Saturday, April 03, 2021

Race, Gender and IQ

In modern-day America, anyone arguing that the difference in average IQ between blacks and whites, or the difference in the distribution of IQ between men and women, at least partly explains the difference between average black and white income or between male and female numbers in some academic fields, risks being accused of racism or sexism. Striking examples of the possible consequence of such an accusation are provided by the cases of James Watson and Lawrence Summers. Watson, who received a Nobel prize for his role in the discovery of DNA, arguably the most important biological breakthrough of the century, was so careless as to tell the Times that he was "inherently gloomy about the prospect of Africa" because "all our social policies are based on the fact that their intelligence is the same as ours - whereas all the testing says not really". He was attacked ferociously for the statement, accused of prejudice, stripped of titles and positions.

Prejudice is belief held without evidence. Watson’s view might well have been mistaken —I now think it was, although at the time I did not — but there was evidence to support it, since the average measured IQ in African countries was strikingly below that in European countries. His attackers, so far as I could tell, had no evidence in their support and were acting on pure prejudice.

Summers, then president of Harvard, commenting in an academic talk on the small numbers of women in elite academic positions in fields such as mathematics, offered several possible explanations. One of them was that although the average IQ of men and women was about the same, female IQ had a tighter distribution than male IQ. That would imply fewer women than men far out on the upper tail of the distribution, where Harvard math and physics professors are located. Summers was fiercely attacked for mentioning that possibility, forced out of his position at least in part as a result. Again there was quite a lot of evidence for the claim, no evidence against.

The result of suppressing arguments for an unpopular view is that nobody honestly knows what conclusion would come out of an open debate, hence whether or not the view is true, although many people may find it prudent to pretend to. Until very recently, the only convincing argument I had seen against the claim of lower African genetic IQ was one offered by Thomas Sowell in his Ethnic America. He observed that the average family income of immigrants to the U.S. from the West Indies reached the U.S. average in one generation. West Indians are blacker than Afro-Americans in both their genes and their skin color, so if Afro-Americans did badly because of their African genes, West Indian immigrants should do worse, and similarly if the reason was discrimination. Sowell offered instead an explanation based on the different cultures produced by differences between West Indian and North American slavery.[1]

I have now found more and even better arguments against the hereditarian explanation of racial differences. Chanda Chisala is a Zambian immigrant who appears, like Sowell, happy to engage in arguments on unfashionable subjects. His main topic is not Afro-American IQ but African IQ. He offers several independent lines of evidence to show that its low measured value, variously claimed to be 70 or 80, must be due to African environment not African genes. His evidence is of two sorts: The academic performance of African immigrants in first world countries, where they are exposed to a first world educational environment, and the performance of Africans on two games, checkers and Scrabble.

U.K. data on student performance is available not merely for racial groups conventionally defined but for linguistic subgroups within those populations. Africans on average do not particularly well, but many of the African subgroups, including both the Igbo and the Yoruba, the two largest Nigerian tribes, substantially outperform the native English, in some cases East Asians as well.

His U.S. evidence is more anecdotal. One year, a single college applicant in the U.S. was accepted by all eight Ivy League schools. He was a Nigerian immigrant. Another Nigerian immigrant is a serial entrepreneur who invented a computer application, founded a company, and sold it to Apple for an estimated billion dollars.[2] Black students in elite universities are African or West Indian immigrants or their children in numbers far out of proportion to their share of the population.[3] In a number of cases where data happen to be available, black refugee immigrants, not native speakers of English, substantially outperform in school native Afro-Americans. That is the opposite of the result one would expect if Africans were genetically inferior in intelligence to whites, since Afro-Americans, unlike Africans, have significant white ancestry.

His second line of argument is that African performance in checkers and Scrabble competition would be impossible if African average IQ were anything like as low as the estimates. Both games are, at the high end, heavily g loaded. While success in Scrabble at low levels depends in part on vocabulary, the critical skill in high level plays is the ability to do the mental arithmetic needed to decide which of alternative plays will give the player the most points and his opponent the fewest. Top white players have very high IQ and many of them are mathematicians. Yet a substantial fraction of the world’s top players of checkers, including some at the very highest level, are African, and a substantial fraction of the top players of English language scrabble, including at least one world champion, are from Nigeria.

In 2015, of the ten top players in the French Scrabble championship, three were from France, three from Gabon, three more from other African countries. Gabon is an ex-French colony with a population of 1.7 million. If one believes Richard Lynn’s figures on its IQ average and standard deviation, there should not be a single person in the country close to the intelligence level required to be a top Scrabble player. Similar arguments make it very nearly impossible that Nigeria could have as many of the world’s top players of English Scrabble as it does if his estimates were close to correct.

Africans do not do nearly as well at chess, although they do not do noticeably worse than other racial groups. Chisala’s explanation is that for chess, unlike Scrabble or checkers, playing at the highest level requires extensive instruction in the literature of the game, so much so that Bobby Fisher found it necessary to learn Russian in order to read the Russian chess literature. Few Africans have the opportunity for that sort of training. Russia has dominated modern chess competition at the highest level not because Russians are smarter than other people but because the Soviet Union chose to put a lot of resources into subsidizing the training of its chess players for purposes of international prestige. They put resources into checkers for the same reason, only to find their dominance challenged by players from Africa.

The evidence Chisala offers does not tell us whether the average African genetic IQ is 95, 100, or 105, but it is clearly not 70 or even 80. That conclusion is one that those skeptical of the hereditarian position will be happy with. Other parts of his argument are not. In the process of arguing that Scrabble performance at the high end requires a high IQ, Chisala  takes on the issue of the effect of differences between the male and female IQ distribution, the same issue that got Lawrence Summers in trouble.

A possible explanation of why top physicists or mathematicians are almost all men is that women are culturally discouraged from entering such fields or discriminated against in them. That does not work for Scrabble, since more women than men play it and nearly half the qualifiers to the North American Scrabble championships are women. Yet only about 5% of the highest rated players are women and no woman yet has won the world championship. As Chisala puts it, “This rising gender disparity as you go higher in expert Scrabble is a big win for the hereditarian corner of the gender-and-intelligence debate.”

He goes on to write:

However, as we have seen many times in this research, a big win for the hereditarian side comes with a hidden pact with the devil: a victory in the gender-and-intelligence debate logically implies a decisive loss in the race-and-intelligence debate (you truly can’t have your cake and eat it in this world). How is it that black Africans, who (on average) are supposed to be about 30 IQ points below white women and supposedly have lower visuospatial or mathematical intelligence and even lower variance in their intelligence distribution, can achieve what has been accepted as statistically impossible for white women – outperforming white men – …

Chisala’s evidence that the genetic IQ of Africans is comparable to that of whites raises the puzzle of why Afro-American IQ apparently is not. One obvious possibility is that, as in the African case, observed lower IQ is due to environment rather than genetics. Chisala rejects that explanation, in part on evidence that the children of wealthy American blacks do less well than the children of poor whites, despite what one would expect to be a more favorable environment, as well as on evidence that African refugees, from a much less favorable environment, outperform American blacks. He offers instead a genetic explanation. His conjecture[5] is that a feature of African genetics makes Africans more vulnerable than whites to unfavorable mutations and that such mutations were imported into the Afro-American gene pool early on by crosses with poor whites. I found his arguments for that conjecture less convincing than his arguments against the genetic inferiority of Africans, which leaves the puzzle of Afro-America IQ, for me, still unsolved.[5]

Part of what I like about Chisala is that he has taught me something I did not know — having read him I am now confident that African genetic IQ is not significantly lower than European. Another thing I like about him is his approach to arguing. He treats Lynn and Jensen, probably the two most prominent of the hereditarian scholars, not as wicked racists but as able scholars who have, for understandable reasons, reached mistaken conclusions. Even when he finds Jensen misstating evidence in a way that makes it appear to support his position, he treats it as a single mistake in the work of a generally careful and competent scholar.

This is connected with a related feature of his work that helps make it more persuasive than other attacks on the hereditarian view of racial IQ — he takes the other side’s arguments seriously. The usual attacks I see on the hereditarian position are ones such as the claim that some races cannot have a lower IQ than others because there is no such thing as race, true in some sense of “race” but irrelevant to questions about the average IQ and average outcomes of races as conventionally defined, or ones that provide some evidence against the position but not very strong evidence, the sort of arguments that might or might not stand up against obvious criticisms.[6]

Each of Chisala’s webbed essays is followed by a long thread of comments, many trying to explain away his evidence. He responds, usually in the next essay, by carefully examining the explanation and showing why it cannot be adequate. It is because he approaches the subject in that way that he does a more convincing job of rebutting hereditarian arguments on race than other critics.

The rebuttals are sometimes entertaining as well as convincing. Responding to the argument that Africans who decide to migrate to the U.K. are a select group, much more intelligent than the African average, he offers statistics showing that many are poor, few have high end careers. He also writes, responding to one critic:

I do not really know how it works in Jamaica, but I am quite confident that realizing that life is better in a very rich country than in your poor country is never exactly the most g-loaded epiphany among Africans.

[1] I gather that Sowell later modified the theory, still attributing the result to culture but with a different explanation of its origin.

[2] Chinedu Echureo, the inventor of HopStop.

[3] “While about 8 percent, or about 530, of Harvard's undergraduates were black, Lani Guinier, a Harvard law professor, and Henry Louis Gates Jr., the chairman of Harvard's African and African-American studies department, pointed out that the majority of them -- perhaps as many as two-thirds -- were West Indian and African immigrants or their children, or to a lesser extent, children of biracial couples.” (Top Colleges Take More Blacks, but Which Ones?, NYT June 24, 2004)

[4] The series starts with and goes on through eight more essays.

[5] One tempting explanation for part of it is vitamin D deficiency. The same adaptation to a high sunlight environment by which blacks are commonly recognized, dark skin, also results in less conversion of sunlight to vitamin D, and there is evidence  linking vitamin D deficiency in pregnant women, to lower IQ of the children. I suspect that if the answer were that simple we would already know it, but it could be a partial explanation.

How might one combine that speculation with Chisala’s evidence on the academic performance of African immigrants to the U.K.? The answer may be that a large fraction of the immigrants were born in Africa in an environment where their mothers were exposed to the level of sunlight they were adapted to. If that is the whole story, it implies that the next generation may not do as well. The British diet appears to compare poorly with the American in that respect: “Unlike most other high latitude western countries, the UK does not fortify any staple food items with vitamin D, aside from a small amount added to margarine.”

[6] One used data on the illegitimate children of Afro-American servicemen stationed in Germany after the end of WWII — ignoring the fact that the fathers were not a random sample of Afro-American males. The other observed that differences in school performance between white and black students could be eliminated by a regression that controlled for differences in parental income, home environment, and the like — all of which are to some degree proxies for parental IQ.


Trey said...

David: Your entire premise here is flawed. You assume that IQ and results in games are "owned" by individuals. There is large and emerging evidence that our cognitive abilities do not belong to individuals but are a part of a global "hive mind", and are distributed and used from time to time, by a much higher global intelligence for it's purposes.

I have named this phenomenon as "the theory of atmospheric existence". The idea is that the entire Earth is some sort of "ant farm" for the higher intelligence and that memories are stored in oxygen atoms. Our brains and bodies are processors of these atoms. So memories come and go as we breathe in and out and circulate oxygen atoms in our body fluids. The higher intelligence somehow chooses who gets what and when.

The goal of the higher intelligence appears to be to gradually increase the overall welfare of the Earth. By allowing different memories in different beings to interact in seemingly spontaneous ways, the "ants" are gradually figuring out to live in peace and harmony and technological progress, even as they must endure horrendous individual outcomes along the way.

So even as the beings have evolved physically, as per the boundary conditions, they are also evolving socially. Perhaps different higher intelligences are in a universal competition to see what boundary conditions increase net happiness faster in their respective planetary experiments?

And some higher intelligences even waive their initial boundary conditions to get involved directly in their personal planetary worlds?

And such is the case here on Earth. Where a higher intelligence, even one perceived as a "God", came to Earth and lived, suffered, and died among His own creation in order to show the Way of peace and harmony and progress. His name was Yeshua, and he came into our small realm about 2000 years ago. And after living, suffering, and dying, He rose to life again, on Easter Sunday! Happy Easter David, may the Lord bless you and keep you, all the days of this life and in the world to come!

SB said...

David: It's certainly gratifying to see a scholar addressing these issues with an open mind, evidence, and the willingness to take opposing arguments seriously and distinguish clearly disproven claims from those that have some evidentiary support.

How much weight do you put on the notion of IQ or g as a real property of individuals (aside from the question of how much of it is genetic)?

For example, height is an objectively-measurable (and largely genetic) property of individuals, which correlates strongly with performance in basketball.

BMI (weight in kg divided by the square of height in m) is objectively measurable, has significant correlations with a variety of socioeconomic and health measures, but is less obviously a "real thing".

Fnorditude (your height in furlongs divided by the square of the number of vowels in your name) is also objectively measurable but almost certainly not a "real thing". Nonetheless, it's somewhat heritable, and is particularly low in people of Thai or Hawaiian origin.

One can reasonably argue that g, the single measure of innate individual intelligence, exists only because that's what researchers went looking for. When you do cluster analysis of multidimensional data on a lot of subjects, you can usually find one cluster of weak correlations, or two somewhat-tighter clusters, or three even-tighter clusters, etc. depending on what you set out to find. How do we know that g really exists as opposed to being an artificial construct or a statistical artifact? And if it doesn't really exist, what sense is there in talking about its correlations with race or performance in various tasks?

David Friedman said...

I am confident that g, imperfectly measured by IQ, is a real property of individuals, just as I am confident that some computers have the same instruction set but faster processors or more RAM than others. Do you disagree with that?

It isn't obvious that there is only one form of intelligence — again consider the analogy to processing power vs RAM — and I think it likely that there is more than one characteristic that feeds into the observable characteristics that we describe as intelligence. But the evidence as I understand it is that there is one characteristic, g, which explains a lot of the variation.

There is, as I understand it, quite a lot of evidence that what IQ measures correlates pretty strongly with a variety of outcomes — for details of quite old evidence see _The Bell Curve_. Height in basketball, similarly, is not the only determinant of success. The only difference is that we understand the nature of height, or of processing power, or of RAM, more clearly than the nature of g.

Anonymous said...

Contrary to what we commonly hear, James Watson did not discover DNA (that was isolated my Miescher back in 19th century), but was part of the effort that lead to solving its 3D helical structure.

One problem I have regarding the checkers/scrabble argument is that you can't really infer so much from just the extreme higher tail of the distribution. North Koreans are really good at the International Math Olympiads, but that's most likely because the government nurtures a few gifted individuals for the sole purpose of winning IMO medals. Something similar could go on with Nigerian scrabble-masters. The other problem is the tails of a bivariate distribution coming apart at the extremes: statistically, if you look at the very top values for something that correlates with g (i.e. scrabble champions), it becomes very unlikely that those are also in the top of the distribution in g itself.

Anonymous said...

Here's a disturbing hypothesis: When the slaves were being taken from Africa, the Africans with the highest IQ found a way to "beat the draft". They got left behind and didn't have to become slaves. The people with lesser intelligence became slaves. This would the explain a situation where African blacks are smarter than African Americans.

I'm also wondering if superior Jewish intelligence is due to a similar cause. All that oppression and all those pogroms down through 1000s of years of history culled out the Jews with lesser intelligence. The high-IQ Jews were able to survive preferentially.

It could be that a history of oppression explains both phenomena.

Eric said...

I would like to be persuaded by Chisala's argument; the consequences of extremely low average IQ in Africa are unpleasant for any libertarian. Alas, I am not.

The problem with it is that in very large ways sub-Saharan African societies look exactly as one would expect if their people have a genetic IQ of lower than 80. Or, at least, an average that is not readily manipulable upwards - it is possible to imagine non genetic explanations for this but genetics is the most parsimonious one.

One thing such a low average IQ level predicts is a short planning horizon and very high time preference. That is exactly what we see, notably in the short-termism and brutality of African politics.

A related point is that sub-Saharan African societies have proven basically incapable of generating their own mercantile elites. They've had to import Indians to fill that role.

One unfortunate thing that the last century demonstrates is that the indications of low average IQ in Africa do not change when Africans get massively better childhood nutrition. Which is the leading candidate to explain the Flynn effect in European populations.

To sum up: if the conventional estimates of SSA genetic IQ average were wrong, we would see large consequences that we do not.

Anonymous said...

I do wonder what percentage of potential immigrants from Africa are successful in finding western citizenship. I believe a Pew poll once said 40% of Nigerians were interested in emmigrating. Is it possible that we are drawing our immigrants from the top 1%?

Of course, this leaves the scrabble championships as unexplained. - Bobboccio

Unknown said...

A quick response on your comment regarding childhood nutrition. As more populations have converged in the urban areas and towns,both the quality and quantity of childhood nutrition has deteriorated. Here's an anecdote I am sure you are unaware of. A typical two roomed shanty compound shack, houses on average 11 people. In these high density low income shanties, it is common for the household to choose for example, that today 5 people will eat,then tomorrow the 6 who didn't will eat. These populations are literally starving. Its common for children as young as 5 to flood the town streets begging for food. Often times these children are mistaken for orphans, but they are not. They are just starving. Nutrition has deteriorated.

Unknown said...

Your comment regarding politicians in Africa can not go unchallenged either. The typical African politician is badly educated and often times too incompetent to make it in the private sector. So the path to riches leads to parliament and cabinet, with terrible consequences as the poverty which corrupts leads to avaricious profligacy. I'm sure you will agree that intellectuals on balance in Africa tend to shy away from politics because the self selected groups of bandits in power, tend to be brutal about protecting their primary source of income which is government. I posit that larceny more so than IQ or genetics, is responsible for our deplorable state of affairs. Trust me. Your libertarian sensibilities can rest easy. There is yet hope.