If Smart Is the Norm, Stupidity Gets More Interesting | Hacker News Reader

If Smart Is the Norm, Stupidity Gets More Interesting

nytimes.com

67 pointstokenadult25 comments

Editing has happened to this comment since the first reply was posted to it. At first I just shared some links to scholarly articles on genetic influences on human intelligence, as I was about to go out the door to have a walk with my son the hacker as this article reached HN's front page. Now I'm back at my computer to provide some more commentary on helpful background reading for the newspaper article from today (which I learned about from a researcher on this issue). Over the years here on HN, I have often seen mistaken conclusions about "heritability" and its relationship to human intelligence from HN participants who haven't kept up with the current research literature. I keep up with the current research literature on human behavioral genetics by weekly visits during the school year to the Readings in Behavioral Genetics and Individual Differences Psychology "journal club" (graduate seminar) at the University of Minnesota.

http://www.psych.umn.edu/ugrad/pastsyllabi.html

(One or two past syllabuses of the seminar are available above. The current course page

http://www.psych.umn.edu/courses/mcgue/psy8935/

has log-in authentication to facilitate sharing current journal articles among seminar participants. The seminar also has an email list from which I learned about the New York Times article submitted here, just after today's seminar.)

An old-fashioned view of heritability estimates for human intelligence, which I see repeated frequently on HN, is that heritability constrains controllability ("malleability") of human behavioral traits. Nothing could be further from the truth.

Johnson, Wendy; Turkheimer, Eric; Gottesman, Irving I.; Bouchard Jr., Thomas (2009). Beyond Heritability: Twin Studies in Behavioral Research. Current Directions in Psychological Science, 18, 4, 217-220

http://people.virginia.edu/~ent3c/papers2/Articles%20for%20O...

is an interesting paper by some of the researchers most familiar with research on human twins, including the director of the Minnesota Study of Identical Twins Reared Apart, which includes the statement "Moreover, even highly heritable traits can be strongly manipulated by the environment, so heritability has little if anything to do with controllability. For example, height is on the order of 90% heritable, yet North and South Koreans, who come from the same genetic background, presently differ in average height by a full 6 inches (Pak, 2004; Schwekendiek, 2008)."

Another interesting paper,

Turkheimer, E. (2008, Spring). A better way to use twins for developmental research. LIFE Newsletter, 2, 1-5

http://people.virginia.edu/~ent3c/papers2/Articles%20for%20O...

admits the disappointment of behavioral genetics researchers. Eric Turkheimer, a leading scholar in the Behavior Genetics Association, points out that the interpretation often given to "heritability" findings that they somehow measure genetic constraint on variation in a human trait is wrong.

"But back to the question: What does heritability mean? Almost everyone who has ever thought about heritability has reached a commonsense intuition about it: One way or another, heritability has to be some kind of index of how genetic a trait is. That intuition explains why so many thousands of heritability coefficients have been calculated over the years. Once the twin registries have been assembled, it’s easy and fun, like having a genoscope you can point at one trait after another to take a reading of how genetic things are. Height? Very genetic. Intelligence? Pretty genetic. Schizophrenia? That looks pretty genetic too. Personality? Yep, that too. And over multiple studies and traits the heritabilities go up and down, providing the basis for nearly infinite Talmudic revisions of the grand theories of the heritability of things, perfect grist for the wheels of social science.

"Unfortunately, that fundamental intuition is wrong. Heritability isn’t an index of how genetic a trait is. A great deal of time has been wasted in the effort of measuring the heritability of traits in the false expectation that somehow the genetic nature of psychological phenomena would be revealed. There are many reasons for making this strong statement, but the most important of them harkens back to the description of heritability as an effect size. An effect size of the R2 family is a standardized estimate of the proportion of the variance in one variable that is reduced when another variable is held constant statistically. In this case it is an estimate of how much the variance of a trait would be reduced if everyone were genetically identical. With a moment’s thought you can see that the answer to the question of how much variance would be reduced if everyone was genetically identical depends crucially on how genetically different everyone was in the first place."

I regret that the recent review article "The neuroscience of human intelligence differences" by Deary and Johnson and Penke (2010), relating specifically to human intelligence,

http://www.larspenke.eu/pdfs/Deary_Penke_Johnson_2010_-_Neur...

has not received more discussion on HN. "At this point, it seems unlikely that single genetic loci have major effects on normal-range intelligence. For example, a modestly sized genome-wide study of the general intelligence factor derived from ten separate test scores in the cAnTAB cognitive test battery did not find any important genome-wide single nucleotide polymorphisms or copy number variants, and did not replicate genetic variants that had previously been associated with cognitive ability[note 48]."

The review article Johnson, W. (2010). Understanding the Genetics of Intelligence: Can Height Help? Can Corn Oil?. Current Directions in Psychological Science, 19(3), 177-182

http://apsychoserver.psych.arizona.edu/JJBAReprints/PSYC621/...

is an article I submitted as a new submission on HN a little while ago. The article looks at some famous genetic experiments to show how little is explained by gene frequencies even in thoroughly studied populations defined by artificial selection.

"Together, however, the developmental natures of GCA [that is, IQ] and height, the likely influences of gene–environment correlations and interactions on their developmental processes, and the potential for genetic background and environmental circumstances to release previously unexpressed genetic variation suggest that very different combinations of genes may produce identical IQs or heights or levels of any other psychological trait. And the same genes may produce very different IQs and heights against different genetic backgrounds and in different environmental circumstances."

The brand new review article

Chabris, C. F., Hebert, B. M., Benjamin, D. J., Beauchamp, J., Cesarini, D., van der Loos, M., ... & Laibson, D. (2012). Most reported genetic associations with general intelligence are probably false positives. Psychological Science.

http://coglab.wjh.harvard.edu/~cfc/Chabris2012a-FalsePositiv...

is especially important for discussing social science implications of what genetic studies have shown for years now. Social scientists are still catching up with how little certainty attaches to genetic studies of human intelligence--except for the certainty, as reported in today's article, that most genetic influences on IQ are small in effect.

"At the time most of the results we attempted to replicate were obtained, candidate-gene studies of complex traits were commonplace in medical genetics research. Such studies are now rarely published in leading journals. Our results add IQ to the list of phenotypes that must be approached with great caution when considering published molecular genetic associations. In our view, excitement over the value of behavioral and molecular genetic studies in the social sciences should be tempered—as it has been in the medical sciences—by a recognition that, for complex phenotypes, individual common genetic variants of the sort assayed by SNP microarrays are likely to have very small effects.

"Associations of candidate genes with psychological traits and other traits studied in the social sciences should be viewed as tentative until they have been replicated in multiple large samples. Failing to exercise such caution may hamper scientific progress by allowing for the proliferation of potentially false results, which may then influence the research agendas of scientists who do not realize that the associations they take as a starting point for their efforts may not be real. And the dissemination of false results to the public may lead to incorrect perceptions about the state of knowledge in the field, especially knowledge concerning genetic variants that have been described as 'genes for' traits on the basis of unintentionally inflated estimates of effect size and statistical significance."

I beg the pardon of the first participant to reply, who was annoyed by my submitting full journal citations with links to full text and relevant quotations of the article conclusions. In my own personal opinion, sometimes a Hacker News comment that links to sound research literature is more useful to participants here than expressions of personal opinion based on individual anecdotes. But reasonable minds can differ on such an issue, and I'm happy to add the commentary here that I once had a much different view of the genetic influences on human intelligence (the first popular articles I read on the subject, back in the 1970s and 1980s, were by Arthur Jensen and the late Richard Herrnstein) but I've had opportunity to learn about new research over the last twenty years, and I now know, and like to tell my friends on Hacker News, about the current consensus view that intelligence is largely a trait fixed in the genome of Homo sapiens, linked to hundreds of genes each of weak individual effect, and that most variance between one individual and another in IQ comes from poorly understood gene mutations and environmental influences that may be much more powerful than was once supposed.

I'm downvoting you because this method of cherry-picking large quotes or abstracts out of citations makes for a lot clutter but you don't actually offer us any synthesis. I would rather read your synthesis and read the abstracts if I want to. I came to HN to converse with you, not Johnson, W. and Chabris, C.F.

Just so the current reply to your comment isn't the only one: I really appreciate that you keep track of this information and make it available to others. I still have the attention span for a few minutes' reading, and I prefer a dense text full of citations over some schmuck's strongly-voiced opinion or an anecdote about that one time they went to that one thing. So, thank you.

I've read enough of the half-baked opinions here, including my own. To paraphrase Wikipedia, [citation definitely needed]. +1.

I beg your pardon, but unfortunately this is not my field, so I tend to look for the abstract/intro/conclusion and presume the intermediate parts are correct.. :-)

If I gather the gist of your essay correctly, current research demonstrates no strong connection between IQ and particular genetic influences.

Heritability is very hard for a layperson (such as myself) to wrap his head around, because the meaning assumed by the layperson ("how much a trait is determined by genetics rather than by the environment") is superficially correct, and accurately sums up the intuitive idea that various technical definitions of heritability are based on, but scientific estimates of heritability usually differ from the intuitive definition in significant ways. For example, one common definition of heritability, the one given by Wikipedia as a matter of fact, has a very important twist: "The heritability of a trait within a population is the proportion of observable differences in a trait between individuals within a population that is due to genetic differences." In this definition, heritability explicitly depends on the choice of population. For example, if North and South Korea are considered as a single population, you will get a different result than if they are considered separately. It's easy to misuse population-specific measures of heritability to make inappropriate cross-population comparisons. For example, if intelligence is highly heritable among native-born Dutch and highly heritable among native-born Croats, it might seem reasonable to conclude that IQ differences between Dutch and Croats must be genetic. Yet the same argument would show that the difference in height between North Korea (where height is presumably highly heritable within the population) and South Korea (ditto) must be genetic.

Heritability of something that so strongly is "nature versus nurture" is very hard to pin down. For instance consider a genetic trait that makes kids uglier. (In honor of this article, I'll state the negative version.) Those kids, being less engaging, will get less interaction. It is known that decreased interaction is correlated with stupidity. Therefore all else being equal, that gene will make you stupider. In twin studies that gene will definitely contribute to the heritability of IQ.

However stop and think about it. The fact that this iota of stupidity is heritable is a testament to the impact of environment on IQ, and the fact that the environment that you experience is correlated with who you are.

I see your edit, thanks[1]. But there are better[2] ways to cite in a forum [3].

[1] citation 1

[2] citation 2

[3] citation 3

Why is this better?

Because it is the universal standard, from when the internet protocols involved mail carriers and the packets involved sheets of paper. It is the universal standard because it interrupts the flow of thought as little as possible. If I decide I want read their articles, I will.

It could be even more complicated than that. Intelligence could be a question of proportion, like beauty.

Intelligence could be a question of proportion, like beauty.

Could you expand on that a little, please? How would a working scientist test the hypothesis you advance here?

Think about different subsystem working together.

They might operate better if they are of similar proportions.

In the simplest case, imagine that there are genes that determine quality x, and others that determine quality y, and intelligence depends on how close x/y is to some constant c.

If intelligence were determined by genes in that way, you wouldn't find it by looking for ("smart") genes whose presence was correlated with intelligence, or ("stupid") genes whose absence was.

I used beauty as an example because it seems unlikely that there is a single "beauty" gene whose presence makes human faces beautiful, but rather that beauty is a product of the interaction of multiple genes that control the dimensions of the face.

I would like to elaborate on that idea by saying that there is a feedback loop between the persons environment and his innate tendencies . I'm specifically referring to the feeling of satisfaction you get as you become better and better at something . It is that feeling of satisfaction and the inner drive that would provide the impetus needed to continue working harder and harder to master the discipline . However if all the components aren't in "proportion" , this feedback loop between practice -> improvement -> satisfaction -> urge to practice more gets broken .

I think PG is mainly talking about how the ratio of certain personality elements and basic strengths might contribute to a higher perceived "intelligence".

i.e. Someone with a high EQ (Emotional Quotient) relative to their spatial reasoning might seem more intelligent than the person with lower EQ and higher spatial reasoning, while someone with an equally high EQ and spatial reasoning might develop some sort of personality disorder. (Hopefully not, but just an example)

This could probably be extrapolated into ratios between language skills, creativity, ability to grasp abstraction, memory, etc. etc.

Why is the idea of innate IQ still taken seriously if there's so little evidence for it? If intelligence is just the ability to think in a structured abstract way, doesn't it make sense to think of it as a skill developed over time? I mean, it's not hard to think of spending most of your time in a different mode of thought. I'd say that for most people their normal day to day tends to require concrete or social/emotional information processing.

If it seems that intelligence (specifically abstract structured thought) seems to stabilize early in life, maybe it's something that starts developing even earlier. A strategy for prioritizing, filtering, and processing information from a world of endless stimuli (of which abstract/structured would be only one such strategy) would start developing the moment we're born. By the time you're 10, you'd already have 10 years of "practice".

You are confusing things.

First and foremost you don't need evidence to look for evidence. A hypothesis is enough.

Secondly, there is evidence that IQ is at least in part genetic. There is no other way to explain geniuses, child prodigies, and mentally retarded individuals whose environments vary greatly.

Sure there is: chemical exposure balance or random variation in the uterine or infant environment.

Fetal alcohol syndrome reliably retards kids, and it isn't genetically inherited. Hypothetically, maybe something like adderall works in utero with lifelong effects.

Also, hypothesis formation should be motivated by data, or else you are wasting time searching through an exponential space with 0 probability of finding a correct one.

Serendipity - just the other day I was replying to a comment about education and IQ and explained I didn't believed in ethnic differences of IQ, but also in IQ altogether - http://news.ycombinator.com/item?id=4680279

At least looking for "dumb genes" is more plausible, because if IQ is an aggregate of multiple functions, some functions could be separately damaged

Copy paste:

The current perspective of IQ and ethnicities show that while there might be slight differences, they are dwarfed by the differences in the education process.

Regarding IQ:

First, I don't believe that IQ exist as a measurable quantity - IQ is too many things thrown together in a single bad. Tests group together various things which may be handled by different subsystems in the brain. If there is something called IQ, it's an aggregate.

Anyway, if there was such a unique quantity, actual data shows a progression of IQ scores in time - therefore if the tests are considered accurate and unbiased, it must be a "quantity" that can evolve based on the society a person lives it. It should then be considered not as a value, but as a function depending on a variable called society.

I do not know if studies have tried to measure the accuracy of repetitive measurement in low education adults enrolled in a learning program. If there are such studies and if they show inconsistent result (ie any change of IQ), then IQ should be considered as a function of 2 variables : f(society, personal experience).

Now, even if we consider that at a time t it could be accurately measured and that societal bias could be removed, considering how other qualities (such as determination, work ethic, consistency, creativity, competitiveness...) influence the outcome of any human activity, it seems foolish to rank people based on just one quality - especially if we don't know the other values, and their individual ponderation in the end result.

This ponderation could also be different depending on the activity, and IQ provide an absolute advantage in some activities (rhetoric?), but say determination would give an absolute advantage in other activities (startups?).

>Anyway, if there was such a unique quantity, actual data shows a progression of IQ scores in time - therefore if the tests are considered accurate and unbiased, it must be a "quantity" that can evolve based on the society a person lives it. It should then be considered not as a value, but as a function depending on a variable called society.

Seriously? Couldn't it just be the subjects getting used to the tests? That's not the point though. It doesn't mean subjects are getting smarter. It just means they are learning, which is natural. However if we have two subjects and subject A gets 115 IQ points the first time while subject B gets 104, 111, 113, 115 IQ points after four consecutive test respectively, it's clear that subject A is superior intellectually in that they can achieve more in less time.

That's not exactly the idea. See http://en.wikipedia.org/wiki/Flynn_effect

Maybe a full society is getting used to the tests, or maybe the tests are biased.

Even with you example (considering repetitive test at the same time on 2 person from the same society), it could show many things: a) that the measurement is flawed b) that you should retest A to see if he also get better scores c) failing that, A has a fixed IQ score while B can improve d) if A improves too what's the limit?

Finally, "it's clear that subject A is superior intellectually " - even if B would get say 120 on its final test while A ends with 118??

Which seed will grow better depends not just of its genes, but also of environment, weather and of how much it was fertilized.

This, for example, is one of the answers why rich getting richer. The difference in environment + variety of experience matters much more than difference in genes, in cases where there is no inherited disorder.

By changing environment, behavior and altering mental models (unlearning the learned helplessness, for example) it is possible to improve even IQ scores.

This is how the Flynn effect works - it is effect of the environment + variety of experience.

During commencement at our medical school, one of the second-years on stage said "I know what you're all thinking: you were expecting all sorts of crazy intellectuals and artists. But everyone is so . . . normal."

'Real' crazy intellectuals and artists don't study medicine, they study the Humanities or don't study at all. The rigor and discipline required of medical students and doctors would actually favour those being 'normal'; I wouldn't, personally, have expected anything else.

Reply on news.ycombinator.com