Data? We Don’t Need No Data

In the July 18 issue of The American Conservative Ron Unz, its publisher, wrote a thoughtful article reanalyzing Richard Lynn’s collection of IQ test data from around the world. His conclusion was that IQ test scores in eastern and southern Europe had increased dramatically along with improving socioeconomic conditions in recent decades and that this response to environmental improvement was also happening in the US Hispanic population. He also pointed out that in East Asian populations IQ scores had remained static and that there and uniquely there no apparent response to environmental improvement occurred.

His article was remarkable for its gracious and even tone with extensive citation to data, not only from Lynn but from “quant bloggers” who follow the literature on ethnicity and cognitive testing. A flurry of responses and critiques soon appeared, and these are detailed and linked in Unz’s followup article. He points out that serious and useful criticisms come almost entirely from “racialist” bloggers while none come from what he might call “anti-racialist” bloggers. There are several interesting aspects of this debate that encourage us to hope for higher grade public awareness and discussion of ethnicity and IQ.

First, the debate is civil and scholarly and polite. Unz feels that Lynn should have noticed what he (Unz) did in Lynn’s data while he acknowledges Lynn’s important contributions to the literature. Second, there is so far no useful analysis from the “liberals” at all. (I use the L word following convention but I don’t think it is right: I am a liberal and most of them are not.) This confirms my long-held impression that diversity deniers are not very numerate and that down deep their assertions are shallow.

The same collegiality is apparent in responses to Unz by Lynn himself and by Helmuth Nyborg at the American Renaissance website. There is a real prospect of useful understanding of the issue to emerge from all of this.

Third, Unz refers unselfconsciously to “racialist” blogs, often called “HBD” blogs, without loading his term with any value. Again, exactly the right touch.

There is plenty of talk about race, ethnicity, and IQ available on the web but it is dominated by sites with lowbrow commentariats that are not very helpful, for example at one extreme sites like this and at the other occasional bouts at sites like this or this. There is hope for us after all.

Unz’s political take in his article is that there are no grounds for concern about low IQ illegal Mexican immigration into the US because within a decade or two the IQ of that population will pop right up to Irish or Dutch levels. I personally favor mass deportation, not on IQ or social capital grounds, but tree hugger that a I am, on the need for us to get our population down and preserve things like wilderness and farmlands. It might not be so difficult: there must be a large number of FEMA trailers that could be used to stock processing centers and in Utah, the site of several WWII Japanese internment camps, plans and blueprints must exist to reconstruct those camps.

This entry was posted in Uncategorized and tagged . Bookmark the permalink.

56 Responses to Data? We Don’t Need No Data

  1. gcochran9 says:

    I think this is nonsense. Unz shops for data outliers, which are probably just wrong. I have lived in New Mexico for almost twenty years now. and I begin to know something about the state. New Mexico has been US territory for 164 years – and every week or so, the local rag has another dumb-as-a-rock article about THE GAP – the mysteriously lower academic performance in Hispanics and Indians, which nobody can budge.

    6, going on 7 generations, and no convergence. Don’t hold your breath.

  2. FOIA says:

    Unz Foundation 990 tax form:

    irs990.charityblossom.org/990PF/200912/207181582.pdf

    $500,000 grant to Greg Cochran, Henry’s coauthor, and a $24,000 grant to Steve Sailer. Probably ought to be disclosed.

  3. j says:

    The 3000 dollar grant to Finkelstein is data that taints the whole gang.

  4. gcochran9 says:

    Make that $600,000. Nevertheless, I don’t believe that Ron Unz is at all correct on this.

    • QU says:

      So why not explain the details of the cherry-picking, etc, if only as a favor to him, since he seems to be requesting it?

  5. DB says:

    The incentives to self-report as Hispanic have changed over the past few decades, in a direction that makes Unz’s reading of the data overly optimistic.

  6. JL says:

    Unz is indeed basing his theories on a few outlier data points. Had he used one of the newer and more extensive national IQ data sets that Lynn has compiled (instead of the old 2002 compilation), he would probably not have been able to make his case at all. Looking at all the countries in the data, there is little evidence of convergence over time due to economic growth and urbanization. The subtext of Unz’s enterprise seems to be that Hispanics are interchangeable with whites. In attempting to make this point, he argues like a lawyer rather than a scientist. It was the same with his Hispanic crime article a couple of years ago.

    However, in light of the information posted above, I must say I appreciate his sponsorship of some of my favorite writers and thinkers.

  7. Kiwiguy says:

    There has been an excellent series of posts discussing the Unz article by the blogger “Chuck”. He has looked at looking at performances across different generations using data from PISA, TIMMS & National Longitudinal Survey of Youth. Ron Unz and Peter Frost have both responded in the comments.

    http://occidentalascent.wordpress.com/2012/07/28/hispanics-the-nlsy-97/

    http://occidentalascent.wordpress.com/2012/07/26/hispanic-performance-by-generation/

  8. Toddy Cat says:

    Personally, I think that Unz is probably wrong about this matter, and it’s pretty clear that he’s biased, but at least he’s putting forward a case, rather than just screaming “racist!”. I suppose we should be grateful for that.

    • harpend says:

      “Grateful” is absolutely right. I am hoping that this is the dawn of a new route to understanding about group differences. Let the numerate come out of the dark and argue with each other while ignoring the Nazis and the screamers-of-racist on the edges.

      • The fourth doorman of the apocalypse says:

        Imagine that Homo erectus was still with us. Would the more intelligent of them appreciate a new understanding of group differences? Would certain individuals/groups within Homo sapiens take the protection and preservation of Homo erectus as a means to extract wealth from the rest of Homo sapiens?

      • typal says:

        No chance of a shared understanding emerging if you can’t even convince someone as independent and intelligent as Ron Unz. It’s naive to think a debate between the numerate and presuppositionless (who don’t exist) is anything but academic to the inevitable real world outcome.

      • gcochran9 says:

        I wouldn’t say that.

      • typal says:

        But if you think that GC, why haven’t you bothered to give a closely argued critique ?

        Ron is the kind of person (a real world success) that ‘academics’ would have to convince to get policies changed.

  9. Ron Unz says:

    Well, it’s since to see “gcochran9″ has now provided a few sentences of negative commentary on my lengthy article somewhat more substantive than just wandering around the Internet, dismissing my 9,000 word analysis as “cherry-picking” based on “crumbs of data” and “not worth discussing,” and myself as basically a silly fool as he had previously done.

    Obviously, his personal experience of having lived in New Mexico for twenty years vastly trumps all the detailed statistical evidence I have provided from Lynn, the GSS, the SAT, PISA, and numerous other highly reputable quantitative sources. Would that all of us could had such infallible insight into all matters great and small acquired from visits to the local grocery store.

    Since Cochran obviously has zero interest in actually examining the data and prefers to rely upon the anecdotal evidence from his home state, perhaps I should reiterate a point I have previously made on several occasions.

    Consider West Virginia. I’ve heard its white population is perhaps the most purely British-ancestry of any state, and certainly has American roots going back centuries. Meanwhile, the social, economic, and academic performance of those whites are absolutely dismal, perhaps the worst in the country. Just to cite one minor statistic, on an age-adjusted basis West Virginia whites are imprisoned at a rate three to four times higher than that of West Virginia Hispanics. Furthermore, most of the surrounding and also very heavily British-ancestry Appalachian region has similarly poor performance numbers.

    Does that prove that British-ancestry whites are generally stupid? Not really—some of the smartest and most successful people I know grew up in West Virginia or their parents did. I’d think the main factor is that the state is very rural and for the last century or two, a good fraction of the most able and ambitious individuals of each generation have moved away to the Big City elsewhere, thereby leaving behind a residual population of much lower quality, which also gradually produced negative cultural consequences.

    And as far as I can tell, New Mexico is essentially the Hispanic West Virginia. For example, its urban crime statistics tend to be dramatically worse than those of other cities with similar (overwhelmingly Hispanic) demographic characteristics. But that’s all “quantitative nonsense”—anecdotal evidence from grocery store visits is far superior…

    Incidentally, I have since added perhaps another 9,000 words of follow-up analysis to my Race/IQ column, all conveniently located at http://www.ronunz.org, while my very detailed 2010 analysis of the crime figures I mentioned above is at http://www.theamericanconservative.com/articles/his-panic/

    • dave chamberlin says:

      A heartfelt thanks for funding the writting careers of Greg Cochran and Razib Khan. If it hadn’t been for your generosity and your damned good intellectual taste they wouldn’t have had the free time to reach me and many others like me with their never ending supply of thought provoking ideas.

    • fan says:

      I second that sentiment.

    • typal says:

      The smartest and most successful person Ron knows is himself, but a little humility may in order about a subject that whole libraries have been written on.

      That said the essential point he made is correct IMO. WV is in a way similar to NM . Enterprising people keep going. Even for the same level of admixture the traits of African Americans from Hawaii are different to those from Louisiana.

    • Pincher Martin says:

      Ron: “Consider West Virginia. I’ve heard its white population is perhaps the most purely British-ancestry of any state, and certainly has American roots going back centuries. Meanwhile, the social, economic, and academic performance of those whites are absolutely dismal, perhaps the worst in the country. Just to cite one minor statistic, on an age-adjusted basis West Virginia whites are imprisoned at a rate three to four times higher than that of West Virginia Hispanics. Furthermore, most of the surrounding and also very heavily British-ancestry Appalachian region has similarly poor performance numbers.”

      “Does that prove that British-ancestry whites are generally stupid? Not really—some of the smartest and most successful people I know grew up in West Virginia or their parents did. I’d think the main factor is that the state is very rural and for the last century or two, a good fraction of the most able and ambitious individuals of each generation have moved away to the Big City elsewhere, thereby leaving behind a residual population of much lower quality, which also gradually produced negative cultural consequences.”

      Culture is important. Even when we control for IQ, cultural differences can still have a major impact on a country or region’s economic growth and social dysfunction. Russia is an excellent example of this. No one claims IQ explains everything. It certainly doesn’t explain why Russia is such a mess today.

      But the topic here is IQ and its correlates, and the evidence that whites in West Virginia have significantly lower IQ scores than do whites in the rest of the country just doesn’t hold up to analysis. They have lower IQ scores, but not significantly lower. According to Lynn’s data using NAEP test results, for which someone created a useful visual aid at this website, the average white IQ in West Virginia is 96.8. That’s the lowest for any state in the country. But to put that score in perspective, the average IQ for whites in California isn’t three points higher (99.2). Whatever economic problems and social dysfunctions exist in West Virginia, blaming them on low IQ scores is a stretch.

      If the typical Hispanic immigrant to the United States showed a similar intellectual capacity as whites in West Virginia, there would probably be less concern about mass immigration. But he doesn’t. And so the fear is that Hispanics will come to represent a permanent and distinct underclass in American society, one that requires constant political attention and resources. As Russia shows, a high group IQ doesn’t guarantee economic success, but failing economically is much, much easier with a low group IQ.

      You approach these problems like the politician you are or once were. You gloss over differences, seeking to bridge the wide gulf between the beliefs of many HBD types and a more mainstream view on immigration by coming up with this questionable notion of convergence. You invert the topic under discussion (like the one on West Virginia above) so that a discussion on IQ is suddenly turned into a discussion on social dysfunction, though not even the least plausible of straw men ever argued that IQ is the secret to everything. You focus on outliers in Lynn’s massive meta-analysis of global IQ research. But Lynn’s compendium is not valuable for its outliers — which you have not come close to exhausting — but for what it shows in broad strokes.

      I suspect your idiosyncratic approach is politically motivated and not a disinterested search for the truth. Being a Republican in California is a tough racket for the ambitious pol. The changing demographics of the state have permanently tilted the balance of political interests away from small government conservatism. What was once a purple state is now a deep blue state. There’s a lesson in there somewhere, but the lesson you seem to have drawn is that we HBD types can pretend it doesn’t matter and that we shouldn’t believe our own damn lying eyes, that convergence is happening even as we speak.

    • hbd chick says:

      @ron – “Consider West Virginia. I’ve heard its white population is perhaps the most purely British-ancestry of any state, and certainly has American roots going back centuries. Meanwhile, the social, economic, and academic performance of those whites are absolutely dismal, perhaps the worst in the country.”

      keep in mind that all british people are not the same (if you’ve read Albion’s Seed you’re probably somewhat aware of that). as i’ve written about quite a bit, the english in the southern part of the island have been outbreeding extensively (comparatively speaking) since the early medieval period, while the welsh and the highland & island scots AND the border populations (the border reivers) — the ancestors of the west virginians — continued inbreeding up until relatively recently. hackett fischer’s backcountry settlers were a people who were very clannish back in northern england/southern scotland and they brought that clannishness with them to the u.s. and maintained it by carrying on their cousin marrying practices. that’s why things don’t work so well in appalachia. clannish groups never manage to build very good civil societies because they don’t manage to cooperate with outsiders.

      (of course, west virginians might also have a lower mean iq than other anglo-americans. i have no idea if that is the case or not.)

    • JohnL says:

      Just to cite one minor statistic, on an age-adjusted basis West Virginia whites are imprisoned at a rate three to four times higher than that of West Virginia Hispanics.

      Any chance that you’ll back up this claim with any data? Real data, not your own data from your own discredited crime report?

  10. j says:

    Ron, Your generosity is commendable. But why Norman Finkelstein?

  11. typal says:

    Ron, you mention your own personal experience living among Hispanics in that video on your site, so it is quite in order for the other side of the argument to give ‘ insight .. acquired from visits to the local grocery store’ . Age-adjusted crime rates are of academic interest only. Actual real life crime is not committed by 50 somethings. The Mexican immigrant community is not aging.

    West Virginia is not pure white at all. The mountainous part of it was first settled by Melungeons. They originated from the mating of white women, and African men on Virginia plantations. There is also quite a bit of Native American. Moreover the whites of West Virginia have much ancestry from colonial convicts. The later immigration from Ulster was white but is not a random sample as those who settled on the mountains ( poor land) would tend to be be less bright than the ones who moved on. Lastly the most troublesome immigrant community in West Virginia was Irish ( eg Molly Maguires).

    The Amish don’t score low on IQ tests. The have a more rural lifestyle than any other population in the US, Or Mexico.

    • gcochran9 says:

      The Melungeons are a tiny group. You are a silly person.

      • typal says:

        Tiny yes but Ron suggested the WV population was pure British. I merely observed that Irish and Africans were in on the ground floor.

        The Ulster Scots weren’t to know that the valleys were not boggy as in Ireland, but not all of them stayed on poor land in the mountains. Ulster Scots were known for moving on and those who did found much better land . Those who stayed put in West Virginia were not a random sample

    • Mark says:

      Ron,

      Regarding the IQs of southern Italians, are the southern Italians scoring so well in the US full southern Italian? Or have they mixed with other white groups?

      I tell people I’m Sicilian, but I have 25% northern European DNA. A lot of Italians are mixed in the US. Though I will say the pure Sicilian side of my family did quite well. And they started out as peasants.

      Anyway my basic point is that mixture with other groups may be confounding regression to the mean.

      • gcochran9 says:

        This is a significant issue. For example, if you’re interested in looking at mestizo IQ in people who been in the US for several generations, you’d want to look at New Mexico, since that’s where most such people live. But there is quite a bit of admixture there. You’re not necessarily looking at the same population, then and now. Or when you look at Americans who self-identify as Irish, how many have > 80% Irish ancestry? Not that many, I would guess.
        You could try and look for unmixed people, say by asking people about their genealogies, but there is a better way: just look at their DNA. if the two ancestral groups are fairly different, say Amerindians and Europeans, you can identify the origin of chromosomal segments with essentially zero error. Do this, regress IQ against the amount of Amerindian ancestry, and Bob’s your uncle. You can do this between siblings, since they too vary in admixture, and that controls for most environmental factors.

        This approach will definitively settle the entire question. That’s why nobody want to do it.

  12. Anonymous says:

    You could do this with blacks too, couldn’t you? They have varying degrees of white admixture. Can’t you give them IQ tests and then see if there’s a correlation between intelligence and white admixture (which I think is what you’re saying wrt to AmerIndians).

    If so, why hasn’t this been done? Is it really just that people are afraid of the answer?? That’s so retarded. Maybe someone ::cough:: with a lot of money ::cough cough:: and a history of funding HBDers ::cough:: could fund such a study so we could have an answer to this question once and for all.

    • harpend says:

      I have reservations about the study design, either for hispanics or for US blacks. There is nothing magical or essential about color: what is important is the traits of the admixed groups. For example, what is the source of anglo genes in the US black population? Thomas Jefferson’s family? Low class white plantation foremen taking advantage of slave women? Catholic soldiers that Cromwell sold in Carolina? We don’t know and without that knowledge we would not know what to make of our admixture study.

      We may have a better idea about Mexicans but interpreting Greg’s proposed study would still require a lot of filling in of blanks.

      • Pincher Martin says:

        Henry,

        “There is nothing magical or essential about color: what is important is the traits of the admixed groups. For example, what is the source of anglo genes in the US black population? Thomas Jefferson’s family? Low class white plantation foremen taking advantage of slave women? Catholic soldiers that Cromwell sold in Carolina? We don’t know and without that knowledge we would not know what to make of our admixture study.”

        I suspect if this experiment were ever done, we would find the exact source of white ancestry in the black population doesn’t matter nearly as much as the degree of white ancestry. We know that both whites and blacks regress to very different means for IQ — enough so that the children of upper middle class blacks do not test as well as the children of lower class whites. We also know that the bell curve distribution for IQ bunches most people in a particular group close to their IQ mean, with approximately two-thirds of a population within one SD above and below that mean.

        So we should probably expect, on average, that the IQ scores of people with both black and white ancestry fall along a continuum that depends less on getting the right white genes than the degree of whiteness. Most people in a group have IQs around their group’s average. And when we look at the one-third of people in a group who are either significantly above or below their group’s average, their offspring still tend to trend in the direction of the group’s mean, anyway.

        In any case, what’s the alternative hypothesis if Greg’s proposed experiment does show the degree of white admixture closely tracks IQ scores in the black population?

      • albatross says:

        Greg,

        Is it really true that an IRB would refuse to approve a study of this kind, relating (say) genetic tests’ indication of African / European ancestry to IQ? This seems really odd to me, because there are already all kinds of studies that have been and continue to be done relating IQ to race and looking for genes related to IQ. Is this hyperbole? Are there examples of studies being spiked for this reason by an IRB?

      • Greying Wanderer says:

        “There is nothing magical or essential about color”

        Except evolution in completely different environments for thousands of years.

    • gcochran9 says:

      Sure, it would work with blacks in the US. It would work with people of varying amounts of Chinese and Malay ancestry in the Philippines or Malaysia. It would likely work with Pygmies, who are an admixed group (part ur-Pygmies, part Bantu). In fact, this has already been done for height in Pygmies, and it turns out that the more Bantu ancestry a Pygmy has, the taller he is. I would guess that increasing amounts of Bantu ancestry would correlate with higher IQs in Pygmies, which with any luck would make someone’s head explode.

      It hasn’t been done with IQ because people don’t want to know. No IRB would approve, nor would the Feds ever fund such a thing. Although, we’re getting to the point where most people really don’t even know that there are any differences to explain, so political correctness may go full circle and let people do such experiments, since OF COURSE there’s nothing to discover. There is hope. In fact, this may already have begun to happen. Ignorance is strength!

      Would your hypothetical Maecenas be interested in funding such an experiment, assuming that it wouldn’t cost much (which is the case). Well, only he knows his thoughts, but I would guess that he would not do it among mestizos, but might consider it in some other group.

      As for Henry’s objections, I think they are no problem. He’s thinking that those accounting for most European ancestry in African-Americans may have been an unrepresentative sample. Possible, but so what? I sincerely doubt they were sufficiently unrepresentative to reverse the large observed differences.

      A weaker, but related, natural experiment has been done in many places. There are social groups with differing levels of admixture in a given nation, say between blacks and Europeans, or Amerindians and Europeans – and the social ranks of those groups almost always rank-order by admixture levels. This certainly happens everywhere in Latin America and the Caribbean. The standard sociological explanation is that this is all enforced by prejudice and social control mechanisms – but that kind of social control never stops Ashkenazi Jews from rising, or Japanese (originally cane-cutters in Brazil) or Chinese anywhere (originally illiterate tin-miners in Malaysia). So the sociologists are wrong. And they’re pinheads, I might add.

      In Mexico, we actually have an admixture study that shows that there is a strong negative correlation between social status and the degree of Amerindian ancestry. Tyler Cowen would undoubtedly say that poor pure-Indian populations in Yucatan have just as much intellectual potential as anyone, but all experience, and every result of psychometric testing, says he’s wrong. Even economists can make mistakes.

    • Jason Malloy says:

      “Can’t you give them IQ tests and then see if there’s a correlation between intelligence and white admixture (which I think is what you’re saying wrt to AmerIndians).”

      There are dozens of older studies showing that proportion of white ancestry is correlated with higher IQ in Native Americans and Mestizos (where ancestry is roughly estimated by number of Indian grandparents, or from physical measurements like skin color), and many similar studies with socioeconomic status, education, etc. But what happens when we link IQ with actual Indian or black genetic ancestry? It wouldn’t tell us much more than we already know, until we can, say, compare mixed race siblings, so that cultural background is equalized and ancestry differences are rendered cryptic. Otherwise the differences could be chalked up to an invisible “Colorist” caste system. Or whatever idiosyncratic theory I’m sure Ron Unz has thought up to explain away these obviously hereditary social patterns.

      We already have modern studies showing that proportion of Amerindian and black ancestry is associated with lower socioeconomic status and education:

      Strong association of socioeconomic status with genetic ancestry in Latinos
      “We determined that socioeconomic status was strongly correlated with European ancestry proportion in Mexicans (33% correlation… ) and Colombians (34% correlation…)”

      Latino Populations: A Unique Opportunity for the Study of Race, Genetics, and Social Environment in Epidemiological Research
      … among Puerto Ricans, ancestry is associated with socioeconomic status. Healthy Puerto Rican volunteers reporting “upper” socioeconomic status had 9.1% lower African ancestry and 9.2% higher European ancestry than healthy volunteers reporting “moderate” and “middle” socioeconomic status…

      • MZT says:

        Even with sibling studies, one could still object that appearance (overall, even controlling for skin color there are recognizable differences) confounded admixture, with ensuing social discrimination. I think it is in fact true that there is discrimination on this basis, although it’s a bit of a stretch to attribute all differences to it.

      • Jason Malloy says:

        “Even with sibling studies, one could still object that appearance … confounded admixture, with ensuing social discrimination.

        This is the difference between a mixed race sibling study and a genetically informed mixed race sibling study.

        In the former category we already see that racial appearance is correlated with social outcomes in full siblings. In his excellent and illuminating book, Race in Another America, sociologist Edward Telles compares mixed race children in Brazilian families who are alternately classified as white and non-white by their parents. The non-white siblings are more likely to have academic problems and have to repeat grades at school (pp 148-150). A similar study found that whiter siblings are more likely to be enrolled in school and attend a higher quality school.

        Both authors use this finding as evidence for parental discrimination against darker children, yet neither one questions whether the whiter siblings actually have higher academic ability.

        Sufficient sample sizes would be a headache, but full siblings would differ in both racial ancestry and racial appearance. Sometimes the latter would be because of the former and sometimes it wouldn’t. Would True Whites and Faux Whites* have the same observed abilities? Any discrepancy or non-discrepancy in social outcomes between the two would be a novel behavioral genetic insight.

        *Of course, the outcomes and racial appearances would both be rated on a continuous or graded scale.

        A gene

      • gcochran9 says:

        An admixture study using modern genetic methods would be simpler and cleaner than those old studies. More than that, it would be modern and high-tech, which makes everything better, just as the Principia would have been far more convincing if set in LaTeX.

        We could compare mixed race siblings. It would be useful.

        Anyhow, you may already know this stuff, but most people don’t. If a modern study gave clear results, it might change the minds of dozens of people worldwide.

      • Jason Malloy says:

        Here’s an example of a recent mixed race sibling study from Brazil.

        Genomic Ancestry, Self-Reported ‘‘Color’’ and Quantitative Measures of Skin Pigmentation in Brazilian Admixed Siblings

        In 35% of 86 full sibling pairs, the siblings report a different race from each other. And the siblings differ from each other slightly in their proportions of racial ancestry (just as full siblings differ, on average, in their relatedness to each other).

        The study finds that proportion of European ancestry is associated with higher socioeconomic status, as is self-reported race, but that objectively measured skin color is not. This implies that ancestry is cryptic (undetectable by society), and yet European ancestry still leads to enhanced economic outcomes.

        So the genetically informed sibling study would appear to contradict the discrimination theories in the Telles and Rangel sibling studies linked above.

      • JayMan says:

        There is a way to probably get around the IRB issue (though not necessarily the funding issue). A database with the necessary data already exists. That is the Add Health study, certain waves of which contain genetic samples of subjects, in addition to IQ tests (among many other data points). Researchers have been mining the data from this to find genetic links for some time (I have discussed one such study). It seems that it would be fairly simple to access the degree of European ancestry in the non-White subjects and determine if there is a correlation between this and IQ. One could even look at siblings, since sibling pairs do exist in the database (don’t know about sample size there, though).

  13. Ron Unz says:

    Mark: Regarding the IQs of southern Italians, are the southern Italians scoring so well in the US full southern Italian? Or have they mixed with other white groups?…Anyway my basic point is that mixture with other groups may be confounding regression to the mean.

    That’s a perfectly valid point, and obviously our dataset is highly imperfect. However, let’s consider the (Catholic) Irish in America. Among all significant white ethnic groups, they have nearly the highest Wordsum-IQ, and are also very near the absolute top in education and family income. On Wordsum-IQ, they are above the English, Scots, Swedes, and Norwegians, and far above the Germans, French, Dutch, and Scots-Irish. Yet as recently as the 1970s, the Ireland IQ was apparently the lowest in all of Europe.

    So if the (self-reported) Irish-Americans are above virtually every other (self-reported) American ethnic group in Wordsum-IQ, then who would they have “mixed with” in order to raise their IQ to that level?

    • OC says:

      Demanding that each contributing factor explain an entire effect is an easy way to dismiss each and every datum, whatever they add up to. Selective emigration and outmarriage reduce the expected IQ gap with older native Ireland scores (when Ireland was suffering massive continual brain drain, as opposed to recent years in which it has had a huge influx of talented people from its diaspora and elsewhere). Then one of a number of modest effects could cover the remainder, e.g. if cultivating and reporting Irish identity is correlated with education and intelligence.

      “So if the (self-reported) Irish-Americans are above virtually every other (self-reported) American ethnic group in Wordsum-IQ, then who would they have “mixed with” in order to raise their IQ to that level?”

      Well, for ethnicities they could have outmarried with Ashkenazi Jews, although I think that did not happen much. However, they could also just have disproportionately “married up” across varied ethnicities. Large cities disproportionately “brain drain” rural areas, concentrating genetic potential for IQ there. Since IQ-based assortative mating is not absolute, proximity allows lower-IQ folks to practice hypergamy in the absence of caste barriers. Irish people were concentrated in big cities with professional classes. You could look in the GSS to see if Irish people outperform other European-Americans in the same cities, or at least in urban areas.

      I would also suggest that you look more carefully at urbanization in general to distinguish between concentrating and increasing cognitive skills.

    • Anonymous says:

      Why are we talking about the Irish all of a sudden? I just want to know what branch of my family tree I should thank for my LSAT score.

      In all seriousness, I think you raise good points, as I’ve said on other blogs. I also think that Lynn is being disingenuous by arguing that southern Italians have lower IQs than northern Italians because of recent Mideastern and North African admixture, when all the principal component analyses I’ve seen show that southern Italians cluster pretty closely with other Europeans…

      I’m open to the idea that Sicilians have an average IQ lower than that of other European populations, much as I may dislike it. But there’s no reason to think that their/our depressed IQ had to have been the result of foreign admixture.

    • Greying Wanderer says:

      “So if the (self-reported) Irish-Americans are above virtually every other (self-reported) American ethnic group in Wordsum-IQ, then who would they have “mixed with” in order to raise their IQ to that level?”

      If those two bits of data are both true i.e. a low average IQ in (some parts of ) Ireland in the 1970s and high average IQ in America in the 2000s then one explanation could be Irish immigrants to America in the 1840s had the same low average IQ but something happened in America during the intervening 100-ish years that increased it in America . This could have been a lot of things – including mixing with other higher average IQ european groups – although personally i think it’s the result of various kinds of inbreeding depression being lifted.

      On the other hand if the Irish data is false then there’s been no convergence.

  14. Ron Unz says:

    OC: Ireland was suffering massive continual brain drain, as opposed to recent years in which it has had a huge influx of talented people from its diaspora and elsewhere

    Well, here’s an exercise to the student. Let’s assume based on recent PISA scores, the performance of Irish-Americans, and various other datapoints that Ireland’s underlying “true IQ” was at least pretty close to 100 a few decades ago. Then what fraction of the Irish population would have had to be “brain drained” away each generation for the reported 1972 IQ figure of 87 to be possible, and what would be the necessary distribution of that emigration “brain drain”?

    Frankly, I think you’re better off going with your alternate theory that Irish-Americans are very heavily Ashkenazi Jewish in ancestry…

    • typal says:

      Irish-Americans may not be very heavily Ashkenazi Jewish in ancestry (apart from John Kerry) but the Tay Sachs mutation spread so rapidly among them that Irish without Tay Sachs must have not been up to scratch in the brainpower department. A few IQ points conferred by Tay Sachs must have led to reproductive success for Irish in an American environment.

      those PISA scores can’t be right as 20% of the population of Eire is functionally illiterate. It is a fact that the in the Irish republic today the protestants (not ethnically Irish) are demonstrably of higher IQ than Catholics (ethnically Irish). Importantly this extends to rural areas.
      “The Protestants of Eire retain an influence out of proportion to their numbers. Although they are not an identifiable political force, 6.5% are company directors, managers or company secretaries compared with less than 1% of Catholics, and about one fifth of farmers working over 200 acres are Protestant.”

      (Mid 18th cent. Ireland was suffering rural overpopulation, people eating up the surplus, which is why the English landlords paid for their Irish tenants to go to America (just as they had earlier paid for Highland Scots to emigrate). A significant amount of the emigration of that time was the opposite of a brain drain .)

    • JohnL says:

      Well, here’s an exercise to the student. Let’s assume based on recent PISA scores, the performance of Irish-Americans, and various other datapoints that Ireland’s underlying “true IQ” was at least pretty close to 100 a few decades ago. Then what fraction of the Irish population would have had to be “brain drained” away each generation for the reported 1972 IQ figure of 87 to be possible, and what would be the necessary distribution of that emigration “brain drain

      Here is a better exercise for the student. Given (1) that the IQ of the urban Irish was 100 in that 1972 study, and (2) that about one third of the population of Ireland was urban at the time, and (3) that the population of Ireland at the time was three million people, what was the IQ of the the two million people living outside the urban areas? The answer is – those two million people would have to have had an average IQ of about 78.

      So here is another exercise for the student – what is the probability that there existed a 22 point difference in IQ between Ireland’s urban and non-urban regions?

  15. OC says:

    “Then what fraction of the Irish population would have had to be “brain drained” away each generation for the reported 1972 IQ figure of 87 to be possible”

    You cherry-picked that as a case of extreme apparent IQ rise from dozens upon dozens of countries, while ignoring the countries where IQ apparently declined while wealth increased, like oil states or Portugal. The winner of that kind of cherry-picking exercise is likely to be exaggerated by the selection bias (since both error and extreme change can contribute to being an extreme outlier). Large samples are no defense against unrepresentative selection methods.

    “Well, here’s an exercise to the student.”
    Well, instead, I’m going to look at what selective emigration could produce.

    Say you send 15% of each generation overseas, and the emigrant group has phenotypic IQ 2/3 of a standard deviation above normal (IQ 110, 75th percentile in an IQ 100 population). Then, in the first generation, the stay-at-homes have a phenotypic IQ of 98.24. With narrow heritability of 0.6, the children of the stay-at-homes will have mean IQ of 98.94, while the children of the emigrants will have IQ of 106, a gap of 7 points.

    And this process can repeat, lowering the genotypic IQ back home by a large amount. If the emigrant population has a faster rate of natural increase than at home (as was the case with Ireland and the US, where the Irish-American population boomed), then the earlier higher-IQ generations will be more represented in the later immigrant population. If population doubles each generation, then later migrants make an exponentially smaller contribution to the population of the new country, so we could widen the genotypic gap to 10-12 points after a few generations. [Remember the similar flows within countries from rural areas to urban ones.]

    Then, at a later time, we reverse the flow of migration, with the onset of peace and relative prosperity in the old country. Now relatively well-off descendants of emigrants return. Not only do they come from a population with genotypic IQ 10-12 points higher than the old country, but the second selective filter picks out relatively higher phenotypic IQs (and somewhat less high genotypic IQs) so when they have kids after returning, those offspring bring up the old country average immensely. If a quarter of the old country population of childbearing age is made up of returning migrants then they can bump up average genotypic IQ in the next generation by 3-5 points.

    Since migrants tend to be young and healthy, that requires much less immigration than a quarter of the existing population. This story would explain huge chunks of higher IQs in the diaspora, depressed IQ at home, and a recent rise as the flow of migration reversed, albeit not everything (and it would be silly to take one cause as the only thing affecting IQs in this period). And the huge migrant flows roughly track this pattern.

    Irish population went from about 6.5 million in 1841, to less than 3 million in 1961, and now is back up to about 4.5 million: http://en.wikipedia.org/wiki/File:IrelandRepublicPopulation1841.PNG

    A 0.66 SD phenotypic advantage, or an average of 75th percentile IQ, for emigrants is somewhat high, but consistent with with other studies, e.g. Lynn cites a study of Scottish emigrants from regions where 15-20% of people had left, with the emigrants averaging IQ of 108.

    Here’s more wikipedia on the Irish diaspora:
    “After 1840, emigration had become a massive, relentless, and efficiently managed national enterprise.[1] Counting those who went to Britain, between 9 and 10 million Irish people emigrated after 1700. The total flow was more than the population at its historical peak in the 1830s of 8.5 million. From 1830 to 1914, almost 5 million went to the United States alone. In 1890 two of every five Irish-born people were living abroad. By the 21st century, an estimated 80 million people worldwide claim some Irish descent; among them are 41 million Americans who claim “Irish” as their primary ethnicity.”

    “Only in the mid-20th century did the Republic’s population start to grow once more, but emigration was still common until the 1990s. The demographics of the country changed significantly from then onwards, as a result of the Celtic Tiger Irish economic boom. After this point, immigration far outweighed emigration and many former Irish Emigrants returned home. Ireland then became an attractive destination for immigrants from a number of nations, mainly from Central Europe, but also from Africa, Asia and elsewhere.”

  16. OC says:

    This is the paragraph from Lynn’s response on magnitude of emigrant selection effects, which is consistent with an “emigrants at 75th percentile rather than 50th percentile” story:

    “Typically, when people emigrate from poor regions to more affluent regions, migrants tend to have higher than average IQs. The reason for this is that a higher IQ is needed to envision the advantages and to plan and find the resources to migrate. This has been shown in the United States by Vigdor (2002), who has found that it was blacks with greater educational attainment (a proxy for intelligence) who migrated from the southern states to the northern states, with the result that blacks in the northern states have an IQ about ten points higher than those in the South. Kaufman & Doppelt (1976) report an average IQ of 90.5 for blacks in the northern states compared with approximately 85 for all American blacks, and around 80 for those in the southern states. Further evidence that migration is typically selective for intelligence is available for Scotland, from a follow-up study of 1,000 11-year-olds whose IQs were tested in 1947. By the age of 30, 17.2 percent had emigrated, and they had had an average IQ of 108.1 (Lynn, 1977b).”

  17. albatross says:

    Is there any evidence available about how self-labeling of someone as hispanic or may be a function of IQ or socioeconomic status or whatever? It seems like this is a kind of big question hanging over discussions of hispanic performance in the US, right? If stuff we care about correlates with whether someone self-reports as hispanic or not, then it becomes much harder to interpret the data.

  18. ml@gmail.com says:

    “I personally favor mass deportation”

    As an investor, I am against this. More people, more land use, more gas and oil are directly increasing my wealth. An investor is very similar to a feudal lord. More people are better. Nationalist movement has its socialist component for underclass. At end, bussiness people should always suppress nationalist or protectionist movement. But I do want have control of what kind people to move on to my territory. If not, more still better.

    • Pincher Martin says:

      “As an investor, I am against this. More people, more land use, more gas and oil are directly increasing my wealth. An investor is very similar to a feudal lord. More people are better.”

      A smart investor today should probably know that, unlike a feudal lord, his success is not dependent on the number of serfs at his command.

      Individual American investors and U.S.-based companies can both invest abroad, which means the serfs need no longer be in an investor’s geographic proximity for him to benefit.

      Quite apart from your flawed investment strategy and demographic determinism (“More people are better”?), it’s telling you see investors such as yourself not as one side of a necessary economic equation for the good society, but as medieval Lords of the Manor; and present-day immigrants not as fellow citizens, but as serfs toiling for your benefit.

      • ml@gmail.com says:

        Best and most stable society is feudalism. Today’s liberal democractic society is self-defeating and degenerative. I doubt i could last.

  19. Kiwiguy says:

    Chuck has another post up on this.

    Hispanic flynn effect in the NAEP

  20. Luke Lea says:

    I like Pincher Martin.

  21. The ecological deporatation argument is ridiculous. The US not a densely populated
    country — in fact it is well down on the list.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s