Disease, Race, and IQ

One common argument for the lower average IQs of African countries is that they are often disease ridden and because of that, cognitive ability is dropped. A solid and plausible theory – but does it hold up?

1: rGE

First, I’ll bring up a topic we’ve discussed in the past: the gene-environment correlation. Because your gene expresses itself into the environment, your environment and environmental effects are partly caused by genes. An example I often use is the kid with higher intelligence. Since intelligence is highly heritable and correlates to life outcomes, there is a lot we can predict about someone’s environment if they have high intelligence:

Their parents passed on the alleles related to higher intelligence, and they share a lot of the same IQ-related DNA. Because they live with the child, they influence the child’s environment, whilst also sharing that DNA. They may encourage him to good things with his intelligence such as read more. They also likely live in a nice neighborhood with other people who share this high intelligence. Once again, this high intelligence shared by people in the nice neighborhood is largely caused by genetics. The child’s peers will hence play an effect on how much he studies and if he makes good and bad choices. He also probably goes to a nice school because of the higher standard of living shared by high-IQ people. Finally, he expresses the genes himself which feeds back into his own habits, reinforcing his own high intelligence.

While I don’t know if this is the standard equation done to calculate such a thing, Michael Levin provided such a way of calculating the heritability once you take into account the gene-environment correlation in Why Race Matters. (Heritability is the phenotypic differences between people due to genetics, so you can apply the term here even though we’re talking about environmental effects.):

Standard heritability * environmental effects = Sum

Sum + Standard heritability = Heritability after rGE

So, I’ll take the heritability estimate of g provided by Panizzon et al. (2014) which is 0.86 (additive genetic influence) and the common environment effect which was roughly 12 percent:

0.86 * 0.12 = 0.10320.86 + 0.1032 = 0.9632

So in actuality, the heritability of IQ is in the high 90s. This is, of course, when the environment is typical. This is also a Western heritability estimate – Lynn (2015) shows that the heritability of IQ in Africa is slightly lower, around 0.7. Regardless, very high.

The genotype-environment correlation has been proven beyond just a theoretical view: see Kendler and Baker (2007); Plomin et al. (1977); Norrington et al. (2016); and even specifically on IQ, Loehlin and Defries (1987)

So how do we apply this to diseases in African countries? Well, we must simply ask why these diseases are so affected. One very plausible cause is the lack of quality healthcare and treatment. And healthcare quality generally must come about through intelligence. Since intelligence is largely caused by genetics, the failure of African countries to develop higher quality healthcare systems (which requires higher intelligence) is in turn, caused by genetics. And beyond just my theoretical model, here’s a bit of data to make the case:

The correlation between socioeconomic development and IQ in the UK is 0.87 (Kirkegaard, 2015).

An analysis of 81 nations shows that 55 percent of the variance in GDP between the nations is due to differences in IQ (Lynn, 2001).

According to Carl (2016), ” regional IQ is positively related to income, longevity and technological accomplishment; and is negatively related to poverty, deprivation and
unemployment. A general factor of socioeconomic development is correlated
with regional IQ at r = 0.72.”

I also give some more on this, specifically related to healthcare efficiency, in the third section of this article.

2: Studies Relating to Disease and IQ

In this section, I’ll briefly summarize a few studies I found defending the effect of disease on IQ:

First of all, we have a critique of the international IQ estimates provided by Lynn and Vanhanen. Wicherts et al. (2010) asserts that Lynn and Vanhanen used lower class people in Africa to make their claims that the average IQ was lower there. This is important because diseases are largely prevalent and can have large effects on the people of Africa. In turn, they took studies of elite people in Africa (an unrepresentative sample) and find these estimates:

Country: Age: N: Test: IQ: Source:
Nigeria 6/11 393 DAM 83 Bakare (1972)
Nigeria 2/6 118 McCarth 89 Ashem & Janes (1978)
Nigeria 13 803 CCF 95 Nenty & Dinero (1981)
Nigeria Adults 28 SPM 89 Morakinyo (1985)
Sierra Leone 8 202 DAM 91 Ohuche & Ohuche (1973)
South Africa 19 228 SPM 97 Crawford-Nutt (1976)
[Reported in Wicherts]
South Africa 24 40 WAIS-3 84 Shuttleworth et al. (2004)
Zimbabwe 8 52 PMA 84 Wilson et al. (1991)
Zimbabwe 12/14 204 SPM 70 Zindi (1994)
[Reported in Wicherts]

Then we come to two studies by Epigg et al. that would help to prove this theory.

The first is Epigg et al. (2010a) which finds strong, negative correlations between parasitic diseases and national IQ estimates globally. This would possibly infer that most of the global IQ differences can be related to the differences in diseases spread in these nations.

The second is Epigg et al. (2010b) which looks specifically at American states to find the effect of diseases on IQ differences. This one is more interesting to me because you have a more controlled, similar environment to test which makes the effect of the rGE less strong. They find a strong, negative correlation between exposure to parasitic diseases and state IQ estimates.

3: And why they’re wrong

First, we can quickly tackle Wicherts et al. This study is fairly simple to refute. For one, as I said, they used studies of elite Africans to prove the point. The problem with this is that these are largely unrepresentative of African countries (see Lynn [2010] for a full treatment and other critiques). Furthermore, it really is telling when the elites of these countries still only have IQ’s in the high 80s. This on its own would represent that the average citizen would have an IQ some amount of full standard deviations lower than this, regardless of the effect of diseases. Of course, this would generally prove Lynn’s theory, if not even show the IQs of average Africans should be lower. Final note on this study: they mostly tested people under the age of 18. This is faulty because IQ is not fully phenotypic until age 18-24. These estimates are elastic and not representative.

Moving on – that study is really stupid.

I was able to find a piece of contradictory empirical evidence of the non-effect of these diseases on intelligence and cognitive function:

Muntendam et al. (1996) finds that exposure to cerebral malaria had no long-term effect on neuropsychological function in Gambian children.

Granted that’s one, limited study, but worth noting. I think studies like this one are important and should be expanded upon, because they are going to better prove causality than current literature (such as Epigg et al.’s). Longitudinal analyses allow us to track the long term effects of diseases of individuals’ IQ rather than giving plain correlations. This may be a key takeaway from this article for what should be researched on this topic.

On to Epigg et al. (2010a).

Epigg et al. (2010a) (the global study) doesn’t take into account how IQ helps develop healthcare structures. Many people cite Epigg et al. (a) using it as proof against any hereditarian hypothesis whatsoever. Chances are they haven’t read beyond the abstract. As they say in the study,

“A nation of more intelligent individuals is likely to produce a higher GDP, but a wealthier nation is also more able to pay for public education, as well as public medical and sanitation services. An indirect link between education and intelligence may also exist, as a better-educated population may be more interested in public health measures–leading to increased IQ by reducing parasite stress–provided that education includes information about germ theory and hygiene. These sources of endogeneity must be considered when interpreting our findings (and see below). It should also be mentioned that we are not arguing that global variation in intelligence is only caused by parasite stress. Rather, variation in intelligence is probably caused by a variety of factors, including those we have mentioned here as well as factors that are yet unknown.”

I figured I would get more in-depth to the argument that IQ is a good predictor of healthcare quality and efficiency. Here is a map of healthcare efficiency by country, and then a map of national IQs globally, just to give a bit of perspective:

WHO healthcare map.PNG


Beyond some exceptions, it very much just looks like a replica. Moving past a fun little graph comparison, we can actually measure the effect of IQ on healthcare efficiency/quality. I took WHO data on the efficiency of 191 countries’ healthcare systems and put it into regression analysis with the IQ estimates by nation presented in Intelligence and the Wealth and Poverty of Nations by Lynn and Vanhanen.

Thanks to a large sample size (over 150 nations I could test), I had statistically significant results at p<0.0005. The regression analysis came up with a moderate (adjusted) r^2 value of 0.419, indicating that IQ can explain 42% of the variation in healthcare efficiency between nations. (In this analysis, VAR00002 is the index of healthcare efficiency and VAR00003 is average IQ.) See all the results below and see the graph of my data (if you want to copy, replicate, have, double check, etc.) in Appendix 1:


Variables Entered/Removeda
Model Variables Entered Variables Removed Method
1 VAR00003b . Enter
a. Dependent Variable: VAR00002
b. All requested variables entered.
Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .649a .422 .419 162.857
a. Predictors: (Constant), VAR00003
Model Sum of Squares df Mean Square F Sig.
1 Regression 3443749.993 1 3443749.993 129.844 .000b
Residual 4720962.785 178 26522.263
Total 8164712.778 179
a. Dependent Variable: VAR00002
b. Predictors: (Constant), VAR00003
Model Unstandardized Coefficients Standardized Coefficients t Sig.
B Std. Error Beta
1 (Constant) -394.915 91.229 -4.329 .000
VAR00003 12.211 1.072 .649 11.395 .000
a. Dependent Variable: VAR00002

These results taken into account, it should be obvious national IQ plays a fair sized role in the development of healthcare systems. This analysis simply looks at efficiency over a given time. Using the more theoretical likelihood I explained in the rGE section, we could assume there is a lot more to it, particularly historically.

In addition, I would recommend reading Steve Sailer’s critiques of the study. He goes over some general methodological problems that make it less likely to prove a specific case.

Final note on this study: race differences in IQ have existed for a while. In response to Wicherts et al., I cited a response by Lynn and in this analysis, he looks over race differences in brain size (correlates with IQ) and other factors over time, and he finds that racial differences have existed for at least the last 10,000 years. This would lead us to the conclusion, racial differences in intelligence, must be, to some degree, causal – NOT the diseases! In case I’ve peaked your interest again, here it is again: Lynn (2009).

Finally, one of the more proving studies, Epigg et al. (2010b) (the one across the United States). I briefly stated why this one was much more telling. Since the environmental differences between states are much smaller than those globally, we are going to have more equality in how much each race is given from the state and how much everyone is protected from diseases. By narrowing the analysis to smaller differences, we’ve made it more accurate at proving the case diseases are causal to some effect. In this study, Epigg et al. find a strong, significant, negative correlation between exposure to parasitic diseases and state IQ estimates – -0.67 across the states.

Rest assured, the study itself will show this is not some end-all case:

In this analysis, infectious disease does not predict average IQ as well as it did in a similar analysis across nations, and education and economic variables have higher predictive power (Eppig et al., 2010). It is possible that this is an artifact of the way average IQ was measured across US states. Although the NAEP test, which was used to calculate average state IQ, is a valid measure of IQ (McDaniel, 2006b), it is likely influenced by education more than tests used cross-nationally which are designed to measure IQ more directly. It is also possible that the zero-order correlation between infectious disease and average state IQ (r=0.67) is lower than the correlation between infectious disease and average national IQ (r=0.76 to 0.82; Eppig et al., 2010) because there is a wider range of IQ and of parasite stress on the cross-national level than there is on the cross-state level within the USA. Despite this, infectious disease is still a powerful predictor of average state IQ, and the best predictor of the variables we examined.”

Summed up? 1) education and economic variables were more predictive than parasitic diseases on an international level, 2) the NAEP test they used is not really a measure of intelligence; it measures what is taken through education rather than g-loaded cognitive ability, 3) this study shows parasitic disease can predict state IQ, but not the IQ we’re really concerned about, which is better measured in the international analysis.

Maybe I’m slightly going back on myself, though we shouldn’t completely forget about the effect less resources is going to have one’s ability to treat themselves in America’s healthcare system. In fact, it may be better to conduct a test like this in a country with single-payer healthcare as rationing and waiting are conducted by severity rather than amount of money one can provide.


More research should be done into the topic of the effect of diseases on cognitive function. But they should not be of this nature. This only proves a correlation, not a causation and is prone to the effects IQ has on one’s life outcomes.

Instead I recommend more research be done like Muntendam et al. (1996) – longitudinal analyses testing for the long term effect different parasitic diseases have on neuropsychotic and cognitive function. These are not hard to do and will be much more proving than the studies we have as of now. This does not deny that diseases will hurt one’s life and poorly affect African nations. But the argument of intelligence still rings strong and this is not some fix-all answer by any means.



Appendix 1:

Country Efficiency Avg. IQ
Afghanistan 327 83
Albania 774 90
Algeria 701 84
Andorra 982 *
Angola 275 69
Antigua and Barbuda 688 75
Argentina 722 96
Armenia 630 93
Australia 876 98
Austria 959 102
Azerbaijan 626 87
Bahamas 657 78
Bahrain 824 83
Bangladesh 675 81
Barbados 808 78
Belarus 723 96
Belgium 915 100
Belize 736 83
Benin 647 69
Bhutan 575 78
Bolivia 571 85
Bosnia and Herzegovina 664
Botswana 338 72
Brazil 573 87
Brunei Darussalam 829 92
Bulgaria 639 93
Burkina Faso 543 66
Burundi 494 70
Cambodia 322 89
Cameroon 357 70
Canada 881 97
Cape Verde 617 78
Central African Republic 156 68
Chad 303 72
Chile 870 93
China 485 100
Colombia 910 88
Comoros 592 79
Congo 354 **
Cook Islands 628 *
Costa Rica 849 91
Cote d’lvoire 517 71
Croatia 812 90
Cuba 834 85
Cyprus 906 92
Czech Republic 805 97
Denmark 862 98
Djibouti 414 68
Dominica 854 75
Dominican Republic 789 84
DPR of Korea 353 106
DR of Congo 171 *
Ecuador 619 80
Egypt 752 83
El Salvador 608 84
Equatorial Guinea 337 59
Eritrea 399 68
Estonia 714 97
Ethiopia 276 63
Fiji 653 84
Finland 881 97
France 994 98
Gabon 511 66
Gambia 482 64
Georgia 615 93
Germany 902 102
Ghana 522 71
Greece 933 92
Grenada 689 75
Guatemala 713 79
Guinea 385 63
Guinea-Bissau 314 63
Guyana 554 84
Haiti 517 72
Honduras 544 84
Hungary 743 107
Iceland 932 98
India 617 81
Indonesia 660 89
Iran, Islamic Republic of 659 84
Iraq 637 87
Ireland 924 93
Israel 884 94
Italy 991 102
Jamaica 782 72
Japan 957 105
Jordan 698 87
Kazakhstan 752 93
Kenya 505 72
Kiribati 495 84
Kuwait 810 83
Kyrgyzstan 455 87
Lao People’s Democratic Republic 356 89
Latvia 630 97
Lebanon 664 86
Lesotho 266 72
Liberia 200 64
Libyan Arab Jamahiriya 683 84
Lithuania 722 97
Luxembourg 928 101
Madagascar 397 79
Malawi 251 71
Malaysia 802 92
Maldives 477 81
Mali 361 68
Malta 978 95
Marshall Islands 504 84
Mauritania 384 73
Mauritius 691 81
Mexico 755 87
Micronesia 579 84
Monaco 943 *
Mongolia 483 98
Morocco 882 85
Mozambique 260 72
Myanmar 138 86
Namibia 340 72
Nauru 647 *
Nepal 457 78
Netherlands 928 102
New Zealand 827 100
Nicaragua 733 84
Niger 337 67
Nigeria 176 67
Niue 584 *
Norway 955 98
Oman 961 83
Pakistan 583 81
Palau 700 *
Panama 656 84
Papua New Guinea 467 84
Paraguay 761 85
Peru 547 90
Phillipines 755 86
Poland 793 99
Portugal 945 95
Qatar 812 78
Republic of Korea 759 106
Republic of Moldova 639 95
Romania 645 94
Russian Federation 544 96
Rwanda 327 70
Saint Kitts and Nevis 643 75
Saint Lucia 740 75
Saint Vincent and the Grenadines 722 75
Samoa 589 87
San Marino 998 *
Sao Tome and Principe 535 59
Saudi Arabia 894 83
Senegal 756 64
Seychelles 773 81
Sierra Leone 0 64
Singapore 973 100
Slovakia 754 96
Slovenia 838 95
Solomon Islands 705 84
Somalia 286 68
South Africa 319 72
Spain 972 99
Sri Lanka 716 81
Sudan 524 72
Suriname 623 89
Swaziland 305 72
Sweden 908 101
Switzerland 916 101
Syrian Arab Republic 628 87
Tajikstan 428 87
Thailand 807 91
The former Yugoslav Republic of Macedonia 664 93
Togo 449 69
Tonga 607 87
Trinidad and Tobago 742 80
Tunisia 785 84
Turkey 734 90
Turkmenistan 443 87
Tuvalu 518 *
Uganda 464 73
UK 925 100
Ukraine 708 96
United Arab Emirates 886 83
United Republic of Tanzania 422 72
Uruguay 745 96
USA 838 98
Uzbekistan 599 96
Vanuata 559 84
Venezuela, Boliviarian Republic of 775 88
Viet Nam 393 96
Yemen 587 83
Yugoslavia 629 93
Zambia 269 77
Zimbabwe 427 66


* = Excluded; not included in both sets of data

** = Excluded; couldn’t find the percentage of Congo that was Braz and that which was Zaire (they are split up in Lynn and Vanhanen’s data) hence possibly leading to some (while likely insignificant) bias in results

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 )

Google photo

You are commenting using your Google 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 )

Connecting to %s