Is there no population genetic 'support' for a racial hereditarian hypothesis?

(10/18/2014 update:  data from two additional studies — Martínez et al.  (2007) and Ruiz-Linares et al. (2014) — have been added.)

Over the last decade, scores of large scale admixture-mapping studies have been conducted largely in an attempt to elucidate the origin of ethnic disparities in disease rates and medical outcomes.  In the simplest type of such studies, researchers determine if there is a robust association between genotypically defined continental racial ancestry (typically: African, European, and Amerindian) and relevant outcomes in admixed populations.  To control for potential confounding effects, measures of educational attainment and other indexes of SES are often included in the analyses.  These variables are often treated as environmental indicators, which is odd, since within populations they are found to be under non-trivial genetic influence.  For example, based on a recent international meta-analysis of biometric studies involving 51,545 kinship pairs, Branigan, et al. (2013) found that educational attainment had a kinship-based heritability of 0.40, meaning that genes explained 40% of inter-individual educational differences; based on a sample involving 7,959 individuals, Rietveld et al. (2013, table S12)  found a GCTA-based heritability, one which takes into account only the effects of population-wide common genetic variants, of 0.22.  These results were replicated by Marioni, et al. (2014, table 3), who found a kinship-based heritability of 0.40 and a GCTA-based one of 0.21.  When genes explain some of the variance in a trait within groups, they plausibly explain an indefinite portion of the variance between groups.  Curious it is, then, that these external outcomes are often assumed to represent environmental influences between groups.

Recently, Piffer (2013, 2014, 2014) and Piffer and Kirkegaard (2014) found that continental races, that is, continental level human groupings where members are arranged by overall genetic affinity,  differ in their allelic frequencies for well-replicated alleles that are causally associated with both educational attainment and psychometric intelligence.   The results are tentative, but they support the position that the well documented regional variation in cognitive ability, and, by way of it, regional variation in levels of education has a partial genetic basis.  Based on the results and, more generally, the typical racial hereditarian position (with respect to cognitive ability), which these results generally support, one would predict that Sub-Saharan African, Amerindian, and Pacific Islander ancestry would be, with high consistency, negatively correlated in admixed samples with educational attainment and other cognitive correlates such as income and occupational prestige.  Conversely, one would predict that European and East Asian ancestry would be positively so correlated.  As educational attainment and SES are substantially environmentally influenced, as the magnitude of associations depend on the degree of range restriction in relevant ancestry components, as statistical significance depends on sample size, and as there are a number of confounding factors, one would not expect a perfect degree of consistency, just a relatively high one.

Other hypotheses could account for a fairly consistent association between continental ancestry and the mentioned outcomes, if such exists.  For example, the associations could be attributed to fairly consistent colorism, to assortative exogamany, or to intergenerationally transmitted environmental advantage.  Regarding the first, it could be said that phenotypic based discrimination mediates the association between genotypic ancestry and outcomes.  Regarding the second, it could be argued that historic racial groups, at least in the Americans, stared out with the same genotypic potential, but due to differences in initial social standing, assortatively outbred such to produce ancestrally correlated genetic stratification in educational and SES related latent traits (see e.g., Valenzuela, 2011).  Finally, it could be maintained that historic racial groups stared out with different shared environmental levels related to the relevant traits and that they have been intergenerationally passing differences on, along family lines, for hundreds of years. Nonetheless, it’s not inevitable that admixed populations will show fairly consistent associations between ancestry and outcomes. Indeed, admixture studies have apparently uncovered some ethnic, in the sociological sense, differences which are unassociated with continental racial ancestry.  For example, despite significant U.S. Black-White differences in the rate of kidney disease progression, Peralta (2006) found no association between disease progression and European ancestry in the African American population.   They reasoned:  “This suggests that the tremendous burden of kidney disease in African Americans may be more attributable to environmental factors than to their common genetic African ancestry.”  Given that such an association is not necessary, a finding of one would represent population genetic support for a racial hereditarian model, since this model would predict such and since some competing models do predict otherwise.

To explore whether or not there is such an association, we conducted a review of the admixture mapping studies published between January of 2004 and September of 2014.  We relied on “google scholar” as a search engine and used search phrases such as “admixture African/Amerindian  socioeconomic/education”.  In total these searches turned up approximately 20,000 hits in descending order of relevance to the search terms.  The first 1,500 abstracts were skimmed.  Approximately 250 papers were identified as potential sources and read.  Of these, 31 reported, for individuals residing in the Americas, associations between continental ancestry (e.g., European, Amerindian, Sub-Saharan African, East Asian, and Pacific Islander) and some index of educational attainment or socioeconomic status.  The following countries and groups were represented: Brazil, Chile, Colombia, Costa Rica, Mexico, Peru, Puerto Rico, Trinidad and Tobago (Black population), and U.S. (African, Hispanic, Native, Asian, Pacific Islander, and Caucasian Americans). The results are summarized in the table below:


(Sources and excerpts from the papers can be found in the excel file. The PDF can be downloaded here.)

Unfortunately, the data did not allow us to compute a meta-analytic estimate of the associations (correcting for range restriction, etc.).  We attach the studies, though, so that readers could peruse the reported results if so desired.

None of the associations went in a direction opposite to that predicted.  Six, all from the U.S. were non-significant in at least one index. Of these samples, the non-significance could readily be attributed to statistical power in two:  Gower, et al. (2003) found that African Ancestry was negatively but not significantly correlated with education among U.S. Caucasian children — this was most likely due to range restriction in admixture, as suggested by the percent admixture reported for this sample; Allison et al. (2010) found that Amerindian ancestry was negatively but non-significantly correlated with educational attainment in a Hispanic American sample — the P-value, though, was 0.11 for education and 0.07 for income. Of the four other samples, all involving Hispanics or Amerindians, two showed significance for income but not education and two show no significance for the single index used, income in one and education in the others.  In total, out of the 9 U.S. Hispanic or Native American samples, in four the association was significant in the one reported measure (in all cases, education), in two, it was significant in one out of two measures and in two it was non-significant in the mentioned measures.  Generally, in the U.S., Amerindian ancestry in admixed populations was inconsistently negatively associated with educational and SES related outcomes.   However, were one to meta-analyze the results, the association would be highly significant in the predicted direction.   It’s not clear what factors are behind the inconsistency;  likely, migrant generational status is a confound.

Generally, the results support a racial hereditarian hypothesis along with others that predict a fairly internationally consistent association between continental ancestry and cognitively correlated indices of socioeconomic status such as education, income, and job prestige.


Branigan, A. R., McCallum, K. J., & Freese, J. (2013). Variation in the heritability of educational attainment: An international meta-analysis. Social forces, 92(1), 109-140.

Marioni, R. E., Davies, G., Hayward, C., Liewald, D., Kerr, S. M., Campbell, A., … & Deary, I. J. (2014). Molecular genetic contributions to socioeconomic status and intelligence. Intelligence, 44, 26-32.

Piffer, D. (2013). Factor Analysis of Population Allele Frequencies as a Simple, Novel Method of Detecting Signals of Recent Polygenic Selection: The Example of Educational Attainment and IQ. Interdisciplinary Bio Central,(1).

Piffer, D. (2014). Simple Statistical Tools to Detect Signals of Recent Polygenic Selection. Interdisciplinary Bio Central, 6(1), 1.

Piffer, D. (2014). Estimating strength of polygenic selection with principal components analysis of spatial genetic variation. bioRxiv, 008011.

Piffer, D., & Kirkegaard, E. O. The genetic correlation between educational attainment, intracranial volume and IQ is due to recent polygenic selection on general cognitive ability.

Rietveld, C. A., Medland, S. E., Derringer, J., Yang, J., Esko, T., Martin, N. W., … & McMahon, G. (2013). GWAS of 126,559 individuals identifies genetic variants associated with educational attainment. Science, 340(6139), 1467-1471.

Valenzuela, C. Y. (2011). Human Sociogenetics. Biological research, 44(4), 393-404.

Studies Included:

Schlesinger, et al.. (2011). African ancestry protects against Alzheimer’s disease-related neuropathology. Molecular psychiatry, 18(1), 79-85.

Leite, et al. (2011). Genomic ancestry, self-reported “color” and quantitative measures of skin pigmentation in Brazilian admixed siblings. PloS one, 6(11), e27162.

Ruiz-Linares, A., Adhikari, K., Acuña-Alonzo, V., Quinto-Sanchez, M., Jaramillo, C., Arias, W., … & Gonzalez-José, R. (2014). Admixture in Latin America: geographic structure, phenotypic

Cifuentes, et al. (2004). DYS19 and DYS199 loci in a Chilean population of mixed ancestry. American journal of physical anthropology, 125(1), 85-89.

Florez, et al.  (2009). Strong association of socioeconomic status with genetic ancestry in Latinos: implications for admixture studies of type 2 diabetes.Diabetologia, 52(8), 1528-1536

Campbell, et al. . (2012) Amerind ancestry, socioeconomic status and the genetics of type 2 diabetes in a Colombian population. PLoS One 7: e33570. doi: 10.1371/journal.pone.0033570

Ruiz-Narváez, et al.   (2010). West African and Amerindian ancestry and risk of myocardial infarction and metabolic syndrome in the Central Valley population of Costa Rica. Human genetics, 127(6), 629-638.

Martinez-Marignac, et al. (2007). Admixture in Mexico City: implications for admixture mapping of type 2 diabetes genetic risk factors.  Human genetics, 120(6), 807-819.

Fejerman, et al. (2010). European ancestry is positively associated with breast cancer risk in Mexican women. Cancer Epidemiology Biomarkers & Prevention, 19(4), 1074-1082.

Pereira, et al.  (2012). Socioeconomic and nutritional factors account for the association of gastric cancer with Amerindian ancestry in a Latin American admixed population. PloS one, 7(8), e41200.

Burchard, et al.  (2005). Latino populations: a unique opportunity for the study of race, genetics, and social environment in epidemiological research. Journal Information, 95(12).

Gravlee, et al. (2009). Genetic ancestry, social classification, and racial inequalities in blood pressure in Southeastern Puerto Rico. PLoS One, 4(9), e6821.

Alarcon, et al.. (2006). Systemic lupus erythematosus in a multi-ethnic cohort (LUMINA): contributions of admixture and socioeconomic status to renal involvement. Lupus, 15(1), 26-31.

Brown, et al. (2009). Effects of ethnicity and socioeconomic status on body composition in an admixed, multiethnic population in Hawaii. American Journal of Human Biology,21(3), 383-388.

Reiner, et al. (2005). Population structure, admixture, and aging-related phenotypes in African American adults: the Cardiovascular Health Study. The American Journal of Human Genetics, 76(3), 463-477

Cheng, et al. (2012). African ancestry and its correlation to type 2 diabetes in African Americans: a genetic admixture analysis in three US population cohorts. PloS one, 7(3), e32840.

Gower, et al.. (2003). Using genetic admixture to explain racial differences in insulin-related phenotypes. Diabetes, 52(4), 1047.

Qi, et al.  (2012). Relationship between diabetes risk and admixture in postmenopausal African-American and Hispanic-American women. Diabetologia, 55(5), 1329-1337.

Piccolo, et al. (2014). The contribution of biogeographical ancestry and socioeconomic status to racial/ethnic disparities in type 2 diabetes mellitus: results from the Boston Area Community Health Survey. Annals of epidemiology, 24(9), 648-654.

Cardel, et a. (2012). Parental feeding practices and socioeconomic status are associated with child adiposity in a multi-ethnic sample of children. Appetite,58(1), 347-353.

Marcus, et. al.  (2010). European ancestry as a risk factor for atrial fibrillation in African Americans. Circulation, 122(20), 2009-2015.

Aldrich, et al. . (2013). Socioeconomic status and lung cancer: unraveling the contribution of genetic admixture. American journal of public health, 103(10), e73-e80.

Ducci, et al.  (2009). Association of substance use disorders with childhood trauma but not African genetic heritage in an African American cohort.  American journal of psychiatry, 166(9), 1031-1040.

Fejerman, et al.  (2013). Genetic ancestry and risk of mortality among US Latinas with breast cancer. Cancer research, 73(24), 7243-7253.

Marcus, et. al.  (2010). European ancestry as a risk factor for atrial fibrillation in African Americans. Circulation, 122(20), 2009-2015.

Aldrich, et al. (2013). Socioeconomic status and lung cancer: unraveling the contribution of genetic admixture. American journal of public health, 103(10), e73-e80.

Fejerman, et al.  (2013). Genetic ancestry and risk of mortality among US Latinas with breast cancer. Cancer research, 73(24), 7243-7253.

Sweeney, et al. (2007). Genetic admixture among Hispanics and candidate gene polymorphisms: potential for confounding in a breast cancer study?. Cancer Epidemiology Biomarkers & Prevention, 16(1), 142-150.

Norden‐Krichmar, et al. (2014). Correlation analysis of genetic admixture and social identification with body mass index in a Native American Community. American Journal of Human Biology, 26(3), 347-360.

Allison, et al. (2010). Genetic ancestry and lower extremity peripheral artery disease in the Multi-Ethnic Study of Atherosclerosis. Vascular medicine, 15(5), 351-359.

Lisabeth et al.  (2011). Ancestral heterogeneity in a biethnic stroke population. Annals of human genetics, 75(4), 508-515.

Klimentidis, et al.  (2009). The relationship between European genetic admixture and body composition among Hispanics and Native Americans. American Journal of Human Biology, 21(3), 377-382.

Molokhia, et al. (2003). Relation of risk of systemic lupus erythematosus to west African admixture in a Caribbean population. Human genetics, 112(3), 310-318.

Sánchez, et al. (2012). Impact of genetic ancestry and sociodemographic status on the clinical expression of systemic lupus erythematosus in American Indian–European populations. Arthritis & Rheumatism, 64(11), 3687-3694.

Martínez, H., Rodriguez-Larralde, A., Izaguirre, M. H., & De Guerra, D. C. (2007). Admixture estimates for Caracas, Venezuela, based on autosomal, Y-chromosome, and mtDNA markers. Human biology, 79(2), 201-213.


  1. Anonymous

    “Recently, Piffer (2013, 2014, 2014) and Piffer and Kirkegaard (2014) found that continental races, that is, continental level human groupings where members are arranged by overall genetic affinity, differ in their allelic frequencies for well-replicated alleles that are causally associated with both educational attainment and psychometric intelligence. The results are tentative, but they support the position that the well documented regional variation in cognitive ability, and, by way of it, regional variation in levels of education has a partial genetic basis.”

    What does ‘tentative’ mean here? What credence that the method is actually capable of detecting the level of genetic difference predicted by the hereditarian hypothesis?

    I didn’t see any ex ante power calculations, and it’s already known that there is a great amount of population structure linked to those IQ figures in neutral variation. Some of the Rietveld SNPs in their larger SNP set go strongly opposite the phenotypes, and I haven’t seen anything using the full set, which had substantially better predictive power than just the top few alleles.

    One of my concerns about power is that the effects reported seem too big. Suppose Steve Hsu is right that there are 10,000 alleles affecting IQ, and variation within a population of perhaps 30-40 common SNPs per standard deviation of IQ. The biggest alleles in Rietveld were about 1/30 of a standard deviation, in line with that.

    But if two populations differ by a standard deviation, then across 10,000 alleles of ~1/30 SD, the average frequency difference per allele would have to be on the order of 0.3%. If the analysis depends on effects that are too large for the hypothesis, then I worry that they are the result of an underpowered analysis. I am more worried when the analysis relies on alleles with r^2 of 0.02%, when polygenic scores are available that explain 0.2%-0.4% (using a weighted sum of ~60 alleles) and ~2% (using all SNPs).

    Given Greg Cochran has written that the sample size is not large enough yet, I’m not ready to put even tentative faith in this until I see sample sizes and power calculations that show results should be forthcoming if the hypothesis is true.

    • Chuck

      There are a couple of issues that should not be confused, but which you seem to (though, it’s difficult to tell):

      (1) The reliability of the SNPs used by Piffer
      (2) The power of the SNPs to explain group differences (in terms of % variance)
      (3) The power of Piffer’s method to detect differential polygenetic selection

      Regarding the first, the alleles used have now been replicated a number of times. Some were even recently replicated in a Han Chinese sample: Zhu, et al. (2015). Regarding the second, Piffer’s method doesn’t concern itself with the magnitude of variance explained (by additive polygenetic scores), but rather with the relative magnitude of polygenetic selection. This brings us to three. The robustness of the analysis depends on the number of alleles used, but not % variance explained. The more alleles covary across populations — which is what differential polygenetic selection predicts — the stronger the support. In the most recent paper, Piffer found that 6/7 identified alleles did. As a basic test of significance, one could use the joint probability, since if there was no differential polygenetic selections, alleles would vary across populations willy-nilly. This is now significant — that is, the covariance is of a degree more than chance would predict. However, significance levels are arbitrary conventions. Thus, I consider the results tentative. They statistically support a differential polygenetic selection model (given conventional standards), but I’m not confident that the next 7 alleles will not show the reverse effect. And this is why, of course, I have been trying to find supporting evidence! But back to your claim: Piffer’s results don’t constitute even tentative support. The claim that they do is consistent with social scientific interpretative standards. You can compare his results to others tentative ones e.g., Matthews and Butler (2011) “Novelty‐seeking DRD4 polymorphisms are associated with human migration distance out‐of‐Africa after controlling for neutral population gene structure”. That said, different people have different standards. But clearly, given what I have said, my qualified claim is justifiable.

  2. JayMan

    Very interesting find!

    While I doubt there’s a problem, hopefully no critics will try to claim selection bias in the studies. It’s also too bad you couldn’t get the data to do an actual meta-analysis.

    • Chuck

      I would be delighted if critics could point me to studies which I missed.

Leave a Reply

Your email address will not be published. Required fields are marked *

© 2024 Human Varieties

Theme by Anders NorenUp ↑