Why Racialism is Sensible: A Pithy Rejoinder to The Prussian

What is Racialism

The Prussian recently wrote a lengthy critique of “racialism”. His central thesis was that:

Biology, and to a lesser extent, genetics, has a powerful influence on individual human life but at the group level it is overshadowed by culture and social institutions.

While he doesn’t precisely define his meaning, he implies that “racialism” is the view that genes condition important outcome differences between racial groups. Here I will briefly add to the points made by Sean Last and Bulbasaur of The Right Stuff and explain why racialism makes sense.

Under What Condition is Racialism True

To show that racialism, so defined, is true one only needs to show that between the said racial groups genetic differences condition practically important outcome differences. Some specific racialist positions might be false but if some are true, racialism as such stands vindicated. In this case, the racial groups under consideration are regional ones (e.g., N.E. Asians, Europeans, West Africans, etc.).

This formulation perhaps does injustice to the situation. After all, there are racialists out there who imagine an endless variety of large regional racial difference (e.g., in personality traits such as conscientiousness, neuroticism, psychopathology, moral looseness, collectivism, and so on) that jointly condition outcome differences. Let us call this view de Gobineauism and tentatively define it as the view that (a) between regional races, individuals differ in a multitude of socially important behavioral traits by at least a moderate amount (per social science standards), (b) that these trait differences condition at least moderate sized outcome differences, and that (b) genes explain a large portion of the trait difference. In this situation race differences would be like individual differences writ large — as individuals do differ in a multitude of traits and as these differences are large, highly genetically conditioned, and practically important.

Let us distinguish this position from narrow racial hereditarianism and define this as the view that (a) regional races differ in some behavioral traits by at least a moderate amount, (b) that these trait differences condition at least moderate sized practically important outcomes differences, and that (c) genes explain a substantial portion of these trait differences. We can also add a third positions called racial geneticism which we can define as the view that (a) regional races differ on the populational level in outcomes by at least a moderate amount due to the aggregate direct and indirect effects of genetically conditioned individual level trait differences. Racial geneticism differs from racial hereditarianism in that it allows for moderate population level differences to emerge from small aggregate individual differences via social multiplier effects.

I will not here attempt to defend de Gobineauism; I will not because I can not; I can not because it’s largely untrue. Boetel and I discussed this in section IV-K of our Nature of Race paper. This section was labeled “Shades of de Gobineau” because de Gobineau’s racialist positions were not altogether incorrect yet also because only a shadow of his views can be rigorously defended.

Here, I will defend racialism qua narrow racial hereditarianism.

But what is race?

Biological races are, and have been thought of as being so since the time of Darwin, subspecific natural populations; natural populations are biological populations delineated in terms of overall genetic (genealogical or genotypic, school of thought depending) relatedness instead of specific genetic characters as in the case of morphs (e.g., sexes) and forms. Subspecific nature populations could be operationalized as: sets of individuals of the same species in which members are less overall genetically related to members of other sets than to members of their own . These populations are basically equivalent to the population geneticists’ (retrospective) genetic populations. Genetic populations are hierarchically nested, with local races nested in regional ones and regional ones nested in continental ones. For example. Aulchenko (2010) tells us:

(Aulchenko, Y. S. (2010). Effects of population structure in genome-wide association studies. Analysis of Complex Disease Association Studies: A Practical Guide, 123.)

As there is no true level of genetic analysis, one can not say that any level of racial analysis is true. Dohzhansky (1946) pointed this out, noting:

One may perhaps question the desirability of applying the term ‘racial differences’ to distinctions as small as those that can be found between populations of neighboring villages and as large as those between populations of different continents. Might one modify the definition of race by specifying that the differences in gene frequencies be above a certain minimum magnitude? Such amodification is undesirable for two reasons. First, since all magnitudes of difference are found among populations, any specified minimum can be only arbitrary. Second, it is most important to realize that the differences between the ‘major’ human races are fundamentally of the same nature as the relatively minute differences between the inhabitants of adjacent towns or villages.

In fact, most racialists have acknowledged the existence of local and regional races and more generally the nested nature of race. For some reason, Mr. Prussian adopts the view that only continental level races exist as races; perhaps he feels that only natural populations with significant genetic discontinuities between them should be called races. This conception, though, is at odds with the majority of the historic race conceptions and with (most) modern biological ones. Much could be said on this matter, but this point has already been discussed extensively elsewhere. In short, one can meaningfully talk about a White (European) evolutionary race in contrast to, say, a Yellow (North East Asian) race. These two races represent natural subspecific biological populations.

What Race differences did you have in Mind, Sir?

If no other form of racial hereditarianism proves correct, racial intelligence hereditarianism alone can vindicate racialism, since global differences in intelligence conditions a large portion of the global differences in quality of life. To give a sense of the significance of these global cognitive differences, I plotted the average of the 52 subcomponent 2014 Social Progress Index scores against National IQs. The correlation between Social Progress and National IQ was 0.82. (For comparison, the correlation between the percent of Muslim population and Social Progress was -0.40.) In short, cognitive ability differences exert a powerful influence on the group level.


There are a number of lines of evidence in support of a racial hereditarian hypothesis for cognitive ability differences (and with them overall quality of life differences). Generally, regional cognitive differences have numerous historic, biological, and genetic correlates; national cognitive differences follow migrants to some degree and transmit across generations in the new regions of origin; within mixed race populations, cognitive related outcomes correlated with racial ancestry.

Phenotypic IQ differences between Biological Races



The Comparative Performance of non-Hispanic White and Blacks in the U.S. by Immigrant Generation


Association between Outcomes and Ancestry in a Mixed Race population

(Cheng, et al., 2012. African Ancestry and Its Correlation to Type 2 Diabetes in African Americans: A Genetic Admixture Analysis in Three U.S. Population Cohorts)

There is very little to argue about here. A substantial global cognitive hereditarian hypothesis is robust yet not established. If global cognitive differences are only 40-60% heritable, we would still have moderate to large congenital cognitive differences, thus vindicating racialism (qua racial hereditarianism).

The Prussian Strikes Back

The Prussian doesn’t consider the racialist case, though. Instead, he moves to dismiss it. He points to, for example, the Flynn Effect, the secular rise in cognitive scores, to call into question the science of mental ability. An example of the Flynn effect is illustrated below. In the figure, age heaping — a measure of numeracy — rates are shown for different ethnic groups across centuries

(Juif, D. T., & Baten, J. (2013). On the human capital of Inca Indios before and after the Spanish Conquest. Was there a “Pre-Colonial Legacy”?. Explorations in Economic History, 50(2), 227-241.)

As can be seen, the ability to correctly report ages, which involves both learning and the ability to learn, has increased over the centuries — because learning has increased. Despite this secular increase, relative differences between ethnic groups (both within and between nations) remain. Chinese were one of the world’s most numerate populations, as measured by age heaping, in the 1600s and they are one of the world’s most numerate populations today as measured by PISA math scores. To be clear: cognitive differences between ethnic groups strongly correlate across centuries.


Now, one can always argue that the general cross century stability is due to stable cultural and or other environmental differences. One can — but one still has to account for the association between genetic differences and ability differences.

The Prussian goes on to point to deviations from well known averages (e.g., that Black African countries have worse profiles than European countries). But since no one argues for racial determinism, let alone racial uniformism, this whole line of argumentation is invalid. He also attempts to dismisses unwanted evidence. He dismisses, for example, the conspicuously low Amerindian and Oceania average abilities on the grounds that these groups were largely replaced by more apt Europeans. This approach allows him to side step the well know intra-national differences. As example of these, self identifying “indigenous” throughout the Americas underperform self identifying “non-indigenous” Europeans. Here is a typical discussion:

The background and test score differences between indigenous and non-indigenous students give additional insight into the distinct challenges that indigenous students face. In every country, the test score gap between indigenous and non-indigenous students was greater in Spanish reading exams than in math exams, and the gaps in both subjects ranged between 0.6 and 1.1 standard deviations. (Hernandez-Zavala, M., Patrinos, H. A., & Sakellariou, C. (2006). Quality of schooling and quality of schools for indigenous students in Guatemala, Mexico and Peru (Vol. 3982). World Bank Publications.)

Moreover, ancestral admixture predicts educational outcomes.

While none of the above proves a causal genetic hypothesis, some explaining is needed. Anti-racialists need to account for the association between ancestry and outcomes (a) between ethnic groups between nations, (b) between ethnic groups within nations, and (c) between individuals within heavily admixed ethnic groups. The Prussian tells us that “cultural institutions trump race – and institutions can be changed” and yet all of the institutional change to date has left the substantial association between ancestry and outcomes largely intact. Apparently, more change is needed — more mixing it up with an extra dose of anti-privilege.

Finally, when not evading evidence, The Prussian attacks the strawview of White Supremacy. We are told, for example, that the Chinese were historically more advanced than Europeans and therefore that Euro-number-one-ism is stupid (which it is, but not for this reason). But if so, this is all the better for a hereditarian cognitive ability hypothesis, as it currently struggles to explain the surprising dearth of East Asians accomplishments.

Significant Scientific Figures and Accomplishments from 800 B.C. to 2000 A.D.


The accomplishment differences are currently attributed to differences in personality factors such as collectivism; while there is a genetic basis to such regional personality differences (e..g, Way and Lieberman, 2010), the connection between these differences and the ones in accomplishments is speculative at best. The Prussian then, at most, simply lifts the burden of accounting for East Asian under-performance.

In summary, The Prussian doesn’t address the evidence in support of racialism, understood as the view that genes condition important outcome differences between racial groups. Instead, he sidesteps the evidence and knocks down strawmen e.g., racial determinism.

This entry was posted in Uncategorized. Bookmark the permalink.

43 Responses to Why Racialism is Sensible: A Pithy Rejoinder to The Prussian

  1. Anonymous says:

    In last few months there has been a paper published by Rindermann on IQ in Vietnam (average IQ of 99), and we are awaiting your, as yet, unpublished work on Southeast Asians. The question is whether average IQ among indigenous Southeast Asians is higher than the Lynn and Vanhanen data suggest.

  2. Anonymous says:

    We are still awaiting your article on Southeast Asians. Is the average IQ of indigenous (non-Chinese) Southeast Asians actually closer to Northeast Asians than Lynn and Vanhanen have suggested? But surely the average IQ of indigenous Filipinos, Malays, and (aboriginal) Taiwanese is, um, unimpressive. Vietnam was, of course, conquered and occupied by Chinese for long periods of time– so who knows what alleles spread through the population? We should not be too surprised if Vietnam scores reasonably well.

  3. The fourth doorman of the apocalypse says:

    Chinese were one of the world’s most numerate populations, as measured by age heaping, in the 1600s and they are one of the world’s most numerate populations today as measured by PISA math scores.

    This is interesting. It selects that much of the selection for higher intelligence happened earlier than 1600 or that there has been little difference in selection for intelligence between the Chines and Europeans, for example.

    However, perhaps it is related to the fact that the Chinese have had a larger single polity for a longer time, and the need to manage larger numbers of people have driven selection for numeracy.

  4. anon says:

    1. “Social Progress Index” is largely derived from highly subjective neoliberal PC nonsense. So that graph is by and large garbage-in-garbage-out..
    2. The PISA graph is much more meaningful. Yet China hasn’t participated PISA as a country since 2003 but cities. So China’s PISA score, as a country, could be somewhat overestimated hence misrepresented there.

    • Chuck says:

      There were 52 components, some were PC goofy (tolerance for immigrants/gays) some were not (access to electricity, child mortality rate). You can see them here: http://www.socialprogressimperative.org/data/spi#performance/countries/com1/dim1,com1,com2,com4,dim2,com5,com6,dim3
      If you make a list of non-“neoliberal PC” ones I will rerun the analysis with that average.

      Also, that progressives themselves value this stuff is the best part….

      • anon says:

        Sorry, can’t get the how data is weighted. Some outragious examples tell clearly some of “social progress” results are completely rubbish:

        1. China is about 58. Phillipines is about 68. I can’t tell China is less socially progress than Phillipine in almost all measures except democracy-related item/items, which means that democracy somehow worth > 10 points in this “sccial progress” scale.

        2. otherwise, more laughble things such as China is lower and/or drastically lower than some sub-sahara African countries and any banana republic ( literally) you can name of. Actually any of these testbook banana republic , usually in their mid 70s to 80 on “social progess” scale , makes China’s desperate high 50s look like coming from Stone Age. Sorry, but that doesn’t fit into any traveller’s common sense no matter how much one hates Chairman Mao, does it?

        3. some other seemingly “objectively & fair” ( but highly troublesome in reality) criteria one can immediately identify is , for example, so called the “standard GDP/per cap” ( in $ term) or its derivatives they deploy. China’s currency is controlled, but hugely undervalued. That explains the bizzare phenominon why most third world hellholes ( i.e. many of the subsahara african countries and most/all north afrcian, non-oil middle eastern countries, and almost all countries in South America and Asia, except some obvious and absolute world’s poorest) have higher GDP per cap than China on paper , yes, Angola on average is richer than China, so are Botswana, Namibia and Gahna, etc — {the world bank/IMF Oxbridge-graduated “chief economists” illiterates who have developed and kept using this “GDP per cap” criterion are somehow deeply racist thinking that the rest of the world are intellectually too retadred to identify this obvious sham :hitwall: } — because, just for the starter, their currencies, called Botswana Dollar, Angola Dollar and Namibia Dollar I guess, are supposely “openly” and “fairly” traded in the “world market” according to Wall Street Crooks Standard.

        Of course, the index works quite well when compraring some relatively similar countries such as Sweden and Ireland, and some countries which sit on the two extremes such as Norway and Zimbabuwe alike that can be identified by any Einstein given some time, with or without the index.

        However for countries which happen to be called or identified as political/economical/ or military or ideological “opponents” or “threats” of good & old US and A govenement that sponsor all this kind of “world’s XXX index ” , this “social progress” is nothing more than an insult to g factor.

        Ditch it for your own good, Chuck.

    • Chuck says:

      “So China’s PISA score, as a country, could be somewhat overestimated hence misrepresented there”

      I didn’t use national PISA scores, I used L&V’s (2012) quality weighted national IQs. The Chinese scores would have been based on IQ test scores (105.5 –quality weight 16) and PISA scores (108.2 -quality weight 2), The quality weighted average came out to 105.8. If there was overestimation it would have been minimal.

      • B.B. says:

        Chuck says:
        I didn’t use national PISA scores, I used L&V’s (2012) quality weighted national IQs. The Chinese scores would have been based on IQ test scores (105.5 –quality weight 16) and PISA scores (108.2 -quality weight 2), The quality weighted average came out to 105.8. If there was overestimation it would have been minimal.

        That depends on how reliable L&V’s national IQ estimates are. From what I understand L&V’s metric of IQ data quality was based on sample size & the number of different studies cited. Whether the studies were geographically representative of the nation as a whole wasn’t considered. From what I can gather of the sources that are available to me, much like the PISA data the national IQ data is biased towards sampling urban regions like Shanghai and Beijing. Here are the IQ data sources that L&V 2012 cited for China:

        Dataset – (Age Range) – Sample Size – Test Type – IQ – Source
        01 – (6-16) – 660 – WISC-R – 107 – Dan et al., 1990
        02 – (5-15) – 5,108 – SPM – 101 – Lynn, 1991
        03 – (14-15) – 297 – Various – 103 – Li et al., 1996
        04 – (6-12) – 269 – SPM – 104 – Geary et al., 1997
        05 – (4) – 60 – Arithmetic – 109 – Ginsburg et al., 1997
        06 – (6-13) – 463 – DAM – 103 – Cox et al., 1998
        07 – (6-8) – 160 – SPM – 107 – Cox et al., 1998
        08 – (17) – 218 – SPM – 103 – Geary et al., 1999
        09 – (19) – 218 – SPM – 113 – Geary et al., 1999
        10 – (6-8) – 300 – BTBC-R – 107 Zhou & Boehm, 2001

        Dan, L., Yu, J., Vandenberg, S. G., Yuemei, Z. and Caihong, T. (1990). Report on Shanghai norms of the Chinese translation of the Weschsler Intelligence Scale for Children – Revised. Psychological Reports, 67, 531-541.
        Lynn, R. (1991). Intelligence in China. Social Behavior and Personality, 19, 1-4.
        Li, X., Sano, H. and Merwin, J. C. (1996). Perception and reasoning abilities among American, Japanese and Chinese adolescents. Journal Adolescent Research, 11, 173-193.
        Geary, D. C., Hamson, C. O., Chen, G-P., Liu, F., Hoard, M. K. and Salthouse, T. A. (1997). Computational and reasoning abilities in arithmetic: cross-generational change in China and the United States. Psychonomic Bulletin and Review, 4, 425-430
        Ginsburg, P. H., Choi, E., Lopez, L. S., Netley, R. and Chao- Yuan, C. (1997). Happy birthday to you: early mathematical thinking of Asian, South American and U.S. children. In T. Nunes and P. Bryant (Eds), Learning and Teaching Mathematics: An international Perspective. Hove, UK: Psychology Press.
        Cox, M. V., Perara, J. and Fan, X. U. (1998). Children’s drawing ability in the UK and China. Psychologia, 41, 171-182.
        Geary, D. C., Liu, F., Chen, G-P., Salts, S. J. and Hoard, M. K. (1999). Contributions of computational fluency to cross-national differences in arithmetical reasoning abilities. Journal of Educational Psychology, 91, 716-719.
        Zhou, Z.339 and Boehm, A. E. (2001). American and Chinese children’s knowledge of basic relational concepts. School Psychology International, 22, 5-21.

        Dataset 01 was from Shanghai as the study title indicates. Dataset 03 utilized students that “attended one of the three participating schools in and around Beijing.” Dataset 04 was from Shanghai. Dataset 08 is “Chinese high school students from Columbia, Missouri, and Shanghai, China, respectively”. Dataset 09 is “undergraduate students from East China Normal University, Shanghai, China.” Dataset 10 is “Three hundred kindergarten, first and second grade children from Beijing, China”. 02, 05, 06, 07 were inaccessible to me. Lynn & Cheng’s “Differences in intelligence across thirty-one regions of China and their economic and demographic correlates” gave Shanghai Municipality an IQ of 108, tied with Jiangsu Province as the highest IQ place in China. Beijing Municipality also had an above Chinese average score of 107. Although keep in mind they collected their regional data from a “Chinese online IQ testing website”. I’m especially interested in Dataset 02 which I couldn’t get a hold of, as it has the largest sample size and the lowest IQ score of the bunch. If it had a wider geographical scope than the studies I had access to, it might represent a more plausible Chinese IQ score.

  5. tankerville says:

    The Atherosclerosis Risk in Communities Study, the one cited in table S2, included a neurocognitive battery. It seems like the data may exist to allow for a more direct comparison.

  6. Nothing but more white supremacy. Why is it that the master race must prove itself the master race over and over and over and over again? It seems you are trying to convince yourself of something. Stop twisting yourselves into confused cavemen and just enjoy the last of what was your dynasty. Because we all know you people will blow this planet into pieces before you bow down.

  7. Henry says:

    It is disingenuous to ask the question, If East Asians are supposedly so smart, why did they accomplish so little? Because most who pose the question simply ignore the fact that well over 50% of the inventions still in use today had their origin in China (scroll to the bottom for a brief list of these). To be sure, many of these things were re-invented by Europeans without little or no prior knowledge of similar devices–I say this as a response to the inevitable pushback that accompanies even the hint that anyone other than Westerners could invent things–or that Westerners owe a great deal to the Chinese, a position that is well backed by evidence compiled by the late Sinologist Joseph Needham (who provided the above estimate) and his research institute at Cambridge. But the point is not whether these Chinese innovations were re-invented elsewhere: the point is that the Chinese were highly inventive and were consistently so over many centuries–despite the lack of cross-pollination resulting from their remote location (they were separated by thousands of miles, and by mountain ranges and deserts, from cultural centers in the Middle East, North Africa and Europe.) Indeed, the noteworthy thing is that they came up with so many innovations despite the lack of economic incentives or the military competition that instigated technological progress in Europe.
    Yet despite their technical innovations, arguably their most influential one was the idea of meritocracy. That is, the best and the brightest should rule, without reference to their birth or social circumstances. This was first promulgated by Confucius c. 500 BC, when he said that the sons of princes and kings, if of poor ability, should become commoners, and that the sons of commoners, if of good ability, should become the rules. This idea was refined over the ensuing centuries, and finally became enshrined in the imperial examinations which were first held in the 2nd century AD. As Europeans sought to make their aristocratic societies more egalitarian and democratic in the 19th century, they looked to China for inspiration: for instance, the Northcote-Trevelyan Report proposed a set of examinations, based on the Chinese “Mandarin” system, for selecting young men for the civil service. Britain’s Mandarin system was then followed by similar initiatives elsewhere in Europe. The idea that the landed elite should at least be partially replaced by those with smarts and character was a revolutionary idea in mid-19th century Europe and even in the U.S.–but it was a very old Chinese innovation.

    NOW for the fun stuff: a short list of inventions and ideas from ancient China, off the top of my head: Paper, printing, paper money, gunpowder, guns, fireworks, cannons, multistage rockets, compass, matches, caliper grips, fishing rod and reel, seismograph, canal locks, rudder, bulkheads, paddle wheel, segmental arch bridges, mechanical clock, deep drilling, use of natural gas for cooking, wheelbarrow, cast iron, blast furnace, chrome, interchangeable parts (sorry, Eli Whitney), seed drill, collar harness for horses, crossbow, understanding of blood circulation, matrix mathematics (methods of finding N unknowns in N number of equations, for example)–critical to modern science, first to predict comet return, differential gear (demonstrated on chariots–all automobiles have these), umbrella, chain drive, binomial equation (“Pascal’s triangle), infinitesimal mathematics (not quite calculus, but methods for determining area of circles/volume of spheres using limits).
    that’s for starters. But that’s already a lot. Especially amazing coming from a country that was relatively isolated for innumerable centuries. Europe and the Middle East and Africa, by contrast, were essentially neighbors.
    It should be noted that the existence of paper money itself reflects a sophisticated understanding of value and money, as well as sophisticated institutions capable of apportioning value to such currency and managing issues of supply and demand. Modern banking systems are rooted in principles that have their origin in 9th century China.

  8. GC says:


    You have cited Christainsen (2013) on more than one occasion. Christainsen classified test-takers according to Cavalli-Sforza’s genetic clusters, with a few modifications for more recent research in population genetics. However, as you know, Cavalli-Sforza looked at alleles for blood groups, blood proteins, etc., and not directly at intelligence alleles. If you start to put Piffer’s factor scores in place of Cavalli-Sforza’s genetic clusters, there are important implications for groups’ genotypic IQs, especially in the cases of Southeast Asians and North Africans/South Asians. Piffer (so far) has found these groups to have intelligence alleles with frequencies close to those for ethnic Europeans. Thus, when you start re-doing Christainsen’s regressions, the phenotypic IQs of these groups are seen. not as reflective of big genetic differences, but of deficiencies in education and nutrition.

    Piffers findings actually imply rather modest differences across groups in genotypic IQ, although sub-Saharan Africans (not just Pygmies and Bushmen) might still end up substantially below Europeans. So far I have Bangladeshi-Bengalis ending up only about 5 points below Brits, and Mexicans in Los Angeles ending up 5-6 points behind. The gap between Brits and East Asians is SMALLER.

    The irony is that Piffer is being seen by many as supplying support for a hereditarian point of view,

    • Chuck says:

      Christainsen’s (2013) results look about right insofar as they show broad geographical patterns such that N.E. Asians have measured scores of about 105, S.E. Asians 92, S.E 85, etc. I was just perusing through: “Relationship between anthropometric indicators and cognitive performance in Southeast Asian school-aged children”. It was striking how similar the scores were across the various S.E Asian countries. The inaccuracy of Christainsen’s racial grouping are of no consequence in this case as the nations within each grouping plausibly faced similar cognitive evolutionary pressures, just as they plausibly currently face similar environmental ones. Also, the groupings are not largely off. At a finer grain of analysis, Caucasoids can more or less be separated into Europeans, South Asians, and Greater Middle Easterners. African, of course, form a distinct continental race. As for East Asians, Northern ones tend to be genetically closer to each other than to southern ones. That is, geographic distance generally predicts genetic distance.

      As for Piffer’s (very tentative results), I know not what you’re piping about. His method doesn’t allow for a direct determination of the magnitudes of relevant genetic differences. One is simply given factor scores for selective pressure. A difference of 1 in the selection score could equal a difference of 0.1, 1, or 2, etc. in terms of genotypic standardized differences. That’s a real weakness of the method. As for a hereditarian hypothesis, I’m not expecting large differences; I’ll be presently surprised is even modest ones are found. I’ve found a lot of anomalies, especially when it comes to migrant scores.

      • You keep dropping hints like this:

        “As for a hereditarian hypothesis, I’m not expecting large differences; I’ll be presently surprised is even modest ones are found. I’ve found a lot of anomalies, especially when it comes to migrant scores.”

        And this:


        “Right now I am reviewing genomic ancestry x education studies. There are dozens of them and they don’t always show the predicted correlations, at least when it comes to Amerindian ancestry.”

        What exactly are you referring to? Have you written about any of this in greater detail? Would you mind sharing the data or links to the data?

        I get the sense that you might be sitting on a wealth of data. The comments I quoted are enticing, but by themselves simply leave me with a lingering sense of curiosity.

        I’m always interested in the anomalies, that which doesn’t fit the predictions. In my own searching for useful data, I’ve found many things that complicate particular theories. This implies there might be something entirely else going on here, that we don’t yet fully understand which factors are involved and how they interact. Obviously, we haven’t figured out how to control for all or even most of the confounding factors, and that is a major problem for anyone making claims of causation.

        I’m for following the data to see where it leads. But following the data won’t be easy.

        • Chuck says:

          I posted on the admixture x SES-education associations over at Human Varieties. Obviously such is open to multiple explanations and the results need to be critically evaluated; they represent an explanandum, nonetheless. Regarding migrant performance, I was thinking of e.g., the performance of British Blacks. See, for example: here, here, and here. Considerations were discussed in the comment sections. The issue is complex. I’m still collecting data on differences in e..g, Canada, France, and the Netherlands. In Latin America, “indigenous” and “afro-descendants” consistently underperform others on cognitive ability tests (e.g., table 3.3 on. But I don’t know much about the social circumstances in the those countries.

          • Thanks for the links! I greatly appreciate that you gave me such a helpful reply. My main interest is the data. I don’t have set conclusions about most things because I think we have a long way to go. I’ll check out what you’ve linked here.

  9. Goku says:

    I bet its epigenetics causing the differences. You do realize the differences between peoples genes are very very tiny right? Its just some difference in the same nitrogen bases that cause a variant of the same gene, its technically still the same gene and those bases are modifiable to the point of silencing the gene entirely or switching them on… or regulating the expression up or down.

    If you take pygmies for example what causes them to not be as big/tall as the average joe in other groups is that they stop producing(or don’t produce as much) growth hormones(and probably some other stuff) sometimes at birth, before birth or after a certain age. There is nothing that says they cant produce the same amount of growth hormones, or whatever else is needed. They technically have the same gene and whatever variants they have technically can as proven in some of the cases can produce the same amount of growth hormone or stuff for being tall… only difference is it stops functioning or isn’t functioning from birth.

    Even Neanderthals seem to show such differences in the SAME GENES.

    It technically could be environment… pretty much all the way too.

    Also in mouse studies(we share 88% of genes with them) it shows that epigenetics, even simple things like diet can have rather large effects which are hereditary.

    Also from what I have read and some other studies on I starvation, the effects were pretty much irreversible until about 2 to 3 generations in of good environment.

    So yeah, there is a possible big chunk of environment in there left. VERY BIG…. possibly.

    • Chuck says:

      This silliness has been discussed ad nauseam. Genome-wide Complex Trait Analysis has already established that common SNPs explain a large fraction of the phenotypic variance (between individuals) in traits like intelligence. Thus, those “very tiny” differences — in coding and non-coding DNA — explain quite a bit. Epigenetics adds little to the debate about group differences. Genomic ancestry is still robustly correlated with relevant outcomes.

      • Here is the main problem of this entire discussion. It turns out we know a lot less than we thought we knew. Many of our fundamental assumptions are being called into question and all the research that was based on those assumptions.

        We can no longer honestly claim percentage estimates about genetic vs environmental influence. It isn’t just that past research wasn’t controlling for all confounding factors. Genetic researchers are beginnning to realize they don’t even know how to control for all confounding factors because quite a few apparently are unknown at present.

        We don’t even know how to attempt to disentangle these factors so as to isolate them all. More importantly, we can’t figure out how to separate genetics from the environmental background of this complex web of confounding factors.

        Researchers have a long way to go before much can be concluded with any certainty.


        The Genius in All of Us: New Insights into Genetics, Talent, and IQ
        By David Shenk
        Kindle Locations 2013-2074

        These histones protect the DNA and keep it compact . They also serve as a mediator for gene expression, telling genes when to turn on and off. It’s been known for many years that this epigenome ( “epi-” is a Latin prefix for “above” or “outside”) can be altered by the environment and is therefore an important mechanism for gene-environment interaction.

        What scientists didn’t realize, though, was that changes to the epigenome can be inherited. Prior to 1999, everyone thought that the epigenome was always wiped clean like a blackboard with each new generation.

        Not so, discovered Enrico Coen. In the case of the Peloria toadflax flower, a clear alteration to the epigenome had subsequently been passed down through many generations.

        And it wasn’t just flowers. That same year, Australian geneticists Daniel Morgan and Emma Whitelaw made a very similar discovery in mice. They observed that their batch of genetically identical mice were turning up with a range of different fur colors —differences traced back to epigenetic alterations and passed on to subsequent generations. What’s more, they and other researchers discovered that these fur-color epigenes could be manipulated by something as basic as food. A pregnant yellow mouse eating a diet rich in folic acid or soy milk would be prone to experience an epigenetic mutation producing brown-fur offspring, and even with the pups returning to a normal diet, that brown fur would be passed to future generations .
        After that, more epigenetic discoveries piled in one after another:

        – In 2004, Washington State University’s Michael Skinner discovered that exposure to a pesticide in one generation of rats spurred an epigenetic change that led to low sperm counts lasting at least four generations.
        – In 2005, New York University’s Dolores Malaspina and colleagues discovered age-related epigenetic changes in human males that can lead to lower intelligence and a higher risk of schizophrenia in children.
        – In 2006, London geneticist Marcus Pembrey presented data from Swedish medical records to show that nutritional deficiencies and cigarette smoking in one generation of humans had effects across several generations .
        – In 2007, the Institute of Child Health’s Megan Hitchins and colleagues reported a link between inherited epigenetic changes and human colon cancer .

        Welcome back, Monsieur Lamarck! “Epigenetics is proving we have some responsibility for the integrity of our genome,” says the Director of Epigenetics and Imprinting at Duke University, Randy Jirtle . “Before, [we thought that] genes predetermined outcomes. Now [we realize that] everything we do—everything we eat or smoke— can affect our gene expression and that of future generations. Epigenetics introduces the concept of free will into our idea of genetics.”

        And that of future generations. This is big, big stuff— perhaps the most important discovery in the science of heredity since the gene.

        No one can yet measure the precise implications of these discoveries, because so little is known. But it is already clear that epigenetics is going to radically alter our understanding of disease, human abilities, and evolution. It begins with this simple but utterly breathtaking concept:
        Lifestyle can alter heredity.

        Lamarck was probably not correct about the giraffe in particular, and he was certainly wrong about inherited characteristics being the primary vehicle of evolution. But in its most basic form, his idea that what an individual does in his/ her life before having children can change the biological inheritance of those children and their descendants— on this he turns out to have been correct. (And two hundred years ahead of everyone else.) Quietly, biologists have come to accept in recent years that biological heredity and evolution is a lot more intricate than we once thought. The concept of inherited epigenetic changes certainly does not invalidate the theory of natural selection, but it makes it a lot more complicated. It offers not just another mechanism by which species can adapt to changing environments, but also the prospect of an evolutionary process that is more interactive, less random, and runs along several different parallel tracks at the same time. “DNA is not the be all and end all of heredity,” write geneticists Eva Jablonka and Marion Lamb . “Information is transferred from one generation to the next by many interacting inheritance systems . Moreover, contrary to current dogma, the variation on which natural selection acts is not always random … new heritable variation can arise in response to the conditions of life.”

        How do these recent findings impact our understanding of talent and intelligence? We can’t yet exactly be sure. But the door of possibility is wide-open. If a geneticist had suggested as recently as the 1990s that a twelve-year-old kid could improve the intellectual nimbleness of his or her future children by studying harder now, that scientist would have been laughed right out of the conference hall. Today, that preposterous scenario looks downright likely:

        “Washington, D.C.— New animal research in the February 4 [2009] issue of The Journal of Neuroscience shows that a stimulating environment improved the memory of young mice with a memory-impairing genetic defect and also improved the memory of their eventual offspring . The findings suggest that parental behaviors that occur long before pregnancy may influence an offspring’s well-being. “While it has been shown in humans and in animal models that enriched experience can enhance brain function and plasticity, this study is a step forward, suggesting that the enhanced learning behavior and plasticity can be transmitted to offspring long before the pregnancy of the mother,” said Li-Huei Tsai, PhD, at Massachusetts Institute of Technology and an investigator of the Howard Hughes Medical Institute, an expert unaffiliated with the current study.
        In other words, we may well be able to improve the conditions for our grandchildren by putting our young children through intellectual calisthenics now.

        “What else is possible? Could a family’s dedication to athletics in one or more generations induce biological advantages in subsequent generations?

        “Could a teenager’s musical training improve the “musical ear” of his great-grandchildren?
        Could our individual actions be affecting evolution in all sorts of unseen ways?”

        “People used to think that once your epigenetic code was laid down in early development, that was it for life,” says McGill University epigenetics pioneer Moshe Szyf. “But life is changing all the time, and the epigenetic code that controls your DNA is turning out to be the mechanism through which we change along with it. Epigenetics tells us that little things in life can have an effect of great magnitude.”

        Everything we know about epigenetics so far fits perfectly with the dynamic systems model of human ability. Genes do not dictate what we are to become, but instead are actors in a dynamic process. Genetic expression is modulated by outside forces. “Inheritance” comes in many different forms: we inherit stable genes, but also alterable epigenes; we inherit languages, ideas, attitudes, but can also change them. We inherit an ecosystem, but can also change it.

        Everything shapes us and everything can be shaped by us. The genius in all of us is our built-in ability to improve ourselves and our world.

        Kindle Locations 1624-1657

        To say that there is much we don’t control in our lives is a dramatic understatement, roughly on the order of saying that the universe is a somewhat large place. To begin with, there are many influences we can’t even detect. In 1999 , Oregon neuroscientist John C . Crabbe led a study on how mice reacted to alcohol and cocaine. Crabbe was already an expert on the subject and had run many similar studies, but this one had a special twist: he conducted the exact same study at the same time in three different locations (Portland , Oregon; Albany, New York; and Edmonton, Alberta) in order to gauge the reliability of the results. The researchers went to “extraordinary lengths” to standardize equipment, methods, and lab environment: identical genetic mouse strains, identical food, identical bedding, identical cages, identical light schedule, etc. They did virtually everything they could think of to make the environments of the mice the same in all three labs.

        Somehow, though, invisible influences intervened. With the scientists controlling for nearly everything they could control, mice with the exact same genes behaved differently depending on where they lived. And even more surprising: the differences were not consistent, but zigged and zagged across different genetic strains and different locations. In Portland, one strain was especially sensitive to cocaine and one especially insensitive , compared to the same strains in other cities. In Albany, one particular strain— just the one— was especially lazy. In Edmonton , the genetically altered mice tended to be just as active as the wild mice, whereas they were more active than the wild mice in Portland and less active than the wild mice in Albany. It was a major hodgepodge.

        There were also predictable results. Crabbe did see many expected similarities across each genetic strain and consistent differences between the strains. These were, after all, perfect genetic copies being raised in painstakingly identical environments. But it was the unpredicted differences that caught everyone’s attention. “Despite our efforts to equate laboratory environments, significant and, in some cases, large effects of site were found for nearly all variables,” Crabbe concluded. “Furthermore, the pattern of strain differences varied substantially among the sites for several tests.”

        Wow. This was unforeseen, and it turned heads . Modern science is built on standardization; new experiments change one tiny variable from a previous study or a control group, and any changes in outcome point crisply to cause and effect. The notion of hidden, undetectable differences throws all of that into disarray. How many assumptions of environmental sameness have been built right into conclusions over the decades?

        What if there really is no such thing? What if the environment turns out to be less like a snowball that one can examine all around and more like the tip of an iceberg with lurking unknowables? How does that alter the way we think about biological causes and effects?

        Something else stood out in Crabbe’s three-city experiment : gene-environment interplay . It wasn’t just that hidden environmental differences had significantly affected the results. It was also clear that these hidden environments had affected different mouse strains in different ways— clear evidence of genes interacting dynamically with environmental forces.

        But the biggest lesson of all was how much complexity emerged from such a simple model. These were genetically pure mice in standard lab cages. Only a handful of known variables existed between groups. Imagine the implications for vastly more complex animals— animals with highly developed reasoning capability, complex syntax, elaborate tools, living in vastly intricate and starkly distinct cultures and jumbled genetically into billions of unique identities. You’d have a degree of GxE volatility that would boggle any scientific mind— a world where, from the very first hours of life, young ones experienced so many hidden and unpredictable influences from genes, environment, and culture that there’d be simply no telling what they would turn out like.

        Such is our world. Each human child is his/ her own unique genetic entity conceived in his/ her own distinctive environment , immediately spinning out his/ her own unique interactions and behaviors. Who among these children born today will become great pianists, novelists, botanists , or marathoners? Who will live a life of utter mediocrity? Who will struggle to get by? We do not know.

        • Chuck says:

          “Here is the main problem of this entire discussion. It turns out we know a lot less than we thought we knew….”

          No, this is stupid.

          BDS: “We can no longer honestly claim percentage estimates about genetic vs environmental influence.”

          Yes, this can be done. For example, epigenetics can readily be incorporated into traditional biometric analyses. See for example: Tal, et al. (2010) “Epigenetic contribution to covariance between relatives”. This isn’t done because convergent evidence from multiple overlapping methods indicate that epigenetics doesn’t account for a significant portion of variance in the traits (here) under discussion, the same applies to GxE interaction These are factors, but they’re on the margin.

          BDS: “We don’t even know how to attempt to disentangle these factors so as to isolate them all. More importantly, we can’t figure out how to separate genetics from the environmental background of this complex web of confounding factors.”

          What you must mean is that you don’t understand the biometric approach. Tell me a variance component e.g., “shared environment” and I will note the method that can be used to disentangle it from other components.

          BDS: “Researchers have a long way to go before much can be concluded with any certainty”.

          According to whom?

          “Experts agreed that the following were sources of reasonable evidence for significant heritability of intelligence: monozygotic twins reared apart, comparisons of monozygotic and dizygotic twins, adoption studies, “patchwork” family studies.

          Asked: Is there sufficient evidence to arrive at a reasonable estimate of the
          heritability of intelligence in populations of developed countries?” 73% said Yes.”

          Assuredness is as much an epistemic as a scientific matter. If you raise the bar high enough, researchers could never conclude anything.

          • I never claimed that epigenetics can’t be included in studies. That would be silly.

            What I said is that we don’t know how to control or even disentangle all of the confounding factors, some related to epigenetics and others to environment,. Yes, some factors have been controlled for and disentangled, but certain studies that I shared demonstrate that many other factors remain uncontrolled for and not yet disentangled.

            That is simply a statement of fact. It may not be a satisfying conclusion, but it is simply where research is at present. Wanting to know more than we presently know doesn’t make it any better known.

            I meant entirely what I said, whether or not you understood what I said. That is a separate issue. The main point of my argument remains true. Simply making statements to the contrary doesn’t refute that.

            We are in the middle of a paradigm change. Much of this most challenging research is relatively new. Many genetic researchers still haven’t come to terms with these confounding factors. I have confidence that they will improve their methods, but that is no reason to dismiss an honest appraisal of the problems researchers still face.

            That mouse study is paradaigm-shattering. They were the genetically exact same mice under the best controlled conditions at several different locations. It was a better controlled study than has ever been done on human subjects. Despite all of these massive controls, the differing results were immense. It was entirely unexpected according to previous genetic theories.

            Nothing you stated or cited offers a satisfying response to that evidence.

      • Goku says:

        The tiny differences are whats changed by epigenetics. IT DOES NOT MATTER WHAT DIFFERENCE THEY CAUSE. See that little chemical that causes the gene to do something? Yeah thats waht gets changed. It does not need to change the sequence, it changes the bases already there.

        Also on top of that its not just the dna bases that get their biological makeup changed its the histones and overall binding around a gene. Epigenetics can change expression without changing the bases which is what you pointed out.

        The point is that the difference between people in their genetic code is changeable by environment and drastically so.

        Also those drastic changes you mentioned are only in rare variants, usually causing diseases, like sickle cell.

        • Goku says:

          Look here, even the messenger RNA(mRNA) that is referred to in that tiny tiny difference link of yours can be epigenetically controlled… and thus even the splicing.


          Here is an example of meditation influencing gene expression. Which demonstrably refutes the notion of “lifestyle does not = genes”.


          • Chuck says:

            “Which demonstrably refutes the notion of “lifestyle does not = genes”

            What are you’re talking about? Behavioral geneticists have long refuted the idea that lifestyle = genes. They’ve shown, for example, that shared and non-shared environment contribute non-trivially to a number of “lifestyle” conditioning traits. No one is a (behavioral) genetic determinist — these don’t exist and probably never have. If you can find one, past or present please reference. If you’re going to play stupid, don’t comment on this blog.

        • Chuck says:

          “Also those drastic changes you mentioned are only in rare variants, usually causing diseases, like sickle cell”

          For highly polygenic traits like height or IQ base pair differences explain the plurality of individual differences. Epigenetics likely explain near zero variance. Could some epigenetic effects induce large differences? If you tinker, you probably could make some people duller and smaller, but the the present epigenetic variance has little to no effect. For other traits it might be a more significant factor. To determine, this needs to be modeled biometrically — variance needs to be decomposed.

          • Goku says:

            “For highly polygenic traits like height or IQ base pair differences explain the plurality of individual differences.”

            Bahahahaha like those genes that have been correlated which the authors have no idea what they do, how they work, the mechanism or even if they actually cause the difference at all? This is what explains the difference hahahahahaha. Or how about those few(mainly 3) alleles that correlate for IQ… bahahahaha.

            Come on just give up already.

          • Chuck says:

            “like those genes that have been correlated which the authors have no idea what they do, how they work, the mechanism or even if they actually cause the difference at all? This is what explains the difference hahahahahaha. Or how about those few(mainly 3) alleles that correlate for IQ”

            Don’t be a retard — if it’s not an act, maybe try some nootropics.

            On the one hand, there’s no evidence of epigenetic influence on these traits. No one has ever shown that epigenetic factors in general or in specific can explain even a minute portion of the variance. On the other hand, it’s established the common variants jointly explain the plurality of variance. Also, for both mentioned traits gene pathways have been identified e.g., “Pathway analysis, using genes selected on the basis of relaxed criteria, revealed notable differences between these two traits. gF appeared to be characterized by genes affecting the quantity and quality of neurons and therefore neuronal efficiency, whereas long-term depression (LTD) seemed to underlie gC” (GWAS-based pathway analysis differentiates between fluid and crystallized intelligence) — or “using a network-based approach, we further reconstructed an IQ-related pathway from known human pathway interaction data. Based on this reconstructed pathway, we incorporated enriched drugs and described the importance of dopamine and norepinephrine systems in IQ-related biological process.” (A systems biology approach to identify intelligence quotient score-related genomic regions, and pathways relevant to potential therapeutic treatments)

            Further, half of a dozen specific alleles have been well replicated. Scores of other alleles and genes have been identified (though not well replicated), and are listed here: http://iqdb.cbi.pku.edu.cn/index.jsp

  10. It seems strange that this kind of paradigm-shattering research doesn’t lead you to question your assumptions and conclusions. When I first came across it, I found it mind-blowing. My response was amazement, excitement, and curiosity. If new evidence could be that revolutionary, I thought to myself, what else might I have thought I understood actually be wrong or less than certain?

    When I first learned about genetics, it was in high school during the early 1990s. That was more than 20 years ago. I wasn’t taught about environmental influences, gene-environment interactions, or epigenetics. I wasn’t even taught multiple gene interactions. I wasn’t taught it because even geneticists didn’t know much of what we now know. What I was taught was a mainstream determinist model of genetics, the same basic model that had been taught for decades. It was presented fairly simplistically as one gene causing one trait.

    For a long time, I never came across research that challenged that old model. I wasn’t interested in the topic until recently and so I didn’t previously go looking for other evidence, even I had known it existed to look for. The mainstream media certainly didn’t inform me of the interesting new discoveries that were happening. These past few years have been an education for me, as I’ve finally started studying the topic for myself.

    How could I have gotten this far in life without coming across so much of this fascinating research? Why are these paradigm-changing studies not being reported and discussed widely in the mainstream? Why do most Americans still believe in the old determinist model of genetics?

    That mouse study is truly perplexing. None of the old theories can explain it. It was always assumed that if researchers controlled for all obvious genetic and environmental factors it should lead to the same results. Slight variances were expected, but nothing to that extreme of differences. It demonstrates possibly very minor differences, so small as to be undectable, can lead to major alterations in end results. It demonstrates how powerful environmental conditions can be, even when they are being controlled for with the best methods researchers know how to use. In the uncontrolled conditions of human lives, the environmental influences would be even more powerful.

    That isn’t just meager evidence that can be explained away. It radically challenges at a fundamental level much of what geneticists previously thougth was true. No one knows what this ultimately means. No one! To me, that makes this period of research to be very exciting. We don’t know where it will lead, what new theories will arise from the evidence. All we know is present theories are lacking in explanatory power. From a scientific perspective, not knowing is a great place to be. It means there is something yet to be discovered. Scientists always have to begin at the limit of present knowledge. We appear to be running up against one of those limits.

    I wish we could have a discussion about what that might mean, without preconceptions, just simply follow the evidence wherever it might lead. These are exciting times. The proper response is to be excited. Let’s spend our time exploring the new, instead of wasting time defending the old.

    • Chuck says:

      “It seems strange that this kind of paradigm-shattering research doesn’t lead you to question your assumptions and conclusions…I wasn’t taught about environmental influences, gene-environment interactions, or epigenetics. I wasn’t even taught multiple gene interactions. I wasn’t taught it because even geneticists didn’t know much of what we now know.”

      The long version of the standard biometric model, which goes back more than one half of a century, decomposes phenotypic variance into that due to additive genetics, epistasis, assortative mating, dominance, shared environmental, unshared environmental, gene x environment interactive, gene x environment correlation, and error. Simplified ACE models (additive genetic, shared environment, and unshared environment) are discussed, when they are, because these fit the data best or because others models aren’t testable given the data at hand. The inclusion of “epigenetics” into the laundry list of variance components is neither a theoretical nor practical problem (except insofar as the particular data set limits decomposition). For traits like IQ, it’s obvious that epigenetics can not explain much inter individual variance. As noted, GCTA provides a direct estimate of variance due to common genetic variants. For IQ, these explain 50% of the inter-individual variance. This leaves 50% to be explained by all other factors mentioned above plus uncommon variants. Error typically explains 5%. Shared environment (directly estimates with adoptions or unrelated sibs reared together) explains 0 to 15%, nonshared environment (directly estimated by comparing MZ twins) explains 20%, non additive genetic variance (estimated with elaborate kinship designs) explains another 10 to 25%. There’s virtually no room for significant variance due to epigenetics at least for this trait. Epigenetics is given a lot of hype in some quarters because it’s thought to somehow call into question behavioral genetic assumptions, showing that genes and environment are conflated. But this “paradigm-shattering research” came one decade too late; behavioral genetics has since evolved. This was from almost a decade ago: http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.0020041 None of this is to say that epigenetics isn’t marginally interesting. And when behavioral epigeneticists come along and demonstrate that epigenetic variance for traits X, Y, and Z (decomposed in a standard biometric model) is nontrivial, they will be worth listening to. It just doesn’t somehow undermine the new behavioral genomics or for that matter classical biometrics. So to reply: regarding the traits which I have been discussing, behavioral genomic research largely confirms the validity of the assumptions; epigenetics has nothing to say on the matter.

      I don’t know what you were taught at school — maybe you could reference works of “mainstream” “genetic determinists” who knew nothing about ..

      • You still refuse to take new evidence seriously. That is almost as amazing as is the new evidence.

        • Chuck says:

          I don’t know what you are talking about. You originally said:

          “I wasn’t taught about environmental influences, gene-environment interactions, or epigenetics. I wasn’t even taught multiple gene interactions. I wasn’t taught it because even geneticists didn’t know much of what we now know.”

          You seem to be confused or ill informed because behavioral geneticists have been discussing environmental influences from the start. Distinguishing between genetic and environmental influences — or nature and nurture — is why Galton established the field in the first place. As for gene-environment and multiple gene interactions these have been recognized for a century. For example, William Bates discussed epistatis in 1910. Likewise gene-environment interactions were recognized in the early 20th century and were referenced in criticisms of early eugenic aspirations e.g., (Myerson et al. 1936) and, more generally biometric, decomposition. Lewontin later continued with the critique in the 60s and 70s. Biometricians, though, have developed sophisticated methods for detecting both and have incorporated these factors into their method. This is why both epistatis and G x E interaction are a part of the formal biometric equation, P = … Thus, the only relatively novel factor here is epigenetics. But as I noted this likewise can be Incorporated into the biometric model. It just adds another term.

          Regarding the focus here, this is not of particular interest because it’s established that common additive genetic variants explain a significant fraction of e.g, intelligence (50%) and a non trivial fraction of other traits e.g., education attainment (20%). Since a non-trivial amount of variance within populations is so explain an indefinite amount of variance between can plausibly be so explained by that same factors.

          Thus, nothing changes. Our model for between groups differences is still:

          P = G (genetics) + Other (all other factors: shared environment, non shared environment, G x E interaction, etc.).

          Adding another sub-component (e.g., epigenetics) to the “other” factors doesn’t change the equation. We still get: P = G (genetics) + Other. Nor does clarifying that G = additive genetics + dominance + epistasis, etc. Ether way, it’s still not established that genes explain a large chunk of the between group variance — so this is an area of interest, especially given the politics of group differences, and the tendency of attributing them to “racism” past or present. Now, the same applies regarding a standard sociological model:

          P = discrimination + Other (all other factors: shared environment, non shared environment, G x E interaction, genetics, etc.).

          Pointing out that there’s another possible contributing component to group differences, doesn’t somehow undermine or call into question the research done. Sure, instead of “uber discrimination” or “colorism” some group differences could, in principle, be due to region of origin epigenetic effects. Indeed, this is what early, pre-Darwin racialists e.g., Buffon and Kant, whose theories of race differences were quasi-epigenetically based, proposed. Ok, but that doesn’t overturn theory, method, or findings. At most it requires a re-evaluation of certain lazy, sloppy interpretations. Now, I have been very cautious when it comes to drawing conclusions. I have always acknowledge that this matter is unsettled — that the cause of group differences is unestablished — indeed, as I have said numerous times, this is how I justify my focus on the matter. So there is nothing for me to re-evaluate,

          Honestly, I don’t see where the confusion lies or why the research mentioned impresses you so. How exactly do you imagine that it changes the dynamics of the discussion. We can take a trait like height, for example. Variance in this trait clearly is substantially genetically conditioned and certain ethnic and racial groups differ in height due to genes e.g., Mbuti Pygmies versus Bantu, ethnic Dutch versus ethnic Sicilians. How has any of this changed?

        • Goku says:

          Yeah thats my surprise too, out of all the people Chuck is also unable to let go.

  11. Boratismio says:

    Hey Chuck, I have a question about the alleles. Piffer in open psych had an article relating to his first study with those 10 alleles(3 of which replicated some of the time) in which he stated that the genotypic IQ for SSA Africans is ~90.

    I went and checked the newer data points using the 7 alleles and the allele gaps are even closer. In the first African score to European is <half eg: British 36 and Yoruba 15. When using the new 7 its more like 47 to 25 polygenic score. Wouldn't that mean even more than 90 for Africans? That would correspond well with the UK data.

    Also that 7th allele is even less reliable than the 6. The sample sizes are too small, or am I wrong?

    Also on top of that most of these alleles have unknown gene associations like LOC(something) in the education studies but they are also associated to other genes that are then associated with things like cardiovascular disease and metabolic things. They are technically all over the place.
    Check out this one for example. The same gene has been associated many times before to many other things it seems.

    • Chuck says:

      Piffer’s method is incapable of estimating the magnitude of genotypic differences; it can only estimate the relative magnitude of selective pressure. (This is why other evidence is needed.) The results would indicate that e.g., Africans experienced -2 factor differences of selection while East Asians experienced + 1. This amount of selection could represent any amount of genetic differences. Perhaps, for example, Africans were selected in g – 0.2 SD, while East Asians were selected + 0.1 SD. To translate the selection factor difference into standardized genotypic ones, Piffer assumed that the Europe-Japanese difference of about 5 IQ points was fully genetic; he noted that this corresponded to 1.0 selection units. Since Africans were 2 selection units below Whites he deduced a genotype IQ of around 90. As I have pointed out in the thread, this deduction involved a number of errors/problems. For one, there is a basic conceptual error: phenotype IQ due to genes is not the same as geneotypic IQ. Imagine that the 5 point phenotypic advantage that Japanese seem to have is fully due to genes; since the correlation between additive genes and phenotypic IQ in both the Japanese and European populations is about 0.77 (SQRT(narrow heritability), the genotypic differences would be 5/0.77 or 6.5. For another, this was an unrelaible way of converting selection into genetic differences. As noted, one should use the regression based on a number of populations. As for the 7 alleles, I recall that Piffer et al looked at this and found roughly similar results. I’m not sure why you’re getting different ones. Note, with the method, ones doesn’t average allele frequencies, rather one uses the principle component of them. So if the frequency of an alleles moves in the “wrong” direction relative to the rest, it will load on a second factor interpreted as “error”. I vaguely recall that one of the 7, which happened to show high frequencies in African populations, did this; consistent with the idea that it represented noise, it didn’t show a positive effect in a recent Chinese study. Anyways, I would wait until another dozen alleles are replicated before trying to guesstimate effect sizes of genotypic differences. In the meantime convergent evidence can be sought.

      It would be helpful to have more immigrant data. Also, keep in mind that smallish individual level effects can amplify on the population e.g., national/sub-cultural level. It’s one thing to be slightly dull, it’s another to be slightly dull and also have slightly dull siblings, peers, parents, teachers, governors, etc.

      • Boratismio says:

        Oh ok.

        As with the the Chinese study it only replicated 3 of the first associations. What happened to the other 4(including the one more common in Africans)? Did they try them? The black gene is rs9320913, I think you meant rs11584700?

        I don’t know, when I look at the data… especially the genes and pathways the alleles are supposed to be involved with its all over the place and some of these have been associated before with things other than cognition.

        How sure are you these alleles are even causing an effect?

        • Chuck says:

          ” What happened to the other 4(including the one more common in Africans)? Did they try them? The black gene is rs9320913, I think you meant rs11584700?”

          They only looked at the three top alleles discussed in Rietvald et al. (2013); three of the other alleles are discussed in a paper only recently published. Zhu et al. (2015) would have been prepared and submitted prior to this.

          “The black gene is rs9320913, I think you meant rs11584700?”

          Whatever the case, the point is that you need to use principle component analysis.

          “How sure are you these alleles are even causing an effect?”

          I’m sure that they’re causally related to IQ; I wouldn’t wager on a specific pathway, though. Generally, I wouldn’t get hung up on these results. They represent tentative allelic support. When e.g, 20 well replicated alleles have been found, we’ll have a much better idea.

          “If you look at the Kalash… Something is off here.”

          Obviously, you have a lot of noise in the data. Nonetheless the non-trivial correlation between phenotypic differences and (the principle components of allele frequencies) indicates that not everything is off, no?

          • Boratismio says:

            I will agree not everything is off, from what I know. However the way its shaping up is why I have increasing doubts.

            Also before I go away.
            The recent paper didn’t detect the first three. Shouldn’t it have? Very good sample size and similar pop.

            Thats the last question I promise, you are right lets not put too much thought into it right now.

          • Chuck says:

            If you looked, the more recent study used the same sample.. Quote: “The first stage of our two-stage procedure consisted of conducting a GWAS meta-analysis on years of schooling in a pooled Education Sample (n = 106,736) using the same analysis plan as in the work by Rietveld et al. (11) and the same cohorts, except for omitting the individuals who we include in the second stage. To obtain our set of education-associated SNPs, we selected all SNPs with P value < 10−5 from the first-stage meta-analysis results and then pruned for linkage disequilibrium." The alleles for the more recent analysis were significantly associated with the cognitive measure at P < 0.00001 The other ones might have been more educational. See table S4 here:
            http://www.pnas.org/content/suppl/2014/09/06/1404623111.DCSupplemental ..
            The p-values are different for cognitive performance and educational attainment. Different but overlapping constructs…

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