be conservative and say that top 2% of SAT takers puts you in the top 1.5% of the population.
(3) Convert this percentile to standard deviations. Top 1.5% is 2.17 standard deviations.
(4) Multiply by 0.7 (or whatever the true correlation is — I don’t remember off the top of my head). In our example that yields 1.52.
(5) Convert to IQ. Here that’s ~123.
Admittedly step 2 requires guesswork, and step 4 relies on the actual correlation, which I’m not sure of (my brain tells me it’s around 0.7, but I wasn’t able to find a source for this after a quick search. It seems a plausible number to me).
Also, this result should not be interpreted as “Scoring 1450 on the SAT provides as much information as scoring 123 on an IQ test”. Just that 123 would be our best bet of what you would score if given an IQ test, but you might score higher or lower than that.
(Also, SAT has a ceiling effect which starts being important in the mid-high 1500s; but then, so do IQ tests.)
Given the above and that the college board doesn’t seem to publish percentiles beyond two digits, we shouldn’t be seeing much higher than 2.33 SD + fudge factor for #2. Say 3 SD (which implies a mean IQ of 110 for SAT testers). In step #4 that becomes 2.1 which means these tables should be topping out at around 131.
SAT is an IQ test. Saying that you should regress by 0.7 is like saying that you should regress a Raven’s score, just because it isn’t WAIS. Or vice-versa, because some people prefer Raven’s. But both of those people are wrong; SAT is a much higher quality test.
ETS publishes nationally representative percentiles [page 6]. (although they might not include drop-outs)
SAT only tests math and verbal ability, though. Not spatial, and not any other category tested on an IQ test.
Regarding Feynman’s book, I very much enjoyed reading his tall tales, but let’s be realistic: Outside of science, he was a bullsh*t artist. So I think that more likely than not, his 124 IQ was an exaggeration just like most of the stories in his book.
I challenge Scott or anyone else here to produce independent third-party verification of Feynman’s supposed 124 IQ.
Eh? I’ve never heard someone claim his IQ was an exaggeration – only discuss whether it was too low due to a test or normalization error. Theoretical physicists run the gamut, and plenty are in the 140-160 area, so it’s certainly not a laughably high number. Unless you’re proposing he fudged downwards to confound expectations?
Anyway, his sister attested it to a biographer (Gleick) and said she used to tease him because she scored higher on the school test they both took. So there’s a source if you trust Joan Feynman (herself an astrophysicist).
Where do you think you saw the story about the teasing, because it’s not in Genius? (It also isn’t really compatible with Faber’s version.)
Gleick does claim to have seen Feynman mention his IQ in a transcript of a speech he gave to his high school in his Nobel year of 1965.
In my opinion, self-reporting is unreliable when it comes to just about anything which might effect the speaker’s self-image or social standing.
When it comes to IQ, an accomplished person would be tempted to understate his IQ in order to make his accomplishments seem more impressive.
If you read “Surely You’re Joking Mr. Feynman” with a critical eye, you definitely get the impression that Feynman is willing to exaggerate or lie in order to make the story better.
So before I accept it, I would like to see a source for Feynman’s IQ which is not ultimately from Feynman himself.
To put it another way, the assertion that Feynman’s IQ was only 124 is a somewhat extraordinary claim and it requires stronger evidence than the say-so of an interested witness with a propensity for telling tall tales.
What I meant was that he probably said it was lower than it really was.
Here is the first definition of “exaggerate” that comes up on Google:
Would you mind providing a quote source and link? TIA
I think one issue is that people conflate strength of evidence with strength of the effect. If we have super-strong evidence that IQ affects future income, they assume it means income is mostly determined by IQ. It is often easier to find strong evidence (say low p values) for large effects, they are not the same thing.
This is one of the best (oblique & probably unintentional) response articles to the James Damore drama that I’ve seen so far.
The distressing thing is that Damore barely discussed IQ at all. Most of the claims he made related to preference and measures of personality. It didn’t stop him from being turned into a pariah, of course. But the general point stands: the value of a human being lies in their humanity, not some arbitrary characteristic.
“But “correlation of 0.66” is also known as “only predicts 44% of the variance”.”
Shouldn’t that be 34%? I have no idea how you work with correlation but when I see six on one side and four on the other, I assume they must add up to one.
No, 44% is correct. While fitting a regression line, finding a correlation coefficient R means that the independent variable (usually on the X axis in plots; here IQ) explains R^2 of the variance of the dependent variable (Y axis; the whatever outcome you are measuring), and
0.66*0.66 = 0.4356, which rounds to 44%.
Thank you.
Wait, what, did you just sneak in a post about James Damore?
Regarding IQ tests in childhood: I think about them as comparable to height and weight percentiles. If my kid is in the 10th percentile in height, they might grow up to be short, or they might just be a little late physically maturing. Saying someone scored 130 on an IQ test at age 11 is just saying they are in the 97th mental ability percentile for their age. But that’s like an 11-year old who is 5’3 — it doesn’t mean they will grow up to be 6’4.
It is not only about earnings that people worry. It is also about social respect. I will give an example close to me.
My wife does not have a high IQ. I love her, and she has many other traits I would never in a thousand years trade for a twenty point IQ bump. She is sort of a hufflepuff.
I am a professor and so we move in intellectual circles, go to parties, etc. where most people are pretty high IQ, or at least put a high value on it. Now, my wife really loves talking to all these people (she finds intelligence glamorous to no end), and mostly it goes great. But now and then, sadly, the women are just horrible to her…trying to draw her out into saying something pretentious or asking her some question they know damn well she can’t understand. Recently a wife of a colleague literally rolled her eyes and grinned at me like “how do you put up with this?” after my wife asked a sort of obvious question to her husband. Never in a million years would they do that if my wife had cleft palette or something. Being mean to the less-than-gifted seems like another “last respectable prejudice.”
To make matters worse, my wife is very, very beautiful—to the point that someone (uber drivers, servers, family) tells me so, unprompted, about once every few weeks. So, at these get-togethers, the men (these are engineering profs, not the most self-aware) will sometimes crowd around her and kind-heartedly explain some arcane interest of theirs at length, ignoring the women. I understand that, for the wives and girlfriends and female colleagues, it isn’t very fun when you go to the party and the guys want to talk to this physical beauty who can’t compete with you intellectually. But I’m still just astounded by the level of subtle meanness coming from people I otherwise know and like.
This also happens at her work. There are a lot of high-performing people at her office, and she gets along with most of them. But sometimes the women sort of gang up on her and put her down in ways too subtle to be, like, outright rude but which are obviously designed to hurt her feelings, put her down, etc. Then she comes home feeling bad about herself. It makes me furious.
Fortunately, in spite of this, and in line with what you’ve written, she does well at work and has a good-paying job. But I do, having seen what life is sometimes like for her, worry about the possibility of our children coming out with “my looks and her brain” (at least the IQ element of her brain, not the wonderful personality and grit).
My seven-year-old son had an IQ test recently. He scored 104. I shouldn’t have been upset — I mean, that’s not dumb or anything! — but I was, because he sure seemed brilliant to me. And I am pretty sure my whole family would have scored above that. I never took an IQ test but my SAT scores were 670 math, 800 verbal.
But the school psychologist said I shouldn’t put too much stock in it, because the reason we were seeing her is because my kid is autistic, and autistic kids don’t score properly on IQ tests. Their talents are so unbalanced that they might score in the 99th percentile on one thing (as my kid did on vocabulary) and a zero score on another (as my kid did on a shape-rotating test). This is made worse by the fact that communicating is hard for him, so on every timed test, he did terribly because he took so long to get the answer out.
If people who are scoring low might be on the spectrum, even mildly, I would be unsurprised about negative results. If they could get a score breakdown, they might find that what comes out as “average” means “wildly brilliant in one area, completely unable to do other tasks.” And while that might hold you back in some ways, it won’t keep you from being a success at the stuff you *know* you’re good at.
Let people study a bit of physics. Physicists, when teaching their dogma to young minds, talk about accuracy and precision. Accuracy is how close you are to the true (that is, thought after) result on average and precision is how well your measurements agree with each other. Usually, there is a picture with an archer and a target, with arrows first hitting all around the bull’s eye (high accuracy, low precision) and then on the other picture the arrows are hitting almost the same spot, but away from the bull’s eye (high precision, low accuracy).
Basically, what OP is saying is that all this IQ stuff is high accuracy/low precision science.
Theoretically, it should be possible to devise some measures that would be highly precise, but measure not exactly “general intelligence”, whatever that is. Something like ELO rating for chess. That would probably require (like in chess) people given enough time to learn a specialized skill, which relies on some, but not other, cognitive abilities, and then measuring them on exactly this task. A bit like what we do with athletes. Then we can talk about precision/accuracy trade-off.
Addendum: Statisticians call the same thing bias/variance trade-off. That is OP claims that IQ is an unbiased measure with high variance.
One of my favorite views on this I’ve seen is “you are never a random sample”.
IQ tells us a lot about e.g. a random person’s probable career, but that’s largely because we don’t know anything else about them. A high IQ correlates with classroom success, high interest in reading, and so on. It’s not just that IQ is a direct input, it’s that without other knowledge it provides a best-guess for all the other success factors.
If you know that you have an IQ of 95, you probably also know all those other things. Do you read science-y books for fun? Did you find high school physics easy? Do you have a strong impulse to work hard at physics? That’s all relevant data! Congratulations, you might be a great physics professor regardless of IQ!
And if none of those things are true? Well, someone confused by physics, who hates reading nonfiction, and doesn’t want to do math for a living, is probably a bad candidate for physics professor quite independent of their IQ.
We are not random samples. IQ is correlated with many things. It is the main correlation for almost nothing, and with the benefit of in-depth knowledge, it’s far more useful to focus on all of the domain-specific things that will actually predict outcomes.
I would note, in addition to all the above points, that IQ at the low end predicts success well, BUT has a real problem with confounder effects from correlated issues. Almost any form of brain damage or mal-development will lower IQ, and will harm most measures of success in life–but it isn’t the low IQ necessarily.
The best-known and most-discussed issue of this sort is low-level lead poisoning. See this Kevin Drum articleIt lowers IQ, but it also seems to directly damage impulse control. So, measures of lots of impulsive behaviour–crime, unplanned pregnancy, just not showing up for work one day–will be correlated with IQ–but it’s the damage to impulse control, not to IQ, that drives them.
You’re talking about IQ as though it were a causal factor for why certain life outcomes occur, which is precisely one of the errors people worry about when they claim that the g-factor isn’t really (I’m thinking about nostalgebraist here). IQ consists exactly of what people score on IQ tests and the traits people with different scores have. In an environment where many people have lead poisoning (and there are no other confounding factors) people with lower IQ have poorer impulse control, and that’s just as much a real fact about IQ as any other trait correlated with IQ. If you try to separate some sort of real effect of IQ from “confounder effects” then you are no longer actually talking about IQ, you’re talking about a new invented concept which may or may not exist and may or may not have the properties you’re familiar with from IQ studies.
The name of the WJ Tests of Cognitive Abilities beggars belief.
As if Woodcock-Johnson weren’t enough, Dr. Johnson’s middle name is Bonner and, really unbelievably, Dr. Woodcock’s first name is Richard.
Can’t believe we expose our children to this smut.
Source
I was wondering if I was the only one who noticed that. In fact the first time I was given the Woodcock-Johnson test it was even labeled the “Richard Woodcock & M. Bonner Johnson Aptitude Test.”
I think that might have been one of the questions: +5 points for getting the giggles on the title page?
A couple of things. First, some terms: fluid intelligence is, roughly, how fast you pick new things up or how many unusual connections you can make. Crystallized intelligence is the things that you know, or know how to do. Your boss probably cares about your crystallized intelligence– can you effectively (diagnose a sleep disorder) (weld aluminum aircraft parts) (translate the manual into French) (install a water heater) (whatever)? But you get crystallized intelligence by combining a certain level of fluid intelligence with experience/practice/work over time, supplemented perhaps with certain “tricks” (checklists, reference materials, work habits, etc.)– it’s not a one-for-one comparison.
Second, for every job/task/hobby, there is an effective “floor” for fluid intelligence. Picture a graph of effort/practice (x-axis) and fluid intelligence/IQ (y-axis). As IQ declines, there is a steady increase in the amount of work needed to achieve the task, but the slope is relatively gentle. But there comes a point where the curve has a “knee”, and below that point the effort required rises dramatically… and rapidly becomes, in effect, impossible regardless of how much work a person is willing to put into it. If you’re to the right of the knee, your IQ value won’t have nearly as much to do with your success as things like work ethic, agreeableness, “luck”, connections, etc. Below that knee, though– no amount of “grit”, coaching, hard work, etc. will do much good. It’s cruel to people below that knee to try to force them to do the task, or to look down on them if they don’t achieve it, or assume if they worked harder they’d achieve it, or whatever.
Fluid intelligence, or IQ, is a tool that lets you do certain jobs. Nothing more, nothing less. If you have enough for the tasks/jobs you want to do, then it’s otherwise irrelevant. Don’t stress out about it. It’s not some sort of numerical representation of your value as a human being.
Topic #2: Related to the previous… there are a critical number of jobs in the modern economy which require a minimum IQ in the 110-ish range– accountants, medical techs, skilled mechanics, programmers, small business owners, etc. There’s another, smaller tranche in the above-120-ish range– physicians, lawyers, engineers, etc. A society which doesn’t have enough of these people to actually cover all of the needed tasks, pretty much can’t have a modern economy. (i.e. an IQ=80 person can learn to use a cell phone and benefit from it; but it takes a group of IQ=110 or higher people to install and troubleshoot the system, run the billing, write the apps, etc., plus a few IQ=120s to engineer it, design the handsets, etc. The IQ=80 people aren’t going to be able to do it, ever, regardless of how hard they work at it.)
Topic #3: Speaking from my own case: IQ, i.e. fluid intelligence, is not simply a static, inborn trait. Like physical capability, it must be exercised and challenged in order to develop to its full extent, at least until the brain reaches full development at age 24 or so. In my case, I was not challenged fully during high school or even undergraduate college. But after getting a BA and going into the work world, I ended up choosing to re-invent myself as an electrical engineer… I took the college STEM (and particularly, math) classes after work, first at undergraduate level, then graduate level, between ages 20-30. I could feel my brain changing as I did so. I could see things better, handle details better, manipulate objects in my mind with ease, work logic problems better, pick up languages easier, even do crosswords easier. I could feel my brain changing in response to the math and engineering work. My SATs at age 16 had an implied IQ almost two standard deviations below my GREs (taken at 32, before starting my PhD). It was a little scary, even– like riding in a car that can smoothly accelerate twice as fast as when you started the trip. But a thrill too– I can do stuff now. Most people don’t put in those last 5-6 years of intellectual work, and so don’t ever “put the cherry on top” of their potential.
Thank you for putting in the effort. You’ve made the world better for everyone (including you).
So … like … a hypothetical person scoring GRE/V+Q of 1600 back in the mid-1970s … who subsequently, by sustained Janet-type STEM-cultivation, gained ~1 IQ-point per year … might by now be very nearly smart enough to appreciate how dumb they were?
Janet, yours is an inspiring — and entirely SSC-compatible — “Message of Hope“! 🙂
I’ve never had a problem with IQ being an arbitrary test that attempts to measure G and does reasonably well in aggregate. The problem is when people are eager to discriminate on these sorts of things.
And they will.
Your great-grandparents had cancer? Now your health insurance is more expensive. You’re Black? Maybe you’ll get turned down on a rental application. There are already plenty of suggestions to increase IQ through genetic engineering. Which is y’know eugenics. And generally bad for the gene pool, in the same way that the suggestions of removing all carnivorous animals from the gene pool (that EAs were floating) would be. Not to mention, politicians will somehow manage to get their hands on it, use IQ tests as a pre-requisite for voting, then game the test to favor establishment views. Sometimes you need the low-IQ proletariat to say “y’know what, that’s a fucking retarded idea” because somehow they often get it right better than us high IQ anomalies.
Experiment: What happens if you have her take melatonin some night to reinforce her body-clock setting, and then take Adderall as soon as she gets up to get out of bed in the morning after a good night’s sleep, and uses sunlight or screens or bright lights or cold or whatever to keep herself awake for at least a couple of hours? If she does this once (or for a week), will Adderall have a stimulant effect if she takes it at some other time of day?
Where the experiment comes from:
These both actually align with my personal experience of taking concerta. Screens are the biggest factor in how consistent my sleep schedule is (using screens past midnight / until I go to bed will frequently shift my sleep schedule forward by about 4-6 hours in a single night). The second biggest factor is whether or not I take concerta at the same clock time that I did the past few days. The third biggest factor is whether or not I took significantly more melatonin than the previous few days before going to sleep (going from no melatonin to 15 mg can shift my sleep schedule backwards by 4-7 hours in one night). (Sunlight is also a pretty big factor, but these days I always sleep with an eyemask, so I don’t know how to describe its effect.) By combining these effects, I can prevent jetlag (take my concerta at 8am in the destination time zone the morning of my flight, use melatonin or screens if I need to sleep significantly earlier or later than normal on the plane, get 5-12 hours of sleep in transit).
Of particular note is that, if I shift my sleep schedule with screens, taking concerta at 8am every morning does not act as a stimulant, but instead pegs my body clock / sleep schedule to whatever I set it to with screens. I don’t know how this effect works (and I would like to), but my guess is that the default effect of amphetamines is to reset your body clock to “waking up”, but if you take them (enough? a couple of times?) when you’re sleepy enough that your body clock refuses to reset, your brain will somehow learn that body-clock setting as what it should reset to when you take Adderall. Is this a plausible hypothesis?
I have always got, consistently, around the 90th percentile in any brain test I’ve done. That’s the brain equivalent of being 6 foot 2. Normally tall.
So, it probably wouldn’t be wise to count on a career in the NBA.
But beyond a few specific niche areas, like getting objects off high shelves without getting a ladder, I probably shouldn’t worry too much.
I think that’s probably the biggest practical advantage of being 6 foot 2. You don’t have to worry. Less nightmare fuel.
Obviously being the tallest must bring it’s own social advantages. Must be inconvenient too.
[Short people can still play amateur basketball if they enjoy it!]
Thanks for writing this.
Anyone have theories about why people can be so strongly affected by numbers? As someone pointed out, people actually have quite a bit of information about their mental skills.
One think I’d like to suggest for people starting out is to look at testing by the Johnson O’Connor Research Foundation.
Broadly, they test your aptitudes for many different things. For example, hand-eye coordination and memory for numbers. When done, they provide a list of occupations which use the aptitudes you score highly in, under the theory that people who aptitudes they are best at are happier in life.
I personally had some interesting revelations as a result as well. I score very low (10th and 30th percentile) for memory for pitch and memory for rhythm. But very high (90th percentile) for pitch discrimination. So despite enjoying singing, I realize that without a lot of work at it I’m not going to be very good at and and I’ll be thoroughly disappointed during the process.
For people who are facing that weird conundrum of very stumbling over things that normally would be expected to be easy for them, all things considered, this kind of assessment might shed light on why, where to expect to excel, and where extra effort will be required.
Disclaimer: I am very stupid. My IQ is failing because of aging, unhealthy lifestyle and probably some early onset dementia. I do many typing errors and I ask to forgive me for it.
One thing I am interested in, is if it possible to increase one’s IQ using some form of life-hack. I once hacked dual-n-back, by finding a way to remember a long string of the presented numbers. (I write them down on the mental screen, thus using visual processing for remembering. It increased my results two times immediately.) There are other tricks, which could increase one’s performance like technic called “innate literacy”, which is claimed to help with one’s grammar skills by using the visual memory of the correct word instead of audio memory.
I hope that the fluid intelligence also turns out to be learnable skill, but we just don’t know what it is like and how to teach it.
As far as we know, it’s not possible to increase someone’s fluid intelligence (at least in adults), despite the marketing claims. I doubt that you actually are stupid, though. As near as I can tell, your concern is not fluid intelligence per se, but rather whether there are techniques to help compensate for specific lost skills/abilities in the event of some sort of brain injury or aging. The news here is somewhat brighter: there are many techniques that can help mitigate the problems.
For example, if I understand you correctly, you are saying that your short-term memory is not what it was– but that your visual memory seems to be either unaffected, or at least less affected. As you’ve found out, a long-established mnemonic trick is to convert what you want to remember into a sequence of very vivid/startling visual images, then mentally “place” these in order in a location you’re extremely familiar with (the ancient Greeks imagined walking through their house, coming across all of these weird/shocking things in sequence as they moved through the building). For other people, auditory memory is better, and so they use rhymes, little songs/jingles, mnemonic phrases, etc.
Another set of approaches revolves around systematically using props or aides– carrying a notebook around always, and getting in the habit of writing things down every time, for example. Calendars, clocks, and so on, to prompt you when it’s time to do something. Having reference books easily available. Tech solutions (phone, Apple watch, etc.) might or might not help, depending on you.
A third type of coping skill is to simplify your environment, so that there’s less remembering to do. That could be getting simpler stuff, or putting “cheat sheets” or sticky notes with things to cue you when you’re doing something, or doing things only one way (or in only one place, at only one time).
The experts in this field are occupational therapists, and they have lots of experience helping people like you. I would suggest that you talk to one, as they’re most effective for you if they catch problems early.
The entire premise confuses me. Having a high IQ is not the same as being An Intellectual (whatever that’s supposed to mean.) Nobody’s kicking you out of your academic or scientific field for failing to meet a number meant to measure overall cognitive potential as applied to generalized populations–if you understand the concepts, have genuine enthusiasm and put in the work necessary to get good at your area of expertise, you’re in. Academia isn’t the G&T program in a public K-12. If you were too stupid to pursue whatever it is you’re pursuing, you wouldn’t be interested in the first place. My IQ is higher than my mother’s, for example. She’s a college professor. I live in a van writing abstruse rants on the internet. The university has yet to oust her and hire me over a statistical measurement within which I’m closer to the far end than she is.
I’m of the opinion that the Weschler is crap. Scoring is too dependent on test-giver bias and the comprehension subtest will artificially skew your score lower if you have a nonstandard moral code and a test-giver who is either simpler-minded, operating under the assumption that you are a fruitcake, or both.
I actually think there’s a neat cultural angle here (just not in the usual way). As I was reading this post, my primary response was “duh, of course”, except for the correlation coefficients, which seemed crazy high to me. Then, when Scott pointed out the r=0.66 explained 44% of the variance, I realized he was talking about r the whole time, while I assumed he was always talking about r^2. I thought this because my primary use and exposure to these correlations is in organismal biology, evolution, ecology and other “whole animal” fields, where everyone always reports r^2, never just r. So maybe my “duh” reaction isn’t reflective of any intelligence, but rather my immersion in a field which just takes for granted that the HUGE levels of variability in things like an entire ecosystem will make simple “X determines Y” statements laughable, and the best we can ever do is say that X explains N% of Y. After all, it’s tough to reconcile “this trait, seen in a quarter of individuals, provides a survival advantage” with “this species has 99% mortality in the first year of life” without realizing the former statement is only a description of a statistical trend, rather than an iron-clad rule.
Are there any other folks from similarly “hyper-variable” fields (for lack of a better term) who have similar experiences with the idea of statistical correlation not defining individual destiny just being “obvious”? I get the impression from the comments that, while I’m not the only biologist here, there aren’t many others who work on non-humans at whole-organism scales.
I have an IQ of around 140 so SHUUUUT UUUP!
OK, more seriously:
I have two children. One has an IQ of 135. What would you expect of such a child? He’s drumming his fingers through math class because he already knows it (and math’s not his thing). He knows who Hobbs and Kierkegaarde are. He can threaten to eat me up in two languages.
The other, they estimate, has an IQ of 40. What would you expect of him? He’s in 4th grade. We’re still working on toilet training. His longest sentences tend to be 2 words. He can dress himself with a lot of help.
Sure, if they’re about 100 points apart, we’d expect such a gap.
And if they were 5 points apart, we wouldn’t be able to tell.
So the question is not whether individual IQ scores mean something. The question is how big do the differences need to be before they mean something. Looks like it’s somewhere between 5 and 100 points.
I think this changes the nature of the discussion. It’s harder to go to extremes (IQ is worthless or supremely important) after looking at extreme scores.
An important thing that I haven’t seen mentioned here: many people take online “IQ tests”. If you happen to be one of them, let me tell you this:
There is no such thing as a serious online IQ test. Those are all fake. Yes, all of them. Yes, even those that say that they are serious. Yes, even those that pretend to be designed or approved by Mensa or whoever. Yes, even IQTest.dk or whatever happens to be your favorite one; even if it is “inspired by Raven matrices”.
If you took an online “IQ test” and you worry about… stop right there. You didn’t take a real IQ test. You have no idea what your IQ is. What you are doing is like freaking out after reading a horoscope.
(I could start explaining why, but my previous experience suggests that this would be futile, because at that moment many people simply refuse to believe that someone could actually go as far as to create a fake online “IQ test” for fun and profit. The short version is that IQ measurement is inherently statistical, which means it has to be calibrated on tons of people, who have to be randomly selected from the population. To do that methodologically correctly is quite difficult and expensive. On the other hand, just making up stuff is very easy and cheap.)
Regarding personal experience, I was always of the impression that other people thought I was smart but only if they knew me at least a bit. But now I’m thinking this might just be one of the things people say about people they like. Is there anyone here with similar experiences? Am I trying to sabotage my self-esteem by lowering my expected IQ or what do you think?
I teach college. I just read your post. Your grammar, spelling, punctuation, and construction are above average of what I see, and I only see those who make it to college. So based on what little I know about you, I think you have nothing to worry about.
I don’t know if I should or shouldn’t point out that I’m German and English is my second language. Kinda feels like bragging. Based on an online vocabulary test, my German is still much better than my English, at least in terms of vocabulary.
I’ve heard from a bunch of places that people whose native language is english actually tend to have worse grammar. Idk if it’s true.
One might not pay as close attention to your grammar if you think you “know” it and therefore are always right. Plus, if you have a kind of dialect with slightly different grammar that can mess up things as well. It will be really hard to prove this either way though. Who are you going to compare? People who had English and school for a couple of years and never used it much out of it aren’t beating native speakers inv English grammar.
@sty_silver
I doubt it. It’s more likely that there is a huge selection effect. Those who are less capable probably tend to stick to forums in their native language, which for less capable Americans/Brits/Irish/Aussies, means English forums.
I also expect that non-natives less confident for an equal level of ability, resulting in more careful writing/proofreading.
Edit: ninja’d
I have heard this too, and suspect it is a myth. But even if it is true, I suspect it is only that, having learned English from a book, their language more closely follows the rules.
My own experience is that, grammatical correctness aside, native speakers are much better able to express themselves.
Similarly: A news article today reports that the effectiveness of a flu shot, in preventing the flu, is no more than 3% and often no greater than a placebo in some years.
This, too, shows how statistics that apply to populations say nothing useful about individuals. If flu mortality is 0.2% (1 in 500) and shot effectiveness is 3% and half the population gets shots, then shots will save nearly 50,000 US lives each year. The odds that it will save your life are less than 0.005% (1 in 20,000).
You are far better off just washing your hands more often. If false confidence in flu shots reduces hand washing by 20% and half of all flu viruses are transmitted by touch, the flu shots then are counter-effective and actually cause up to 10,000 deaths each year.
Odds of a shot being effective improve for the elderly, and it would be nice to do a randomized, controlled, double blind study of that. Alas, under NIH rules, it would be unethical to do so. Meanwhile, flu vaccine manufacturing is a huge (although concentrated) industry and US health care providers can bill $4 billion each year just to give the shots.
Argggh, matey. See what reading SlateStarCodex has done to me?
The flu shot isn’t very effective and it gives you
the fluflu-like symptoms. The CDC is gung-ho about it and there are plenty of proponents who like to claim flu-shot refusers are the same as anti-vaxxers; they also like to mumble about herd immunity which as far as I can tell is not going to happen with the current effectiveness of the flu shot. I’ll continue to skip it.The CDC claims an effectiveness of roughly 50% most years; the 3% appears to be for the UK in 2015.
Anecdote, but I’ve gotten the flu shot for years and have never experienced any flu-like side effect. The same for my wife. My kids have always gotten the shot (since they’ve been old enough) and we haven’t seen them get shot-sick. So I can offer a personal endorsement, for whatever it’s worth.
Matthew Tobin Anderson’s hilarious new SF novel Landscale With Invisible Hand (2017) is a Swiftian send-up of high-IQ/low-empathy civilizations that calamitously espouse ultra-propertarian/ultra-libertarian values. Not a few high-IQ SSC commenters, perhaps, will recognize core rationalist values in Anderson’s dystopian parable.
Intractable IQ-related worries? The human-vuvv civilization has got `em … by reason that Scott’s reassuring arguments fail human families.
Thanks, Willa! I only come here for the book recommendations.
Reciprocal thanks are extended, tvt35cwm!
Please let me say, in particular, that your admirable avatar-picture calls to mind neuropsychiatrist Eric Kandel’s Reductionism in Art and Brain Science: Bridging the Two Cultures (2016), specifically Mondrian’s painting “Composition No. III, with Red, Blue, Yellow, and Black” that appears as figure 6.7 of Kandel’s chapter 6, “Mondrian and the radical reduction of the figurative image.”
Hmmm … beginning with Leibniz, no few logicians, mathematicians, and rationalists have striven toward this same visionary goal.
If rationalism can (and should) inform the arts, then perhaps the arts can (and should) inform rationalism … and thereby artistically inform, too, all of the disciplines to which rationalism lays claim … including medicine, economics, mathematics, science, and morality? Metimur non fingo! 🙂
This cognitive reciprocity is a crucial aspect of Eric Kandel’s visionary long-term cognitive research program … and it was telling (to me) that the acronym “IQ” appears neither in Kandel’s Reductionism in Art and Brain Science nor in his Nobel Lecture.
IQ MINI-TEST On what lexigraphic grounds might we be piqued at the assertion that there is no “IQ” in Kandel’s uniquely antiquated and oblique research techiques? 🙂
So in light of Scott writing a whole article about how uninformative IQ is at the individual level, i’m going to go ahead and completely disregard that in order to talk about my individual IQ.
Not going to brag about its size though, unlike some smart people, i do not attach any value to my quantity of smartness points. If anything it’s somewhat embarrassing. That said i have been tested multiple times as part of various psychological evaluations. Until recently though, these were all when i was a child, and since i know that childhood IQ scores do not necessarily reflect adulthood ones, i thought of myself as not knowing mine. In fact i preferred it that way.
So when i recently had a full psych eval, i actually asked the evaluator if the IQ portion was strictly necessary, hoping we could skip it. She said it was diagnostically useful, and to my surprise she was right, and i’m glad for it. You see, while i do not give emotional valence to my exact smart number, i do attach emotional valence to being a little weird, and my results are definitely a bit weird. As it turns out i don’t actually have a valid IQ score.
The Wechsler Adult Intelligence Scale (Fourth Edition) is normed to have a mean of 100 and a standard deviation of 15. It arrives at its Full Scale Intelligence Quotient by summing the scores of ten subtests grouped into four indices. The indices are Verbal Comprehension, Perceptual Reasoning, Working Memory, and Processing Speed. Looking at them individually gives you a better picture of the a person’s functioning than the composite score, but the full scale is still useful because the indices tend to correlate. However, if any two indices thought to be sufficiently far apart – meaning more than 1 to 1.5 standard deviations – the FSIQ is generally considered to be an invalid measure. At the very least, it fails to tell the whole story, since it will overestimate functioning in some respects and underestimate it in others.
In my case, three of the four indices are over a standard deviation apart from each other. The smallest difference is between Working Memory and Processing speed, just barely managing to be inside a standard deviation at 14 points. The largest is between Verbal Comprehension and Processing Speed, which are 56 points apart. That’s 3.7 standard deviations! My FSIQ isn’t just invalid, it’s so far from valid it actually somehow wound up nine points outside its own 90% confidence interval. That shouldn’t be possible, but i choose to believe it’s because my weirdness broke the math on the WAIS-IV.
So yeah, i didn’t want to have an IQ score and it turns out i don’t. There’s a number, but it doesn’t mean anything. Sometimes the world delivers! Also thanks to that i know why my mind always feels like a goddamned glacier. It’s not just all in my head! Or well it is, but it’s a legit problem! A test designed and administered by fancy-pants people with PhDs has indicated that i have the architecture for good depth and complexity of thought, but not enough clock cycles to take advantage of it in a timely fashion. Or as my Boyfriend put it, i’m really smart when given time to think, but will BSOD and reboot is forced to handle too many things at once.
My son is like this. His processing speed is abysmal; it may have been too low to test, I can’t remember. At any rate several of the tests he failed to answer at all in the time given. Yet his verbal skills are far above average. He happens to be autistic, but this may happen with others.
In theory, his intelligence is “average.” In reality, his scores are not valid. Depending on what you need him to do, he may be a star and surprise everyone — or he may need careful, one-on-one help to complete the task at all.
Your son sounds like he has it a lot worse than i do, since my processing speed is merely below average. Also i don’t meet the diagnostic criteria for autism. That said yes the disparity definitely causes a lot of uneven performance. People find it hard to deal with someone who is brilliant in some respects, very much not in others, they don’t know where to place their expectations.
It’s interesting that doctors have such a high IQ (higher than college professors!). Does anyone think this is anything beyond a reflecton on how hard it is to get into medical school? As a doctor myself, it’s really hard to believe that a high IQ is needed to be a doctor.
This consideration applies to most of the other occupations. Is a certain IQ necessary for holding the job/studying for the job (in absolute terms) or is it necessary to compete with everyone else who wants he same job?
I think that memory may be more useful in everyday life and more important for employment and general success in life than IQ, and that may account for the “spread” in professions (i.e. someone with a lower IQ can still do well in a job if they have a good memory).
I myself have a fairly good analytical intelligence, I think, but an abysmally bad memory – except for odd things I’ve been obsessed with, and have, over the years, accumulated some useless knowledge about. So I can rotate images, complete sequences, do all those little puzzles you get in IQ tests till the cows come home, but because I’ve never been very good at memorizing, I’ve never been able to learn languages, and the standard higher-paying professions (medicine, law) have never been an option. (I’m 57 now and have spent most of my working life as a professional musician – and even there, memorizing pieces for classical, or remembering chord sequences for pop and jazz, was always the hardest part for me.)
Life for me is like a drawing on water – I spot the beauty in the pattern, and then it’s gone, and the next shiny object appears.
For pre-med students, organic chemistry is often the make or break class. It requires a whole lot of memorizing.
IF you’re talking about memory, then I might even agree with you. I don’t believe (as I stated in my comment above) that being a doctor requires a high IQ (maybe being a lawyer or judge does, I don’t know). But being a doctor certainly require a good memory. To be even a passable doctor the amount of information you have to read from large books and know by heart is enormous. Medical textbooks are often very bad, and that makes it even harder to memorize, but even the best written condensed summary boils down to a huge amount o information.
But again I wonder. How much information is reaaaaaly necessary? Often, the hardest things to memorize are irrelevant biochemistry or other basic science details which you won’t have any use for: they don’t help in diagnosis, treatment or prognosis. Some people might argue that they might help you memorizing (or explaining) stuff, but that’s not been my experience. In my experience, such explanations only work after the fact. That is, they are made up reasons for the fact to be true. If you’ve forgotten the fact you want to explain, good luck in trying to “derive” it from the basic information you know about.
For example, you have a patient having a heart attack (technically, a ST segment elevation myocardial infarction). Should you treat high blood sugar aggressively? Let’s try to “reason” it out: Hm… Lowering blood sugar is probably not very smart in a patient with a heart that has the blood supply partially blocked… I mean, the heart is starved for ocxygen and nutrients (the heart rarely uses sugar directly, but it can turn it into other things). Should we deprive it even further? We often give oxygen to these patients. Why starve it for sugar? But wait… High blood sugar might do some mumbo-jumbo with free oxygen radicals and stuff like that. And might raise the heart’s metabolism more than the supply of oxygen that’s available. That might kill the heart even faster. Help, what do I do?!!
This is medicine for you. For every question, there are always arguments in favor of Yes and arguments in favor or No, because the human body is Just That Complex. So you can never reason from first principles (it’ll always be close to 50-50, and that’s not even information in the Bayesian sense). You need to memorize the answer. Yes, during Medical school we teach all those “explanations” and students nod in agreement, clearly “understanding how everything is put together, which will help enormously during clinical practice”, but the truth is you just memorize the answer. If you ever find yourself reasoning from first principles, God help us all.
If you want a simple example, pick up an physiology book (Boron’s, Guyton’s, Ganong’s), look at the pictures of the heart and understand the cardiac cycle. That’s simple, it’s just a pulsatile pump with some valves to ensure unidirectional flux of blood. Then pick up a textbook on clinical medicine (Harrison’s will be fine), or clinical examination (Bates, Moore, etc.) and read the chapter on hear sounds, which Totally Make Sense Given The Physiology You Already Know. Reach the end of the chapter. Try to recall something by reasoning from first principles. Fail spectacularly. Call the authors on their bullshit :p
So, I do believe we could cut the amount of memorization (less irrelevant details! Yeah!) a lot and still have good doctors. The main problem is that I think such doctors might be easier to fool by mischievous players, like big Pharma and stuff. But they might practice medicine perfectly well.
This doesn’t change the fact that it’s still an enormous amount of information, but it might reduce it somehow, and this might actually be needed in the future, as the amount of possibly useful medical knowledge keeps growing without bound.
I wonder if the high reliance of medical work on memory may be a reason why a fair amount of diagnostic work is likely to be replaceable by AI (or so I’ve read). Perhaps that will allow IQ and experience to have a relatively higher weighting in medical work.
(Years of experience and a truckload of real-life cases I should imagine is the irreplaceable thing – i.e. you have that mass of memorized stuff; it comes alive through years of experience and gets knit into a more solid structure that the analytic aspect, IQ, can get to work on, and more quickly find solutions to problems.)
Anatomy is ungodly complicated. The body is not organized on a grid system. Everything has funny names, like the humerus.
One doctor told me that medicine doesn’t take high intelligence. It takes the ability to remember a lot and make short chains of inference.
And a friend with some medical problems said it can take a merely decent doctor to treat, but a genius to diagnose.
This is a good text when it comes to people worrying about their IQ. However, I’m pretty certain people anxious due to whether their IQ is high enough quite often suffer from insecurities regarding their intelligence in general. No amount of explaining will do away with those feelings, especially if the people in question are already of the opinion that their achievements up to today are lousy in comparison with their smart friends.
Take, for instance, the chart showing how IQ results are scattered across professions. As I’m aware of, IQ still matters as the productivity of a person in a given profession is relative to their IQ. High IQ individuals are more productive as long as the profession has to do primarily with cognitive tasks; this makes it dubious, in my opinion, to point out that non-high IQ people can take up theoretical physics – indeed they can, but they’ll suck at it when compared with more intelligent people.
All of this obviously shows that it’s not, well, smart to worry about one’s intelligence, but my point is that as long as the source of insecurity is one’s intelligence, tackling IQ testing won’t help.
How about I stress out that with a my high IQ (measured by SAT scores), I haven’t accomplished enough in life? 😉
I wrote up a long-form response to this: https://seekingquestions.blogspot.com/2017/10/iq-scores-what-are-they-good-for.html
Basically, while I agree with the intent behind this post and with the main arguments, it ignores ways in which estimating your own g can be useful. IQ/g are obviously very relevant information for lots of things in life, so presumably it should be possible to extract value from that information. My long form response explains one interesting way to do so.
I’ve been tested >130 so this is always an interesting topic to me. For me a high IQ simply means I’m able to process information faster, which also gives me an advantage in understanding complex systems.
But interestingly it also gives me an advantage when i try to control fast things, which funnily this research also shows: http://www.motoiq.com/MagazineArticles/ID/2140/What-is-a-Rally-Driver-Part-1.aspx
Somewhat late to the party, but IQ tests are BUNK. They were developed long before we knew anything significant about the human brain, in the era of phrenology. We have no idea what they actually measure.
IQ tests should be rendered obsolete. And hopefully, we will be able to test peoples’ mental abilities through an MRI scan instead in the future.
They predict outcomes, so they are not bunk.
If I have a black box with a button that gives you a number & people with high numbers do significantly better by other metrics, then the black box is a useful tool. It then doesn’t matter if we don’t know how the black box arrives at the number.
Very good post. I appreciated the end with the notation about radically opposite responses to the belief of a communist takeover demonstrate the same underlying misconception.
Keep in mind the IQ debate has been tarred by the right, IMO, more than the left. Unfair reactions to Charles Murray has spun off a cottage industry of would be pundits who think IQ is some ‘settled science’ that have proven all unequal outcomes in life are due to inherent IQ differences and any attempt to compensate would not only be ineffective but they would also be futile leaving us all less well off than we could otherwise be.
As the author noted IQ tests do not seem very helpful when applied at the individual level but are they helpful at the collective level? If I was hiring an office of 30 people, for example, and I have 500 applications would selecting the top 30 IQ’s give me the best odds of selecting the highest quality team of 30 that could be formed from any combination of those 500? In sports the answer seems to be a clear NO. Take football. Does running, physical strength and eye hand coordination correlate to better football performance? Yes. Would anyone form an NFL team by simply taking players that max out on those measures (or even worse a ‘football IQ’ type measure that combines those elements into a single score)? Almost certainly not since the results they are seeking are not just the sums of top performers but the synergy that comes from combining compatible players into a superior team.
Almost every aspect of our economic lives are less in the mode of the individual performer (like the Olympic weight lifter) and more in terms of an economic team. IQ’s usefulness IMO is probably real but almost entirely offset by the damage it does from our natural inclination to overestimate just how real it really is.