Modern Racecraft

By now you’ve heard that a right-wing scholar by the name of Jason Richwine has resigned from the Heritage Foundation as a result of his previous work on race and IQ. Richwine produced a paper on the costs of immigration, which brought him into the public eye and led some to check his credentials. This inevitably led to the discovery of his Harvard dissertation on low IQ in latinos.

Some folks on the right, including the inevitable Andrew Sullivan, weigh in. They speak as if this in unplumbed territory that is being carefully guarded by political correctness and leftist thought police. I think the best response comes from Ta-Nehisi Coates, “The Dark Art of Racecraft“:

It is almost as though the “dark arts of race and IQ” were an untapped field of potential knowledge, not one of the most discredited fields of study in modern history. We should first be clear that there is nothing mysterious or forbidden about purporting to study race and intelligence. Indeed, despite an inability to define “race” or “intelligence,” such studies are one of the dominant intellectual strains in Western history. We forget this because its convient to believe that history begins with the Watts riots. But it’s important to remember the particular tradition that Charles Murray and Jason Richwine are working in.

Coates shows off the advantages of the blog format by blockquoting selections from previous racial scientists like Lothrop Stoddard. Coates points out that their success rate with predictions is on par with Christian apocalyptics:

One might oppose the Stoddard tradition strictly on its tendency to birth suffering, misery, and catastrophe. But one can oppose it for simpler reasons — its practitioners have a nasty habit of being wrong.

Just as an aside, it’s amusing that this should all happen just as The Great Gatsby is coming out it theaters.

“Civilization’s going to pieces,’ broke out Tom violently. ‘I’ve gotten to be a terrible pessimist about things. Have you read The Rise of the Colored Empires by this man Goddard?’

“Well, it’s a fine book, and everybody ought to read it. The idea is if we don’t look out the white race will be — will be utterly submerged. It’s all scientiic stuff; it’s been proved.”

“Tom’s getting very profound,” said Daisy, with an expression of unthoughtful sadness. “He reads deep books with long words in them. What was that word we ——””

“Goddard” is probably a reference to Stoddard, possibly by joining it with Madison Grant, another eugenicist.

  • http://www.facebook.com/kevin.r.cross.9 Kevin R. Cross

    I wonder, is it a good thing or a bad thing that the first thoughts I had on seeing the headline were about Minecraft and formula one?

    • Msironen

      My guess was powerboat racing!

  • Michael

    The problem is not that Richwine wrote about race and IQ, which as you say has been a frequent topic of study and even now pervades the academic literature. The problem is that he attempted to justify inhumane policies by claiming Latinos were less intelligent than whites (who I guess he thinks make up 100% of Americans).

    Race is of course a very slippery concept, but when we look we do find some consistent correlations between race self-identification and other factors we identify with race and IQ. When we control for more variables, these differences tend to shrink but not go away. What can we conclude? Well it may be a bit early to conclude anything, but it seems that different populations of people do have slightly different median intelligence.

    Assuming it is a fact, this is a mere statistical fact, along the same lines as saying that the mean life expectancy in Switzerland is slightly higher than in Hong Kong. The notion that small variations in aggregate statistics should have any significance in the way we treat people is astonishingly simple-minded.

    (I would also like to point out that I recognize racial distinctions are not biologically justified. Combining this with the fact that it is impossible to control for all external variables when it isn’t even clear which variables constitute “race” as we mean it, and the fact that IQ is notoriously unreliable when taken across populations, and we can begin to see the difficulty with these sorts of studies. So I do not take the position that current data on the subject are reliable. Even so, I would not be surprised if there were small racial variations in IQ, much for the same reason there are national barriers, religious barriers; etc. Specifically: There is a degree of cultural and reproductive isolation between races.)

    • f_galton

      He didn’t “claim” average Latino IQ is lower than whites’, he reported that fact.

      • Michael

        He did far more than that. He declared that giving amnesty to Latino immigrants would be too “costly,” specifically because Latinos are less intelligent than whites. I cannot see any way to understand his position other than outright racism, or his proposed policy other than racial discrimination.

        • f_galton

          Latinos are, on average, less intelligent than whites.

          • kessy_athena

            And do you have anything at all to back up such a claim other then pure, naked racism, bigotry, and hatred?

            • f_galton

              The government has data on welfare usage, you can look it up. While you are at it, look up the data on Hispanic educational attainment.

            • kessy_athena

              LOL You seriously expect me to do your research for you? What was that you were saying about people being lazy?

            • f_galton

              I’ve done the research. You were doubting the fact that Hispanics use welfare at higher rates than whites do, because you make things up.

            • Michael

              You were doubting the fact . . . because you make things up

              That doesn’t even fucking make sense. You can’t “make up” doubt.

              Latinos receive substantially less in welfare than whites or blacks in the US, but by percentage of the population they do receive somewhat more than whites. However, this is clearly consistent with their economic status. You are effectively pointing out that poorer people receive more in welfare than richer people.

            • f_galton

              She was doubting the fact Hispanics have a higher rate of welfare usage than whites. It’s nice of you to admit they do. The fact that Hispanics tend to be poor and use welfare is a good reason to place limits on them immigrating here.

            • Michael

              A rational person would say that the fact that the poor receive more in welfare may be a reason to limit the immigration of the poor. However, you seem to be absolutely bent on basing policy on racial categories, so because race is correlated to income, you think it would be better to discriminate based on race than the relevant underlying variable.

            • f_galton

              Fine with me. Of course the result would be a more restrictive immigration policy than Richwine was arguing for, and that being said, there is nothing wrong with basing policies on racial differences in situations where such differences matter.

            • Michael

              As I understand it, Richwine was arguing for the status quo, in which it is nearly impossible for Mexicans and other Latinos to immigrate to the U.S. So no, I don’t think my suggestion would be more restrictive.

              I also don’t think it is necessary. There is not much evidence immigration represents an overall significant or unaffordable cost to the U.S.

            • f_galton
            • Michael

              These are legal immigrants. So not only are they biasing the sample by only considering people already eligible to get visas, they are biasing it by only considering people who do not need to become citizens because there is no legal requirement. And of course, these are only legal immigrants who are eligible for citizenship; many are ineligible.

              Furthermore, if you actually READ anything you cite, rather than just looking at the pictures, you will find that there clearly ARE significant barriers. 17% of the respondents to the study say they need to learn English while another 6% say the citizenship test is too hard. So language requirements and the test requirement represent barriers to some legal immigrants. More significantly, 17% say they cannot afford the application, which at the time of the poll was $680 per person. So the cost represents another barrier. And so on.

              You can simply declare that Mexicans are lazy and stupid, or you can recognize that people who are already legally residing in the U.S. don’t want to spend hundreds of hours learning a new language and culture and hundreds of dollars to take a test on it.

            • f_galton

              It’s obviously grossly inaccurate to say it’s “nearly impossible” for Hispanics to legally immigrate when so many are doing it, or it least it should be obvious.

            • Michael

              It is nearly impossible for many people. It is trivially easy for some. You gave a study that only considered people who were eligible for citizenship to prove how easy it is to be eligible. I can’t believe you don’t see how useless that is.

            • f_galton

              The fact that millions of Hispanics have been allowed to immigrate here suggests that it’s not “impossible”.

            • Michael

              I did not say it was impossible.

              However, for some, it is indeed impossible, and for many it is effectively impossible.

            • f_galton

              “There is not much evidence immigration represents an overall significant or unaffordable cost to the U.S.”

              Vastly expanding the lower class has all sorts of costs.

          • Nox

            Anecdotal, but at least one white person here is a complete fucking moron.

    • f_galton

      Race is biologically real. If you want to refuse to recognize reality that’s your business, but you shouldn’t get upset at people who don’t share your delusions.

      • Michael

        Race is a biologically meaningful distinction in many organisms, but not in humans. If you want to understand why this is the case, I could direct you to good resources. If you just want to reinforce your own predetermined conclusion and don’t care what science has to say, then I’m afraid I can’t help.

        • f_galton

          That’s a political statement, not a scientific one.

        • f_galton
          • Michael

            That chart isn’t even labeled. I don’t know what it’s supposed to represent. Is it some sort of cluster analysis?

            And anyway, you don’t understand the issue. Nobody is arguing that
            there is not localized diversity, just that humans cannot be
            meaningfully grouped into any particular number of defined “races.”
            Consider this
            statement
            from the American Anthropological Association, which
            points out that 94% of genetic variation occurs
            within racial categories, not between them, that the
            other 6% varies gradually across geography, making any specific
            divisions arbitrary, and that our current notions of race originate from
            medieval superstition and encompass far more than just physical
            differences.

            If you have access to a library, you can check out this article by Weiss and Fullerton which makes the salient point that while you can of course find genetic differences between distant populations and thus define all individuals as some mixture of these groups, you can do the same with ANY arbitrary choice of distinct populations. All this reflects is the initial hypothesis (or bias) of the researchers. There are no obvious natural divisions.

            • f_galton

              Humans can easily be grouped into races, and it’s not arbitrary at all, look at the population structure.

              “you can of course find genetic differences between distant populations”

              No kidding. Humans living in different places evolved differently, so there are racial differences.

              “you can do the same with ANY arbitrary choice of distinct populations. All this reflects is the initial hypothesis (or bias) of the researchers. There are no obvious natural divisions.”

              Look at that image again. Examine enough loci and the misclassification probability approaches zero.

            • Michael

              You still haven’t sourced that graph, which is unlabeled and out of context. However, I was able to find its source here.

              When taken in context, the data does not support your claim of racial essentialism. First, note that the study performed a sort of cluster analysis, in which the number of groups must be determined beforehand. In fact, they performed separate analyses on each k=1 through 7, What you have shown is the graph for k=7. This cannot be used as evidence to suggest there are seven races, because this number was chosen arbitrarily before the results were produced (indeed, in order to produce those results).

              At any given k, we can group human diversity roughly by geographic ancestry (though if you read the study, you will see that most of the listed groups in fact have genetic contributions from many regions beyond the inferred component). At k=5, we get continental groups. At k=7, we see SE Asia, the Middle East, and Europe being separated. And at k=6, we get an intermediate between these two.

              Furthermore, the regions that appear homogeneous in that graph actually demonstrate a substructure of their own upon individual regional analysis. I’ve attached an image (again, from this same study) of a principal component analysis showing clear differences within the European and Middle Eastern populations. Are we to say that Italian, Sardinian, and French are all different races?

              And then there is the issue of “mixed” or “misclassified” individuals. As I said, some individuals’ genetics can be primarily explained by the inferred component, but many cannot, and appear to be a mixture of multiple components. As this study points out, in many cases this cannot be explained in terms of recent admixture. It gives the example of the Middle Eastern population which has had continuous gene flow with the rest of Eurasia for most of its existence.

              What you are seeing here is the classical example of clinal variation. Humans exist in a genetic continuum, not in isolated populations. Even if we suppose that nobody from Eurasia ever mated with anyone more than fifty miles away from their birthplace, there would still be no natural barriers to genetic markers spreading across the continent. Genetic drift in the far East would take a long time to propagate to Europe, and you could determine how far East someone lived by their genetic makeup, but you could never determine at which point on the EW axis people stopped being Asian and started being European.

              The issue of racial essentialism is not that there are not genetic differences between geographically distant populations, it is that there is no natural way to determine group them. What defines a “race”? If there are three human races, then I am the same race as a native American. If there are five, then I am not, but I am still the same race as an Arab. If there are seven, then again I am not, but I am still the same race as a Russian. If there are 30, then again I am not, but I am still the same race as a Scott, and so on. Furthermore, even for a given number of races, it is arbitrary where we draw the line. As I pointed out above, the Eurasian supercontinent gives the best visual example of this, as there is no point on the EW axis at which a discontinuous change occurs in the genetic structure.

              It’s also important to note that this graph only considers markers on noncoding DNA (perhaps it is thus slightly inaccurate to call them “genetic” markers, as they are not in genes but between them). These markers are not subject to selective pressure, and thus the variation you see is only due to genetic drift. But when we talk about human races, we are talking about phenotypic differences, and these can only come from genes, which are in fact subject to selection. When the researchers in this study performed an analysis of molecular variance on strictly noncoding DNA, they found that 9.0% of variation was between groups, while when they performed the analysis on microsatellite markers (which affect gene expression and are thus subject to slight evolutionary pressures), they found that only 3.7% of variation was between groups, and an astonishing 94% was within populations. I have attached that image as well (F3, left side).

              Nowhere in this study do the researchers suggest these statistics support meaningful racial distinctions between human populations.

              Next time, please do not present unlabeled charts out of context alongside an assertion of their meaning that does not fit the data they represent. And please do not presume to know me, whom you have never met.

            • f_galton

              “they found that only 3.7% of variation was between groups”

              To put that in perspective, the human-chimp difference is 4%.

            • kessy_athena

              (facepalm) Oh for crying out loud… About 4% of the total genome of humans is different from chimpanzees (more or less). About 0.1% of the total genome of humans varies by individual. 3-7% **of that 0.1%** constitutes variations between groups. That means that 0.003% – 0.007% of the total human genome varies by group.

            • f_galton

              “the regions that appear homogeneous in that graph actually demonstrate a substructure of their own upon individual regional analysis”

              No kidding. You can subdivide the major groupings.

            • Michael

              yes, you can subdivide the “major groupings” no matter how large or small you define the “major groupings” to be. There is also overlap between the groups and no natural boundaries. We can choose groups such that they are well-defined and we can easily identify people within them, and we can furthermore define them in a statistical manner so as to maximize the between-group differences, but these are still arbitrary and do not reflect any underlying biological truth. This should be evident when you consider than 96.3% of variation occurs within these groups. Thus the genetic difference between two people of the same “race” is barely distinguishable from the genetic difference between two people of different “races,” ultimately coming down to the individual.

            • f_galton

              You’re saying discernible biological differences do not reflect underlying biological differences.

            • Michael

              You are correct that human genetic diversity is similar to human height diversity. There are short people and there are tall people but there is no natural way to divide the population into “the short” and “the tall, because in fact most people are near the middle. So it is with race, where there are clear geographic differences, but they vary clinally and most people are in between any arbitrary groupings.

              No, we can detect groups by analyzing DNA
              Did you read anything I just wrote? The study itself clearly explains how you can create arbitrarily many divisions using the statistical analysis, and how the inferred component from their PCA was not responsible for most of human diversity. Yes, there have been bottlenecks created by migration that lead to a sort of human pedigree, but this is only responsible for a portion of the 3-4% of diversity that is localized at all. Furthermore, using this as a definition of “race” should imply that all populations are separate races, or else that they are all the same race.

              You’re saying discernible biological differences do not reflect underlying biological differences.

              That’s not what I said. I said that the fact that you can perform this sort of statistical analysis on human diversity does not reflect any underlying biological truth. That is to say, just because you can group people doesn’t mean those groups are meaningful.

            • f_galton

              “most people are in between any arbitrary groupings”

              Actually, no, most people belong to five major populations that align with the observable, historical conception of racial groupings.

            • Michael

              No they fucking don’t. I don’t know how much clearer the evidence could be. First of all, the “observable, historical conception of racial groupings” is based entirely on superficial characteristics like skin color that are of almost no biological consequence. Second, historical groupings are not based on actual human migration patterns, and so place Australian aboriginees in the “Negroid” race, and Native Americans in the “Caucasoid” race, and so on. Third, the individuals in the study were not all representative of their assigned group; only the aggregate showed meaningful patterns (so a given “Asian” person might be genetically closer to “white” people). Fourth, the majority of populations studied were mixed, not only because of recent admixture, but also because the groups themselves were not isolated. Fifth, the suggestion that five is somehow the magic number for races is completely unsupported; the study suggests seven, although every K gives similarly convincing results. Sixth, none of this had anything to do with intelligence, except that it demonstrates your lack.

            • f_galton

              “Scientists studying the DNA of 52 human groups from around the world have concluded that people belong to five principal groups corresponding to the major geographical regions of the world: Africa, Europe, Asia, Melanesia and the Americas.

              The study, based on scans of the whole human genome, is the most thorough to look for patterns corresponding to major geographical regions. These regions broadly correspond with popular notions of race, the researchers said in interviews.”

              http://www.nytimes.com/2002/12/20/health/20GENE.html

            • Michael

              Please stop citing news articles as sources.

              Unsurprisingly, the NY Times doesn’t provide the name of the study they are discussing or provide a link. Nevertheless, I was again able to find it here. Imagine my surprise when I found that the study did not find five major groupings of humans but rather six, something the Times article failed to mention.

              The article points out that 95% of genetic variation was again found within groups, and that only 7% of alleles found were exclusive to a single group. Moreover, these alleles were extremely rare, not exceeding a frequency of 1% in their region.

              In fact, the results here are broadly consistent with the last study you cited. It also examines results from k=2 through k=6, finding that each gives geographic regions until k=6, which is not geographical and instead reflects cultural barriers in Pakistan. It also gives this quote, nearly the same as the last article:

              In several populations, individuals had partial membership in multiple clusters, with similar membership coefficients for most individuals. These populations might reflect continuous gradations in allele frequencies across regions or admixture of neighboring groups

              Look, if you split humans up into five ancestral groups, you will find they are fairly geographically localized. You find the same thing if you split them into three or ten or fifty. This does not imply there are three, five, ten, or fifty races. It does not change the clinal nature of the variation; humans exist in a racial continuum.

            • f_galton

              Because racial categories lack sharp boundaries does not mean the categories do not exist.

            • Michael

              Racial categories clearly do exist, but they do not have a biological basis. They are arbitrary divisions of a continuum.

              Arbitrary categories still exist.

            • f_galton

              Physical, mental, and DNA differences are all biological, and the divisions are not arbitrary.

            • f_galton

              The sixth group was the Kalash, who live in northwest Pakistan.

            • f_galton

              The genetics of the Kalash is quite interesting, at least for people who don’t reject science.

              http://en.wikipedia.org/wiki/Kalash_people#Genetic_origins

            • f_galton

              Historically Native Americans were not considered Caucasian. The populations are not mixed, and there isn’t lots of recent admixture, and the groups were isolated.

            • Michael

              You can’t just make a bunch of unsourced claims and expect people to believe them.

              I already pointed out where in the study YOU referenced it says that most of the populations were mixtures of groups, and you can see this by looking at the figure you posted. It explains specifically that in some cases this is due to recent admixture and in others it is due to shared origin between multiple groups (such as the Russians and the Yakuts in the case of Native Americans).

            • f_galton

              Everything I’ve said is accurate. Unlike you, I’m not in the habit of denying science or making things up.

            • f_galton

              You are detecting evolved differences. These differences are meaningful if you think things like intelligence and health are meaningful.

            • f_galton

              “We can choose groups such that they are well-defined”

              No, we can detect groups by analyzing DNA.

              “these are still arbitrary”

              No, it’s shared ancestry

            • f_galton

              “There is also overlap between the groups and no natural boundaries”

              Using that logic there are not short or tall people, either.

            • f_galton

              “Nowhere in this study do the researchers suggest these statistics support meaningful racial distinctions between human populations.”

              Among other things there are observable physical differences between populations, or races.

            • f_galton
  • f_galton

    You think the “best response” comes from a blogger at the Atlantic with no background in statistics and science and psychometrics? That’s just dumb.

    • Fred

      I think the best response comes from the person who addresses the points of others and then explains themselves better. You’re free to continue to gripe about this or you could step up your game a little better than just posting a bare link to a website.

  • Noelle

    Meh. Doing well on IQ tests doesn’t make you intelligent. It only means you do well on IQ tests.

  • f_galton

    “I think the popular conception of intelligence is really mostly a myth based on cultural factors, not biological ones”

    Because you don’t what you are talking about.

    • kessy_athena

      Bravo, such an eloquent ripost truly showcases your command of the facts and ability to construct rigorous arguments. Although you did forget to call me a stinky doody head. Jabster, I think you’ve found a kindred spirit.

      • f_galton

        I do have command of the facts. You name call and assert things based on your feelings.

  • f_galton

    “The vast majority of humans can become competent at almost any skill if they are willing to invest the time and effort”

    So you’re saying Hispanics are lazy.

    • kessy_athena

      “So you’re saying Hispanics are lazy.”

      This statement merely demonstrates what I think everyone here already figured out – that you believe that Hispanics are inferior to whites, which therefore makes you a racist.

      • f_galton

        You asserted that the “vast majority of humans can become competent at almost any skill if they are willing to invest the time and effort”. If that’s true, then the low educational attainment among Hispanics is a choice by them.

        • kessy_athena

          (sarcasm) Why yes, because obviously every human on the planet has exactly the same opportunity and access to American style education, regardless of income or geographic region. (/sarcasm) Let’s see, just how full of holes is that argument? You need to actually show that there is a significant difference in the educational attainment of Hispanics and other groups. That includes controlling for factors like income and the educational level of parents. In the case of immigrants, you have to look at the education available in their place of origin. Then you have to demonstrate that this has something to do with laziness. Then you need to somehow connect this back to your original argument that Hispanics have lower IQs and are less intelligent then whites.

          • f_galton
            • Michael

              Galton, That table pretty clearly shows that the performance difference cannot be primarily attributed to genetics, seeing how different generations performed substantially differently. There are clearly cultural barriers.

              Kessy, if there is no general intelligence factor, then how do you explain the psychological fact that all cognitive abilities are positively correlated?

            • f_galton

              You should look up what scientists have to say about the heritability of intelligence.

            • Michael

              I don’t know how you can look at that chart and say the results are genetic. Why do intermediate generations perform better (and indeed comparably to white populations) in school if they are genetically the same “race” as their parents and children?

              And yes, the results are very bad. In order to improve the situation, we should look at what the evidence suggests is the actual cause of the difference, rather than using outdated assumptions.

              And of course IQ is heritable, but that does not mean there are significant differences between geographical regions. Even if there are, that does not justify racially discriminatory policies.

            • f_galton

              Hispanics have lower average IQ’s than whites, but for cultural reasons they under perform their IQ’s. No one is advocating racially discriminatory policies. You think we should so something to “improve the situation”, but nothing has worked so far. It’s silly to expect you will be able to make the US more Hispanic, and then change things.

            • Michael

              The difference between white and Latino median IQ is less than one standard deviation. Stating that we should not give amnesty to Latino immigrants specifically because they are Latino and thus stupid is blatantly racially discriminatory.

              Improving academic performance of minorities is a separate question and not one to which I claim to have any answers.

            • f_galton

              The gap is significant. Over 60% of Hispanic high school graduates do not score high enough on screening tests to enter the military, for example. It’s not “discriminatory” to not give a particular group amnesty.

            • Michael

              It’s not “discriminatory” to not give a particular group amnesty.

              That is literally the definition of “discrimination.”

            • f_galton

              The US is not obligated to accept or amnesty any immigrants, there’s nothing prejudicial or unfair about it.

            • Michael

              It is discriminatory to exclude a particular group. It is racially discriminatory to exclude that group because of its race.

            • f_galton

              Even if the outcomes were primarily because of culture, or cultural barriers, which they are not, the results are dismal.

            • kessy_athena

              Well, I’m not real familiar with the state of research on intelligence, and I’m speaking mainly from personal experience, so I’m pretty uncertain about it and am simply speculating at some points. I’m inclined to think that the appearance of a general intelligence factor from correlations between different skill tests is a statistical artifact, largely arising from cultural biases both in how the tests are designed and in the sort of life experiences people have (by which I mean it’s a case of nurture over nature.)

              I’m not at all clear on exactly which skills are being measured or how strongly correlated they are. However, my impression is that the tests are mainly geared toward memory, reasoning, math and verbal skills. Do you know if that’s right? If it is, it leaves out a bunch of skills that I think are important, particularly emotional and interpersonal skills. That’s one of the big cultural biases that I see – we’re only defining intelligence as including a certain subset of skills that are culturally favored as being more important then others.

              I also think a lot of what we perceive as intelligence is related to behaviors driven by social roles and pressures. “Geeky” kids are expected to be smart, and so are more inclined to learn and practice the sort of skills that get tested for measuring intelligence. Of course there’s also the question of access to education, and how much different subcultures value academic achievement. It’s very hard to really fully disentangle measuring ability from measuring achievement.

            • Michael

              The results are pretty resilient against various tests for bias, and anyway I think it is a fairly obvious fact that smart people tend to be smart in many ways. It is true that the tests focus primarily on memory, reasoning, math, and verbal skills, because these are what are generally considered to constitute “intelligence.” If you think, say, empathic or musical abilities should be considered “intelligent” too, then you are just using a different definition; but that definition does not allow for a coherent understanding of intelligence as a singular property, because things universally considered to be indicative of intelligence are not well correlated with them.

              The question is not whether we focus on specific “culturally favored” aspects, but whether we focus on those that are apparently cognitively related. IQ tests are not an attempt to find the “best person,” and the notion that more intelligent people are somehow superior to or more fully human than less intelligent people is a dangerous misconception.

              IQ tests are not measures of education. Although it is more difficult to test people who are very uneducated, there are nonetheless ways to do this, and IQ does not change substantially after education.

              ­

              I’m not trying to say that intelligence is not a complex or hotly debated issue, but there are some things we can say about it, and simply dismissing it as a “statistical artifact” is unwarranted.

            • kessy_athena

              I actually mostly agree with you here, and I think our differences on the final conclusion are mainly about definition of terms. I do have to kind of wonder, though… When you say that, “That definition does not allow for a coherent understanding of intelligence as a singular property,” aren’t you essentially saying that intelligence is a singular property because you’ve explicitly defined it to be so? Not that there’s anything wrong with that; it just makes the existence of a general intelligence factor a starting assumption instead of a conclusion.

              My impression has always been that the main reason to be interested in the idea of how intelligent individuals are is because it serves as a predictor of things like an individuals ability to perform a demanding job. Since almost all jobs involve working with others, wouldn’t the ability to correctly read other people’s non-verbals be as important as, say, math skills? Just as an example. My feeling is that the way you seem to be approaching intelligence is overly narrow to reflect an idea of general competence that I think is the core of the informal conception of the term.

              You’re absolutely right that intelligence is *not* about finding the “best person,” but unfortunately in the real world that seems to be how it generally winds up being used. I don’t think it’s a coincidence that high IQ tends to correlate with high social status. And in my personal experience, people seem to have a tendency to use being intelligent as a way to look down on others. I think in some ways there are parallels with how the idea of race is used. Race is problematic not really just because of the idea of grouping humans into general categories of common descent, but because of all the stuff that goes along with it. Stuff like the racist human garbage that’s wandered in to the site in this thread.

              Dismissing intelligence as a statistical artifact is a bit stronger of a statement then I really intended. I just think that intelligence is a much more complex and multidimensional thing then just a single scalar measurement.

            • kessy_athena

              Actually, I don’t think I really explained well what I meant by the influence of social roles and education. You’re right, IQ tests are designed to test natural ability, not education. But as a practical matter it’s difficult to fully disentangle the two. And I would suggest that education and experience can effect ability. For example, the conventional wisdom is that learning a second language is more difficult then learning third, fourth, etc languages. Exercising a set of mental muscles makes them stronger, as it were. Kids who are labeled as being smart are encouraged to work harder at school, read more, do more extra credit assignments, do more academically oriented activities (Physics Olympics FTW!) and take more challenging and advanced courses. I think there’s an aspect of a self fulfilling prophecy to it.

            • Michael

              I mean, you can say these things, but the evidence does not support it. The test-retest reliability of IQ tests is very high, even with intervening education. For example, the WAIS-IV for children has a test-retest reliability coefficient of about 0.9, even with three years of school intervening. The subscores have slightly lower retest reliability, as you would expect.

              IQ tests are excellent predictors of scholastic aptitude, job performance, and salary later in life, among many other things. What few tests exist for other proposed sorts of “intelligence” are not (or at least not yet). Current intelligence testing is clearly not perfect, but it is pretty good in a number of ways.

            • kessy_athena

              IQ tests are normalized to have an average score of 100 with a standard deviation of 15, and are different for different age groups, correct? That means that having consistent scores on multiple tests over time indicates that individuals tend to maintain their relative levels of performance on the tests – it doesn’t measure how an individual’s aptitudes change over time.

              And in my view, things like scholastic aptitude and most especially salary primarily reflect social status.

            • Michael

              Obviously they are not attempting to assess absolute intelligence during development. Every child gets smarter as they age. The only relevant question is if they are of the same relative intelligence, which it seems they are.

              And yes, it does measure individual “aptitude,” because aptitude of a child should reflect his or her ability to develop.

              And of course intelligence is beneficial for social status. That’s one of the reasons humans became intelligent in the first place.

            • Alex

              IQ tests are supposed to do One thing, and they do it well. They rank you against your peers of age vis-à-vis RAW INTEL.

              Scores are **supposed** to follow a Normal distribution: It turns out intelligence is excellently Normally distributed within an age group. Some come out high (for various reasons), some come out low (for various reasons). But it doesn’t matter why: IQ only measures INTELLIGENCE, not the reasons. (And with certain assumptions, it’s trivial to show via probability theory why everyone’s test scores should be normally distributed)

              And IQ tests show aptitude **relative to peers,** not your improvements relative to an older, crappier you. Individuals move up and down IF AND ONLY IF they advance *faster* than their immediate peers. It’s not meant to show any other kind of improvement. It’s a convenient well-ordered comparison of people.

              And it’s correlated far stronger with parent’s IQ than parent’s personal income. Smart people raise smart children, b/c they have better genes, and they do more things right. They eliminate confounding variables (e.g. beating the child, religifying the child, etc), and their children end up higher on the distribution.

              Moreover, IQs (and other tests) are NOT fully causal to personal income. One’s IQ is only slightly correlated with one’s income (r^2≈0.5). Much more important is geographical location and parent’s income. These variables are all mildly correlated, but it’s Lightyears from complete causality.

              @Kessy, don’t go spraying garbage about how “scholastic aptitude… reflects social status”, just because of your own experiences. There’s a mountain of education literature that contradicts this. On a numerical basis, the “scholastically apt” children of the poor vastly outnumber the “equivalently scholastically apt” children of the rich.

              >>”salary primarily reflect[s] social status”
              Well no shit. Personal income is almost the definition of social class. A meaningless statement.

            • kessy_athena

              re IQ tests, that’s what I was saying. They only rank you relative to your peers. They tell you very little about how intelligence changes over time, if it changes over time, or even really the nature of what exactly it’s measuring. They don’t tell you important things you’d need to know to investigate those sorts of questions, such as whether the distribution gets wider or narrower over time. IQ test scores just aren’t a very good tool for investigating the nature of intelligence.

              What IQ tests actually measure is relative proficiency in a certain set of skills. The selection of which skills are in that set is a cultural choice. Some of those skills have little real world utility. Most notably skill at taking standardized tests. And there are lots of skills that are essential in the real world that are not measured at all. Such as social skills. The favoring of those skills tested by IQ over those skills that aren’t is purely a cultural bias.

              And how exactly do you define “raw intel” anyway? And how do you distinguish it from learned skills? Especially since long term patterns of brain activity actually alter the architecture of neural connections. Neural pathways that are used a lot tend to get reinforced, and ones that aren’t used tend to atrophy. And what makes you think that raw intel is genetic? Just because something is inherited doesn’t necessarily mean it’s genetic. For example, there are a wide variety of epigenetic factors that influence gene expression. And heck, for that matter, culture is inherited. And if intelligence were a single, simple attribute that’s very strongly genetic, you’d expect the genetics to reflect that. But instead we find that the genetics of intelligence are very complex indeed. Wouldn’t that suggest that intelligence is also similarly complex?

              And I’m calling you on abuse of statistics. Of course there are more “smart” poor kids then “smart” rich kids. There are a lot more poor kids then rich kids. Raw numbers in this context are pretty meaningless – it’s rates of occurrence that would be relevant here. And social status is a lot more complex then just money. There are many, many different subgroups withing society, all of which have their own status hierarchy. Those groups often overlap, and furthermore there’s a complex dynamic of the relative status of different groups. For a lot of people there’s a tremendous allure to belonging to a group (such as “smart” people) that confers a sense of worth and value. I completely reject the notion of any sort of personal worth that is based on anything other then a person’s actions. Value that isn’t earned by what you do isn’t value at all.

            • f_galton

              “I’m not real familiar with the state of research on intelligence”

              I agree.

  • f_galton

CLOSE | X

HIDE | X