Monday, November 22, 2010

IQ and Infectious Disease

http://rspb.royalsocietypublishing.org/content/early/2010/06/29/rspb.2010.0973.full
http://isteve.blogspot.com/2010/07/reader-writes-in-reference-to-new-paper.html
http://isteve.blogspot.com/2010/07/guardian-lower-iqs-found-in-disease.html
http://www.economist.com/node/16479286

Science as the Evasion of Science by Means of Political Correctness
The first link is to a recent paper on the link between infectious disease rates in different nations and their IQ levels. They found a correlation of r = -.82 (the closer to -1, the stronger the negative correlation). They then mention, in passing, this curious inconvenience: 
Lynn (1991) and Rushton (1995, 2000) proposed that temperature and climate provide important Darwinian selective pressures for intelligence, with cold climates selecting for higher intelligence, because low temperatures provide more fitness-related problems for humans that must be solved through cognitively demanding means, and through more complex social organization. Templer & Arikawa (2006) tested and supported predictions of this proposal in a cross-national study and found that average IQ correlated significantly with winter high temperature (r = −0.76), winter low temperature (r = −0.66)...Templer & Arikawa (2006) also found that average IQ correlated significantly with average skin darkness (r = −0.92). The authors offered little explanation of why this trend exists, except that they believed skin colour was related to exposure to certain climates over evolutionary time.
Yes, skin pigmentation is a stronger predictor of international IQ differences than any other factor yet discovered. The second best predictor is infectious disease rates, third average winter temperatures--both of these two factors, however, may be considered good proxies for skin pigmentation since infectious diseases are more prevalent in the environment (and not only in the populace) in equatorial regions and higher latitude strongly correlates with the lighter skin pigment of peoples who evolved at those latitudes. This means that when skin color is used as a predictor of national IQ, it is highly accurate in its predictions and essentially eliminates the supposed predictive powers of infectious disease rates. To be fair, it is also true that when this process is inverted and infectious disease rates are first used as predictors of IQ, skin color has reduced predictive power beyond that gained by looking at infectious disease levels. These two characteristics almost cancel each other out--though, again, skin color has a bit more power. This makes the problem of distinguishing the causal factors from the merely correlative factors much more difficult. In dealing with various other potential influences on IQ expression, they used statistical analysis to clarify degree of overlap in the correlations compared. This way the proportion of causation to be assigned to each factor may be estimated more accurately. But, in the case of the most powerful correlation (and presumptively the most important), skin color, here's how the study's authors approached their challenge:
Although Templer & Arikawa (2006) found a positive relationship between IQ and skin darkness, we will not use skin darkness in our analyses for three reasons: (i) although evidence suggests that skin darkness is a measure of historical infectious disease intensity over evolutionary time, it is unclear exactly what kind of infectious diseases it is indicative of (see discussion); (ii) Templer & Arikawa (2006) argued that the relationship between skin darkness and IQ is not causal; and (iii) Templer & Arikawa (2006) did not sufficiently explain why the association between intelligence and skin darkness exists. Without a reasonable theoretical framework for this association, we did not feel it was appropriate to compare it with other variables for which there is a better theoretical rationale.
What's this? Inconvenient data? Will the reviewers let us toss it? Yes? And they agree with us that political correctness must here prevail over science and reason? Yes? Then let it be tossed!

Truly, it must be a bit discouraging for scientists to have to write such weak stuff. But, if you want to be published in "mainstream" journals, the proper political mask must be donned--howsoever ridiculous it may be. After all, in our time, this does not count as stupidity among one's peers. It's only convention, habit, a matter of rote--and barely distracts the conscious mind for a moment. Enough.

My rebuttal to their "reasons" for not analysing the skin color factor:

It scarcely matters what specific infectious diseases skin color may be a marker of--darker skin is known to be associated with enhanced protection from infectious disease in general. This leads to the inference that the pressures exerted by Darwinian selection in equatorial regions favored those with stronger immune systems, the added strength of which may have diverted resources from the highly resource intensive human brain. Those who allocate their energetic resources optimally for their environment survive and reproduce most successfully, enacting natural selection. Dark skin is genetically determined. The same forces which caused the evolution of dark skin may very well, by this logic, have caused a genetically determined reduced allocation of resources to the brain development of dark skinned people. It looks as though some IQ is given up at the genotypic level and, for those still exposed to equatorial disease rates, still more IQ is given up at the phenotypic level when children must fight off diseases. The different degree of exposure to infectious disease in childhood may provide part of the explanation for the IQ differences between Africans raised in developed nations and those still living in Africa. African-Americans (and African-Dutch and African-English) run at about 83-86 on IQ tests, indigenous Africans manifest, for various reasons, a large range of 64-82. In contrast to the evolutionary pressures faced by the dark skinned equatorial dwellers, those who left Africa 60,000 years ago and migrated north confronted a harsher environment which may have constrained them to evolve more intelligence than their tropical ancestors had required. Reduced UV exposure coterminously caused evolution of lighter skin. I think this constitutes a "reasonable theoretical framework." It's called The Theory of Evolution.

Another sign that their chosen leading predictor of IQ may be second best is the major group of outliers that can be seen in Figure 1 of the paper. These countries have IQs about 15-25 points lower than their disease rates would predict. They happen to be a group of African-populated Caribbean countries with relatively low disease rates, yet IQs consistent with those found in disease-ridden Africa. The authors' "explanation":  "Because these outliers are in the same geographical location, it is possible that local parasites that are not included in the DALY owing to infectious disease variable are causing these outliers." That's it. No notice that these countries are not outliers when skin color and IQ are graphed. I might add that the authors are quite lucky that American blacks never founded their own nation in N America. If you add American, Dutch, English blacks to their graph these groups would also constitute a group of major outliers about 15 IQ points below their disease-predicted levels. By comparison, East Asian nations end up 11 IQ points above their disease-predicted levels (there does appear to be a latitudinal or racial gradient for IQ in East Asia, which, when accounted for, leaves little if any predictive power to disease). Europeans are close to their predicted levels. Another point is that, in comparing Sub-Saharan African nations, no IQ-disease relationship visible in the data. Given the skin color similarity, this is perhaps unsurprising. The virtual absence of an IQ-disease relationship once nations are grouped according to skin color is rather striking. This absence is true in groups comprising Sub-Saharan Africa, N Europe, S Asia, NE Asia, and SE Asia.  The lack of relationship in S America (inclusive of the Caribbean) is so obvious that the authors admit to it themselves. There seems to be some disease-IQ relationship in S Europe (which would make them an anomaly, the exception that proves my proposed rule that the authors thesis lacks a sound basis in the data presented).

What do we have left of the original thesis? This: infectious disease rates may be considered a proxy for skin color in most places. Those cases where they are not a proxy (eg, blacks in rich northern countries or healthy caribbean ones) may reveal the racial component in genetic IQ determination. Since both infectious disease rates and skin color are latitudinally determined in most cases, the clearest way to reveal the relative causal power of these two factors would be to find instances where either disease or skin color are not in line with what is typical of a given latitude. In Singapore, we have a case of equatorial latitude, but low disease prevalence (2.7) and light skin (most being of Chinese extraction). Result (with both factors in its favor): average IQ 108, the highest in the world and 17 points higher than disease levels predict--as much as a 17 point racial bonus, though some of this is certainly non-racially determined. Surrounding Singapore, Malaysia is almost half lighter-skinned Chinese, the rest darker indigenous Malays, with a disease level (3.2) somewhat higher. IQ is 92, 6 points higher than disease-predicted. Another Chinese-based racial bonus?

Nations in the following graph (I made it based on the paper's data) are all of the 18 non-African nations with the highest disease burden (3.5-4.4, compared to the 4.2-4.8 range in black Africa), the closest disease-based comparison to sub-Saharan Africa.

High Disease Non-African Nations:
https://www.google.com/fusiontables/DataSource?snapid=S290610_-A9

First, and once again, there is little disease-predicted IQ relationship to be seen here. Second, on average these unhealthy nations have IQs 5.5 points higher that predicted by the disease-IQ correlation. By comparison, black Africa is 4.5 points lower than predicted. In other words, these unhealthy non-African nations had IQs measuring 85, the black African nations measured 68--a 17 point difference--whereas according to the disease-IQ theory the numbers should have been 79.5 and 72.5, which would be a 7 point disease-created difference. Could it be that black Africans have a more serious case of melanin over-determination? Amusingly, the only two non-African nations actually within the black African disease range (4.2-4.8) are Cambodia (4.2, IQ 91) and Afghanistan (4.4, IQ 84). These are the two most direct comparisons with black Africa and they clock in at 16 and 23 points higher than black African average IQ. It's difficult to think of any IQ relevant advantages those two have over Africa. Except skin color, of course.

So, when we compare black African IQs to the 18 globally scattered nations with similar disease burdens, African IQs fall well short of the expected level, insofar as disease is at all determinative. And when we compare blacks in various rich countries with low disease burdens to other races in such places we find that blacks fall 10-15 points short. The consistency of this shortfall in every study and every environment of which I am aware is quite impressive.

The ironic thing about this post is that I've followed the data in a direction that surprises me. As an a priori matter, I expected infectious disease to be a major factor in creating IQ deficits. But, once race is accounted for, it looks as though the impact is limited. With uncontrolled studies like this (ones without good control groups), one can only estimate the degree of causation behind disparate IQ levels. And different causes will have more influence in some nations than others. In America, infectious disease levels are low and therefore cannot explain much of any deficit that may exist. For the black American deficit vis-a-vis white Americans other causal factors must be sought. Besides, as the authors commendably note, there are multiple other factors, known and unknown, that influence the phenotypic expression of the genetically-determined IQ potential.  Nutrition, education, wealth, pollution exposures, stress levels (like growing up in a war zone), mental stimulus in the environment, familiarity with testing--it's likely that all of these elements (and probably others) play a role in achieving one's genetically-determined potential IQ. Disentangling the relative causal power of each factor will only ever be an exercise in estimations. All of these factors have improved in those places (all rich countries) where the Flynn effect (IQ scores rising 2-3 points per decade over the last 100 years) has been observed.

Various studies have come up with an IQ heritability number of about 80% for adults (even the red-tainted NY Times assents to this established figure). The main qualification of this number is that it presumes a similarly favorable environment for the parents and children. The Flynn effect makes for a secondary qualification, though the IQ increases seen in the developed world are diminishing. These stable, favorable conditions are found in the rich world, not in the third world. If you have black African parents with typical 68 IQs, roughly 30% of IQ variation would be genetic. Black Americans with typical 84 IQ parents might have 80% of their IQ genetically determined--though this percentage might be lower for the extremely poor. Basically, the worse your environment, the less influential genes are in finalizing your IQ and the more influential is the environment.

On the whole, analyzed critically, this paper is useful in diagnosing some of the problems causing low IQ. It would have been rather better had it presented a proper statistical analysis of the race-IQ correlation compared to the disease-IQ correlation. This would have provided a clearer view of the proportion of IQ lost to disease. Nevertheless, disease certainly takes its toll on a phenotypic level. And, whether the worst effected lose 3 points or 15 points (I suspect it's about 6-8 points in the worst areas, though the paper implicitly claims something like 12-15 points), it's worth funding corrective measures. Bill Gates has explained that this is part of the rationale for his multi-billion dollar effort to cure the major diseases of black Africa. On a return on investment (ROI) basis, even if African lives are valued at relatively low dollar amounts, this area has been hugely underfunded for decades. Gates knows this too. Unfortunately, most NGOs are mismanaged by mediocrities and timeservers ducking the challenges of private sector competition. The non-profit sector of the economy is flabby and incompetent precisely because they are not forced into fitness by the competitive forces of capitalism--the only thing (some of them) compete for is handouts--and this is the only thing most are good at, coincidentally--as Gates also knows and even, delightfully, says. Not only is the money for poor countries' assistance misspent, but too little is allocated to begin with. The other source of funds, rich country governments, also have the very same competency problems common to all non-competitive enterprises and have historically done little good (though there be exceptions, esp work on eradicating small pox and polio). If black Africa were at the IQ levels of black Americans (84), they would then be at the same level as these countries: Brazil, Colombia, India, Indonesia, Mexico, Morocco. All are in better shape than black Africa--Brazil and Mexico have progressed so far as to escape into the second world. There are also a couple of failed or failing states at that IQ level (Afghanistan and Pakistan), nor is IQ anywhere near the sole determinant of national success (or any other success). My contention: a reasonable national IQ is necessary, but not sufficient.

A few graphs I created from the paper's data, broken down into world regions (and skin color), showing, except in S Europe, limited IQ-disease correlation once region/race are accounted for:
Sub-Saharan Africa and Black Caribbean nations:
https://www.google.com/fusiontables/DataSource?snapid=S290505LvL5
NE Asia:
https://www.google.com/fusiontables/DataSource?snapid=S290517b9r1
SE Asia:
https://www.google.com/fusiontables/DataSource?snapid=S290519Fue6
S Europe:
https://www.google.com/fusiontables/DataSource?snapid=S290611C5lb
N Europe:
https://www.google.com/fusiontables/DataSource?snapid=S290612tTHr

No comments:

Post a Comment