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?
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 = 0.1032; 0.86 + 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:
|Nigeria||2/6||118||McCarth||89||Ashem & Janes (1978)|
|Nigeria||13||803||CCF||95||Nenty & Dinero (1981)|
|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)|
[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  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:
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:
|Model||Variables Entered||Variables Removed||Method|
|a. Dependent Variable: VAR00002|
|b. All requested variables entered.|
|Model||R||R Square||Adjusted R Square||Std. Error of the Estimate|
|a. Predictors: (Constant), VAR00003|
|Model||Sum of Squares||df||Mean Square||F||Sig.|
|a. Dependent Variable: VAR00002|
|b. Predictors: (Constant), VAR00003|
|Model||Unstandardized Coefficients||Standardized Coefficients||t||Sig.|
|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.
|Antigua and Barbuda||688||75|
|Bosnia and Herzegovina||664|
|Central African Republic||156||68|
|DPR of Korea||353||106|
|DR of Congo||171||*|
|Iran, Islamic Republic of||659||84|
|Lao People’s Democratic Republic||356||89|
|Libyan Arab Jamahiriya||683||84|
|Papua New Guinea||467||84|
|Republic of Korea||759||106|
|Republic of Moldova||639||95|
|Saint Kitts and Nevis||643||75|
|Saint Vincent and the Grenadines||722||75|
|Sao Tome and Principe||535||59|
|Syrian Arab Republic||628||87|
|The former Yugoslav Republic of Macedonia||664||93|
|Trinidad and Tobago||742||80|
|United Arab Emirates||886||83|
|United Republic of Tanzania||422||72|
|Venezuela, Boliviarian Republic of||775||88|
* = 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