Hacker News Reader

Comment by dzdt | Hacker News Reader

It is interesting to me that they use years of education as a measure of intelligence. As an American I am more used to thinking of intelligence as an innate, almost fixed characteristic (IQ). In that view education increases knowledge and skills but isn't expected to much affect general intelligence.

But to this article's authors, "intelligence" is increased by education, so harmful effects of pollution can be measured in years of education lost.

Is there a general cultural difference in how intelligence is thought of in China as opposed to the US?

View on HN

I must admit that from perspective of the someone from the UK I've never really understood the US obsession with the desire to represent intelligence as a single number.

“Obsession” is not the right word. I don’t think most Americans have an opinion on it (and they generally shouldn’t unless they’re psychologists or willing to do a lot of personal research on the subject).

I’m also pretty skeptical that there’s a meaningful difference between what the modal person in the UK and US (respectively) believes about IQ. I could probably be persuaded, but I don’t see any reason to think your opinion has anything to do with your country or nationality.

About the only place I encounter references to IQ regularly is here on HN - I've never heard of it in an educational or employment context in the UK.

I live in America and I still never see IQ mentioned in an educational or employment context. Only on the internet.

HN is basically the furthest thing I can imagine from an accurate representation of mainstream American culture.

If you draw any conclusions about what Americans are obsessed with from HN comments, you will almost always be wrong.

Yes, I realised I was over generalising as soon as I made that comment - I was mostly (over)reacting to the parent comment that did imply that it was a part of American culture.

It's an interesting phenomenon, perhaps worthy of study by propagandists or advertisers. If you tried to claim that a person's physical strength could be measured with a single number people would think you're mad, and probably most people in the world would think the concept of "IQ" is mad, and yet, in some countries, so many people believe in it.

Eh...

> If you tried to claim that a person's physical strength could be measured with a single number people would think you're mad

I wouldn’t think you’re mad. Many others wouldn’t either. Powerlifting competitions are already close to this. Take your performance across the four major lifts (benchpress, squat, shoulder press, deadlift) and devise a formula to combine it into a total number measuring the “general” strength of a person. You can compete on individual lifts, and if you break out that score some will be better than others at individual lifts, but the larger number remains indicative of well-rounded achievement in orthogonal areas.

That being said, I don’t think physical strength is comparable in any case.

> probably most people in the world would think the concept of "IQ" is mad, and yet, in some countries, so many people believe in it.

I think most people either “believe” in IQ or don’t think about it often enough to have an opinion. In fact skepticism about IQ is, in my experience, a distinct phenomeon on message boards and not at all representative of society in general.

"If you tried to claim that a person's physical strength could be measured with a single number"

Its consistently measured as a single number by the US military - it's just a combination of your performance on several exercises - traditionally pull ups, distance run times, swim times, number of crunches, etc.

So, again, fitness is consistently measured as a single number by one of the world's largest organizations. Don't postulate it's absurd. The same aggregation can be done for intelligence.

Yes, you can do that: you aggregate the results from some arbitrary tests, using some arbitrary weights, and out comes an arbitrary number. The trick must be to give that number a catchy name, so that people think it's something highly meaningful, almost akin to mass, distance or time.

Fluid and crystallized intelligence (which is what IQ test measure) are the most robust ways of measuring intelligence, and IQ is the single most reliable metric ever discovered in psychology, in terms of making predictions. If you're interesting in measuring how smart somebody is, there's no better way of doing it than with that single number.

Which IQ test? IQ as a number alone mean nothing, as you can't make a standardized test for it. And IQ number is a comparison (you calculate the standard deviation). Also, even in clinical tests, multiple scales can be used, and this mesure evolve. And there is no "single number", multple subjects are tested, and now we have another limitation: those tests don't mesure everything, or mesure thing that can evolve. I "scored" 123 on the weshler when i was 8 (or 9) with a really bad spatial visualisation score of 90 (89?). Since i'm now able to play blindchess or to chart a course on a marine map, either my spatial visualisation skills have improved, either people i know hide a lot. And i'm sure i would score way less on the other subjects too, looking back i was a pretty interesting kid, while now i'm an average adult.

I'm not saying it's useless, but to me it is way overvaluated by some people. Maybe you can approximate intelligence with an IQ test, but the number itself is meaningless imho. The classification can be used though.

The hypothetical g which IQ tests are trying to find is the extent to which every test of mental abilities that anyone has ever devised plus some things like reflex speed all seem to be fairly strongly correlated with each other and all seem to be correlated via a common factor. You can certainly be more skilled at one test than another. These tests are also fairly noisy when applied to individuals.

For individuals the results of IQ tests probably are oversold as you say given their noisiness. But for public health problems like air pollution where the results are averaged over populations they're probably undervalued if anything.

>IQ as a number alone mean nothing, as you can't make a standardized test for it

You don't need standardisation for IQ tests, because 100 will always be the mean. Get people to do your test, set the mean at 100 and you've got an IQ test.

>And IQ number is a comparison (you calculate the standard deviation). Also, even in clinical tests, multiple scales can be used

It measures deviation from the mean, and 15 points is used to show a standard deviation. This is how any clinical test will be administered.

>those tests don't mesure everything

Fantastic point. They ONLY measure intelligence.

>I'm not saying it's useless, but to me it is way overvaluated by some people. Maybe you can approximate intelligence with an IQ test, but the number itself is meaningless imho.

The problem with saying this is that IQ tests are far and away the best way of measuring intelligence. They do a pretty good job of it, as shown by their relative success in predicting outcomes. If you don't put any weight in IQ tests, then there's really nothing else you could rationally give any value to when it comes to measuring intelligence.

>but to me it is way overvaluated by some people

Personally, I think they're overvalued to the extent that intelligence in general is.

I assume it’s exactly for this same reason there’s an archetype of an incredibly intelligent person, who ends up with a cocaine addiction, because he was never understood...

I can't understand and I would like to. Could you explain yourself?

I assume that the parent poster is referring to Sherlock Holmes.

... as long as you're limiting yourself to measurements that result in a single number?

IQ is a single number for a very good reason. If you take all the questions psychology researchers have come up with to try to measure intelligence, take the pass/fail statistics of test-takers on them, and run a principle component analysis on it, you wind up with a single dimension.

There's no forcing the data to be single-dimensional regardless of the underlying truth about intelligence. People genuinely made an attempt to measure intelligence, and the measures people came up with happened to measure something best described as a single number.

Psychology researchers have tried to do the same with personality, and the principle component analysis spit out five dimensions rather than one. That's why the "Big 5" personality traits (Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism) are a thing.

I find the unwillingness to accept that general intelligence could be measurable (especially at the macro level) similarly perplexing.

Actually, I have no problem with measuring it - just that the idea of commonly representing it as a single value that I'm a bit sceptical about.

It's typically referred to as "general intelligence," or "g," and the reason this single value is significant is because of the strong correlations it has with other life outcomes. I don't think anyone claims it represents all intelligence, and presumably we could come up for numbers regarding other aspects of intelligence. It's just the best number we've got so far for what people are trying to measure.

It's definitely useful but with caveats. Your measurable iq isn't constant, people can have impairments that lower it temporarily and there are also certain drugs that can raise test scores.

> It's definitely useful but with caveats. Your measurable iq isn't constant, people can have impairments that lower it temporarily and there are also certain drugs that can raise test scores.

What!? Where can I get a hold of those drugs? :-)

I don't recommend naively swallowing pill cocktails in pursuit of IQ test scores, but informally people have taken stimulants and anti-narcolepsy drugs and gotten better scores, doing a reasonable job controlling for other factors.

According to this study, one year of education translates to 1 to 5 IQ points:

https://psyarxiv.com/kymhp/

I read the abstract - supposedly they are aware of, and try to eliminate correlation, but I am unconvinced: We know that maintenance keeps intelligence from declining, could be what's happening here.

It does not matter here whether it is correlation because whatever the pollution study is referring to, it will likely involve the same correlation.

Are IQ tests truly generalized and accessible enough that education doesn’t affect one’s performance? Personally, I’ve been exposed to years of standardized testing in the educational system and developed skills specifically at taking tests.

Yes, as a young kid I had an older sibling who was studying educational psychology so I spent a lot of time doing "intelligence" tests - I just thought they were interesting puzzles and didn't realise I was being continually tested.

The fact that you developed those skills shows that you've got a few IQ points... Some other may not have them.

Which is besides the point. Yes, having higher intelligence often means you'll benefit more from practice, but GP was asking of the tests were good enough that there's no practice benefit at all.

Which is false.

IQ tests are not generalised measures of innate intelligence at all and your IQ absolutely increases with education.

My girlfreind is currently doing a doctorate psychology study, of which a small part is a standardised IQ test (though I dont know which standard; there are several). She asked if she could practice the process on me.

There are several sections, but to pick 2.

For language acuity, I scored OK. This section basically asked you to define a series of given words. Your knowledge here absolutely would change depending on your environment, job and education.

For pattern recognition, I scored fantastically. But I know this is because of years of studying geometry in math classes. I know exactly what my thought processes were, "Count the vertexs, count the edges, count the bisections, are some things odd, are some things even, do some things add together, do the shapes mirror once/twice, are they rotated". I know exactly where I picked those ideas up.

In the UK you can drop Maths and English/Creative Writing at 17. If you finish a pHD at 26 you havent regularly explored those ideas in near on 10 years, but you absolutely could still achieve it, so I do think its a better measure.

> your IQ absolutely increases with education.

Then I am not sure if it is actually measuring intelligence.

Only if you want "intelligence" to be a fixed characteristic of a person. Which would be a strange requirement, since other characteristics can and do change a lot, especially with a targeted effort to change them.

> since other characteristics can and do change a lot, especially with a targeted effort to change them.

Some attributes can be changed, some, on the other hand, not so much.

"The new work [...] analysed language and arithmetic tests conducted as part of the China Family Panel Studies on 20,000 people across the nation between 2010 and 2014. The scientists compared the test results with records of nitrogen dioxide and sulphur dioxide pollution."

While I admittedly did not read the science paper, the newspaper indicates that they used test scores as the actual number, and the "cognitive drop equivalent to one year of education" phrase is just a way to explain the results, presumably anchored by how people in different school grades score in those tests.

What gave you the impression that IQ is fixed or innate? It's neither nature nor nurture, it's both.

It is (at least after a certain young age). There's been a lot of research around this because the incentive is huge. Military has spent quite a bit of money on that too (IARPA). So far, science tells us that it is fixed. I mean, you can lower it with some brain injuries but you pretty much cannot increase it[1].

I know it ruins some American Dream ideas, people with IQ below 80 are pretty much not employable, but that's just how it is. Some people have much harder time achieving things that others do with ease and we don't know any way to help them. For some reason people see it in a very different light than e.g. somebody having some disease that he has to organize his life around it or e.g. being paralyzed.

1. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3950413/

edit: btw I love and highly recommend Brain That Changes Itself book, by Norman Doidge, but this is something different

Give people modafanil, and they score higher on IQ tests. Get them drunk and they do worse. It's really not fixed.

Any relevant studies regarding this?

According to this paper[0]:

the cognitive benefits of modafinil seem particularly marked in tests of vigilance and speed, in which sleepiness would be an important factor. Furthermore, the results indicate that high IQ may limit detection of modafinil's positive effects.

[0]https://www.ncbi.nlm.nih.gov/pubmed/16140369

Unfortunately no, I don't have any concrete evidence on that point, just some amount of anecdata from people who are "nootropic" enthusiasts that I found convincing enough.

Edit: This might be interesting to you if you want some scientific literature. https://doi.org/10.1016/j.euroneuro.2015.07.028

Give them the test one more time after modafinil wears off and they score even higher ;)

Some references are welcome. Pure guess, but we do know that most society is sleep deprived and that sleep deprivation affects cognitive performance. As far as I understand it, modafinil can alleviate that.

That may be what was happening --It makes me wonder, if we say that someones higher baseline intelligence is hidden behind some semi-permanent condition like chronic inflammation or poor sleep patterns, is that really their baseline? How many other factors may be "artificially" decreasing someones measurable general intelligence for most or all of their life?

I'm not sure about this, but i have no study, only anecdot.

I scored 123 at 8 or 9 on the weshler test, and 90 on the spatial visualization part. (that mean that around 60% of the european population was better than me at spatial visualization when i was a kid). However, i'm able to play blindchess and chart a course since i'm a teen, i was probably one of the best at understanding geometry/vectors in highschool and the only math course i wasn't bad at in college were topology and n-dimmension vectorized spaces. If i "blundered" the IQ test, i'm not aware of it (i know i wasn't very cooperative, but still). Maybe i misundertand what spatial vizualisation mean, but i don't think so.

Oh and also one part of the test was general knowledge, and i don't believe it can't be trained (like the puzzles/tangrams, symbol recognition or verbal expression tests).

If you're right, then it's reassuring for me since i feel like i'm more dumb on multiple subject than when i was a kid.

I don't believe in that science. The brain can learn how to approach problems more efficiently, how to break them down, how to think in the way these IQ tests are designed.

We put way too much confidence into science in general. It's our God we don't question.

> We put way too much confidence into science in general. It's our God we don't question.

On the contrary, it's almost a definition of science that we question, challenge, and evolve it all the time.

Sometimes this is true of the scientists but not the general populous, they can tend to hold onto ideas that were once accepted science and turned out inaccurate.

It's both fixed and innate. Education can stretch it a little, but it benefits more high IQs than low IQs. (It's fixed, but stretchable, and you can squeeze it, with stress, hunger, poverty, etc).

I'd like 20cc of study with that statement, given that brain plasticity can change radically the way your brain works all your life. Also, accidents can alter the brain and lower your IQ, so are you saying you can go down but not up ?

> so are you saying you can go down but not up

You have a specific amount of neurons and a lot of them die every day. The body doesn't produce them like other cells after a certain age threshold (age 21 IIRC, also depends on the individual). - Although this theory is currently on the edge of being disproved, it's clear that adult brains are not producing many of them.

Training and experience can compensate this loss, but the general cognitive capacity gets reduced each day.

Pretty depressive to think about it, but maybe we'll be able to use stem cell research to change this in the future.

edit: I don't want to state this as definitive facts. The field of neurogenesis is currently heavily researched and there are some hints that adult brains might be able to produce new neurons. See https://www.scientificamerican.com/article/does-the-adult-br... for a discussion.

What I think is most likely is that we lose 10x times the amount of neurons that we create each day, so even if it's producing new ones, the rate of death is higher.

Heritability of IQ is the term you're looking for. Wiki has a page on it here [1]. The numbers have not been pinned down and it was previously thought to be as low as 57%, but recent studies tend to show heritability rates upwards of 80%. There are a lot of complex issues that make study difficult. For instance there's the paradoxical issue that childhood IQ can be somewhat meaningfully influenced by environment, yet that notwithstanding, as children approach adolescence and especially adulthood their IQ becomes ever more a product of genetics.

That said environment obviously does play some role, but that role is already a minority player and seems to be getting smaller as we learn more about genetics.

[1] - https://en.wikipedia.org/wiki/Heritability_of_IQ

From the article: "Heritability measures the proportion of variation in a trait that can be attributed to genes, and not the proportion of a trait caused by genes." If e.g. schools get better, and the whole population would have access to these better schools, everyone would be more intelligent, yet the heritability would still be at 80%.

You've gotta be careful with Wiki at times. In that same paragraph you'll also see the sentence "Contrary to popular[citation needed] belief, two parents of higher IQ will not necessarily produce offspring of equal or higher intelligence. In fact, according to the concept of regression toward the mean, parents whose IQ is at either extreme are more likely to produce offspring with IQ closer to the mean (or average)." That literally made me laugh. Whoever wrote that has no absolutely clue what regression towards the mean is or means, yet felt qualified to write on it. The reason Wiki is useful is as a collection of sources to meaningful information and statistics. But much of what is written by Wikipedia editors themselves is often very bad, and exponentially more so when you enter into any area that could be in any way controversial -- which includes this topic. It should be taken with a grain of salt.

In this case, however, the statement you mention is technically correct. But one thing you have to account for is that 20% encompasses all environmental differences. Put another way, if we all had a 100% identical upbringing and environmental factors then the heritability of any given observation would obviously be 100%, since any deviations from the norm would be primarily due to genetic factors. So that means that all of the wild variations in our upbringings and backgrounds are constrained to as little as 20% of any variation in IQ.

> Also, accidents can alter the brain and lower your IQ, so are you saying you can go down but not up

What sort of an argument is this? You can lose an arm in an accident, does this mean you can also gain an arm?

If you can lose mobility in an accident by making your muscle stiffer, you can game some by making it more flexible.

Funny how people around here are usually rather skeptical about any results in psychology apart from intelligence tests!

If there's one thing in the field of psychology that is actually science, it's IQ tests. Empirically, IQ has been shown to not change much at all over time, and is the single most reliable factor for predicting outcomes that psychology has ever produced. The big 5 personality model is the next best tool psychologists have, but it's not nearly as reliable as IQ.

That said, it's still not an excellent measure for basing predictions on, by the standards of "traditional" science. But if you don't put any weight in IQ, then there's nothing in the entire field that you should have any faith in.

> 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 [1], 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.

Quoted from http://slatestarcodex.com/2017/09/27/against-individual-iq-w...

Ref [1] is https://www.huffingtonpost.com/scott-barry-kaufman/intellige...

34, not 44, right?

The parent is using the common definition of R^2, i.e. 0.66^2 = 0.44

Does a correlation of 1 mean it predicts none of the variance?

> It is interesting to me that they use years of education as a measure of intelligence.

I had a similar reaction when I read that line. I mean how does one correlate number of years of education to intelligence, which is already a somewhat abstract concept. That is like saying "If you wear this makeup, your partner will love you as if they were married to you twice as long"..

Most psychologists wouldn't operationalize intelligence as years of schooling but its common practice in economics and some other fields.

In social science research there is often a tradeoff between measurement quality and sampling quality. Do you want to measure your variables really well in a small sample, or measure them crudely in a large sample.

> In social science research there is often a tradeoff between measurement quality and sampling quality. Do you want to measure your variables really well in a small sample, or measure them crudely in a large sample.

Wouldn't it be honest to just accept the fact that both methods don't provide data that's relevant and trustworthy enough for the research at hand? It sound irresponsible to draw conclusions (that will be picked up by newspapers and eventually policy makers) based on data that does not merit them, only because no beter data is available.

I think the thinking is that getting many years of education is desirable for almost everybody (it generally leads to better careers later in life, etc), but that only those who possess sufficient intelligence get the chance to pursue higher education.

So there's a high correlation, and it's much easier to measure than IQ.

Yeah that struck me as odd too (and I'm not American). I think a year of school makes me more educated, but not more intelligent.

No, not really: they're using genetic engineering to breed higher intelligence. http://www.dailymail.co.uk/sciencetech/article-3721991/China...

I guess it's more an expression of the point of view of the society, who doesn't care about persons, but about what they can bring to the enterprises. Education is formating workers, and if pollution makes you lose one year of education, that's a lot of money lost.

Reply on news.ycombinator.com